Sei sulla pagina 1di 288

Practical Methods in Low Temperature Methods in

ELECTRON MICROSCOPY BIOLOGICAL


ELECTRON MICROSCOPY

Volume 10

A. W. ROBARDS
Department of Biology
University of York
Edited by York, England
AUDREY M. GLAUERT
Strangeways Research Laboratory U. B. SLEYTR
Cambridge Zentrum fiir Ultrastrukturforschung
der Universitat fiir Bodenkultur
Vienna, Austria

1985
ELSEVIER ELSEVIER
AMSTERDAM · NEW YORK · OXFORD AMSTERDAM ·NEW YORK · OXFORD
© 1985, Elsevier Science Publishers B . V. ( Biomedical Division)

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system
or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or
otherwise without the prior written permission of the publisher, Elsevier Science Publishers B. V.
(Biomedical Division), P.O. Box 1527, 1000 BM Amsterdam, The N etherlands.

Special regulations for readers in the U.S.A .: This publication has been registered with the Copy-
right Clearance Center In c. ( CCC), Salem, Massachusetts. Information can be obtained from
the CCC about conditions under which the photocopying of parts of this publication may be made
in the USA . All other copy right questions, including photocopying outside of the USA, should
Previous volumes in the series
be referred to the publisher.
Practical Methods in Electron Microscopy
ISBN Volume: 0-444---80684-9 (Paperback )
0-444-80685-7 ( Hardback)
ISBN-Series: 0-7204-4250-8

Published by:
Volume I Part I Specimen preparation in materials science
ELSEVIER SCIENCE PUBLISHERS B.V. (BIOMEDICAL DIVISION) by P.J. Goodhew
P.O . Box 2I I Part II Electron diffraction and optical diffraction techniques
I 000 AE AMSTERDAM by B.E.P. Beeston, R.W. Horne and R. Markham
THE NETHERLANDS
Volume 2 Principles an~ practice of electron microscope operation
Sole distributors for the U.S.A. and Canada:
by A.W. Agar, R.H . Alderson and D . Chescoe
ELSEVIER SCIENCE PuBLISHING COMPANY, INC. Volume 3 Part I Fixation, dehydration and embedding of biological
52 VANDERBILT AVENUE specimens
NEW YORK, NY IOOJ7 by Audrey M. Glauert
U.S.A.
Part II Ultramicrotomy
Library of Congress Cataloging in Publication Data by Norma Reid
Volume 4 Design of the electron microscope laboratory
Robards, A. W. (Anthony William) by R.H. Alderson
Low temperature methods in biologica l electron Volume 5 Part I Staining methods for sectioned material
microscopy.
by P.R. Lewis and D.P. Knight
(Practical methods in electron microsco py; v. 10)
Part tr X-Ray microanalysis in the electron microscope
Bibliography: p. by J .A. Chandler
Includes index. Volume 6 Part I Autoradiography and immunocytochemistry
I. Electron microscopy--Technique. 2. Cryobiology-- by M.A. Williams
Technique. I. Sleytr, U. B. (Uwe Bernd) II. Title.
Part II Quantitative methods in biology
III. Series.
QH212.E4P7 vol. 10 [QH201] 502'.8'25 s 85-15872
by M.A. Williams
ISBN 0-444-80685-7 [578' .45] Volume 7 Image analysis, enhancement and interpretation
ISBN 0-444-80684-9 (pbk.) by D.L. Misell
uNrvYR'SITY OfAUCKLXND ~ Volume 8 Replica, shadowing and freeze-etching techniques
Printed in The Netherlands I h ~I oJO jo[§(
.t by J.H.M . Willison and A.J . Rowe
Volume 9 Dynamic experiments in the electron microscope
BIOLOGY by E.P. Butler and K.F. Hale
518.
nn -,
/~
General Editor's preface
to the series

Electron microscopy is a fundamental technique with wide applications .in


all branches of Science and Technology, and every year a large number of
students and research workers start to use the electron microscope and re-
quire to be introduced to the instrument and to the many and varied meth-
ods for the preparation ,of specimens.
Many books are available describing the techniques of electron micros-
copy in general terms, but when the authors of Practical Methods in Elec-
tron Microscopy first met in 1970 they agreed that there was an urgent need
for a compfehensive series of laboratory handbooks in which all the techni-
ques of electron microscopy would be described in sufficient detail to enable
the isolated worker to carry them out successfully. Now, some fifteen years
later and with the fourteenth book in the series in production, Practical
Methods in Electron Microscopy has an international reputation as a
unique source of practical information for all electron microscopists. This
achievement largely results from the great care that has been taken in the
selection of the authors, since it is well known that it is not possible to de-
scribe a technique with sufficient detail for it to be followed accurately un-
less one is familiar with the technique oneself. This fact is only too obvious
in certain ' one author' texts in which the information provided quickly ceas-
es to be of any practical value once the author moves outside the field of
his own practical experience. Perhaps the best compliment that Practical
Methods in Electron Microscopy has received was a statement in a review
of the book by Alan Agar, Ron Alderson and Dawn Chescoe: " The book
leaves the reader with the impression of not having read a textbook, but
having talked to a practising microscopist".
Our aim continues to be to help the worker in the laboratory (and parti-
cularly the isolated worker) to be a better electron microscopist. Each book
VIII Editor's preface

of the series starts from first principles, assumuing no specialist knowledge,


and is complete in itself. The books are planned to be guides through the
often bewildering choice of techniques that are available, and every attempt
is made to simplify this choice. Consequently only well-established techni-
ques which have been used successfully outside their laboratory of origin
are described in detail. We do not provide descriptions of the latest exciting
'break-through', although each book hopes to be able to indicate the most Authors' preface
promising of future developments.
This series of books will eventua lly cover the whole range of techniques
for electron microscopy, including the instrument itself, methods of speci-
men preparation in biology and materials science, and techniques for the
analysis of the image. In addition, as envisaged by the authors in their origi-
nal plan, each book will be revised , independently of the others, at such time This book grew out of a long friendship and scientific collaboration with
as the authors and editor consider necessary, thus keeping the series of a special mutual interest in the use of low temperatures in ultrastructure re-
books up-to-date. Consequently a number of the earlier books are now in search. Looking back some years, to when preliminary conversations were
the process of being revised, and for most of them this revision will be so made in the appropriate setting of the ice caves under the Great Dachstein,
extensive that they will emerge effectively as new books and will therefore we now realise that we were totally unaware of the vast amount of work
be given new titles. that would be involved In surveying the techniques in such a rapidly devel-
This whole enterprise would not have been possible without the fullest oping fielc;I. In common with many of our fellow biologists, we have often
support and collaboration of our publisher. North Holland/Elsevier have found ourselves in the position that, while realising the immense potential
continuously helped us to maintain a high standard in all aspects of the importance of low temperature methods, satisfactory practical means for
books, from the accuracy of the text and figures to the high quality of the achieving acceptable results have not always been available. Indeed, despite
final volumes. I look forward to a future in which authors, editor and pub- much progress over the past few years, this is a circumstance that, to a con-
lisher continue to contribute to the Science and Art of Electron Microscopy. siderable extent, still prevails. The present trend towards the collaboration
of biologists, physicists, engineers and, not least, commercial people means
Strangeways Research Laboratory AUDREY M. GLAUERT, Sc. D . that low temperature methods will become more effectively, more widely
Cambridge , England General Editor, March 1985 and more easily available.
This book does not and cannot exclusively present well tried and tested
methods. Rather, it attempts to pursue a middle course whereby the reader
is not only made aware of what can be done but is also informed of back-
ground details so that, as the subject develops further, he or she will have
the scientific basis upon which to build. The purpose is to paint the back-
ground against which future developments will take place. The book tells
the reader both how to chose and implement different methods as well as
why. It also attempts to provide an adequate gateway into the literature so
that those who wish to extend their knowledge further will be able to do
so.
Low temperature methods are already used in very many electron micros-
cope laboratories and it seems highly probable that, sooner or later, all such

TX
x Author's preface

laboratories will call upon these methods as they become better established.
They are often ignored for a variety of reasons, including that they are com-
plex and expensive. This book will serve both to help the absolute beginner
to progress into this fascinating field as well as to provide additional infor-
mation to those who are already familar with some, if not all, cryotechni-
ques . Acknowledgements
York and Vienna, February 1985 A.W. ROBARDS and U.B. SLEYTR

We are very much indebted to our many friends and colleagues who have
helped in the production of this book by reading through various Chapters,
in whole or in part, correcting errors and suggesting changes and improve-
ments. It is impossible to express our thanks to each and every one of you
and we hope that it will suffice to know that the willing help that we have
uniformly received when it has been requested has been greatly appreciated.
We should, however, specifically like to thank Dr. John Baker, Prof. Stan
Bullivant, Dr. Hugh Elder, Dr. Diana Harvey, and Dr. Walter Umrath for
their invaluable help with the manuscript. Despite all the help that we have
received we remain , of course, fully responsible for any errors that have not
been eliminated.
In addition to those who have contributed illustrations, and are acknow-
ledged in the accompanying text, we would like to thank Dr. P. Messner,
the late Hilary Quine and Meg Stark for help with some of the diagrams
as well as E. Pohoralek and Ch. Robien, who prepared the photographs for
Chapters 5, 6 and 7. A special word of thanks goes to Sue Sparrow who
has skilfully prepared the final version of the many diagrams presented here:
her expertise and good humour, even when diagrams appeared yet again for
still further changes, has been greatly appreciated . For additional general
assistance with typing and word-processing we are most grateful to Joan
Chambers and Wendy Crosby.
We should also thank the members of our laboratories for their unceasing
support and understanding when we have, perhaps, been more occupied
with this major project than they would have wished!
Finally, it is not inappropriate to say a word of special thanks to our Edi-
tor who, with expert guidance and encouragement, has undoubtedly caused
us to produce a much better book than would have resulted if we had been
left entirely to our own devices.
XI
Nomenclature

A =area
r:x = thermal diffusivity
Bi = Biot modulus (dimensionless)
C = capacitance
cP = specific heat c~pacity at constant pressure
D =length
d = thickness
L1 T = temperature difference
E = total emissivity
1:: = dielectric constant
g = acceleration due to gravity
h = surface heat transfer coefficient
r; = coefficient of evaporation
I =ice
I0 = cubic ice
Ih = hexagonal ice
Iv = vitreous ice
I · = intensity
= current
J = flux
J. = actual sublimation rate
J ct = departing flux
Ji = impinging flux
J, = absolute rate of sublimation of ice
j = J- 1
k = thermal conductivity
A = mean free path
XIV Nomenclature

Ar = latent heat of fusion


Av = latent heat of vaporisation
M = molecular weight
m =mass
µ =viscosity
NNu = Nusselt's number
NRe ~ Reynolds' number Contents
NPr Prandtl's number
n =number
P pressure
Pc partial pressure
P, saturation vapour pressure
TC pi (3 .14) Editor's preface . VII
Q heat flux
A uthors' preface IX
R outer radius
r inner radius Acknowledgements . XI

p density Nomenclature XIII


s concentration
surface tension
temperature Chapter 1. Introduction . 1
temperature cold body
freezing temperature Chapter 2. Freezing 5
temperature of homogeneous nucleuation
initial temperature 2.1 Principles of freezing 5
melting temperature 2.1.1 Ice nucleation 5
2.1.2 Phase separation and eutectics II
equilibrium temperature between solid surface and specimen
2.1.3 Warming frozen specimens. . 20
temperature of hot body 2.2 Cooling mechanisms . . . . . . 21
Tat distance x after time t 2.2.1 What is the actual cooling rate? 22
temperature bulk fluid 2.2.2 Measurement of cooling rates . 25
2.2.2a Thermocouples and temperature measurement 25
time
2.2.2b Electrical capacitance method for determining freezing rates . 35
<p gas constant (Stefan Boltzmann constant)
2.2.2c Evaluation of cryofixation in frozen samples 38
v potential 2.2.3 Heat transfer in cryogenic fluids . . . . . . . . . . . . 38
v velocity 2.2.3a Characteristics of liquid cryogens. . . . . . . . . 47
critical cooling velocity :f.2.3b The effects of specimen surfaces and shapes on heat transfer . 50
(J) frequency 2.2.4 Heat capacity . . . . . . 53
2.2.5 Cooling rates in liquid cryogens 53
x distance 60
2.2.6 Heat transfer at cold surfaces .
2.2.7 Comparison of cooling methods . 65
2.3 Treatment of biological specimens prior to freezing 66
2.3.1 Cooling rates in relation to structure and viability 66
Contents XVII
XVI Contents

Chapter 4. Freeze-sectioning 201


2.3.2 Principles of sampling a nd pretreatment 71
2.3.2a Cryoprotectants 72
4.1 Principles and problems 201
2.3 .3 Practical methods of sampling and pretreatment. 79
4.2 Equipment 205
2.3.3a Pretreatment of isolated cells and cell components 79
4.2. l Basic requirements of cryoultramicrotomes 206
2.3 .3b Pretreatment of plant material. 83
4.2.2 Cryoultramicrotomes 207
2.3.3c Pretreatment of animal tissue . 86
4.2.2a Cryostat ultramicrotomes . 208
2.3.3d Ballistic methods of sampling and freezing. 88
4.2.2b Ultramicrotome cryoattachments 210
2.4 Freezing methods . 91
4.3 Sectioning . 218
2.4.1 Freezi ng with liquid cryogens . 91
4.3. l Specimen preparation for sectioning . 218
2.4.la Immersion method 96
4.3. la Pretreatment 218
2.4.1 b Propane jets 110
4.3 . lb Mounting specimens for freezing . 219
2.4. lc Spray-freezing 114
4.3. lc Freezing specimens 220
2.4.ld High pressure freezing 116
4.3. ld Mounting specimens on the ultramicrotome 223
2.4.le Liquid paraffin method 119
4.3.2 Dry sectioning . 225
2.4.lf Droplet method 120
4.3.2a Cutting the sections 225
2.4.2 Freezing on cold surfaces . 121
4.3.2b Section collection 227
2.4.3 Comparison of freezing methods. 127
4.3.2c Section mounting . 230
2.5 Subsequent processing and storage of specimens 128
4.3.3 Wet sectioning 233
2.6 Cooling and viability after thawing 130
4.3.3a Contrast enhancement 234
4.4 Applications of freeze-se~tioning 235

Chapter 3. Direct viewing and analysis of cold specimens ..--. ·. 147


Chapter 5. Freeze-drying 243
3.1 Principles and problems . 147
3.1.1 Types of specimen 150
5.1 Introduction . 243
3.2 Instrumental requirements for low temperature electron microscopy 151
5.2 Basic principles of freeze-drying 247
3.2.1 Microscopes 151
5.2.1 Theory of drying by sublimation . 247
3.2.2 Transfer devices . 155
5.2. la Rate of sublimation of ice 247
3.2.3 Other ancillary equipment . 156
5.2.1 b Rate of freeze-drying of biological specimens . 251
3.3 Frozen bulk specimens 158
5.3 Practical aspects of freeze-drying . 256
3.3.1 Frozen-hydrated bulk specimens. 159
5.3. l Basic principles 256
3.3.2 Frozen-fractured bulk specimens. 163
5.3.2 Freeze-drying apparatus designs . 261
3.3.3 Preparation 164
5.3.2a Simple vacuum systems for freeze-drying 262
3.3.4 Fracturing . 165
5.3.2b Modern vacuum systems for freeze-drying. 262
3.3.5 Sublimation 170
5.3.2c Simple freeze-drying devices for use in vacuum systems 264
3.3.6 Coating 171
5.3.2d Simple freeze-drying device for immersion in liquid nitrogen . 266
3.3. 7 Transfer . 173
5.4 Freeze-drying from non-aqueous solvents . 268
3.3.8 Viewing and analysis 175
5.4.1 Procedure for freeze-drying 270
3.4 Frozen thin specimens. 177
5.5 Freeze-drying procedures for different types of specimen 272
3.4.1 Frozen thin sections . 178
5.5. l Freeze-drying of small biological specimens . 272
3.4.2 Frozen thin layers . 183
5.5. la Specimen mounting and freezing . 273
3.4.3 Preparation 184
5.5. lb Freeze-drying and subsequent handling. 277
3.4.4 Thin layers . 185
5.5.2 Freeze-drying of bulk specimens . 278
3.4.5 Cryosections 187
5.5.2a Examination of bulk specimens in the SEM 279
3.4.6 Coating . 189
5.5.2b Specimen replication for the TEM 285
3.4. 7 Transfer . 189
3.4.8 Viewing and analysis 193
XVIII Contents Contents XIX

5.5.2c Freeze-drying and embedding for thin sectioning. 287 6.4.2 Replica strengthening methods . 385
5.5.2d Freeze-drying of cryosections . 292 6.4.2a Silver technique 386
5.6 Freeze-drying artefacts and problems of specimen interpretation . 293 6.4.2b Other reinforcement (and replica-stabilising) methods . 386
5.6.1 F ixation artefacts . 293 6.5 Etching 388
5.6.2 Adsorption artefacts . 293 6.5.1 Physical conditions for sublimation and condensation. 388
5.6.3 Washing artefacts 293 6.5.2 Etching methods and contamination phenomena 392
5.6.4 Air-drying artefacts 294 6.5.3 Etching of non-aqueous systems . 396
5.6.5 Freezing and recrystallisation artefacts. 295 6.6 Thawing and replica cleaning . 397
5.6.6 Dehydration artefacts . 295 6.7 Mounting of replicas on grids . 402
5.6.7 Rehydration artefacts 299 6.8 Interpretation of freeze-fracture replicas. 403
5.6. 8 Shadowing and/or replication artefacts. 300 6.8.1 General considerations . 403
5.6.9 Freeze-drying: structural comparisons with other dehydration procedures 300 6.8.2 Interpretation of membrane structures and nomenclature. 409
6.8.3 Artefacts 414
6.9 Cytochemical and qu antitative methods . 418
Chapter 6. Freeze-fracture replication . 309 6.9.1 Freeze-fracture autoradiography . 419
6.9. I a Freeze-fracture autoradiography of bulk specimens . 420
6.1 Principles of freeze-fracture replication and freeze-etching techniques 309 6.9. lb Freeze-fracture autoradiography in combination with the sectioned
6.2 Specimen preparation . 313 replica technique . 424
6.2. l Sampling and pretreatment 313 6.9.lc Freeze-fracture autoradiography of monolayers . 424
6.2.2 Specimen supports and specimen mounting 313 6.9.2 Labelling with morphologically detectable markers for replica techniques 428
6.2.2a Single cells and cell components 315 6.9.3 Labelling with electron-opaque markers for the sectioned replica technique. 430
6.2.2b Bulk specimens 316 6.9.4 Specific pre-freezing alterations of the molecular organisation of membranes 430
6.2.2c Monolayer cul tures . 318 6.9.5 Post-fracture labelling procedures 431
6.2.2d Specimen mounting for complemen tary replica methods 325 6.9.5a Thin sectioning of post-fracture labelled specimens . 431
6.2.3 Freezing and storage 327 6.9.5b Replicating critial point-dried, post-fracture labelled specimens . 434
6.2.4 Transfer of frozen specimens 328 6.9.5c Replicating freeze- or critical point-dried, . post-fracture, post-
6.3 Fracturing 329 sha<low labelled specimens 434
6.3. l General considerations . 329 6.9.5d Thin sectioning of post-fracture, post-shadow, labelled and repli-
6.3. l a Fine topography of the fracture plane 329 cated specimens 435
6.3.1 b Gross topography of the fracture plane . 330 6.9.6 Lipid extraction from fracture faces . 436
6.3. lc Conditions for freeze-fracturing 332 6.9. 7 Specific decoration of fracture faces in vacuo with water vapour 437
6.3.2 Vacuum units for freeze-fracture replication 333 6.9.8 Molecular and particle weight determination . 438
6.3.3 Instruments for freeze-fracturing. 336 6.9.9 Freeze-fracturing and freeze-fracture replication in combination with
6.3.3a Freeze-fracturing under vacuum . 336 negative staining . 439
6.3.3b Freeze-fracturing under improved vacuum conditions . 347 6.9.10 Quantification of structural details and interpretation aids . 440
6.3.3c Freeze-fracturing outside the coating unit under a liquid gas . 361
6.3.3d Freeze-fracturing outside the coating unit at atmospheric pressure . 365
6.4 Shadowing and replication . 367 Chapter 7. Freeze-substitution and low temperature embedding . 461
6.4.1 Evaporation and evaporators . 367
6.4.1a Formation of thin films 367 461
7. 1 Introduction .
6.4.1 b High resolution shadowing and contamination problems 369 465
7.2 Substitution media .
6.4.1 c Methods of resistance heating . 375 469
7.3 Freeze-substitution apparatus
6.4. ld Methods of electron-beam evaporation. 378 474
7.4 Freeze-substitution procedures .
6.4. le Monjtoring film thickness 380 474
7.4.1 Ultrastructural studies .
6.4.1f Rotary shadowing 383 7.4.1 a Freeze-substitution for thin sectioning 475
6.4.1g Bi-directional shadowing 385 7.4.1b Freeze-substitution for scanning electron mjcroscopy 484
xx Contents

7.4.2 Localisation of water-soluble substances


7.5 Low temperature embedding techniques .
485 Chapter 1
486
7.5. 1 Low temperature resins . . . . . . 487
7.5.2 Fixation and dehydration . . . . . 487
7.5. 2a Maintaining low temperatures 489
7.5.3 Infiltration and polymerisation . . . 491
7.5.3a Preparation of resins and infiltration. 491
7.5.3b Polymerisation .
7.5.4 Sectioning and staining .
493
495
Introduction

Chapter 8. Future outlook . 501

8. 1 Freezing 502
8.2 Direct viewing methods 503
8.3 Cryo ultramicrotomy 504
8.4 Freeze-drying 505 Life on earth began in water, evolved from water, and remains critically de-
8.5 Freeze-etching 505 pendent on water. Most cells of living organisms contain more than 70%
8.6 Freeze-substitution 506 water. The interior of an electron microscope is evacuated to a very low
8.7 Conclusion 506
pressure of less than 1 x 10- 5 Torr (1.33 x 10- 3 Pa) . This is approximately
76 million times less than normal atmospheric pressure. It follows from this
Appendix 1. Safety . 507 that, if a living cell is put into the vacuum of an electron microscope, the
water usually evaporates extremely quickly with disastrous consequences
A 1.1 Direct contact with cold gas, liquid or solid . 508 fo r the preservation of structure. If we want to observe or analyse biological
A 1.2 Problems of inhalation and asphyxiation with cryogenic gases 509
cells and tissues in electron-beam instruments it is therefore necessary to
Al.3 Expansion associated with decompression of liquefied gases 510
A l .4 Inflammable nature of some cryogenic gases 510 take steps so that such cataclysmic artefacts are avoided . In principle, there
A l .5 Toxic nature of some cryogenic gases . 512 is a number of possibilities: the specimen could be contained at approxima-
Al.6 Concl usions . . . . . . . . 512 tely atmospheric pressure within an electron-transparent envelope (the so-
called 'environmental cell'); the material could be preserved (fixed) so that
subsequent dehydration produces minimal interference with the in vivo
Appendix 2. List of suppliers . 513
structure; or the water could be frozen so that it would not evaporate in
Index to list of suppliers . 529 the electron-beam apparatus. Environmental cells (Butler and Hale 1981)
pose many different problems and, while they have made valuable contribu-
Subject index . 531 tions in some areas of electron microscopy (mainly non-biological), they do
not provide a routine means for the observation and ana lysis of biological
specimens at high resolution. Fixation, dehydration and embedding do, of
co urse, together constitute the major pathway by which the vast majority
of biological specimen processing is carried out. Numerous books and pa-
pers have been devoted to this subject (e.g. Glauert 1974; Hayat 1981) and
more will not be said here beyond noting that such techniques almost al-
ways involve the removal of water. Using 'wet' processing methods, the
withdrawal of water inevitably leads to changes in the native biological
2 Robards and Sleytr Low temperature methods in biological electron microscopy Introduction 3

Pressure ( Pa)
structure; many soluble materials will be extracted, and, further, the in vivo 10-20 10 2 10 4 10 6 10 8
distribution of ions and other mobile components will be radically changed. 400
Such considerations have led to much greater attention being paid to an 650

alternative possibility for preparing biological specimens: that of immobilis- 300


550
ing the water by freezing it. In principle this may sound straightforward but
~ g
in practice it is not! This book is devoted to the practical assessment of low - 200
~ 450
temperature techniques for the preparation of biological specimens for elec- .
.a ~"
l 100 i"'
tron microscopy.
In learning how to freeze cells so that they are preserved in as life-like
E
..
....
350 ~
....•
a state as possible, we could do much worse than to look at the responses 250

of cells to natural chilling in Nature. Reduction of the ambient temperature


-100 150
below the norm for a particular cell type is accompanied by a reduction in
the rate of energy-requiring processes - 'life' slows down. More severe (but -200
above 0°C) cooling will, in many cells, result in a metabolically inactive state 10-25 101 103 10 5
PreHure (b8')
in which life can appear to be suspended. Ultimately, some cells in Nature
are subjected to such low temperatures that water would normally freeze. Fig. I. I. Pressure-temperature phase diagram for water. There are many different units used
for the citation of pressure. We generally use the pascal (Pa) which is the appropriate SI unit.
Under these circumstances there can be a number of different responses: However, because pressures are frequentl y cited in torr or bar, these are sometimes also quoted
many cells do not possess cold tolerance at all and will die; others may con- in this book. 1.0 Pa= 7.5 x 10 - 3 Torr= 1.0 N m - 2 = 1.0 x 10 - 5 bar =9.87 x 10- 6 atm.
tain relatively small amounts of water, thus minimising the problems caused
by ice formation; and still others can actually produce chemicals that will
act as antifreezes and will thus protect the cell and, in some situations, allow effect; the rate of change can also be important. Indeed, as we shall see later
it to function normally at sub-zero temperatures (e.g. Zachariassen and (§2.3 .1), the rate of cooling is absolutely critical in determining the nature
Husby 1982). Only by appreciating the structure of liquid water, the struc- of the final, frozen cell. Although a sound understanding of the structure
ture and formation of ice and the response of cells to different temperature and behaviour of water is important, it should also be appreciated that the
regimes can we hope to develop the optimum methods of artificially impos- reactions of the pure liquid are not necessarily reflected within living cells.
ing freezing techniques for the preservation of cells. This is a point that is frequently overlooked and, as yet, inadequately stu-
Water is a curious and special substance. Many of its properties are quite died. For example, while parameters such as the recrystallisation tempera-
dissimilar to any other molecule (Eisenberg and Kauzmann 1969; Franks ture (§2.1.1) or cooling rate to achieve vitrification can be relatively easily
1972- 1982). In order to understand the interrelationships between the pro- and precisely determined for the pure substance, they will certainly be quite
cesses of life and water we should thus attempt to gain the best possible in- diffe rent for water as a solvent within cells. Moreover, such parameters will
sight into the structure and behaviour of water itself (Eisenberg and Kauz- vary greatly from cell type to cell type or even between adjacent cell com-
mann 1969; Fletcher 1970; Franks 1972- 1982, 1983; Hobbs 1974). Only partments.
then can we logically attempt to adjust the normal relationship between a As this book progresses through the description of low temperature tech-
cell and its (liquid) water without creating excessive artefacts. niques, the relevance of these points and the need for further work will be-
The pressure/temperature phase diagram for water (Fig. 1.1) illustrates come evident. Nevertheless, there is now an adequate platform of sound
the relationships between the three physical phases of the molecule. Water theoretical and experimental knowledge so that many techniques can be de-
within cells is usually liquid but changing the ambient conditions of temper- scribed with complete confidence. In other instances (and cryoultramicro-
a ture and/or pressure can bring about a change to either the vapour or solid tomy is a good example, see Chapter 4), there remains much to be done be-
phases. Not only does a bso lute change of pressure or temperature have an fore we can express, with conviction, an adequate understanding of the
4 Robards and Sleytr Low temperature meth ods in biological electron microscopy

physical principles of the technique which would lead to its use as a more
Chapter 2
routine method than is the case today. Although low temperature techniqu-
es create their own artefacts, these are often more easily understandable
than those arising during chemical treatments because only physical pro-
cesses are involved.
Low temperatures can carry some degree of risk in their use which may
not be encountered in other environments. For this reason a complete Freezing
Appendix (Appendix 1) has been devoted to the topic of safety in the use
of low temperature techniques. It is recommended that this is read by new-
comers to the field before any practical work is attempted.
Finally, in this brief introduction, it is worthwhile pointing out that low
temperature methods neither supersede nor replace conventional prepara-
tion techniques for electron microscopy: they should rather be thought of
as augmenting them. There is now a vast range of microscopical methods. The process fundamental to the whole subject of low temperature micro-
The most effective worker will be the one that has the knowledge, ability scopy is freezing: the conversion of water from a liquid to a solid state. This
and (it is to be hoped) equipment to select the most appropriate course of apparently simple phase transition is, in fact, a complex series of events
action to achieve his particular end . which is modified by s1,1ch factors as cooling rate, presence of solutes, and
pressure . In biological systems, the physicochemical changes during freezing
remain poorly understood and many of our present practical methods are
References based more on empirically judged success than on a true understanding of
the changes taking place in the cooling tissue. The questions that need to
Butler, E. P. and K. F . Hale ( 198 1), Dynamic experiments in the electron microscope, in: Prac- be asked in the context of cryobiological research are primarily: what cool-
tical methods in electro n microscopy, Vol. 9, A. M . G lauert, ed. (North-Ho ll and , Amster-
ing rate is optimal and over what temperature range; can chemical additives
dam).
Eisenberg, D. and W. Kauzman n (1969), T he structure and properties of water (Oxford be used to improve the quality of the frozen specimen withou t producing
U niversity Press). unacceptable side-effects; and what is the highest temperature to which we
F letcher, N. F. (1970), The chemical physics of ice (Cambridge University Press). can safely subject the frozen specimen? To answer these questions it is
Franks, F. (1972- 1982), Water - a comprehensive treatise (Plenum Pub lishing Corpora tion, necessary to understand something of the principles of freezing and thawing
New York).
and we shall therefore consider these processes before going on to deal wi th
Franks, F . (1983), Water (Royal Society of Chemistry).
Gla uert, A. M. (1974) , Fixation, dehydration and em bedd ing of biological specimens , in: Prac- more practical methods.
tical methods in electron microscopy, Vo l. 3, A. M . G lauert, ed. (North-Holl and , Amster-
dam).
Hayat, M . A . ( 198 1), Fixation for electron microscopy (Academ ic Press, New Yo rk). 2. 1 Principles offreezing
Hobbs, P. J. (1974), Ice phys ics (Clarendon Press, Oxford).
Zachariassen , K. E. and J . A. H usby ( 1982), Antifreeze effects of therma l hysteresis agents pro-
tects highly supercooled insects, Na ture , La nd . 298, 865.
2. 1. 1 Ice nucleation

When pure water is cooled below its equilibrium temperature, Tr, the temper-
ature at which two phases are in equilibrium (for water and ice Tr = 0°C,
273 K), it does not freeze immediately but remains in a metastable super-
cooled state. Super-cooling cannot continue indefin itely and eventually
6 Robards and S leytr Low temperature methods in biological electron micrmcopy Freezing 7

spontaneous crystallisation occurs, accompanied by release of the latent


heat of fusion which warms the water back towards its melting point (Fig ..
2.1; see Stephenson 1956, 1960). The degree of super-cooling depends on the
purity of the water but cannot fall below approximately 0.8 Tr (measured
'i
in Kelvin; Fletcher 1970). The reason for this is that, as the water cools, "'
small ice crystals form but, close to the equilibrium temperature, they imme-
diately dissociate (melt) due to their energetically unfavourable surface/
volume ratio. However, as cooling continues, these crystals become more
stable until they eventually grow and act as the seeds for crystallisation of
the whole volume. Such freezing is referred to as homogeneous nucleation.
The subsequent rate of growth of ice crystals is temperature-dependent (Fig.

10 1 ~-:-----:-J__~_l_~~_J_~~L_~_L~-=~~~
270 250 230 210 190 170 150 130
250 Temperature (K)

Fig. 2.2. Rate of recrystallisation of ice in relation to temperature. The rate of ice crystal
grow th is temperature-dependent with a max imum rate at about 260 K ( - 13°C). (Curve I re-
225 drawn from Riehle 1968a; curve 2 adapted and redrawn from Moor 1973.)

~
~ 2.2), with a maximum at about 260 K, and falls as the temperature decreases
:J
n; 200
iii and the liquid viscosity increases until , at about 140 K, ice crystals no longer
c.
..
E
f-
grow . Fra nks ( 1980) has emphasised the contrast between the behaviour of
wate r in the normal physiological temperature range (0 to 100°C) and super-
175 cooled wa ter. For example, the heat capacity of liquid water in the normal
temperature range is essentially constant at 75 J mol - 1 K - 1; at -35°C
su per-coo led water has a heat capacity of I 05 J mol - 1 K - 1, whereas the heat
150 ca pacity for ice at the same temperature is only 33 J mol - 1 K - 1 (Fig. 2.3).
Such factors affect the rate of cooling and , therefore, the morphology and
size d istribution of ice crystals that appear at the super-cooled limit. Yannas
125'--~'----'"---~"---~-'-'--'-....__~_._..__....__~
( 1968 ) gives the vitrification temperature of water (the temperature above
0 10 20 30 40 50 60 70 80
Time (ms)
which vitrified water would transform into the crystalline state) as 127 ± 4
Fig. 2.1. Theoretically computed cooling curves for water al different rates of heat remo val K . The significance of this is that ice crystals will form and grow at all tem-
(Q) (GJ = gigajoules = I x I 09 Joules). The lower two curves illustrate partial vitrifica ti on of the pera tures between the homogeneous nucleation point (Th) and the recrystalli-
water; the upper three curves illustrate complete transfonnation of water to crystalline ice. sation point (T,): i.e. between about 230 and 130 K. Only if this range is tra-
Thus, only for the two highest ra tes of cooli ng is the heat of fusion removed more quickly th an
it is produced. In the other three cases some degree of rewarming is evident. (Redrawn from
versed sufficiently rapidly (i.e. faster than the critical cooling velocity, vc),
Stephenson 1956.) so that the heat of fusion is removed more quickly than it is produced , can
Freezing 9
8 Robards and Sleytr Low temperature methods in biological electron microscopy

critical interval from approximately 270 to 180 K is probably still too large
Temperature (K)
230 250 270 290 310
to be bridged in the super-cooled state (Fig. 2.4). In homogeneous nucle-
ation, ice crystals themselves form the seeds for crystallisation but the stabi-
120
lity of a 'crystal embryo' is considerably enhanced if it grows in association
Amorphous ice at -150°C (123K)
I I
I
with the surface of some insoluble foreign particle ~20 nm in diameter
l<:
7 /
/ (Fletcher 1970). In such a system super-cooling is much less and the freezing
0 Liquid: water process is known as heterogeneous nucleation. It will now be evident that
E 80
2 maximum super-cooling can only take place in extremely pure water and
!' that small droplets can be super-cooled further than equivalent large ones
·;;
....
Q.
Water vapour
because, although the total number of 'seeding' particles per total volume
"
«i
•:r
40
at
100°C (373 K) 270 ~
...... ... l
•..l•
- 0

•• ••

••
•••
-20
250 f-

- 40 -20 0 20 40
S-+• • •
••
5 ..1

Temperature (°C)
• -40

·~1
230
R•'
Fig. 2.3. Heat capacity of liquid water and ice as a function of temperature. (Redrawn from
- -60
Franks 1980.) 210
Q 0
t..
~ R .. •
crystallisation be totally avoided. The virtue of maximum super-cooling pri- z. 190
- -80
z.~
or to freezing becomes more apparent from the knowledge that, whereas the 4i ~
Q. Q.

latent heat of fusion at 0°C is 334 J g- 1, it is only about half this value at ..E
I-
170 f-
- -100 E

I-
the homogeneous nucleation point (approximately - 40°C; Franks 1980).
-120
150 ~
With the possible exceptions of the work of Briiggeller and Mayer (1980),
· Mayer and Briiggeller (1982) and Dubochet and McDowall (1981), such
130
••
R ..
-140
rapid heat extraction has never been achieved for pure water where the criti-
cal range is at least from 230 K to 130 K (Fig. 2.4). Indeed, Rasmussen
- -160
110
(1982) concludes, from a re-evaluation of the energetics of homogeneous
nucleation, that " the nucleation of ice cannot be avoided simply by rapid 2 3 4 - -180
90
quenching" . However, the addition of solutes and the consequent reduction
in the total proportion of available water for freezing both Jowers the freez- Fig. 2.4. Variat ion in freezing and recrystallisation behaviour for water and different cell
ing point and raises the recrystallisation point, hence reducing the critical types. ( 1) Pure water melts a t 273 K (*)but may be super-coo led to abo ut 235 K (S) . Recrys-
tallisation of ice can take place down to temperatures as low as abo ut 130 K (R). Crystal
temperature interval through which very rapid cooling must take place. growth is therefore possible anywhere within the range 273 to 173 K. (2) In active, li ving cells
Work such as that of the Dubochet group (Dubochet and McDowall 1981 ; the melting poi nt is depressed and the recrystallisation point is raised, thus reducing the critical
Dubochet 1984) gives real hope that, at least for some very small specimens, temperature interval in which crystallisation can occur. (3) The si tuation is further improved
in frost -hardy or cryoprotected cells, where the melting point can be lowered by about 1O K
true vitrification may be achieved. and the recrysta llisa tion point raised to approximately 230 K. Under some circumstances, the
In 'typical' biological cells containing more than 80% water, the freezing super-cooli ng may be so great that ihe cells can be solidified without crysta lli sa tion . (4) In dry
point is only depressed by about 2 K below that of pure water, although cells, where no wa ter is presen t, ice cannot form and , therefore, there can be no crystallisation
damage.
the recrystallisation temperature is raised by about 50 K; nevertheless, the
---- -----------...-------~----------------

10 Robards and Sleytr Low temperature methods in biological electron microscopy Freezing 11

of liquid will be the same, the chance of having a seeding particle in any 2.4). Under these circumstances the ice crystal size is inversely correlated
one droplet becomes less as the droplet size is reduced. For example, with cooling velocity; that is, the slower the cooling rate, the larger the size
Heverly (1949) found that water droplets of about 0.4 mm diameter could of the ice crystals formed (Fig. 2.5). This poses one of the basic problems
be super-cooled to 257 K, whereas droplets of 50 µm diameter could be of low temperature biological microscopy.
super-cooled by a further 14-19 K. It might be expected that biological spe-
cimens would freeze primarily through heterogeneous nucleation . This ap- 2. 1.2 Phase separation and eutectics
pears not to be so as most cells super-cool well below the predicted level
(Mazur 1966; Rebhun 1972). The explanation for this probably lies in the The process of ice crystallisation automatically involves phase separation:
fact that up to 15% of cellular water is bound to macromolecules and does as ice crystals grow, so solutes become more concentrated. This is an ines-
not freeze at low temperatures, thus removing a large component of the capable fact , no matter what the ice crystal size. Phase separations can have
potential heterogeneous nucleating sites. The number of water molecules serious repercussions on structural interpretation and can render microana-
arranged in a crystal-like form to provide a seed for homogeneous nucle- lytical methods effectively impotent. Ideally, the solution to this problem
ation is inversely related to temperature so that the probability of nucleation would be to vitrify the specimen . Vitrification is the formation of a glassy
increases as super-cooling continues until , at about 235 K ( -38°C), sponta- or amorphous state without the production of ice crystals. (Strictly, we
neous freezing occurs even in absolutely pure water. In the presence of het- sho uld not refer to 'vitrified ice' as 'ice' signifies a crystalline state. However,
erogeneous nuclei, freezing occurs at higher temperatures during cooling at the term is used relatively indiscriminately in the literature and its meaning
moderate (i.e. not ultra-rapid) rates at atmospheric pressure although the is clear enough. Furthermore, the distinction has sometimes been made
super-cooling temperature may be lowered by the presence of additives (Fig. between 'vitreous ice', only obtained by ultra-rapid freezing, and 'amor-
phous ice', which is formed by condensation at low temperature in a
vacu um. Some properties of these two forms , for example the density, do
1 2 3
differ but need not concern us here. A review of the properties of vitreous
water is provided by Pryde and Jones (I 964).) If such 'amorphous' solid
water could routinely be produced in specimens, then many of the major
limitations to low temperature ultrastructural research would immediately
be overcome . However, critical cooling rates (vc) to achieve true vitrification
I have rarely been reported (Briiggeller and Mayer 1980; Dubochet and
I
I McDowall 198 I; Mayer and Briiggeller 1982) for pure water nor, under nor-
I mal circumstances, have they been adequate to vitrify untreated hydrated
I
I biological specimens at normal atmospheric pressure (Table 2.1; Fig. 2.5).
I Menold et al. (1976) have reported that cooling rates of approximately JOO
10 --+- - - - - - - - - - - - - - - - - - - +- - ---~- 1
K s- give ice crystals of 4 to 5 µm while rates of several thousand K s- 1
I' II give crysta ls of 1 to 2 µm in aqueous suspensions. Biologists have fre-
Cooling rate
quently, and erroneously, used the term 'vitrification' to describe freezing
Fig. 2.5. Ice crystal size in relation to cooling rate in biological cells. (l) At very low ra tes
of cooling (K min - 1) , water is withdrawn from cells; there is thus less water to crystallise and w_here the ice crystals are smaller than the resolution of the measuring tech-
the ice crystals formed are small. (2) At intermediate cooling rates (K s - 1 to > 1000 K s - 1), n~que employed. For example, Riehle and Hoechli (1973) consider the
large intracellular ice crystals are formed. (3) Above a critical cooling velocity (which will va ry vitreous state to be "a solid modification of the water with crystals less than
from cell type to cell type, depending on, among other things, intracellular concentrat ions).
ice crystals have inadequate time to grow and , therefore, are very small. (Ice crystal dimen sions
10 nm in size" . Such ice crystals, while of small dimensions (5 to 10 nm in
and cooling rates shown here are relative and not indicative of precise measurements fo r any diameter), nevertheless indicate that phase separation has occurred and
one cell type.) theJr presence is not to be confused with the real attainment of a vitreous
12 Robards and Sleytr Low temperature methods in biological electron microscopy Freezing 13

TABL E 2.1 co
I
Cooling rates required for vitrification I

Rate
(K s- 1)
Specimen Reference

Fletcher 1970
t
0.732nm

l
l x 10 10 Micrometre-sized droplets of pure wa ter
l x 106 Pure water Riehle l 986a
2x 104 Pure water at 2.05 x 108 Pa (2045 bar) Riehle l 968a T
5x 10 3 Hydrated specimens Stephenson 1956
1x 104 Hydrated specimens Riehle l 968a
1x 102 Hydrated specimens at 2.05 x 108 Pa Riehle l 968a

These data provide a very approximate indication of the cooling rates required to vitrify water Fig. 2.6. Lattice structure of hexagona l ice (1 1,). Oxygen atoms are shown wh ite; hydrogen
or hydrated specimens. However, it is important to appreciate (as stressed by a number of auth- atoms are shown black. The lattice parameters are temperature-dependent. Those given apply
ors) that the rates cited depend heavily on the size of the specimen being cooled and may vary at 110 K (- 163°C). At thi s temperature, the density of ice is 933 kg m - 3 and the linear coeffi-
by a few orders of magnitude. cient of thermal expansion is approximately 2 x 10 - 5 (Dubochei-et al. 1982). The oxygen atoms
lie in crinkled sheets which are orientated normal to the c-axis and each pos ition is surrounded
tetra hedrally by four other oxygen positions, thus keeping four molecules per unit cell.
(Redrawn from Fletcher l 970.)
state. It might be Jess confusing if ice with crystals of< 10 nm was referred
to as microcrystalline. Whenever phase separation occurs, other physiologi-
cal effects within the cell, such as increase in solute concentration and As the freezing of biological cells and tissues is usually a process of ice
change in pH, must necessarily follow . Ice crystals of about I 0 nm or less crystallisation accompanied by phase separation, it is apparent that this
do not usually interfere greatly with ultrastructural observations and it may physical method of'fixation' can produce a spectrum of artefacts of its own.
well be that such small crystals still allow preservation of viability (Nei Extracellular crystallisation may lead to dehydration and shrinkage, so dis-
1973). Amorphous (vitreous) ice (Iv) has so far only been produced by turbi ng the osmotic balance across cell membranes; proteins may be dena-
vapour deposition at very low temperatures (Burton and Oliver 1935; Lons- tured, and the conformation of polymers may be changed. In particular,
dale 1958; Dowell and Rinfret 1960; McMillan and Los 1965); by using spe- mechanical rupture of cell membranes caused by ice crystallisation may lead
cial conditions for extremely small droplets (Briiggeller and Mayer I 980; to an irreversible loss of compartmentalisation (§2.3.1 ).
Dubochet and McDowall 1981; Mayer and Briiggeller 1982; Dubochet et When pure water is frozen then, as already explained, ice crystals form
al. I 982; Dubochet 1984; see §3.4.4); or by melting ice under pressure below by hom ogeneous nucleation. Adjacent crystals abut against each other with
its glass transition temperature (Mishima et al. 1984). Whether true vitrifi- the crystal edges (or 'grain boundaries' , to give them their correct nomencla-
cation has been, or can be, achieved within normally hydrated cells remains ture) in close contact. However, in biological samples, the likelihood of hav-
an open question. For the moment we must usually content ourselves with ing a single componen_t system is remote: we are usually dealing with com-
the goal of producing extremely small, intracellular ice crystals. If these plex mixtures of solutions and suspensions. In such specimens, the crystalli-
crystals are those of normal , hexagonal ice (Ih), then they are likely to be sation of ice produces a phase separation such that the dissolved salts (and
relatively large (of micrometre-sized dimensions as opposed to crystals of macrom olecules) become concentrated into the liquid phase surrounding
cubic ice with dimensions in the 0.1 µm range; Dubochet et al. 1982; Dubo- the growing crystal. This may be ·exemplified by considering a simple two-
chet 1984) (Fig. 2.6). Even if vitrification of water is achieved within cells, component (binary) system consisting of salt and water. If a solution of, say,
some structural alterations during freezing will still occur because phase common salt in water is cooled, then ice crystals eventually form . The ice
transitions can take place in liquids other than water; for example, in the crystals grow, excluding solutes from the crystal lattice and so concentrating
fluid lipid components of the cytoplasm and membranes. the remaining, unfrozen salt solution until it reaches a critical concentration
Robards and S leytr Low temperature methods in biological electro11111icroscopy Freezing 15
14

TABLE 2.2
Eutectic temperatures of some so luti ons

Substance Eutectic tempera ture Eutectic concentration

coq (K) (%)

K N0 3 - 2.9 270.3 10.9


KC I - l l. l 262.l 19.7
K Br - 13.0 260.2
N H 4C l - 15. 8 257.4 18.6
NaC l - 21.8 251.4 23.6
MgC l 2 - 33.6 239.6 21.6
CaC I, - 55 218.2 29.8
MgS0 4 - 3.9 269.3 19 .0
Glucose -5 .0 268.2 32.0
Sucrose - 13.5 259.7 62.5

Taken from Meryman ( 1966) and Lange ( 1961 ).

- the eutectic concentration - at which point the salt and remaining water
freeze as a single component (Fig. 2. 7). The phase around the ice crystals
is referred to as the eutectic (a mixture of two or more substances with the
lowest melting point); a specific eutectic will have its own constant freezing
or melting temperature - the eutectic temperature (Fig. 2.8) (Rey 1960). If
cooling rates are higher than those which allow equilibrium conditions to
prevail, then supersaturation of the system can occur so that the eutectic
point is moved to the so-called 'hypoeutectic point' (Menold et al. 1976; Fig.
2.9). As, during cooling, the eutectic freezes last so, during warming, it will
melt before the ice crystals. Eutectic temperatures of biologically important
salts are usually relatively high; for example, NaCl, 252 K; KCl, 262 K;
CaCl 2, 21 8 K (see also Table 2.2). Many solutes do not form true eutectics
(Meryman 1966; Luyet and Rasmussen 1968). For example, glycerol sol-
utions show progressive ice formation with lowering temperature until a
concentration of 60-70% has been reached, beyond which point further
cooli ng res ults in the formation of a glass; i.e., it vitrifies (Fig. 2.10). In fact,
Rasm ussen and Luyet (1969) determined the borderline between freezeable
and non-freezeable glycerol to be at a concentration of 73%, the remaining
soluti on then solidifying at 216 K . When a binary system is considered, such
as H20/NaCI, the eutectic temperature may be lowered (from 252 to 193
K) by the addition of glycerol to produce a three-component (ternary) sys-
16 Robards and Sleytr Low temperature methods in biological eleclron microscopy Freezing 17

G
I Under sat uration
I
I
I
I
Pure I solution
.... ....
~ ....
::J -:.,.._-
;
~
a.
E Unatable supersaturation
Q)
.... 100%8
Composition 100%A
0%A 0%8
+crystals of 8 +crystals of A

Fig. 2.9. A_t higher rMes than those tha t all ow equilibrium conditions to prevail (Fig. 2.8),
the sohdus lme (saturation curve) is crossed without solidification of either component A or
Crystals of A B._ Afte r crossing the solidus line and passing the zo ne of metastable supersaturation, nucle-
Crystals of 8+eutectic
+eutectic atio n occurs which leads to crystallisa tion. Thi s means that homogeneous nucleation is a
s retarded process, the magnitude depending on both the composition of the system and on the
100%8 Composition 100%A cool ing rate. Co nsequently, in such a phase diagram of a binary system, there is a line below
0%A 0%8 the solidus line which is less sharp a nd which connects all points of spontan eo us nucleation.
This li ne is termed t_he supersaturation line or curve. When the system is in a state correspond-
Fig. 2.8. A simple pha se diag rnmJor a binary system (e.g. water a nd sodium chloride) fro zen mg to a point within the zone between the satu ra tion and the supersaturation lines, crystals
very slowly. The two substances (e.g. A, sodium chloride; B, water) form a homogeneo us may grow provided _the system c,ontains crystallites th a t are sufficiently la rge. If the original
liquid , are intersoluble in all proportions but have negligible solid-so lid solubility. When cool- binary system co ntains no nuclei , there will be no spontaneous nucl ea ti o n and consequently
ing such a system containing a relativel y small proportion of substance A, the line from G to no _crystal growth. The system is then in some sort of metastable sta te. Beca use the supersa tu-
Fis followed until, at F, the system starts to deposit crystal s ofB. As further coo ling continues, ration curve runs parallel to the saturation line, the eutectic point will be found a t another
more and more crystals of B are deposited so that the remaining solution beco mes more con- concent ration than that corresponding to equilibrium conditions. This change in the eutectic
centrated in A. Thus, the line from F to E is followed until, eventually, E (the eutectic poin t) pomt can be seen here. The modified point (HJ is termed the hypoeuleclic point. (Redrawn from
is reached. At this point, there is a eutectic mixture which , if cooled still furth er, freezes like M enold et al. 1976.)
a pure element or compound. Freezing along the line F- E is termed primary crysta llisa tio n;
that along the line E--S (the eutectic mixture) is called secondary crystallisation. If the propor-
tion of component A is relatively high , then crystals of A (sodium chloride) will be depo sited
first. The A crystals will push a so lution in front of them in which the concentration of B be- (Rasm ussen and Luyet 1970; Fig. 2.10); the relationship between the lowering
comes higher and higher. When the solution consists of a eutectic mi xture from the very begi n- of Tr(Tm) a nd Th is linear (Rasmussen and MacKenzie 1972). Thus, whereas
ning, it solidifies as a whole as soon as the eutectic temperature is reached . The situation de-
the homogeneous nucleation temperature (Th) of pure water is about 235
scribed here only prevails if the system is allowed to pass through states of equilibrium where
the cooling rate is slow enough, and there are sufficient nuclei for crystallisation. If these cond i- K, the addition of glycerol to give a 40% solution lowers Th to 203 K. Ab-
tion s are not satisfied, the solidus line (sa turation curve) will be crossed without solidification sence of nucleating sites in both the solute and solvent phases leads both
of either component A or B (see Fig. 2.9). (Redrawn from Menold et al. 1976.) to super-cooling and to supersaturation of the solution. If the maximum de-
gree of supersaturation corresponding to homogeneous nucleation runs
t~ ro ugh the lowest value of the eutectic point in the glycerol/water phase
tern (Shepard et al. 1976; Boutron and Kaufmann 1979). Clearly the com- dia gram (Lusena 1960; Fig. 2.10), then the most favourable situation for
plexity of such multi-component systems is considerable and little informa- achie_ving the amorphous state would be by freezing a 60% glycerol/water
tion relevant to biological studies is available. Nevertheless, the magnitude solu t10n. However, as we shall see later (§2.3 .2a), biological considerations
of the changes that can be induced by adding further components is clearly do not usually allow the use of such high concentrations of additives to faci-
evident. litate such enhanced super-cooling.
The addition of a solute to water serves to lower the temperature not only When water is placed under pressure, its melting point is depressed (Fig.
of heterogeneous nucleation but of homogeneous nucleation as well 2-11 ). This is an unusual phenomenon shared by water, bismuth and gal-
Robards and Sley tr Low temperature methods in biological electron microscopy Freezing 19
18

3
Pressure (x 10 b ar)
0 270 1 2 3
270
Liquid
-20 250
250

-40 230 ~ -40 ~


!! 230 !!
\
\

'' ;z
.
;?
~
.
;?
~
-60 210 !
::>
Q.

..E -60
210 ~
Q.

f- f-
'la

190
•..
Q.
E -80
I
190
f-

-100
-100 0 2 3
170
8
Pre ssure (x 10 Pal

Fig. 2. I I. Part of th e pressure/ temperature phase d iagram fo r wa ter (see F ig. I. I) showin g
-120 150 melting ( Tm) and nuclea tion (T.)·temperatures of ice as a function of press ure. Roman numerals
indicate the different stable ice polymorphs. The cross-h atched area shows the region within
which super-cooled water can exist.(Redrawn from Franks 1980.)
-140 130

0 20 40 60 80 100
% (w / w) glycerol during the change of water from a liquid to a solid state (Fig. 2.12;
Ushiyama et al. 1979).
Fig. 2.10. ' Phase diagram ' for the glycerol/water system. The temperatures at which different
phenomena occur for different concentrations of glycerol are indi cated . Tg. glass tran siti on It wi ll be apparent from the above statements that untreated biological
temperature; T, , devitrification temperature; r., homogeneous nucleation temperature; Tm, specimens normally freeze somewhere within the range 273- 223 K . How-
melting temperature; R, recrystallisation . (Redrawn from data in Luyet 1970 and Rasmu ssen ever, irrespective of the original cooling rate (and hence ice crystal size), ice
and Luyet 1970. See La ne 1925 for thermophysical data on glycerol solutions.)
crystals continue to grow if the specimen is maintained within this relatively
warm temperature range. The temperature below which pure ice does not
lium. At sufficiently high pressures, different forms of ice are produced but recrystallise is as low as 130 K but this is temperature-dependent and recrys-
these are not of general biological interest (Fig. l. l). However, the fact that tallisation probably does not take place at significant rates until the temper-
the ~ritical cooling rate (vc) required to achieve vitrification of water may ature is above 180-190 K (Moor 1973; Nei 1973). However, few detailed
be reduced by the application of 0.21 GPa (2, 100 bar) pressure, which low- studies have been made, and this is an area where much more work is
ers the melting point to about 251 K, has been used by Riehle (1968a) as required. If it is accepted that the frozen specimen must be retained below
the basis for a high pressure freezing device (§2.4. ld). 180 K if ice crystal growth is to be avoided, then it is usual and convenient
The conclusion to be drawn from these facts is that frozen biological spe- to use liquid nitrogen as the storage cryogen. Having frozen the specimen
cimens normally contain ice crystals of a size related to the composition of and maintained it at low temperature (77 K), the effect of once again bring-
the cell fluids and to the local cooling rate. Various artefactual changes will ing the temperature above the equilibrium freezing point for water can be
have occurred, not the least of which is an increase in volume of about 9% considered.
Freezing 21
Robards and Sleytr Low tempera ture methods in biological electron microscopy
20
Temperature (K)
Temperature (Kl 120 130 140 150 160 170 180 190 200
300 350 8.0
~~~~~~~~~~

250

285 7.0
19.6 275 280 Cubic
270 Hexagonal
18.028 5 .0
] 6.0
18.026 QI

19.4
18.024
~ 5.0
.;,
0 4.0
18.022 -' 4.0
Amorphous Cubic
19.2
18.020 3.0 1=1 . 2 2 X 1 0 3 0e-0. 4 SST

18.018 t =1.55x 1014 e-0 ·126T 3.0


2.0
19.0 -150 -140 -130 -130 -110 -90
18.016 -160 -70
I
0 4 6 8 10 12 Temperature (°C)
-4 -2 0 2
E
M Fig. 2. 13. The time necessary for th e conversion of vitreous (!,) to cub ic (!, ) ice and cubic
E 18.8 ice to hexago nal ice (1 1,) as a function of temperature as dete rmined by X-ray diffraction. (Re-
E
M
0 drawn from Dowell and Rinfret 1960; see.al so Dubochet et al. l 982.)
~

~
QI
E 18.6
~
0
> continue to grow at increasingly high rates. Finally, the crystals undergo a
18.4 phase transition, with concomitant absorption of latent heat, and revert to
the original liquid state. The condition of the thawed cell, however, greatly
depends on the processes that have occurred during both the cooling and
18.2
warming (thawing) sequences. In general, low temperature microscopy has
Super-cooled
H20 not been greatly concerned with cells after they have been thawed (but see
18.0
§2.6) and the following sections only consider the processes and
practicalities of freezing biological specimens prior to microscopical exami-
60 80 100
20 40
- 40 -20 0 nation and analysis.
Temperature (°Cl

· f water and ice (lh) in relation to temperature. The inset shows


Fig. 2. 12. Specific volume o l to o•c (273K). (Redrawn from De Querva in
specific volume in relation to temperature c ose 2.2 Cooling mechanisms
1975.)

As will have been seen from the previous section, the rate of cooling of a
2.1.3 Warming frozen specimens biol ogical specimen has an enormous influence on the nature of the final
frozen specimen. It is therefore important to know what rate of cooling
Were the specimen to be vitrified, warming would produce a devitrificati~_n should be aimed at, over what temperature interval, and at what position
transition as the glass (Iv) transformed into cubic ice (fo~ pure water t ~
1
in the specimen. In a later section (§2.3.1) we will consider cooling rates in
would be at about 135 K; Fig. 2.13). Subsequently, the cubic form (le) woul relati on to both viability and structure but, for the moment, we shall assume
Cha nge to hexagonal ice (Fig. 2.13; Dowell and· Rinfret
d
1960; Dubochet et
· t a kes the spe-
that we usually require rapid cooling rates so that the critical range is tra-
normal crystalline specimens contmue warmmg versed as rapidly as possible.
a 1. 1982) · In ' h. h · rystals
cimen through the recrystallisation temperature, after w 1c ice c
Robards and Sleytr Low temperature .methods in biolof(ical electron microscopy Freezing 23
22

In considering cooling rates (i.e. the extraction of heat from a specimen) the specimen; the size and shape of the specimen; the nature and size of the
we enter the subject of heat transfer (§2.2 .3). This is a field that has been monitoring element (usually a thermocouple); the rate of entry into the
intensively studied but remains inadequately understood, particularly in cryogen; and the position of the monitoring element within the specimen.
relation to heat transfer through and from complex biological specimens. Thus, to say that "sample x was cooled at n K s- 1" is really not in itself
Nevertheless, it is instructive to consider some of the phenomena involved a meaningful statement. Suggestions have been made that some standard
and to establish how well our understanding will allow improvemen ts in the for citation of cooling rates should be adopted (Bald I 978b) but the enor-
practical treatment of real biological specimens. Before discussing further mous variability of biological samples and freezing methods makes the se-
the theoretical background to and practical attainment of rapid cooling lection of a universally acceptable standard virtually impossible. The best
methods, it will be useful to consider the rates of heat extraction that are that can be hoped for is that authors will make the methods for recording
required and how they might be measured. cooling rates as sensitive and reliable as possible and will then cite the fullest
possible details of how these rates were attained.
It will be helpful to take an example of the difficulties encountered in
2.2 .1 What is the actual cooling rate?
attempting to use a single figure to cite a cooling rate over a substantially
The range of cooling rates that is relevant and practicable for biological cells large ( > 100 K) range when cooling in a liquid cryogen. Most biological spe-
1
4
and tissues is from extremely slow ( < 0.1 K s- 1) to ultra-rapid ( > I 0 K s- ). cimens to be frozen are small ( < 1.0 mm 3) and the monitoring element must
1
The literature is full of citations of cooling rates expressed in K s- that, therefore also be so small that it does not itself modify the cooling rate . In
without further information, are almost worthless in either absolute or practice this means that cooling rates are almost always determined using
comparative terms. Cooling rates are modified by, among other factors , the thermocouples (but see t.he electrical capacitance techniques used by Heuser
non-linearity of many cool-down curves (Figs. 2.1 and 2.14); the nature of et al. (1979) and Van Harreveld and Trubatch (1979) in §2.2.2b, as well as
the fluorescence label method of Aurich and Forster (1984)). When two dis-
similar metals are joined together, an electromotive force (voltage) that is
proportional to the temperature is produced across the junction (the See-
beck effect, which is the reverse of the Peltier effect). By measuring the open
I
I circuit voltage it is possible to determine the temperature with considerable
·O Film boiling : accuracy (§2.2.2a). However, the thermocouple itself modifies the apparent
.;.
0
..... :Transition cooling rate by increasing the heat capacity (§2.2.4) if it is large in relati on
1 boiling I to the size of specimen within which it is positioned. Thi s phenomenon is
Nucleate
, boiling more clearly demonstrated if bare thermocouples of different sizes are
Convection
cooled in the same manner (Table 2.3). The greater heat capacity of the
Log. temperature difference (6Tl betwe en surf ace and coolant larger thermocouples reduces the perceived cooling rate. This indicates that,
so far as possible, small thermocouples should be used when dealing with
Fig. 2. 14. Poo l boiling hea t transfer regimes. Th e hea t tra nsfer rate (Q) is. expressed in rela-
biol ogical samples.
tion to th e tem pera ture difference (ti. T) between the surface of an object bemg coo led and t_he
temperature of the coolant. When the surface tempera ture of the sol id object in co_ntact wi th A further difficulty in citing cooling rates arises from a simple thermo-
the liquid is at a relatively hi gh temperature in relation to the temperat ure of the 11qu1d, a sta ble dyn amic consideration of what happens when a warm object is placed into
vapour film forms at the interface. This film in sul a tes the object (spec.im en) a nd reduces the
rate of heat withdrawal (coo lin g ra te); as L'i.T becomes smaller, a t ra ns1t1on region is enco un-
a liquid cryogen such as liquid nitrogen at its boiling point (Fig. 2.14). When
te red in which the film becomes un stable and a maximum tran sition boiling region is reached. there is a large temperature difference (A 7) between the object and the coo-
Subseq uentl y, bu bbles o f gas form a t specific sites on the specimen surface, grow and detach lant, the heat exchange is so great that a stable film of insulating gas forms
themselves: a phenomenon known as nuclea te boiling. (See Westwater 1959 fo r a general.d is-
cussio n of boiling heat tra nsfer and Ma rlo et al. 1968 fo r a di scussion of nucl ea te po ol boilin g
around the specimen and serves to retard the cooling rate: this is known as
of nitrogen with different surface co nditi o ns.) the 'Leidenfrost' phenomenon and arises fromfilm boiling. As the tempera-
24 Robards and Sleytr Low temperature methods in biological electron microscopy Freezing 25

TABLE 2.3 specimens undergo thermal histories (records of temperature against time)
Some cooling rates in relation to thermocouple size similar to that shown in Fig. 2.14 and the only way to be aware of this is
Reference
to follow the temperature accurately and rapidly as the specimen cools.
Coolant Diameter Cooling Cooling rate
range
(µm) (K) (Ks - 1) 2.2.2 Measurement of cooling rates

25 248-198 39,000 Luyet and Kroener 1960


Liquid nitrogen Although the importance of high cooling rates for the preservation of biolo-
70 273-173 16,000 Costello and Corless 1978
Glover and Garvitch 1974 gical specimens has been well established for a long time (Luyet and Gonza-
240 273- 173 1,880
273- 173 437 MacKenzie 1969 les 1951; Stephenson 1956; Luyet 1960) and the absence of appropriate stan-
250
500 273- 173 80 Robards and Crosby 1979 dards for measuring or reporting rates has not gone unnoticed (Meryman
1966; Bald 1978b), it remains a fact that relatively few authors have
25 279- 176 167,000 Luyet and Gonzales 1951 ,
lsopentane attempted to consider the whole topic of biological freezing as opposed to
51 296- 196 111 ,000
looking at one or two limited, specific aspects.
76 296-196 83,000
As explained earlier, the small possible size of thermocouples makes them
These data , culled from various sources in the literature, serve to show the general dependency the best means for monitoring rapid cooling rates and some techniques for
of measured cooling rate on thermoco uple size. making and using thermocouples will now be considered. It is also possible
to make an indirect measurement of the rate of freezing of a hydrated speci-
men by measuring the change in its electrical capacitance during ice forma-
ture of the cooling object moves towards the temperature of the _coolant tion (Heuser et al. 1979; Van Harreveld and Trubatch 1979; §2.2.2b). Aurich
pool, it passes through a phase of transition boiling before movmg mto the and Forster (1984) have used a novel method for determining temperature
phase of rapid nucleate boiling where individual gas bubbles are formed at during the rapid cooling of microlitre-sized droplets. The fluorescent dye,
the surface of the coolant but do not have time to form a stable layer. The umbelliferone, fluoresces at two wavelengths, the intensity ratio of which
temperature of nucleate boiling depends upon the cryogen being ~se~ an? is proportional to temperature. Thus the system allows sensitive and accur-
the temperature difference (LIT) between object and coolant. For hqmd ni- ate monitoring of cooling rates without mechanical interference with the
trogen this temperature difference extends only 11 K above the tem_perature specimen.
of the coolant and is hence of little significance in improving coolmg rates
over the range relevant to biological specimens. (Incid_entally: _the on~et ?f 2.2. 2a Thermocouples and temperature measurement
nucleate boiling is the explanation of the violent, terminal b01hng of liquid
nitrogen as an immersed object finally achieves the coolant te_mperature - Junctions between dissimilar metals produce voltages related to the temper-
a phenomenon commonly observed when cooling anticontammator plates ature of the junction. This is the principle of temperature measurement
in electron microscopes). It should now. be clear that, if the temperatu~e of using thermoc.onples. A range of different junctions can be used in practice
a specimen as it cooled were to follow a changing profile such as depicted (Table 2.4) but copper/copper-nickel (copper/Constantan; Type V) junc-
in Fig. 2.14, then any arbitrary cooling rate between temperature I and tem- tion s have been much the most widely used for temperatures between 77 and
perature 2 would merely give a misleading integration of what happe~s. For 273 K, although Chromel/Alumel (nickel-chromium/nickel-aluminium;
specimehs of significant size and relatively low th~r~al c?n~uctivity, _ th~ Type K) is also often used. Thermocouples can be bought ready-made in
cooling rate will vary greatly in relation to the posit10n withm .the object. a range of metals, wire diameters and sheathing forms (see List of Suppliers
at the surface the rate might be quite high but within the specimen much in Appendix 2) but it is often more convenient (and less expensive) to make
lower rates will prevail. If extremely high cooling rates are achieved , then up thermocouples as required. For the moment we shall confine our atten-
it is possible to obtain curves of much greater linearity but many coohng tion to copper/Constantan (Cu/Con) thermocouples.
26 Robards and Sleytr Low temperature methods in biological electron microscopy
Freezing 27
bl)


u The accuracy and reproducibility of thermocouple measurements
"c::c:: -0
1;l ~
depends primarily upon the purity and homogeneity of the wire and the ab-
0 ;::l
;::l
~ >. '§ sence of any further interfering junctions between the thermocouple junc-
""'~ .':l
.5 -0c::
c0 "0.
·c:u E E tion itself and the terminals on the monitoring instrument. If a series of mea-
2 "'"'
'-
E .':l
"'
:5"' "0
bl) >. -
0 "0. 0
u ~
_g
surements is to be made, it is best to obtain enough wire from the same
i:'
u
-
-.;
;::l
0
u t: "'0 .5 batch . In the U.K., thermocouple wire is manufactured to British Standards
<S
E
.5
Oj
bl) 0
E "'
.,g
.,
E t:
0
"'t:
E .2
specifications (see Table 2.4; BS 4937, output tables; BS 1041 , output toler-
"'t: '5" <=E
bl)
t: E Oj ances; BS 1834, colour codings). A number of methods for making junctions
2
'-
15.~
" ;::l
~ 0
u .>!
-.; E " "
0. v; '- -0 "' a.
"
;8 u
0
~ .5 ~ ""' 0.
....l
"' Argon g a s -

"'t:
2 <: :::
<:;j 2<J V) f- ~ M -.
Oj ;:;; 0 """0. """0. "0. 0- 0
.>! ~ ~ 0. ">.0. -r- r- >. r-
0. ">.
0. -r-
a.0. <:;j
'- -{5 "'"'<.> r- - M
r-
M M - M
r-
M - M
"' "' Vi z"""
°' °' °' Vi """ °' Vi """
"''- "'"'
\.'.)
'-
~ """ z """ """ z z """ z z
..,"
Cl'.) Cl'.) Cl'.) Cl'.)
;::l ill <( Ci ill ill <( Ci ill <( Ci
'-
"0. 0 0
V)
0
E 0 V)
r-
,..._; .':l
""" "';::'- !::'-
"""+ "' +
Plastic tube
U-l ~
_g <; ~ +
....l .... u >:: 3 3 3
ill '- <>. 0.._, 0 0
<( 2
"'

!:: "'
"'
~
V)
N
0
00 00
Wires twisted Thermocouple
f- -.;"' ~ .9 I together
<:; Carbon rod---+--- connector
·c
..," "".... ·-,"'
.:;
-'<!
;::
I
"' ~"'"<>
0 0
E 0 0 0

a." "' "' ";:: M 0 r-


~
"';:: + +
;::l
0
u "'~
.... <:;j
'- .!';
;:
3
V)
+ 3
0 00 3 0
E "<:
c3 0
N

+
"
.c
f-


E
I . ~~ 10000fF 40V DC
.9 ] .5 E
;; ""' u E ;::l
.!'; E
-c c::
!:: "' z~~
"'"'- ·E"
"' "'E
<:;
~ .':l
~ ~ E 12V DC

"'"'
(.) "'
-~
'-
"0.0.
'-
Q) ' "§ E
;::l
"
·z;
"'
Cl.. u
0
u
0.
0.
0 "z
.,.;

~
2
2 .c
.>! .c
u
t:
2
Fig. 2.15. A simple apparatu s for preparing thermocouples from thin wires. The two wires
(e.g. 0. 1 mm diameter copper and 0. 1 mm diameter Constantan are twisted together at one
end and connected to a capacitor. This capacitor is then discharged across a connection made
between th e twisted thermocouple wires and a carbon rod. The discharge tak es place in an
atmosphere of ine rt (e. g. argon) gas to avoid oxidation of the thermocouple wires. A simple
' f- >
arrangement for achieving this is illustrated here.
28 Robards and Sleytr Low temperature methods in biological electron microscopy
Freezing
29

has been described: some involve ' butt-welded' junctions using solder; oth- .. 70

ers do not use any additional metal component. The use of solder is not to ~
.. 60

be recommended if it can be avoided . The simplest form of junction is the ~ 50


'bead-welded' thermocouple that can be produced very simply by discharg- . ! 40
ing a capacitor across the pair of wound wires as they are held in an atmos- ..
£ 30

phere of argon gas. A simple device for doing this is illustrated in Fig. 2.15 . ..g'
.r.
20
10
Such a system can be used for thermocouples down to about 100 µm dia- ()
?f.
meter but smaller junctions pose more of a problem and , in addition, often 0
10 20 30 40 50
cannot easily be obtained from commercial suppliers. Time (fs)
Some methods of producing very small thin-film junctions have been de- Fig. 2.17. Response of thermoco uples. The stipp.led area indicates th . f
e range o computed
scribed by Clark et al. (1976), while Escaig et al. (1977) have also used very temperature changes at the centre of 5.0 µm thick slabs of different materi·als ft b. ·
. .
th em t o s tep- f unc t 10n temperature changes m the surro u d. . Th · a er su 1ectmg
thin film thermocouples, produced by the deposition of about 200 nm of b races t he responses of copper/Consta ntan and C hromel/n Almg water. e stippled area em
1 h -
copper onto 500 nm of nickel and supported by a 12.5 µm -thick polyimide . . ume t ermocouples. (Redrawn
from data provided m Balko and Berger 1968 .)
layer (Bullis 1963). Escaig has subsequently used the same method to pro-
duce very small Chromel/Constantan thermocouples (Fig. 2.16). The res- couple is connected to a direct-reading thermometer, then the absolute
ponse time ('thermal inertia') of such thermocouples is less than 1.0 ms and accuracy should be checked both with the internal and exte rna 1 re1erence
.-
. . .
allows the accurate monitoring of cooling rates of the order of 105 K s- 1• JUnct10ns 1f available. In the former system the thermometer contains an in-
This is important as the size of the thermocouple determines its heat capa- ternal electronic circuit generating a constant voltage as a reference whereas
city and, therefore, the response time (Balko and Berger 1968; Fig. 2.17). in the latter, a second, separate external thermocouple is maintained in ~
Having once produced a thermocouple, the question of how accurate it
is has to be considered. It is necessary to calibrate the junction against some
known standards (such as the melting point of ice, the sublimation poin t
of dry ice (solid C0 2) or the boiling point of liquid nitrogen) and to deter-
mine the precise voltage produced (with a calibrated DC voltmeter) (Fig. Plug and socket
Digital
2.18). The results can then be compared with the relevant voltage/tempera- voltmeter
ture tables published for thermocouples (e.g. Weast 1971). If the thermo-
UV
oscillograph

Oscilloscope

Reference
junction

Fig. 2.16. Preparation of thin film th ermocou ples (a fter Escaig et al. 1977). The thermocoup le
is made by sputtering a layer each ofChromel and Constantan (Table 2.4), each 100 nm thick ,
onto a 6.0 or 12.0 11111 thick polyimide (§3.4. l ) film. The termal conductivity is 3.1 2 x 10 - '
Fig. 2.1 8. A typica l ~ ·
cal s- 1 K - 1 ( 1. 3 mJ s- 1 K - 1) and the thermocouple is suitable for operation between 4 a nd 673 Co11J1ect · h 1 arrangement or measuring tempera tures with a n accuracy of + 1 0 K
K. The junction surface a rea is approximately 2.0 mm 2. Between 293 a nd 173 K , such a thermo- ion s s ou d be made · · 1 1 . - · ·
ence junction sho uld b . using spec1a p ugs and sockets of compatible materials. A refer-
couple has an average sensitivity of 20,µV K - 1• (Diagram redrawn from a n original ki ndly pro- e incorporated as shown. Examples of different monitoring devices are
vided by J. Escaig.) illustrated.
Robards and Sleylr Low 1empera1Ure me/h ods in biological elec/ron m icroscopy Freezing 31
30

constant temperature medium (again, such as liquid nitrogen at it_s boiling The direct measurement of temperature now poses little problem down
point or melting ice) so that a precise reference temperature 1s available. In to liquid nitrogen temperature (77 K) and there are ma ny inexpensive and
matters of calibration everything depends on the accuracy requ1red . Many efficient electronic thermometers commercially available (see Appendix 2)
direct-reading thermocouple thermometers now give good accuracy with lit- to carry out such a function. These are also able to monitor relatively slow
tle attention but it is important that the need for calibration is understood , ·cooling rates and some can be connected to a chart recorder so that a conti-
particularly when thermocouples are connected to 'non-~tandard' i_nstru- nuous thermal history can be obtained. However, once the range ·o f fast/
mentation or are used in rapid temperature change expenments. It is also ul tra-fast cooling rates is entered, an entirely different approach is required.
important that the input impedance of the measuring device is very high. The problem is basically one of finding a system that allows a permanent
Useful publications describing the practicalities of temperature measure- record of temperature fluctuations in the millisecond and even microsecond
ment are Hall (1966) and White (1979). range. This means that recording instruments must have an adequately large
One aspect of thermocouple use in cool-down experiments that is of pa_rti - bandwidth. Chart recorders with mechanically moving pens are generally in-
cular importance is the possibility of cooling rate modification by_ heat with- capable of responding at the rates required. Indeed, the thermocouple itself
drawal along the thermocouple wire (Fig. 2.19). This effect is likely to be may not be able to respond adequately if its thermal capacity is too large.
greater at low cooling rates than at high ones (Costello and Corless 1978). Furthermore, it may be necessary to use some sort of'trigger' for the record-
For this reason it is essential that the thermocouple is insulated as well as ing apparatus so that it starts operation in synchrony with the process to
possible right up to the monitoring junction. be monitored.
A well-tried and -tested system for recording cooling rates is the sequence:
thermocouple; high impt;:dance, high sensitivity ( < 1.0 mV) DC amplifier;
oscilloscope (e.g. Luyet and Gonzales 1951; Rebhun 1972; Glover and Gar-
1.5
[~ Lead wires (copper)
vitch 1974; Menold et al. 1976; Escaig et al. 1977). An alternative arrange-
ment, which avoids the problem of triggering the oscilloscope sweep at the
appropriate moment, is to feed the amplified thermocouple output into an
Lead diameter = d
ultra-violet (UV) recording oscillograph in which very high paper speeds
1.0
can be used (>400 mm s- 1) and to record the trace by a beam of UV light
which is deflected onto UV-sensitive paper by a miniature mirror galvan-
ometer (e.g. Bald and Robards 1978). Further advantages of this system are
that a number of thermocouple outputs can be monitored simultaneously
0.5
(UV recorders can usually accept 12 or more channels) and, by using a dif-
ferentiator in association with the amplifiers, it is possible to feed to the
recorder simultaneous temperature <n and cooling rate (dT/dt) signals in
relation to time (t). However, for very high rates(> 1,000 K s- 1) the res-
50 75 100
25 ponse time of galvanometric mirrors may be limiting.
CI D
With the advent of modern semiconductor devices it is now possible to
fig. 2.1 9. Conduction losses in thermocouple lead wires. Thi s fi gure illustrates graphically record rapid cooling rates even more easily. This can be done by using a
an equation (Bald 1979) which shows that, if thermoco upl e lead losses a re to be kept bdow
5'/;; of the hea t flow through the sample itself (i.e. Qi/Q«0.05), then the length of well-msu-
transient recorder or with a purpose-designed cooling rate meter (Robards
lat~d wire mu st be quite la rge. Although these curves have been derived for a s~hencall~ 1980; Appendix 2). This instrument can be programmed with the tempera-
shaped specimen of diameter D , simi lar curves w<;iuld be expected for b1ol og1cal _samples ~­ ture range over which it is desired to monitor the mean cooling rate . Once
cylindrical shape. Qi. heat flow through the leads; Q,, heat flow through the sampl e, D,_ sphe
cal specimen diameter (m); d, lead wire diameter (m); C, length of in sulated lead wire (m).
the thermocouple. has passed through this temperature interval, the mean
(Redrawn from Bald 1979.) rate is immediately displayed . Moreover, the whole time/temperature his-
32 Robards and S/eytr Low temperature methods in biological electron microscopy Freezing 33

tory during the cooling process is stored in a Random Access Memory so connector are maintained at a constant temperature so that no fluct uating,
that it can be displayed or used for subsequent analysis. Such a device temperature-induced variations arise. This can be particularly impo rtant if
enables cooling experiments to be carried out both rapidly and accurately junctions are positioned close to liquid nitrogen containers where cool gas
and is capable of monitoring rates up to about 10 5 K s- 1• It is not necessary will drift across the wires and keep the spurious voltages in a constant state
to describe in detail the various amplifier/recording arrangements: commer- of fluctuation . All junctions other than the recording junction should be well
cial systems are available (see Appendix 2) or suitable combinations of protected from such changes in temperature. It can be seen that, as with stan-
amplifiers and recorders can be put together in the laboratory. However, dardisation of thermocouples, ideal theory and actual practice can be some
it is necessary to make some comments on the possible pitfalls that may be distance apart. Once again, the degree of accuracy required must be consi-
encountered while attempting to make accurate determinations of cooling dered a nd, more impo rtantly in this context, the reproducibility of measure-
rates with thermocouples. ments should be taken into account. If the ideal situation cannot be
Assuming that suitably standardised thermocouples are available, these attained, then the system should be constructed with careful attention to the
should, ideally, be connected to the amplifiers without any further junc- elimination of random or contin uous fluctuations so that, once assembled,
tions. (Additional junctions between dissimilar metals also produce voltag- it can be standardised in rela tion to known temperatures. Quite frequently
es, thus giving rise to inaccuracies.) Often, however, it is necessary to make it is necessa ry to ta ke thermocouple leads over considerable distances to the
joints (as, for example, when taking a thermocouple lead out of a vacuum monitori ng instruments: in such circumstances the appropriate compensat-
apparatus) and then particular care must be taken not to introduce errors ing wire sho uld be used .
in monitoring. It is possible to make vacuum-tight continuous thermocou- In a later section (§2.4) the practical methods of freezing will be consi-
ple lead-throughs (Fig. 2.20) but these often cause inconvenience in assem- dered but it is necessa ry here to consider how the method of freezing can
bly and thermocouple replacement. Special thermocouple connectors are affect the accuracy of the determina tion of cooling rates, particularly when
commercially available so that the copper wire has a copper connecting pin/ using liquid cryogens (i.e. when a specimen is immersed or 'quenched' in
socket and the Constantan wire has Constantan connectors (see Appendix a coolant) . It is well established that the nature and rate of movement of
2). This minimises the introduction of 'extraneous' voltages. If thermocou- a specimen into a cryogenic liquid will modify the Reynolds number (NRe)
ples must be soldered to connectors, then it is best to use low melting point and hence the turbulence and thus may significantly affect its cooling pro-
solder and high quality connectors (gold-plated connectors are recom- file. For example, Costello and Corless (1978) noted a 21 % increase in cool-
mended). Even here, an additional electromotive force will be created but ing rate (from 36,800 to 44,400 K s- 1 over the range 273 to 173 K) when
the effect can be minimised by ensuring that all leads passing through the a bare 50 µ m thermocouple was plunged into propane at 0.74 m s- 1 as com-
pared with 0. 39 m s- 1• Robards and Crosby (1983) found a similar linear
Flange relationship between cooling rate and entry velocity over the range 1.0 to

·~··~
6.0 m s- 1•
The depth to which a specimen is plunged into a cryogen can also have
·····~-- '.: ... .~
... an important effect on cooling velocity since it influences the amount of
exposure to ' new' cold liquid . Costello and Corless (1978) cite an improve-
Thermocouple lead1 • Thermocouple men t in cooling rate from 4,400 to 4,900 K s- 1 when a copper sandwich was


' """""""""'''"''"::::.·~
.
junction
plunged at a velocity of 0.49 m s- 1 into liquid propane to depths of 5 or
·.-:::>
15 mm, respectively. These authors have suggested that the optimal depth
of plu nge can be simply calculated as the product of the time that it takes
Fig. 2.20. Simple method of making a jointless _lead-through for. thermocouple wires. ~ihi~
for the specimen to cool to a prescribed temperature (e.g. 173 K) and the
method is suitable for using thermocouples to monitor temperature m apparatus at low or g entry velocity. If the entry velocity were 0.49 m s- 1 and the cooling rate
vacuum. 4,900 K s- 1 then the time to cool from 293 K (ambient) to 173 K would
Robards and Sleytr Low temperature methods in biological electron microscopy Freezing 35
34

be 24.5 ms and consequently the depth of plunge needed would be 12 mm. method may be satisfactory for non-quantitative work. A number of differ-
To this might be added a 25% increase in depth to allow for deceleration ent reproducible systems based on plunging the specimens into liquid cool-
and other effects not directly taken into account. It should be noted that ants or apposition onto cold blocks (see §2.4.2) have been used and are de-
this calculation is only likely to be of practical use if high cooling velociti~s scribed in detail in §2.4. There is some indication that rates in excess of 5.0
are applicable and relatively slow entry velocities are used . The combination m s- 1 may be needed to take the maximum advantage of plunging into
of a slow cooling rate with a high entry velocity will produce an impractica- liquid cryogens (Fig. 2.21), in which case compressed air-actuated devices
ble depth of plunge. In fact, the relevant cooling rate is probably the rate may be more convenient.
Using arrangements similar to those described above, precise information
at the specimen surface.
If reproducible results are to be obtained, then it is essential that a stan- rel ating to cooling rates has been obtained, making it possible to select the
dardised method of specimen insertion is used. The shape of the specimen preferred method of freezing for different specimen/freezing combinations.
as well as the angle, velocity and depth of plunge can all have a pronounced
influence on the cooling rate and hence render determinations from speci-
2.2.2b Electrical capacitance method for determining freezing rates
mens cooled in hand-held forceps totally meaningless, although such a

15
Although thermocouple measurements of temperature and cooling rates are
straightforward, for very high rates and where it may not be desirable or
14

I
I
J' possible to have a thermocouple within the specimen, some alternative
means of monitoring the· rate of freezing must be explored. Heuser et al.
13 I (1 979) and Van Harreveld and Trubatch (1979) both described methods for
I
I measuring the change in electrical capacitance of a hydrated specimen as
;:- 12 I
'., I
liquid water changes to ice (Fig. 2.22). The hydrated specimen is treated as
lo: I a dielectric between two conducting plates of a capacitor: one of these is
11 I
"'0
~ I an electrically conducting specimen support, the other is a metal block
~
10 J against which the specimen is to be thrust. Both Heuser et al. and Van Har-
~" I
I
reveld and Trubatch describe the method for use in conjunction with cold
·=0"' block freezing, but it is possible that it co~ld be adapted for other cooling
I
9
0
() I
I methods, although this has not been reported.
8 I The capacitance of the specimen can be calculated from:
I
I
7 I
I C= eA (2. I)
I 4nd
6 I
where
4 5 6 7 8 C = capacitance
0 2 3
Entry velocity (m s- 1 ) A = area (cm 2)
e = dielectric constant
Fig. 2.2 l . Coolin g rates of model specimens plunged into liquid propane at various vel ocities.
The copper/Constantan thermocouple (60 µm diameter with 25 µm leads) was pos1t10ned
d = thickness
between a pair of 50 pm thick copper plates 2.0 mm x 5.0 mm. The means (±standard error)
of 25 determination s at each entry velocity are indicated. The thin, dotted line shows the rela- Before freezing, the liquid water of the specimen has a very high dielectric
tionship between entry velocit y and cooling rate obtained by Costello ( 1980) for a smaller spe-
cimen assembly than the one used here. (Redrawn from Robards and Crosby 1983. )
constant (approximately 80) but this cannot be measured directly because
Freezing 37
Robards and Sleytr Low temperature methods in biological electron microscopy
36
capacitor (s~ecimen) during freezing is monitored using an oscilloscope. The
the conductivity of the dilute electrolyte solution of the specimen provides
moment of impact (first formation of an ice layer) is determined from the
for the flow of a large resistive current between the two plates. However,
abrupt cessation in the flow of resistive current (Fig. 2.22). From these data
as soon as freezing commences, this flow is terminated and any subsequent
it is possible to construct a detailed record of the pas.sage of the ice front
current flowing must be capacitative and will decline as the ice thickness
in to a freezing specimen. Heuser et al. calibrated their results by using vary-
grows because the dielectric constant of ice is only about 2. Thus, as th.e
mg thicknesses of a plastic film having approximately the same dielectric
ratio of these two layers (water and ice) changes, so will the total capaci-
tance. Taking an example from Heuser et al. ( 1979): if the tot~! sa~ple
constant as ice, while Van Harreveld and Trubatch used a specimen support
that ensured that the total thickness was quite accurately controlled (50- 60
thickness is 0.05 cm and its area is 1.0 cm2 then, when the water is entffely
µm ) and obtained a direct record of the progress of freezing from their re-
in the liquid form, its capacitance is:
sults.
80 x 1.02 127 (in CGS units)
4x3.14x0.05
As 1.0 Farad (F)=9 x 10 11 CGS units, this becomes 141 pF.
However, when a 10 µm layer of ice has formed, the capacitance will be:

I I I
- = - -+ --=79 pF
Ctotal 177 ice J441iquid
Thermocouple
and when a 20 µm layer of ice has formed it is: 4 ma 20 ~ oc
0
2 ma
1 ma -----""""'"'"'·'''"11""•;;;;
I -20
I 1 I
- =- + - - =55 pF.
Ctotal 88;ce J471iquid
Capacitance

The conclusion is that the change in capacitance ( C) will be dominated by


the change in thickness of the growing ice layer. . . .
Since the thickness of the layer of frozen material ts a function of the reci-
procal of the current passing across the ~pecimen and:
C=-t- (2.2) L11~ -.:~(·
1.5 v -
0.5
1.0
J
Jv
Vjw battery 'W 1.5

Fig. 2.22. Schematic diagram of th e circuits used to meas ure the rate of freezing on a cold
where block as described by H euser et al. ( 1979). The biological tissue (muscle) is shown as circles
C =capacitance between the wedge-shaped copper freezi ng block (below) and the flat a luminium specimen
=current mount (above). The simples t circuit is a thermocouple placed in contact with the surface of
the muscle but it is difficult , if not impossible, to use a thermocouple to record what happens
j =J-1 m the first 10-20 µm thickness of tissue where freezing is best. A lternatively a simple battery
w =frequency circuit is used to pass current th rough the muscle after contact with the copp~r block and thus
V = oscillating potential to measure its resistance. As soo n as a thin layer of ice is formed on the muscle surface the
resistance becomes so high that the circuit is essentia ll y broken. A better indication of fre~zing

it is possible to obtain a direct indication of the course of fre~zing by apply-


1 giv.en by follow ing the changes in the capacity of the specimen by applying a 100 kHz oscillat-

mg signal to the muscle mount and recording transmission through the muscle with a virtua l-
ing a high -frequeµcy oscillating potential ( V) and measunng the vo1ta7~ e.arth connected to the copper block. Once the surface has frozen and the relatively huge resis-
tive curren t is ove r, th
. e d ec1.me m
· capacitance
· is a measure of the subseq uent growth of ice
transmitted. Heuser et al. used a frequency of 100 kHz while Van Harreve
mto the muscle. (Redrawn from Heuser et al. 1979.)
and Trubatch used 20 kHz. The change in voltage passing through the

Freezing 39
Robards and Sleytr Low temperature methods in biological electron microscopy
38
peratures.
· Forced convection arises from mo t.!On t h at is
. pnmanly
. . due to a
2.2.2c Evaluation of cryofixation in frozen samples supebnmposed flow. The flow characteristics depend on the Reynold ,
num er (NRe): s
While the experimental determination of act ual cooling rates, or rates of
fusion, can be very helpful in optimising cooli ng procedures, ultimately it
(2.4)
is the sample itself that will show whether or not the cooling method was
satisfactory . For most frozen, uncryoprotected specimens, the ice crystal
where
size is itself a sufficient indicator. This can be observed visually or by low
NRe = Reynolds' number
temperature X -ray diffraction (Gulik Krzwicki and Costell o 1978). Costello =relative velocity of specimen and fluid
v
et al. (1982) have described a method of estimating the cooling rate by = leng~h (chara~teristic dimension of specimen)
D
observing the morphology of freeze-fracture faces of a 30% dilauryl lecithin / =density of flmd
p
water system. At high cooling rates the lamellar phase of the lipid is pre- =viscosity of fluid
µ
served, leading to smooth fracture faces; at lower rates, an easily identifiable
worm-like or rippled texture is seen. Costello et al. found that cold block Take, for example, a 100 µm diameter (D = 100 ) th
freezing, rapid plunging into propane and propane jet-cooling all gave rates, in at 2 0 m s- 1( ) . . . µm ermocouple plung-
1 g . . . v mto hqmd propane having a density (p) of 0 5 k -J
measured using a thermocouple, of more than 10,000 K s- and all were andv1scos1 ty(µ)of5xJ0 - 3Ns - 1m -2Th N 20 . . · g m
b 1 h · en Re= , which IS very much
able to maintain the lipid in its lamellar phase. e ow tf e value (approximately 2,300) above which turt..ulent fl ow exists
.
Th · lT

th us, or
h 1. ·dmost practical purposes ' smooth-sided .
specimens plungmg . .
2.2.3 Heat transfer in cryogenfcfluids roug iqm cryogens experience laminar flow .
. While the thermodynamics of natural convection have been relative! well
As previously mentioned, consideration of how a specimen cools (i.e. how
heat is removed from it), involves the subject of heat transfer: this is usually
mvest1gated, it is forced convection heat transfer that is applicable w~en a
speci men is thrust mto a coolant (or when a coolant is projected a ainst a
expressed in units of W m - 2 s- 1. As a first approximation, the heat flux (Q)
spe~1men)f A theoretical appraisal of forced convection in relation ~o rapid
can be determ ined from Newton's rate equation : coo mg 0 b10log1cal specimens has hardly been made at all ( s·1
(2 .3)
et al. 1982) W h fl · . . . . see 1 vester
. en a md m lammar flow impmges on the surface of a speci
men , a tem perature gradient begins to become established as do 1 . -
grad· t h fl · , es ave oc1ty
where ferre~ ~ as t e md moves downstream . The rate at which heat is trans-
Q =heat fl ux point a;~: ~~:/o) the surface var_ies with the distance from the starting
h =surface heat transfer coefficient
heat tra n s~ b ace smce t~e Founer rate equation states that the rate of
A =area er y conduction JS proportional to the temperature gradient:
Tw =temperature of hot body
Tc =temperature of cold body Q
A
=k(dT)dx
(2.5)

The complexities of fluid flow, thermal conductivity and other paramet- where
ers are combined together in the single factor h, the heat transfer coefficien t.
Q = heat flux
The value of h is sign ificantly changed depending upon whether hea t A = area
transfer is by natural convection or by forced convection, and both of these k -- t hermal conductivity
can be subdivided according to the flow characteristics (whether laminar, d T/dx = temperature gradient.
transition or turbulent). In natural convection, motion is primarily due to
gravitational forces acting on differen t densities of flu ids at different tern-
Freezing 41
Robards and Sleytr Low temperature methods in biological electron microscopy
40
ing the incidence of film boiling and experiments have been carried out
Laminar forced convection systems are among the _most difficult to_ an~~ where the surface has been suitably modified to encourage this (e.g. Reb-
artl because the boundary layers may chang~ m_ form ~rom pomt hun, 1972; see §2.2.3b). This is ofless significance in cryogens such as pro-
lpy~:~;andyalso because any disturbance in the impmgmg flmd lmday cad~ste pane where the tendency towards film boiling is much less marked.
d. d than theory wou pre ic · So far, the discussion has centred on the surface heat transfer characteris-
eddy formation closer to the lea mg e ge . d t oflimited qualita-
Further theoretical treatments can thus only provide ~ a 1 h (Schenk tics. However, the quantity that determines the overall ability of an object
However, for a circular cross-sect10na s ape to transfer heat is the Biot modulus (Bi). The Biot modulus is a dimension-
tive significance.
(2.6) less number determined from the equation:
1959):
N -
Nu-
0 .
023 N Re
o.8 Np r
o.3
Bi=hR (2.8)
k
where
where
NNu =Nusselt's number (=hD/k)
=instantaneous surface heat transfer coefficient
NRe =Reynolds' number (Eq. 2.4)
=Prandtl's number (=µcp/k) R =characteristic dimension of the specimen (outer radius of a sphere
NPr
=surface heat transfer coefficient . . or cylinder; semi-thickness of a slab)
h
=length (characteristic dimension ?f specimen: diameter if the k =thermal conductivity.
D
specimen is of circular cross-sect10nal shape)
=thermal conductivity If Bi<O.I, the specimen is classified as microscopic whereas ifit is >0.1
k
=viscosity it is considered macroscopic (Bald 1978b). Microscopic samples, having a
µ
=specific heat capacity at constant pressure low characteristic dimension (R) are generally cooled rapidly and hence
have relatively high surface heat transfer coefficient (h); Bi can be > 0.1
hD (v D p)o.8 x ( µcp)o.3
NNu=l(=0.023 -µ- k and hence different parts of the specimen will cool at different rates. This
is an important conclusion, because it implies that any specimen of large
(2.7)
02 03 x k0.7 Bi will cool at different rates at different points within the specimen; hence
h=0.023(vp)o8xµ-o 5 x D - . xcP·
even the citation of a precise cooling rate from a single thermocouple will
with terms as in Eqs. 2.4 and 2.6. ~ ffi only define the thermal history of that particular point and no other. This
From this it is evident that, to obtain a high surface heat trans er coe .;- problem has been well reviewed by Bald (1975) who has suggested a method
. t v p c and k must all also be high. The last three terms are ea~i y for defining the thermal history 0f any point within a specimen of cylindri-
cien , , , P . ·deration of conduct10n
understood as they all derive from the prev10~s consi .ll . ve the heat cal, spherical or flat geometry. Bald (1979) and Bald and Crowley (1979,
and it also seems obvious that increased flmd flow wi ~mpro b th µ and 1981) have carried this type of theoretical analysis still further and demon-
transfer coefficient. However, it is also seen :hat~ depe~ s on as~re of the that the cooling rate profile into a homogeneous specimen is not lin-
. h i"deally would both be small. The viscosity,µ, is a me 1 and, moreover, may be higher at the centre than at points intermediate
D wh ic , , h t1 f the coo er
'fluid inertia' and a high fluid inertia would ~ean : a ess othe tendency tween the specimen surface and the centre. This theoretical prediction
th cimen High µ mcreases · 'ves some support from the work of Venrooij et al. (197 5) and Elder et
liquid would be able to reach e spe . fl d this
of a fluid to laminar flow. Thus, h increases. fo~ turbul~n~n ~:l;;s (e.g: 1982). It is clearly important that h, as a function oft, can be precisely
could be an important consideration when designmg spec m nt bounda 'oduced and this is, in effect, the rationale for using devices that allow
by including ridges or a ro~gh surface to ~cour~ge~ t~:1~:1~980) to ha bducible cooling rates to be achieved for this sort of work. A further
layer). However, it is also important (ass own. y os or restrictin .etical analysis of ice crystal size distribution in frozen specimens has
an overall streamlined shape to discourage trappmg of v~pour f 1 for redu 1:nade by Kopstad and Elgsaeter (1982), while Jones (1984) has made
the free flow ofliquid around the specimen. Turbulence is use u
Freezing 43
Robards and Sleytr Low temperature methods in biological electron microscopy
42
point of view, it offers no significant advantage over the use of sub-cooled
a detailed estimation of the time taken to freeze at different depths within
nitrogen at atmospheric pressure and it has the disadvantage of requiring
a tissue block in relation to the method of cooling. . pressurisation of the specimen (although only to relatively low, non-damag-
Bald and Robards (1978) investigated some characteristics of specimens
ing, levels). Nevertheless, it is of interest in comparison with other methods
of high and low Bi when subjected to a range of pool boiling heat transfer
and its use derived from a suggestion by R. Thaine that freezing cells of high
experiments. Some results from these and later experiments are instructive.
As model systems, 6.0 mm-diameter, high Bi cylinders (e.~. _polytetra-
turgor pressure (e.g. plant phloem elements) under pressure might alleviate
some of the difficulties associated with preserving such specimens: a hypoth-
fluoro-ethylene, PTFE ~ 2. 5) and low Bi (e.g. high conductivity copper
esis as yet untested. It should be noted that many authors refer to the use
~ 13 x 10- 4) were plunged into: of liquid nitrogen 'slush' as a coolant: i.e. a mixture of solid nitrogen and
the sub-cooled liquid. It would be impossible to achieve reproducible cool-
liquid nitrogen at its equilibrium boiling point (77 K at atmospheric
(i) rates in such a cryogen because of the uncertain locations of solid and
pressure_ 191.325 kPa), giving a typical surface heat transfer coeffi-
fluid phases within the mixture; even sub-cooled liquids have substantial
cient (h) of 180 W m-2 K -1; . . temperature gradients throughout the liquid depth.
nitrogen sub-cooled by heat conduction into a surroundmg mtrogen
(ii) The measured cooling rate can be related to the thermocouple size (Table
container pumped to the triple point of liquid nitrogen (about 13.5
2.3 in §2.2.1). If rates are to be monitored within specimens then very small
kPa) to 63 K; . thermocouples (in fact, of low heat capacity) in relation to the size of the
pressurised (hyperbaric) nitrogen, where both specimen and cryogen
(iii) 6 specimen must be used (see §2.2.4). The difference in cooling rates between
were simultaneously subjected to pressures of up to l x l 0 Pa; and
that of a high thermal conductivity specimen (low Bi) plunged into liquid
Arcton 12 (CC1 F 2 ; Freon 12) at its freezing point (approximately 118
(iv) 2 nitrogen at its equilibrium boiling point at atmospheric pressure and the
K). same specimen plunged into hyperbaric nitrogen at 650 kPa is illustrated
in Fig. 2.23. A number of important features are evident:
As has already been explained (§2.2.1), liquid nitrogen at its equil_ibrium
boiling point is not a good coolant. However, by pumping liquid mtrogen
(i) the total cool-down time is very much less in the 'sub-cooled' cryogen;
down to its triple point it can be sub-cooled by about 14 K to 63 K. If _a
warm object is then inserted into the cryogen while this is _at atmosp~enc
(ii) the cool-down curve is non-linear; and
(iii) the peak cooling rate is a function of either the 'entry phenomenon'
pressure in a well-insulated container (so that it does not gam heat rapidly)
(mechanical action of the specimen moving into the cryogen) or the
the heat removed from the specimen raises the temperature of the pool b~ck
towards the equilibrium point rather than boiling the liquid pool ~wh'.ch
transition to nucleate boiling (which is at far too low a temperature

produces the disadvantages of film boiling). T~e s~me thermodynamic tnck


to have any biological relevance).

can be played in the opposite direction by subJectmg_ both cryogen ~nd sp~­
The cooling rates in liquid nitrogen sub-cooled by evacuation ('nitrogen
cimen to pressure (isentropic compression) in a well-msulated contamer pn-
or to plunging the specimen int? the ?~es~urised. ~oblan~. If: for ~xampl;~
slush') are very similar. Some representative cooling rates for low and high
Bi specimens are listed in Table 2.5. Perhaps the most significant feature
liquid nitrogen at its atmosphenc eqmhbnum bo1lmg pomt 1s subJe_cted
a pressure of 1 x lQ6Pa (10 bar) then a new equil~brium po_int appr~x1mately
here is the small difference in rate of cooling of the PTFE specimen whether
•. •vvu.•ul'. liquid nitrogen or sub-cooled cryogens are used. This brings us to
30 K warmer will result. This cannot be immediately achieved by mflow of
of the main problems concerning cooling of biological specimens: if the
heat from outside the pool if the pool is well insulated and, consequently,
al resistance of the specimen is large compared with the solid/fluid
if the specimen is inserted into the effectively sub-cooled liquid, the he~t.
·1·b · botl- al resistance as heat is conducted away from the surface of the speci-
removed will again raise the pool temperature towards the eqm 1 num •
ing point. This method of hyperbaric freezing has not been used to a~y great
then any increase in surface heat transfer will not be manifested in the
imen centre (or detected by a thermocouple in that position; see Bald
extent in the freezing of biological specimens because, from a coolmg rate
Freezing 45
Robards and Sleytr Low temperature methods in biological electron microscopy
44
TABLE 2.5
Some representative cooling
. rates (Ks-I) from m o d e1 specimens
.
of high and low Biot number

a) Coolant
Sample

Liquid Sub-cooled liquid Hyperbaric liquid Halocarbon 12


nitrogen nitrogen nitrogen
! I

0 1 2 3 4 5 6 7 8 9 101112131415161718
Bare 0.5 mm 80 317 319 714
Time (s)
thermocouple

b) 2.4 mm diameter 34 79 116 63


phosphor-bronze
bead

2.4 mm diameter 43 79 46 37
PTFE bead
0123456789
This
h Table lists
1 some. . 273-223 K cooling rates for a bare o.5 mm d"iameter copper-Constantan
Time (s)
.
t ermocoup e; a similar thermocouple embedded in a 2 ·4 mm d"tamet er h"igh conductivity
· · speci
men
J phosphor-bronze;
( d · · · B10t number approximately 0.002)·' and another· m a . mm d"iameter
2 4
Fig. 2.23. A comparison between a copper specimen cooled in liquid nitrogen at 100 kPa (1.0
atm) (a) and the same specimen cooled in sub-cooled liquid nitrogen at a hyperbaric pressure ow con uct1V1ty specimen (PTFE; Biot number approximately 2.5). All coolin
determmed
h as described in Bald and Robards (1978) · It can b e seen th at the coolmg
g rates
ratewere
}
alt~ough ~e~~
of 650 kPa (b). In each case the temperature vs. time curve is shown (T) as well as the cooling
rate vs. time (B). (Redrawn from Bald and Robards 1978.) t ese large samples are very low the higher conductivity phosphor-bronze
cooled more rapidly m the more efficient cryogens.

and Robards 1978). The only ways of achieving an improvement in cooling


rate are either to improve the thermal conductivity of the specimen (not mous amount of time has been spent in the practical pursuit of the theoreti-
practically possible for most biological specimens) or to reduce the total cally unobtainable!
thermal resistance (heat capacity) of the specimen by making at least one Assumin~ that. we do have a specimen; of significant size (and, unfortuna-
of its dimensions smaller (the main method available for biologists). The sig- t~ly, most b10log1cal tissue specimens will need to be at least 0.1 mm in each
nificance of the results with the PTFE specimens becomes the more appar- dimension), t~en it is possible to calculate the relation between the cooling
ent when it is appreciated that, size for size and under the same conditions rate and the time .(thermal history) for any point within the specimen if a
of cooling, the Biot modulus of a 6.0 mm diameter cylinder of PTFE (2.5) few basic assumpt10ns are.made (Arpaci 1966; Bald 1975).
is very close to that of ice. Ice is an extremely poor conductor of heat (Bald
k(;r~fT0) ;h (R- ~) ~
2
1975; Fig. 2.24) and, therefore, if a specimen of any appreciable bulk is con-
sidered (e.g. > 1.0 µl), the cooling rate (except for the surface layer of about
[ - ~ 1
{ /2 +In ( ~)}J (2.9)

10.0 µm) will primarily be determined by the thermal conductivity of ice and
not by the surface heat transfer characteristic ofthe specimen. In spite of =time
a literature full of cooling rate citations, this point has not been fully appre- =density of specimen
ciated. In so far as the production of very small ( < 10 nm) ice crystals an =latent heat of fusion
still more a vitrified state, throughout untreated, hydrated biological sped =thermal conductivity
mens of significant size (all dimensions > 0.1 mm) is concerned, an enor
-
I
i

Freezing 47
Robards and Sleytr Low temperature methods in biological electron microscopy
46

Four conditions of h (wm-2 K-11

h=840

20
h=4000

~ '.§
'en .... h=8400

'e
3
10 ..
N

.;
,., .2 0.01
h=a:
:!:::
.::
-;:; ..
c

-g"' ""!!
0
5.0
..
E
-;;;" j::
E
:;; 0.001
.c
....
2.0

1.oL-~_L~j__~_L~_L__j__L_L-'----1...L~~-'--__1~___, 0 •0 00 1 o:--C.-'--"---'::o-':.o::-:5L..L_,_Lo.L.1_,_L.L-'0-.L15.L..J.--'--_J_o_j._2L..l-'---Lo.J12 5
10 20 50 100 200
Temperature (Kl Depth from surface (mm)

The thermal conductivity of polycrystalline ice. (Data simplified and diagram re- Fig.. 2.25.. The time taken to freeze (i.e. to pass 0°C) at different depths within a biological
Fig. 2.24. specimen rn relat10n to four different values for surface heat transfer coefficient (h) calculated
drawn from Fletcher 1970.)
usrng Eq'. (2.9). The lowest value of h approximates to that obtained by cooling in liquid nitro-
gen; the intermediate ones represent cooling in halocarbons or propane. Even at an infinite
rate of heat withdrawal, points deeper than about 0.1 mm from the specimen surface take a
significant time to freeze.
Tr =freezing temperature
T, =bulk temperature of cooling fluid
h =surface heat transfer coefficient even if only the outer few micrometres of material are frozen well and for
R =radius of specimen very small specimens (droplets or thin films) the correct selection and use
=distance of cell from central axis of specimen ('inner' radius).
r of cooling procedures is critical. Some further attention should therefore be
to question of maximising surface heat transfer. Let us first con-
This formula has been used effectively by Johnson ( 1978) to calculate val-
the situation for liquid cryogens.
ues oft for various values of h, R, r, and T and we have used it to construct
a series of curves showing the time taken to freeze, at different depths within
Characteristics of liquid cryogens
a specimen, using different surface heat transfer coefficients (Fig. 2.25).
The fact that the low thermal conductivity of ice (Fig. 2.24) so greatly
·1e sub-cooled nitrogen is a convenient cryogen (§2.4. la), other liquid
modifies the cooling rate within any biological specimen of significant size
nts have also been used. In the main these have been selected on the
does not mean that we should totally neglect further efforts to optimise the
that have a low melting point and high boiling point so that the
heat transfer characteristic at the specimen surface. It is still often useful
Freezing 49
48 Robards and Sleytr Low temperature methods in biological electron microscopy

TABLE 2.6
phenomenon of film boiling is minimised. These cryogens will be considered Some characteristics of liquid cryogens
in detail later (§2.4.1) but suffice it to say for the moment that dichlorodif-
Melting Boiling Specific heat Thermal conductivity
luoromethane (CCbF 2 - halocarbon 12, Arcton 12, Freon 12, Genetron 12, Liquid
point point
etc.) is typical of this class of compound: it has a melting point of approxi- (J g-1 K-1) (mJ m- 1s-1 K-1)
°C (K) °C (K)
mately 118 K and a boiling point of approximately 243 K. (The generic
name 'halocarbon' is used here for the halogenated methane compounds, Ethanol -117.3 78.5 1.89 206
such as dichlorodifluoromethane, which are sold under various trade C2 H 50H (156) (352)

names.) I so butane -159.2 - 11.7 1.68 180


The liquids suitable for rapid cooling must have a number of essential
CH 3CH(CH 3h (114) (261)
characteristics: they must be liquid at a suitably low temperature (at least
140 K) and must thus have a low melting point at atmospheric pressure; Isopentane -159.9 27.8 1.72 182
they should ideally have good heat conduction, heat capacity, fluidity, and (CH 3),C 3H 6 (I 13) (301)
high density; and their boiling point should be well removed from their
-189.6 -42.1 1.92 219
melting point. It is also preferable that they should be safe to use, relatively (231)
(84)
inexpensive and should not contaminate specimens to be used for microana-
lysis with extrinsic elements. A limited range of potential cryogens is avail- But-1-ene - 185.4 -6.3 1.94 222
able at present but some cryogenic liquids are as yet untested. Liquid helium C1H6= CH 2 (88) (267)
II, with its very low temperature ( <2.19 K), very high heat conduction
n-Pentane -129.7 36.l 1.89 177
capacity (8.03 mW m- 1 K- 1 at 4.0 K) and superfluidity has been suggested (109)
CH 3(CH 2)JCH 3 (143)
as a potentially excellent cryogen and tested by Fernandez-Moran (1960).
However, experiments on heat transfer in superfluid helium (Rinderer and Ethane - 183.5 -88.8 2.27 240
Haenseler 1959), as well as actual cooling rate determinations (Bullivant CCH 1CH 3 (90) (184)
1965) show that a stable film of gaseous helium insulates the specimen from
Halocarbon 12 -158.0 -29.8 0.85 138
the coolant and thus retards the cooling rate. This, together with the diffi- (115) (243)
CC1 2 F 2
culty and expense of obtaining liquid helium II, as well as the fact that the
liquid helium (and the specimen) must be subjected to a vacuum of 70 µm Halocarbon.22 -160.0 -40.8 1.08 152
Hg (9.33 Pa), has meant that little further use has been made of liquid heli- CHC1F 2 (I 13) (232)
um II as a direct liquid coolant (but it has a use in cold block devices; 153
-210 -195.8 2.0
§2.4.2). (63) (77)
Rebhun (1972) has given a useful assessment of some different cryogens
using a simple 'ballasted' thermocouple (a thermocouple inside a piece of oxygen -218.9 - 183.0 1.7 147
thin hypodermic tubing) as a test probe. He stresses the very great difficulti- (54) (90)

es involved in using biological materials in comparative experiments: the 18.1(2.2K)


- 271.4 - 268.9 4.5
size and geometry of tissue and its surface characteristics, as well as the (1.75) (4.25)
influence of the thermocouple itself, can all seriously affect measured rates.
Table 2.6 gives some physical characteristics of a range of quenching agents. ·This
S for Table
S ]"has .been
d p repare d to I.11 ustrate the general comparable thermophysical para-
For many fluids the temperature of the object to be cooled exceeds the tem re I ome iquli cryogens. There are many discrepancies in citations of such data in the
re · n part1cu ar ' the thermoph ysica · 1 properties
. of hquid
. helium vary sharply with tem-
perature to which the cryogenic fluid may be super-heated and this resul
ver a small range
rs. 0Furthermore th . For. th·sI reason, t h e fiigures have often been rounded to whole
in bubble formation at the object surface as the liquid boils. Rebhun (1972 es c/oc t th ' e_spec1fic heats and thermal conductivities have been quoted for tem-
~c. · t__ b ut_ not fo~ ~ny s!:_ecific temperature. Data drawn from
..•e o . e• meltmg _porn
1 / ~ rv·-.n\ "'!;""! T - - ~.i.. f 1 0.""11 \
Freezing 51
Robards and Sleytr Low temperature methods in biological electron microscopy
50
boiling, with the retardation in cooling rate that this necessarily involves.
considers that this will occur with liquid nitrogen, liquid oxygen, halocar- It was _report~d by Cowley et al. (1961) that coating the surface of a speci-
bon 14 and halocarbon 13, and probably also with halocarbon 22 and pro- me~ with an msu_lating material such as Vaseline or glycerol actually leads
pane. As indicated in Fig. 2.14 in §2.2.1, the highest cooling rate can ~ften to mcreased coohng rates in liquid nitrogen. This effect was further consi-
be obtained in the presence of nucleate boiling and, in fact, the maximum dered by Rebhun (1972) who also studied the influence of adding particles
nucleate boiling point is that point which separates the phase of transition or powders to the surface, which had previously been shown to increase
boiling from that of nucleate boiling (Fig. 2.14). At this point the. heat cooling rates in liquid nitrogen (Luyet 1961; Cowley et al. 1961). Rebhun
transfer (expressed as heat removed per unit surface area of the specimen) used insulating coatings such as glycerol, dibutyl phthalate, Vaseline, Form-
is given by Rebhun (1972) as: var and a number of oils; and powders such as copper, carbon, protein,
Q=_JC_ [ga(p1-Pv)Jo.2sx[ Pi Jo.s (2.10) pollen grains, spores, rosin and activated alumina. In general, insulat-
A 2ivPv pl Pi - Pv coatings and nan-metallic powders improve the heat transfer when speci-
mens are plunge~ into cryogens under conditions where nucleate boiling
would not otherwise occur. The insulating coating appears to provide a lay-
where
=heat removed er between the ~pecimen surface and the coolant such that the outer layer
Q
of the msulator 1s at, or below, the maximum nucleate boiling point, where-
A =area
=latent heat of vapourisation as the inner layer is at the temperature of the object. The net effect is that
Av
=density of vapour the additional surface heat transfer more than outweighs the extra thermal
Pv resistance imposed by the insulating layer and thus the specimen cools more
P1 =density of liquid
=acceleration due to gravity quickly. The powders act by providing nucleating sites for the initiation of
g
=surface tension. nucleate boiling at a higher temperature than normal. It should be noted
a
that, although improvements of an order of magnitude or so may result
when cryogens that would not normally allow nucleate boiling within
This equation assumes that the liquid is at its boiling point. Cons~dera~ly
the relevant cooling range (e.g. liquid nitrogen), no improvement (or even
greater heat fluxes can be obtained if the fluid is cooled belo"'. t~1s pomt. a reduction in rate) occurs if the cryogen already permits nucleate boiling
One aim of rapid cooling techniques might, therefore, be to max1m1se nucle-
(e.g .. propane). These techniques have not been widely applied to biological
ate boiling at a temperature relevant to biological specimens (e.g. not at the
~pec1me~s but provide a useful insight into how mechanical/thermodynamic
nucleate boiling point of liquid nitrogen at atmospheric pressure, 88 K,
mteractwns can modify measured cooling rates. Other surface modifica-
which is far too low to be of any significance). Bald (1984) has made a
tions (e.g. coating with PTFE; Umrath 1974b) have been used to enhance
detailed theoretical study of the relative efficiency of cryogenic fluids in rap- the 'wettability' of the specimen. Of more direct interest to the biologist is
id cooling. He concludes that liquid nitrogen, maintained near its melting
:he fact, referred to earlier (§2.2.3a), that the initial cooling rate can be
point of 61 K at a pressure in excess of the crit_ical va~ue of 33.5 b~r
improved maximising the exposure of the specimen surface to new cold
One simple way of carrying this out is to have a deep bath ~f the
(3.39 x 1Q6 Pa) will produce the quickest cooling. This remams to be expen-
mentally verified. . ola?t. and project the specimen in at a moderate velocity (1-5 m s~ 1 : too
h. m3ect1on velocities may mechanically damage or deform naked bulk
2.2.3b The effects of specimen surfaces and shapes on heat transfer c1mens); this is one of the favourable aspects of the spray-freezing techni-
(§2_.4.l_c). The propane jet method (§2.4.1 b) also maximises this feature
Although heat conduction through the specimen and the necessit_Y for heat OJ~ctmg the coolant at the specimen. An interesting alternative ap-
transfer across the specimen/coolant boundary have been mentioned, th h is to keep the coolant in constant motion (e.g. by ultrasonic means)
nature of the specimen surface has not been considered in detail. Frequentl
the large temperature difference between specimen and coolant leads to fil
____ , __________ - -

Freezing 53
Robards and Sleytr Low temperature methods in biological electron microscopy
52
equivalent · volume cooled even more rapidly·· 11 ,900 K s -1"ior a 0 1 mml
or to vibrate the specimen at about 50 Hz as it is plunged into the coolant vo
·. 1ume ma 20 µm-thick
. layer sandwiched between copper stnps . and. cooled
(Umrath 1974b). m propane. Despite. the above comments ' the b a1ance b etween heat mput .
Costello and Corless (1978) and Costello (1980) have given considerable to
, · an d h· eat withdrawal. from a cooling specimen
. must be considered in rela-
attention to the size and shape of the specimen and its support in relation , 1to(Iits overall cooling
non · rate. This has led p sc h er.d et al. (1981) and Knoll
to cooling rate. With some supports Costello and Corless reported improved el a . 982) · to stress the advantages 0 f usmg . a propane Jet . against
. a thin
freezing at lower injection velocities and attributed this to an optimisation copper specimen support under which is the (th. ) . . . .
turn backed b · . m specimen which 1s, m
of the interaction between coolant and specimen. At high rates of immersion ' . ya thermally msulatmg, plastic layer (§2.4.lb). Usi th.
some designs of specimen support caused the coolant to splash away from method, cooling rates of> 18 ' 000 K s-l h ave b een o b tamed. . ng is
the specimen, so retarding the cooling rate. The relationships between cryo-
gen viscosity and surface tension and between specimen shape and entry 2.2.4 Heat capacity
velocity are paramount in determining the actual cooling rate. These rela-
tionships are ones that, with the exception of the work by Costello and Cor- The
. heat capacity of a specimen is that quanftI y o f h eat reqmred
. to increase
.
less, have received rather little attention. The shape of the specimen can also its temperature by 1 K. Essentially ' the heat capacity
. te11 s us how much heat
influence cooling rate insofar as the relative surface area exposed to the
coolant is concerned (Moor 1973). If a plate is cooled from one side only
(half an infinite body) then the extent of any particular cooling rate into the Temperature (°C)

specimen can be given by the expression of distance 'x'. x can, in fact, coin-
cide with the limit of formation of ultracrystalline ice (so-called 'vitrifica-
tion'). At all events, if the limit of such a cooling rate is the distance x in
cooling a body from one side only, then it will be 4x if the body (platelet)
4.0
can be simultaneously cooled from both sides; 6x ifthe specimen is cylindri-
cal; and 7 x if it is a sphere (Fig. 2.26). The theoretical advantage of using
'
a specimen of cylindrical or spherical geometry for rapid cooling is clearly "' 'i.:
'E 4.0
evident and has also been emphasised by Costello (1980), who showed 1
that .., '
spherical samples of about 1.0 mm3 cooled at rates of < 500 K s- whereas "'
1
such samples of 0.001 mm 3 had rates of > 50,000 K s- under identical cool-
~
.::
u
..
3
II>
.::
ing conditions in liquid propane. The exception is that very thin samples of ::i

"c: 2.0 ~
0 ·c:;
"
'" 2.0
II>
Cl.
I/)
E
:;;
.c:
I-

o'--~~1*;:;---~---::~~__j_-----1.__Jo
100 200 300
Fig. 2.26. The influence of the shape of an object on the maximum thickness of the well-fr Temperature (K)
zen zone. For a half-infinite body (i.e. cooled from one side only) we may consider the de
of well-frozen material to be distance 'x'. Cooling from both sides of a platelet will prod The specific heat(Rand
d therm a1 conduct1V1ty
. . of water as a function of temperature
a well-frozen zone of 4 x while the values for a cylinder or sphere will be 6 x or 7 x respectiv e rawn from Bald and Fraser 1982.) ·
(Adapted from Moor 1973.)
Freezing 55
Low temperature methods in biological electron microscopy
54 Robards and Sleytr

. .. d volume can 'store'. Materials of high expressed in terms of specific heat (cp) in J kg- 1 K- 1 and is itself tempera-
an object of specific com~os1ti~n ~~e heat than those of low heat capacity. ture-dependent (see Fig. 2.3 in §2.1.1 for the heat capacity of water and ice
heat capacity are able to hold m . ·11 cool more quickly than high and Fig. 2.27 for the specific heat and thermal conductivity of water and
h t ·ty specimens wi
Conversely, low ea capac1 1 d ct1.vity The snecif!c heat of ice). Thus, the total heat capacity of a specimen derives from both its speci-
· f · alent therma con u · ,, .
heat capacity ones o eqmv . .t At constant pressure it is fic heat and its actual volume and this is the reason why, for example, ther-
a substance is the heat capacity per um mass. mocouples of identical composition show cooling rates that are proportion-
TABLE 2.7 al to volume (i.e. to heat capacity).
Specific heats of some liquids and solids at low temperatures If we were able to specify the nature of the specimen that we wanted to
(J g-1 K-1) freeze, then it would have high thermal conductivity, low density and low
Specific heat close to melting point heat capacity. Unfortunately biologists are usually unable to modify the
Melting point
Substance thermal conductivity and density and can only reduce the heat capacity by
°C (K)
reducing the volume. It is, however, possible to consider the requisites for
4.2 the cooling medium which, ideally, would have a high heat capacity as well
0 (273)
Water 2.1
Ice
0 (273)
2.06
as a high thermal conductivity. Some representative heat capacities of mate-
Liquid nitrogen - 210 (63) rials encountered in low temperature biology are cited in Table 2.7.
4.41
- 271.4 (l.75)
Liquid helium 1.89
-117 (156)
Ethanol 0.854 2.2.5 Cooling rates in liquid cryogens
-158(115)
Freon 12 1.088
- 160 (113)
Freon 22 2.287
Isopentane
- 159.9 (113)
1.92
The number of citations of cooling rates in the literature is extremely large
- 189.6 (84) and, as explained earlier, many of the citations are oflittle use for compara-
Propane 2.27
- 183.5 (90)
Ethane tive purposes. However, Table 2.8 lists some of the rates that have been
Specific heat at a temperature ofC C (K)) quoted when cooling in liquid cryogens. As will be seen from this table,
- 253 (20) - 269 (4) actual measured cooling rates from different laboratories are not always in
- 100 (173) - 196 (77)
0 (273) good agreement and it has not always been entirely clear which liquid cool-
0.33 ant is 'best'. Rates have been measured using different sizes of thermocouple
1.04 0.69 0.000008
Plastic (PTFE)
0.402 0.063 0.002 and, whereas improvements of 2.6 (Luyet and Kroener 1960, from 248 to
Sapphire 0.00837 0.000261
0.89 0.77 0.37 198 K), 5.8 (MacKenzie 1969), and 4.0 times (Umrath 1977) have been
Aluminium 0.010 0.000091
0.7 0.33 found when sub-cooled nitrogen is used instead of boiling nitrogen, the
Copper (crude) 0.008 0.000091
0.35 0.21
Copper (ultra-pure) 0.38 0.0125 0.000124 improvements recorded by Glover and Garvitch (1974) and Costello and
0.23 0.146
Silver 0.23 0.016 0.00016 Corless (1978) were only about 1.2-1.3 times. Absolute rates cited for liquid
0.121 0.096
0.126 0.000317
Gold
0.418 0.23 0.006 11itrogen at 77 Kare: 39,000 K s- 1 (Luyet and Kroener 1960, using a 25 µm
Titanium 0.54
0.47 0.39 0.17 'thermocouple); 437 K s- 1 (MacKenzie 1969, using a 250 µm thermocouple);
Stainless steel 0.22
Glass (Pyrex) 0.71 0.5 K s- 1 (Robards 1979, using a 500 µm thermocouple); 1880 K s- 1 (Glover
Quartz 0.75 d Garvitch 1974, using a 240 µm thermocouple) and 16,000 K s- 1 (Cos-
. .
·
ust be considered as approx1mat10ns.
· There are man and Corless 1978, using a 70 µm junction). In general, it appears that
N .B. The data provided m this Table m . . ·11 b c d Furthermore, the cited te erences in rate are mainly ascribable to different thermocouple sizes and
. d hence vanat10ns WI e ioun .
discrepancies in the literature an ' ' . f ·nterest to low temperature
d · ·d with temperatures o 1 . erent immersion velocities but the comparative differences between the
P eratures have been selecte to comet e . the region o' the temperatures specz
.. h . ly to temperatures m ~ . cited above for nitrogen at 77 K and at 63 K cannot be so dismissed.
roscopists but the specific eats app . d f om numerous and vaned sour
arily precise. Data rawn r llo and Corless (1978) discuss the problems arising from the thermo-
and are therefore not necess . 1 (1970) Weast (1971) and Vargaftik (1975).
including Lange (1961), Touloukian et a. '
;

Freezing 57
Low temperature methods in biological electron microscopy
56 Robards and Sleytr
;;, dynamic nature of sub-cooled nitrogen. Depending upon how the sub-cool-
~
-
,..,~

t- .~ I
'1: "'ci
otj" """ ing is effected, either by direct pumping as in Fig. 2.41 (§2.4. la) or by indir-
&; ~",. '---'
"'8 ~ ""
<.!::: ect conduction as in Fig. 2.42, we may be dealing with a true mixture of
solid and liquid nitrogen at 63 K ('slush') or we may simply have a sub-
cooled liquid at some temperature close to (but not actually) 63 K. In either
'"§<a"' ~,._ case there will be strong convective temperature gradients within the pool
0 0 0

!
0
"a ""E:':l 0 <-- 'rl 'D
and some of the discrepancies in cooling rates probably result from this.
E 'B "' "'
Furthermore, the problems of immersing the specimen in a reproducible
""
~ way in the sub-cooled nitrogen can also lead to inconsistencies that do not
arise when the container need not be evacuated.
0
ff we continue to examine absolute cooling rates from 273 to 173 K (Ta-
00

-
<--
"'
"'"'
-
00
"'0
'()
ro
ble 2.8), then Costello and Corless (1978) find that propane is significantly
better (98,000 K s- 1) than any of the halocarbons (with halocarbon 22,
"0
-.::
...
...
N
u "
E--< K s- 1, rather better than halocarbon 12). Rosenkranz (1975) calcu-
-
00
"O
"' "O 0
<:I lated 273 to 133 K cooling rates for a 1.2 mm diameter water droplet and
-...
00
§ ro
~ "'"O
:::-- """.: 0 "
~ cited rates of 110, 460, 1410 and 2060 K s- 1 for Freon 22, Freon 13, propane
~
I
"' ""
~ "w..."
"O
~0
u
0
"
N
~ and sub-cooled nitrogen, respectively. Once again propane appears better
"' ""
"<
(/)
than the halocarbons but this time not as good as sub-cooled nitrogen,
~... although Costello and Corless (1978) have made some criticisms of the cal-
Oil
00
0i ;§ .: 0 00 culations and methodology of Rosenkranz. Glover and Garvitch (1974) also
0 '.::l
0 0
0 N s;
""""a
0 <--
""<!'.
.--l
~
0
(.)

">
"'
.::r
t:
>4 '£
00 0

'°-
compared calculated and measured cooling rates under a variety of condi-
tions. The measured rates always considerably exceeded the calculated rates,
E--< ·~
...ro a fact that the authors attributed to the occurrence of forced convection
""8
0 .: during immersion, which they had not taken into account in their calcula-
g""""
(.)

"80 0 0
0
0
0
tions. Using a 240 µm junction they obtained actual 273 to 173 K rates of
0
(/) '£ ""'~"' <--. o.
;:::;
o.
2,120 K s- 1 (Freon 12), 1,880 K s- 1 (nitrogen at 77 K), 2,320 K s- 1 (pumped
'.::l
nitrogen) and 1,350 K s- 1 (Freon 22). Rebhun (1972) cited his cooling rates
"'
.::r
>4 in a different form, giving the final rate at 191 K. He obtained rates of over
~ 5,000 K s- 1 for propane and propylene, with lower rates for the halocar-
"""'.:""
a
0
00
'rl
0
0
0
c-'.'
0
otj"
<--
"'
0
'rl

"'
-
"'<--
"O
bons, Genetron 23 (5,410 K s- 1). Once again, Freon 22 (3,976)
proved better than Freon 12 (2,940). Umrath (1977) found propane and
G; "" '° otj" §
"'<--N "'u.',-"''""cunitrogen to be better than Freon 22 and isopentane. Ethanol has
g shown to be an extremely good cryogen down to about -100°C (173

"""".: 0
0
0
0
0
0
-
N
'D
"
~
.0
at which point it becomes very viscous prior to freezing (Silvester et al.
2) while these same authors also reported that ethane gave cooling rates,
a
G;""
a::

'D "' "...ro in 3.0 µl samples, approximately twice those obtained using propane.
"O
~ virtues of ethanol as an efficient cryogen have also been described by
'()
"' hesse-Ragona (1984).
"".:':l 0 0 0 'D ~...
sa; 0
0
0
oo. summary, sub-cooled nitrogen does not uniformly produce the high
~ ~
,._
i:i.. - 00
"'
...£
Freezing 59
Robards and Sleytr Low temperature methods in biological electron microscopy
58
bal. I 1981)h that the addition of20-50 olo1 isopen t ane a ll ows propane to sub-cool
rates that some authors have attributed to it. On the other hand, it is safe,
e ow t e te~perature ofliquid nitrogen (Fig. 2.28) so that th b th .
easy to produce and does not leave a contamination of solid coolant on the constantly liquid; the cooling properties are stated not to bee ada rem] amfs
specimen if this is transferred for storage in liquid nitrogen. The halocar- rected Th · ·· · verse y. a -
1 · e maJor cntlc1sm of the use o f 1·1qm·d propan · ·t
bons are also relatively easy to use but have the disadvantages that they hazardous nature: it is highly inflammable c . e is 1 s potentially
form solids at liquid nitrogen temperature; they may give contaminating as low as 2% [20 000 ]· . ) . ( ombustible at concentrations
o ' ppm m air ; and it can be sub-cool d b .
peaks if the specimen is subsequently used for X-ray microanalysis (Echlin faction point of oxygen so allowing th d . e eyond the hque-
. ' e con ensat10n of oxyge t f
and Saubermann 1977; Gupta 1979); and they may have adverse health and a potentially explosive mixture (Stephenson 1954) · N evert h eless handled
n o orm
. ·
environmental effects (Taylor and Harris 1970). Of the halocarbons, halo-
sma11 amounts usmg properly constructed apparatus and proced ' h.in
carbon 22 is better than 12 but the relatively unused halocarbon 23 has been should not be allowed to detract from th e use o f propane as a ures,
ffi .t is
found by Rebhun (1972) to give particularly high cooling rates. Umrath cryogen. Provision of a flammable gas sensor (A d" . very e ic1ent
(1977) and Rosenkranz (1975) obtained very good rates with Freon 13 sible precaution to take in areas where pr . p~en ix 2) is a further sen-
- . . opane is m use.
(Rebhun did not) but this is not a particularly easy cryogen to use 6in view
~os~s~~:~:hie ~eed t~
1 he confus10n m the citation of cooling rates em h .
of its high vapour pressure at room temperature (35 bar (3.5 x 10 Pa) at duce an acceptable standard in this field. Because pro-
25°C (298 K)). Costello and Corless (1978) discussed the particular virtues are heterogeneous and cool at a rate related to the thermal g ca spe.c1~ens
of Freon 13 as a cryogen and concluded that, using biological specimens the ice layer as it forms ' it is probabl Y s1mp . Iest and mostconduct1v1ty
· t · of
of normal size for freezing work, the rates obtainable were less good than obtain an estimate of surface heat transfer properties simply bms Y c1 mg ratesto ~t~ctive
those obtained with Freon 22. Menold et al. (1976) carried out some experi-
ments on freezing small droplets (approx. 2.0-20.0 mm diameter) of water
in liquid nitrogen on Freon 13 at about 93 K. The cool-down rate from 273 300
to 173 K was measured using 50 µm wire thermocouples. The results show I
I
that rates of up to 100 K s- 1 were obtained using liquid nitrogen, but rates I
I
in excess of 1,000 K s- 1 resulted from quenching in Freon 13. The relation- 273 /
/
ship between droplet size and cooling velocity was linear. In fact, the techni- /
/
que used was to re-cool ice spheres that had been warmed to exactly 273 b.p:.- /
/

K and hence the latent heat of crystallisation was not taken into account. 2., .....

However, the authors consider that this would not reduce the actual rates ~ 223 -- -
<;;
by more than a factor of two or three. a;
c.
While not all past workers have experimented with liquid propane, the ~ 200
I-
most recent work by Costello (1980), Elder et al. (1982) and in the authors'
laboratories tends to reinforce the undoubted superiority of liquid propane
as the best liquid cryogen (Table 2.8). This might be expected from its low 100
melting point and high boiling point as well as its good thermal conductivity m.p.
and specific heat. It must be concluded that, for the highest cooling rates
77 --------
in liquid cryogens, propane offers the best current possibility although the 70
lesser-studied ethane may well be equally effective (Silvester et al. 1982). Propane (%) 100 80 60 40 20 0
Propane has the added advantages that it is readily available and is relati- lsopentane (%) 0 20 40 60 80
228 .. 100
vely inexpensive. While pure propane melts at about 83 K, some commercial . ·30--40%
· Boi!mg points in
isopentane (b.p.) and meltin g pomts
· .
(m.p.) of different propane/isopentane mix-
forms remain liquid at temperatures between 83 Kand liquid nitrogen tern~ nit~:og~an(e~o;ers thfe meltmg pomt below the temperature of liquid
n. e rawn rom Jehl et al. 1981.)
perature (77 K) for prolonged periods. It has recently been shown (Jehl et
Freezing 61
Robards and Sleytr Low temperature methods in biological electron microscopy
60
that the liquid does no~ r~main as a pool. This is largely because of the poor
TABLE 2.9
Some criteria for the citation of cooling rates using thermocouples heat transfer
. charactenstic
. resulting from film b 01·1·mg. H owever, we would
~ot thmk of placmg our hand on the surface of a piece of metal even if it
(i) Thermocouple Small size (25 µmjunction would provide a good standard). were far less cold than the temperature of liquid n·t1 rogen, smce
· a severe cold
Cite metals (Cu/Con is satisfactory for most purposes).
Cite junction diameter and wire diameter.
Cite how junction was made (better welded than soldered). TABLE 2.10
Ensure wires are well insulated up to the junction, and carry out tests to Thermal conductivities of some liquids and solids at low temperatures
ensure that heat is not transferred via wires. (Jm-Is-IK-1)
Standardise thermocouple and use the same standardised thermocouple
for comparison. Substance Melting point Thermal conductivity close to
Immerse at a reproducible velocity and cite entry velocity and deceleration °C (K) melting point
(ii) Immersion
profile. 0 (273) 0.5551
Immerse to constant depth and cite depth. Water
Ice 0 (273) 2.2
Cite nature of specimen immersion mechanism: whether specimen is
Liquid nitrogen - 210 (63) 0.1322
rigidly held, vibrated, etc. 0.0181
Liquid helium - 271.4 (1.75)
Cite volume of cryogen, depth of liquid, temperature and thermodynamic -117 (156) 0.2064
(iii) Cryogen Ethanol
state. Freon 12 - 158 (115) 0.129
Measure and cite temperature variation within liquid pool. -160(113) 0.143
Freon 22
With freezing cryogens (e.g. halocarbons) ensure reproducible conditions - 159.9 (113) 0.1
Isopenlane
(e.g. use a liquid pool at point of refreezing or a co'ntrolled temperature - 189.6 (84) 0.1
Propane
device to maintain the cryogen close to its freezing point). - 183.5 (90) 0.24
Ethane
State whether cryogen was stirred or agitated in any way.

If a specimen is used, cite: Substance Thermal conductivity at a temperature of(°C (K)):


(iv) Specimen
location of thermocouple in specimen;
method of attachment of thermocouple to specimen; 0 (273) - 100 (173) -196(77) - 253 (20) - 269 (4)
data relating to thermocouple as specified in (i) above;
precise volume, dimensions and shape of specimen; Plastic (PTFE) 2.5 2.5 2.4 1.4 0.5
orientation of specimen during immersion; nature of specimen (so far as Sapphire 46 82 960 15,700 410
possible if biological materials are being used); Aluminium 210 210 410 15,000
method of attachment of specimen to immersion apparatus. Copper (crude) 400 400 460 1,200 380
Copper (ultra-pure) 398 413 570 10,500 11,300
Cite temperature range over which cooling velocity was calculated 471 5,100 14,700
(v) Presentation Silver 427 430
(273-173K is appropriate). Gold 315 327 352 1,500 1,710
Cite whether T vs t curve is linear; preferably show graph of T vs t or 34 28 5.76
22 25
dT/dtvs t 15 13 9.4 2.0 0.3
1.0 0.9 0.55 0.15 0.1

of bare thermocouple response. The criteria provided in Table 2.9 would 1.38

go some way towards allowing comparable data to be cited. . . provided m


The data · th.IS T a ble must be considered
. as approximations. There are many
epancies m the literature and, hence, variations will be found. Furthermore the cited
2.2.6 Heat transfer at cold surfaces atures
. have been se1ec t ed to comc1 · ·de with
· temperatures of interest to low temperature
'
.fi d but the thermal con duc t'1v1t1es
op1sts · · app1y to temperatures in the region of the tempera-
Those who have spilt liquid nitrogen onto their skin will know that the re~ . 1 ;e d_and are therefore not necessarily precise. Data drawn from numerous and varied
me u mg Lange (1961), Touloukian et al. (1970), Weast (1971) and Vargaftik (1975).
sult need not be disastrous, despite the temperature of the coolant, provide
I
Freezing 63
Robards and Sleytr Low temperature methods in biological electron microscopy
62
However, when a warm specimen comes into contact with a cold surface
'burn' would result (see Appendix 1 on hazards in low temperature biology).
the surface
· must necessarily warm up ' th us ch angmg
. o: and therefore th
This is a simple demonstration of the greater heat transfer across a solid
coolmg rate. Heuser et al. (1979) approached this problem, by attem , tine
interface than across a liquid one. This phenomenon is made use of in 'cold
to calculate
rd) ld the equilibrium temperature (T ) 0 f th e b oundary between p theg
block' freezing (§2.4.2). The use of cold blocks for freezing biological speci- 0

(so 1 co
t d" t surface
· and the specimen. They calculated that th t
e emperature
mens was first described by Eriinko (1954), while Feder and Sidman (1958)
~ a is ance x mto a body at time t after initial contact will be:

.!~=·
and Van Harreveld and Crowell (1964) later constructed systems for effi-
cient freezing in such a manner. A number of authors have subsequently
Tx = Ti erf __
described variations of the same technique (see §2.4.2). i x_
j(4o:t) (2.12)
Some metals increase considerably in thermal conductivity as the temper- 1
ature is reduced. For example, the thermal conductivity of high purity cop- •
where
per is more than ten times greater at 10 K than at room temperature (Table J•.· =temperature at distance x after time t
2.10). This characteristic may be compared with that ofliquid cryogens, the •
absolute thermal conductivity of which is both lower (Table 2.6 in §2.2.3a) j x
=initial temperature
=distance into body
and decreases with falling temperature. Similarly, the heat capacity of met- ·l
=thermal diffusivity
als is much higher than those of liquid cryogens (Tables 2.6 and 2.7). For ;Ill
=time
these reasons, provided that a specimen can be brought rapidly into contact
=error function.(Carslaw and Jaeger 1959).
with the cold surface of a high conductivity metal, heat can be withdrawn
at a very high rate indeed. It will be clear that the surface of the metal needs
Applying the boundary conditions for their model (Fig. 2 29) H
to be extremely smooth so that good surface contact can be achieved and al. denved: · , euser et
must also be well protected from contamination (e.g. with condensing water
vapour) prior to the freezing process if the full benefit of the improved heat
To= A1T1 +A 2 T2 (2.13)
transfer is to be obtained.
The cooling rate achievable with cold blocks is not, as is usual with liquid Ai+A2
cryogens, limited by such factors as film boiling or other effects of that kind
where
on heat transfer. Some comparisons of the cooling characteristics of differ-
To =equilibrium temperature of the bounda b t .
ent liquid cryogens and solid cold biocks have been restricted to thermal cold surface and the specimen ry e ween the (sohd)
conductivity alone (which, for example, suggests both copper and sapphire
A =J(kpcp)
as excellent materials to use as blocks for rapid cooling). However, this
k =thermal conductivity
grossly oversimplifies the situation. It can be considered that, for thermal
p =density
properties independent of temperature but with temperature varying with
Cp =specific heat.
time, the important parameter is the thermal diffusivity (o:) where:
It was argued that To is the only influence
k the specimen and since this d of the metal _on what happens
o:=-
pcp '1?roperties should be c ' "d d epends on A, then this combination of
ons1 ere . Thus for a I T 1
a~d _A2 are necessary. ' . ow o, ow T1 and T2 and high
where H is mstructive to consider th . . .
=thermal conductivity ~stituting A: e md1v1dual effect of each of the properties
k
p =density If (thermal conductivity) must clearly be as high as pqssible since this de-
•...
=specific heat.
---------·-
- --- ---- --- -- --·------ - -

Freezing 65
Robards and Sleytr Low temperature methods in biological electron microscopy
64
far higher cooling rates than liquid cryogens, a fact that is supported by the
theoretical estimations of Jones (1984), although a more equivocal conclu-
sion was reached by Kopstad and Elgsaeter (1982). Whether this method
in fact, used must, therefore, depend primarily on the nature of the speci-
/
' Time (=t)

Specimen (01}
men to be frozen as well as what is subsequently to be done with it. ·

22.7 Comparison of cooling methods

From the above considerations the choice of the important parameters for
a good cry~gen can be made. As an effective method of rapid cooling, cold
Fig. 2.29. The theoretical distribution of temperatures in a metal freezing block and specimen block-freezmg clearly should have advantages over liquid quenching meth-
just before (time= O) and at any time after (time= t) the two make contact. It has been reasoned ods although it remains to be demonstrated that this theoretical advantage
that at the interface where the two make contact, the temperature at the surface of the metal can be translated into practical fact. Cold block-freezing produces frozen
blo;k (which starts at T ) must very soon become the same as the temperature of the surface
2
of the tissue (which starts at T 1). This interface temperature may be called To· (Redrawn from specimens that have a relatively large area of flat, well-frozen (albeit only
Heuser et al. 1979). to a depth of 10-20 µm) tissue, so providing a good surface for subsequent
cryoultramicrotomy (§4.3). Cooling using liquid cryogens also has advan-
termines the rate at which heat is conducted into the coolant (liquid or solid) tages, not only on account of simplicity but also because, by moving either
and away from the specimen. . the specimen or the coolant, the temperature of the coolant at the interface
cP (specific heat) determines the quantity of heat absorb.ed per umt te~- can be maintained essentially static, thus maximising AT. Furthermore, be-
perature rise. It is desirable that a cold block should have high hea: capac.1ty cause of the mechanical deformation or damage that must necessarily ensue,
so that it will absorb much heat without a large temperature nse which not all specimens lend themselves to the technique of being thrust against
would reduce the temperature gradient (AT) and hence the cooling rate of a metal surface and, finally, such a freezing method is essentially unilateral,
the specimen (Eq. (2.3) in §2.2.3). Furthermore, the diffusion equation .(rx) i.e. cooling can only occur from one side of the specimen. Consequently,
predicts that low cP will result in a high rate of temperat~re change w.h1c~, although heat withdrawal may be extremely good at one surface, the rate
for the coolant, would mean a high rate of temperature mcrease, which 1s will rapidly diminish as the freezing front penetrates into the tissue. For very
clearly undesirable. Thus, cP should be high for the coolant (block) but l?w si:iall specimens where better bulk freezing is required (as opposed to a very
for the specimen (although this is not usually a parameter that can be vaned high rate at one surface) it may be better to contemplate spray-freezing
when dealing with biological materials). (§2.4. lc), propane jet-freezing (§2.4.1 b) or even simple quenching in a liquid
p (density) of the coolant should also be as high as possible (thus cooling bath. All of these methods improve the overall cooling capacity by with-
properties of solids are better than those of liquids which are better than drawing heat over the whole specimen surface.
gases) and, as with cP, the density of the specimen would ideall_Y be l~w. . In summary, when first contemplating how best to freeze a particular spe-
Bald (1983) has re-evaluated the theoretical aspects of freezmg usmg ~old cimen, the following aspects should be considered. What is the basic infor-
blocks, commenting that Eq. (2.13) above neglects the latent heat of fus1?n, ·1fiation to be obtained from the specimen and what maximum size of ice
which is released during freezing and is only valid provided that the density, . ta! is compatible with this objective? There is, for example, no point in
specific heat and thermal conductivity of both the block and specimen re- ng to produce 10 nm-diameter ice crystals ifthe specimen is to be studied
main constant. Using the method of Finite Elements, Bald has concluded n SEM with an effective resolution far worse than this. How small can
that the ultimate in quick-freezing methods is obtained when using a hi~h­ specii:-ien reasonably be made and is it 'solid' or could it be 'sprayed'?
purity copper block cooled to 19.1 K, a temperature tha: ":'~ll give cooling possible to use well-frozen material at the periphery of the specimen
rates approximately 50% higher than if the block were m1tially cooled .to ust deeper parts of the specimen also be studied? If a relatively large
5.0 K. At all events, it remains clear that, in theory, cold blocks should give
67
Freezing

66 Robards and Sleytr Low temperature methods in biological electron microscopy


·
'good' cell structure even though the cell will not b e via
. ble i"f b rought back
( > O. l mm in all dimernions) specimen block is to be well frn"'n thrnugh- to its no~mal growth temper_ature. The reasons for this distinction are relati-
out, then a cryoprotective is likely to be essential. Furthermore, such large stra1ghtforward. and anse from a consideration of wh a t h appens when
specimens cannot be well frozen deeper than about 10-20 µm from the sur- are frozen at different rates (see Luyet 1966" N . 1973· F
Williams and Hodson 1978). , e1 , arrant et al.
face and there is therefore little point in utilising ultra-rapid methods. For
such 'bulk' specimens, simply freezing in sub-cooled nitrogen or a halocar- . At slow ( < 1.0 K.s-1) rates of cooling, crystallisation of ice commences
bon bath will be equally as effective as other methods. Where the specimen m the external. medmm . surrounding the cells an d , b Y furth er slow crystal
can be made very small ( « 0.1 mm in at least one dimension), or where only · ~- h11, water is withdrawn
growt . across the membranes o f the ce11 s so raismg
.. the
the surface need be well frozen, then cold block-freezing, propane jets, .
imrace u1ar concentration of solutes . This causes sh rm . k age and deforma-
plunging into prnpane m sprny-fre°'ing (if the specimen is 'sprnyable') all tion of the . .cell structure (Figs. 2.30 and 2·31 ) b u t re d uces the available
. free
serve equally well and the final choice will primarily depend on the nature water w_1thm the cell ~o that, when the cell contents eventually freeze the
of the specimen, the equipment to hand, subsequent processing steps and do so w>th th_e focmatwn of small, non-injmious ice c;ystals. Such cell; may
surnvo and'. cf thawed undec a pprnpriate conditions, show high viabilit y
the information ultimately required. . .
Intermediate rates of cooling (1.0 to 1,000 K s -1) are not so slow y.that
there ·is. time for mass. withdrawal of water across the ce11 memb rane but are
2.3 Treatment of biological specimens prior to freezing certam1y not sufficiently rapid for the critical ternperat ure range to be tra -
- .

~co~;
versed so qmckly that small, non-damaging crystals are produced Th

~e
So far, we have almost exclusively considered model systems and cooling sequence rn that celativcly lacge intcaccllulac ice ccystals fonn so fuat
rates obtained using bare thermocouples or simple specimens. The situation is badly damaged morphologically (Figs. 2.30 and 2 ·31 )., so1u te re1ocat10n
e
becomes far more complicated when dealing with highly heterogeneous bio-
logical specimens. Different cells of a tissue may freeze at different times -10°C
and, depending on the nature of the contents of various intracellular com-
partments, these may freeze with entirely different patterns of ice/eutectic
-formation. In addition, the low temperature biologist has a fundamental
choice to make: is the specimen being frozen to provide the best preserva-
tion of ultrastructure and retention of solutes in their in vivo positions; or
is the aim to achieve the best viability of frozen cells after they have been
thawed? Rarely are these two criteria of successful freezing met simulta-
neously from the same cooling method. The following Section considers
what happens when biological cells and tissues are cooled at different rates.

2.3.I Cooling rates in relation to structure and viability

The acme of freezing biological specimens would be to do so in such a way


that, without any pretreatment, the cells would remain potentially viable
and would also maintain their in vivo ultrastructural relationships and
solute gradients. Such a goal is at present not achievable for the vast major-
ity of cells. It is usually necessary to make a choice between using a freezing
prntocol that cbang" the internal st<uctme and solute conccn trn ti om of th<
cells but allows them to survive after thawing and a method that preserves
Freezing 69
1ods in biological electron microscopy
Lo w temperature me tl
68 Robards and Sleytr
is considerable; membranes are disrupted; and the cell will certainly be dead
when thawed. Such rates of cooling are of no use for either ultrastructural
purposes or preservation of cell viability when applied to unprotected cells.
Perhaps their only virtue is for the purposeful disruption of cell contents
as a biochemical homogenisation method.
At very high rates of cooling(> 1,000 K s- 1) the ice crystal size within
cells becomes much smaller (Figs. 2.30 and 2.31 ). If the critical temperature
range can be traversed extremely rapidly ( > 104 K s- 1) then it should be pos-
sible to freeze unprotected cells in such a way that they are both viable and
also show good structural preservation. This has been achieved using ultra-
rapid freezing techniques on suspensions of isolated cells (Moor 1973) but
it is no t possible for large specimens and tissues. For these, it is almost al-
ways necessary to protect the cells from freezing damage by the addition
of a cryoprotectant (§2.3.2a).
The relationship between the structure and function of cells frozen at dif-
ferent rates has been discussed by a number of authors (references below).
In relation to freeze-etching, Moor (1973) carried out experiments on the
survival rate of yeast cells frozen at different cooling velocities (Fig. 2.32)
and found that high viability could be achieved with either very low or very
high cooling rates. Yeast, in fact, appears to be a rather tolerant organism

100

*
iii 50
>
>

"
(/) I
I
I
I
I
o~~~-'-~~--'~~~~~~--~~-'-~~-'-~~-'-~~~

0.001 0 .01 0 .1 1.0 10 100 1000 10000


Cooling rate (K s- 1)

C f at cool- Fig. 2.32. Surviva l of yeast cells in relation to cooling rate. The solid lines (taken from Moor
. d ·f~ t rates (a-{;) Yeast cells rozen 1973) demonstrate the cryoprotecti ve effects of glycero l at different concentrations in enhanc-
F ig 2 31 Micrographs illustrating coohn gd altOOOi (er)eK
n s- 1 The lowest rate produces a shrun- ing viability of yeast cell s after cooling at different rates. The dashed line (from Bank and
. · · · l 1 o( ) 1O (b) an
ingrates of approximate y . a , F c2 30) while· · t leads to
the intermediate ra e . Mazur 1973) shows yeast cell viabi lity when cooled in distilled water and corresponds quite
ken cell as a result of w.ater wit~t~~alT~:\i~;~st . rat~ gives relatively good preservau~nO~~ well wtth the equiva lent curve from Moor's data . (Redrawn from data presented in Moor 1973
some large intracellular ice crysta s ed cardiac myocytes fro zen by propa.ne iet at about ~hilc and Bank and Mazur 1973).
seen at this magniflcatron. (d4:) l~~lat at the lower rate (d) shows clear ice crystals (~id and
(d) and 20,000 (e) K s- 1. The eel roz~n s extremely fine ice grain structure. (Fi gs. 2.
the more rapidly fro zen myocyte (e) s ow. f Robards and Severs 1982.)
-1 _ .,...,...,...l ' " 't\, n P rn11SS10fl. [Qffi
Freezing 71
thods in biological electron microscopy
70 Robards and Sleytr Low temperature me known to remain viable (e.g. spermatozoa; Koehler 1970, 1972). As com-
.t. much more difficult to obtain high mented by Farrant et al. (1977), most cells which are able to recover on
so far as freezing is concerned a~d i is h Us Nei (1977 1978) also thawing are grossly shrunken at low temperatures, but since they are poten-
. ates with most ot er ce . ,
viability at fast freezmg r . f t cells together with that of functional they are of interest structurally. Further work in this area
. t and funct10n o yeas .
investigated the struc ure H found that the freezmg pat- be rewarding.
. 11 d human erythrocytes. e d
mammalian ce s an . d "th different cooling rates an If it is decided that the best retention of cell structure can be achieved
·val of cells vane wi
tern and post-th aw survi . f t and red blood cells. He high cooling rates, then the question of ice crystal size in relation to
. t re different or yeas
that the optima_1 ra es we ell distortion at low cooling rates but g~od cooling rate in a particular subcellular location becomes of paramount im-
recorded the typical pattern of c . t l f rmation at high rates, part1cu- portance. Some methods of freezing can produce extremely high cooling
structural preservation with little ice crys a o f rythrocyte membranes rates provided that the specimen can be made small enough (§2.4). If, however,
sed The structure o e
larly if cryoprotectants were u . d. d by FuJ.ikawa (1981) while the specimen is in any sense a 'bulk' specimen ( > 0.1 mm in all dimensions)
. te has also been stu ie .
in relation to coo1mg ra . function of cooling rate m then the cooling rate throughout the whole object will be inadequate to pro-
t died ice format10n as a
Leibo et al. ( 1978) s u 11 h d"fferent optimal cooling rates duce uniformly small ice crystals. This results primarily from the low ther-
t that different ce s ave i
mouse ova. The f ac . db ber of workers including Mery- mal conductivity of liquid water and ice (§2.2.3; Fig. 2.24) and the conse-
. lh b emphasise ya num '
for surviva as een 1 (1977) (Fig. 2.33). quent slow rate of heat withdrawal so that the critical range (effectively 273
) M (1977) and Farrant et a. . §2 6
man (197 4 ' azur b" t f viability after thawing m .. to 173 see §2. l) cannot be traversed sufficiently rapidly to give small ice
We shall return briefly to_ the s_u ~ecf t~ processes of freezing for ultra- crystals. This is a fundamental feature of hydrated biological specimens and
However' f or a further considerat10n o e t that fast cooling rates wi·11 a1-
can only be surmounted by:
structure research it should now be apparen 11 not give cells that are
f 1 and that these may we
most always be essen ia h d slow-frozen cells for ultra- (i) increasing the cooling rate (impossible in bulk specimens due to ther-
. bl Only a few authors ave use b
potentially via e. . 11 b en where the cells have een modynamic considerations; §2.2.3), or
structural examination and this has usua y e (ii) by treating the specimen with a cryoprotectant (§2.3.2a) that will by
some means reduce the ice crystal size for a given cool rate. It is this
second method that becomes a necessity for many specimens.
Red blood cells
70 stem cells
!2.3.2 Principles of sampling and pretreatment
60
ideal method for freeze-fixation is simply to take the specimen and
50 e it directly in such a way that it retains good morphology, ultrastruc-
# ' solute distribution and viability. For some isolated cellular specimens
~ 40
iO goal can become a reality if optimal cooling methods are used (§2.4.1 ).
>
~ 30 ver, most biological specimens have to receive some pretreatment
"
(/) immediately poses problems relating to damage and artefacts. For in-
20 to freeze most tissues it is necessary to excise a small sample. This
leads to mechanical damage at the edges of the specimen. Even some
10
100
test methods currently available for freezing only give good preser-
0 o.o 1 o. 1 a depth of about 10 µm in hydrated tissue (Fig. 2.34; Table 2.11 ).
1
cooling rate (K s- 1 about 25 µm in diameter then clearly good freezing will not extend
·rr t cooling rates
. s of cell when frozen at d l eren
Fig. 2.33.
The survival of different typ.e d"fferent rate. (Redrawn from Mazur et e depth of mechanical damage. Furthermore, the necessity for
oc Each cell has a maximum survival at a I
- 196 . 1970J
l

ds in biological electron microscopy Freezing 73


Robards and Sleytr Low temperature me tho
72
TABLE2.ll
Depth of ice crystal-free zone in tissue without cryoprotection
.....a------------,c--
900 '\
I \ Depth Method of quenching Reference
I I
I \ (µm)
800 I \
I \
I \
10-15 Impact on Ag block (66 K) (1.0 m s- 1) Van Harreveld et al. (1974)
700 I I
B Impact on Cu block (77 K) (unspecified rate) Dempsey and Bulivant (1976a,b)
E
I \
/
12
\
.=. I
I / 15 Impact on Cu block (4 K) ( < 0.8 m s- 1) Heuser and Reese (1976)
.. 600 I ' - - - - - - -C
/
/
20 LN 2 (77 K); motor-driven entry Saetersdal et al. (1977)
.
N
I
I
/
/
5-8 LN 2 slush (65 K); 'rapid-entry' Seveus (1978)

-...,
500 I /
I A / 25-30 Freon 22 ( 113 K); spring-driven entry Handley et al. (1981)
I /
I / 10 Freon 22 (109 K) Frederick and Busing (1981)
~ 400 I ,,, / 5-10 Freon 22 (109 K); compressed air-driven entry Somlyo and Silcox (1979)
..
CJ

CJ 300 I
I
I
(0.6 m s- 1)
I 5-10 Propane (83 K), Freon 22 (113 K); gravity Elder et al. (1982)
I
I entry
200 I
I 10 Propane jet (83 K) Moor et al. (1976)
I
I <l Propane jet (83 K) Pscheid et al. ( 1981)
100 I
I 500 Hyperbaric LN 2 at 2.1 x 10 8 Pa(77 K) Moor et al. (1980)
I
I
80 90 100
30 40 50 60 70
0 10 Augmented and reproduced from Elder et al. (1982).
Distance from surface !pml

. . s from the surface of rapidly frozen cells. (A) Cold


. 2 34 Ice crystal size at different depth
F 1g. B 11. t 1976a.) (B) Rat kidney cells ex-
· · . (F D mpsey and u 1van . (C) p
block-frozen pea root tip cells. rom e . F om Frederik and Busmg 1981.) ro-
amined by different low temperature techn'.{u~~.8~ )r The significant feature of these curves is
pane-frozen rat liver cells. (From Elder et a . s th~ all show that small ice crystals .are only
that, despite being from very different s~~c:t,al. f~und that the peak.in ice crystal size co
found in the outer. l!l-20 µm of tissue. E ths (fine horizontal dotted !me) and it may well 2.3.2a Cryoprotectants
occur over a relatively wide range of dep ·mum if the tissue had been exammed
that curves A and B would also have sh:e:;e: l:,:~:. The routine method of overcoming the problem of damaging ice crystal for-
in specimens which have to be frozen in pieces larger than the critical
is the use of a cryoprotectant, with or without previous chemical fixa-
(see e.g. Nash 1966 and Skaer 1982 for reviews). Substances which al-
. er o rotectant often leads to other con w cells to be frozen with the formation of smaller ice crystals and with
specimen pretreatment with a yl p 'f 1 cerol is used as a cryoprotect damage than would otherwise occur are now generally referred to as
· · ts For examp e, 1 g Y d oprotectants. It has sometimes been sought to distinguish between cryo-
quential reqmremen . . ble to this compound an m
it will be found that many cells are ~mp;rme~ render them more permea tectants (additives used at low concentration where the objective is
be treated in some way (usually by ix~ ion)li~le further from the ideal sit e-thaw survival) and cryofixatives (substances used to assist in the pres-
Each of these steps move~ freeze-fixation ant However, for the majority on of ultrastructural and physiological features in the solid hydrated
tion of direct freezing w1t~~ut J?!~:::~~:e th~ reasons for using them m (e.g. Franks 1977). This is not a useful distinction because the same
specimens they are necess1t1es. . t be sou ht and their c unds may serve either function. Furthermore, to attribute 'fixation'
be understood, the optimisation of their ~se ~rn~orne in !ind (Table 2.1 rties to cryoprotectives is misleading as no formation of covalent
city for producing artefacts must be cons an y
Freezing 75
ds in biological electron microscopy
Low temperature me tho
74 Robards and Sleytr
it
(d) Ultrastructural changes in biological and artificial (model) membranes such as
TABLE 2.12
l' nd specimen pretreatment
f•. aggregation of membrane-intercalated particles (Plattner 1970, 1971; Mcintyre et
al. 1974; Sleytr 1974; Niedermeyer and Moor 1976; Groot and Leene 1979;
~~~~~P~o~s~si~b~le~ar~t~e~fa~c~t~s~a~s~a~c~o~n~s~e~q~u~en~c~e~o~f:sa:m:':p=1=n:g~a~~:__~~~~~~~~~~~~1 Arancia et al. 1979)
(e) Masking of fine-structural details due to phase separation phenomena (as visua-
(J)
Gravitation forces (centrifugation)
(a) Dislocation of cell components (organelles);
I lised by etching in most cryoprotected samples) (Staehelin and Bertaud 1971;
Miller 1979)
especially in eucaryotic cells . I ted particles (e.g. in bovine chro-
. ) f membrane-mterca a (j) Changes in frequency of exposure of potential fracture planes (e.g. cell envelopes
(b) Segregation (clustenng o . 1 ntact (Schuler et al. 1978) of Gram-negative bacteria) (Gilleland et al. 1973; Thornley and Sleytr 1974, 1978)
maffin granules) by forced phys1ca co brane structures (e.g. mesosomes in procar-
(g) Glycerol-induced membrane fusion (Chandler 1979)
(c) Generation of vesicular or tubular mem
yotic cells) (Higgins et al. 1976) (8) Chemical fixation with glutaraldehyde
· t ·mming or freezing . (a) Slow fixation artefacts seen as changes in distribution of membrane-intercalated
(2) Mechanical stress durzng n . . . of the tissue block with a razor
fi . 1 ll layers durmg tnmmmg particles;
(a) Damage of super 1C1a ce . of fracture planes of specimens frozen on
blade (consequences for the evaluat10n e.g. in systems which have a poor permeability for glutaraldehyde (fungal spores,
mycelium) or systems which require fixation by perfusion (Schmalbruch 1980)
cold metal surfaces) .h t l surface before freezing (cold block
. . d · impact wit a me a (b) Shrinkage of organelles (in plants) (Fineran 1970c, 1978)
(b) Specimen d1srupt10n unng
. d s·1 and Kachar 1980) · I I (c) Reduction in frequency of membrane-orientated fracture planes (Furcht and Scott
freezing) (Pmto a Iva . l d pray- or jet-freezing. Part1cu ar y
. . etting of matena an s f th 1974; Nermut and Ward 1974)
d
(c) Shear forces urmg pip d . th emulsification necessary or e
forces can be expected unng e (d) Changes in fracture behaviour of membrane constituents (e.g. changes in size 'mor-
strong sh ear . heim 1972) phology', frequency, and distribution 'partition coefficients' of membrane-interca-
liquid paraffin-freezing techmque (Buch
lated particles (Dempsey et al. 1973; Staehelin 1973; Nermut and Ward 1974;
Hydrostatic pressure ce of the high hydrostatic load Furch! and Scott 1975; Parish 1975; Bullivant 1978; Van Deurs and Luft 1979;
(3) I d ages as a consequen hi.
Lethal effect and structura am f . techniques (Moor and Hoec I Willison and Brown 1979; Satir and Satir 1979; Arancia et al. 1979; Kachar et al.
( > 2 x l os Pa) necessary for high pressure- reezmg
1980; Sleytr et al. 1981)
1970; Riehle and Hoechli 1973) (e) Generation of particle-free membrane blisters (Hay and Hasty 1979; Shelton and
Mowczko 1979)
Solute concentration . d. during trimming and mounting.
(4) d. ( s ens1on) me mm
Evaporation of the surroun mg su .P . th may provoke cell and/or organelle Chemical fixation with.osmium tetroxide
of pH or wmc streng s·1 1972'
Osmotic effects, ch anges . ted particles (Pinto da 1 va ' (a) Reduction in frequency of membrane-orientated fracture planes (generally to a
. membrane mterca 1a
shrinkage or aggregat10n or larger extent than with glutaraldehyde) (Nermut and Ward I 974)
Staehelin 1979; Raviola et al. 1980) (b) Destroys natural fracture planes of membranes (especially with membranes that
Changes.in the gas composition in the specimen environment contain significant quantities of unsaturated fatty acids) (Meyer and Winkelmann
(5) 1970; James and Branton 1971)
Anoxia (Raviola et al. 1980)
(c) Morphological changes of the nucleus of bacteria and generation of artificial mem-
I ges before freezing . h
(6)
Specimen temperature c wn . . f m the liquid crystallme to t e brane structures in bacteria (mesosomes) (Lickfeld 1968; Nanninga 1973; Fooke-
. f b ane lipids (trans1t10n ro . df
Phase separat10n o mem r . t lated particles during sampling an re Achterrath et al. 1974; Higgins et al. 1976; Ghosh and Nannings 1976)
. f mbrane-m erca (V k
state) and aggregat10n o .me. db. l ical or artificial (model) membranes er
ing as seen in native and isolate 10 .og d Plattner 1976· Maul 1979; Kachar et Chemical fixation with glutaraldehyde and infiltration with crypoprotectant (generally
and Ververgaert 1975, 1978; Zmgshe1m an ' a combination of (7) and (8))
(a) Modification in the organization and partition of membrane-intercalated particles
1980)
( ·thout chemicalfixation) . (Pinto da Silva and Miller 1975; Hasty and Hay 1977; Lefort-Tran et al. 1978)
(7) Infiltration with cryoprotectants wz . h 1968a b· Richter and Sleytr 1970; Fine (b) Masking of fine structural details due to phase separation phenomena (visualised
. 1 1 Jysis of cells (Ric ter ' '
(a) Irrevers1b e P asmo . illison and Brown 1979) . by etching)
1970c, 1978; Willison 1975, 1976, W rowth in glycerol-containing media (Fm
(b) Cell or tissue degenerat10n dunng g . . Treatment with non-penetrating cryoprotectants (e.g. polyvinylpyrrolidone, hydroxy-
ethylstarch)
1978) wellin of mitochondria, vesiculat!O
(c) Changes in the form of organelles (e.g.. ~oor 1~71; Zingsheim and Plattner l
. Shrinkage of tissue (caused by a decrease of cytoplasmic water; 'osmotic dehydration')
ER cisternae) (Moor and Hoechh 1970, XAJien and Weatherbee 1979; Robards et al. 1980)
Freezing 77
Robards and Sleytr Low temperature methods in biological electron microscopy
76
. that it should not be unacceptabl Y mJunous
. . . to the cell
bonds or any fixation in a conventional sense is involved. Other terms have srve artefacts. No alien substance 1s . comp1etely w"th 1 t s nor
f cause exces-
also been used (e.g. antifreeze agents) but cryoprotectant is probably the its very
. presence and cryoprotecti· ve ac iv1ty is bound t h actual
t. . . ou arte h
effects:
.
most useful and apt generic word. The purpose of such compounds is, liter- !og1cal and/or biochemical activity of the cell. All t o c ange t e _phys10-
ally, to protect the cells during the freezing process. The ways in which this tempt to understand what the cry t . hat can be done is to at-
opro ectant is doi d h
can be accomplished vary but, in general, the useful effects of cryoprotec- to interpret the results that arise from its use. ng an t us to be able
tants are to: lower the equilibrium freezing point (Tr); lower the temperature The range of cryoprotective substances is .t .
of homogeneous ice nucleation (Th) (increase the super-cooling capacity); few have been used for rapid fre . . lqm e wide, although relatively
raise the recrystallisation point (and hence reduce the critical cooling range 2.13). A number of r . ~zmg m u trastructural work (see Table
ivmg orgamsms have develo d
(§2.l.1; Fig. 2.4) and the critical cooling rate (vc)); lower the eutectic point; natural cryoprotective substances and th Id pe extremely efficient
remove 'unbound' water; and remove nuclei for crystallisation. For useful investigation to evaluate their us e m
. be.sel co~
10 og1cal well
stud·benefit
( from further
general reviews on cryoprotection, see Meryman (1971) and Skaer (1982). Edwards 1979· Kappen 19 . z h . ies1982
e.g. Baust and
' 79 , ac anassen and H b
Different cryoprotectants may have different combinations of these proper- 1982 (a bibliography)). Some auth ors (e.g. A nanthanarayanan
us Y ; Baust
and et al.
Hew
ties. In addition to its specific cryoprotection properties it is, of course,
necessary that the cryoprotectant should have other virtues and, in particu-
290

TABLE 2.13 270 ---


Properties of some cryoprotectants

Melting Boiling 250 T,-------- // /


Molecular /
point point
Substance and weight Th /
(°C (K)) /
formula (° C (K)) /
30 /
/
ID
6 /
/
20 (293)
" 210 /
Penetrating 92.11 ~ /
/
Glycerol "0.
E
/
HOCH 2CH(OH)CH 20H /
190
18.5(292)
"'
1- /
/
78.13 Tg
Dimethylsulphoxide
170
H 3COSCH 3
- 11.5(262) PVP
62.07
Ethylene glycol (1,2 ethane diol)
HOCH 2CH 20H

30 40 50 60 70 80
Non-penetrating % (w/w) PVP
44000-700000Ka
Polyvinylpyrrolidone e a 'Phase diagram' for po1yvmylpyrrolidone
. (PVP)/ .
lgla~s recrystallisa~
r ture of homogeneous nucleation- T .. water. Tm, meltmg temperature·
450000 rature. Solid lines are for PVP of •• trans1t10n temperature; T,,
Hydroxyethyl starch ,0,000. The lower limit for vitrifica~o ecu a~ weight (MW) 44,000; dashed lines are
68500 1cal wavy line. The molecular struct~~eonf tp~ curve for_ MW 44,000 is indicated r•.
Dextran m. (Redrawn from data provided . Ro p IS shown m the lower right corner
m asmussen and L t 1970
(C 6H100s), Rasmussen 1972 and F k uye ; MacKenzie and
ran setal.1977.)

"See Fig. 2.35.


Freezing 79
Robards and Sleytr Low temperature methods in biological electron microscopy
78

Franks and Skaer (Franks and Skaer 19~~~- n a_ senes of papers by Echlin,
cryoprotection of erythrocyte ultrastruct I .
1977) have attempted to emulate the effects of such natural cryoprotectants
by synthesising polypeptides. If a distinction is to be made between different 1977; Skaer et al. 1977 .F ' Echlm et al. 1977; Franks et al.
r . ' 1978 ' 1979 , ranks 1980· Skaer 1982) th ·
classes of cryoprotectant, then it is probably simplest to divide them into 01 these high molecular weight h d h. . ' , e vlftues
two groups: those which exert their effect by working within the cell (pene- method by which these cryopro;ectyanrtop ihcpho~yme~s were explored. The
s exert t elf action · ·
trating cryoprotectants such as glycerol (Fig. 2.10 in §2.1.2) and dimethyl tely understood
. . although it is th oug t t at retardation 0 f t mcomple-
h h . remams 11 .
sulphoxide); and those which operate without crossing the cell membrane nucleat10n is one important aspe t 0 . . ex race ular ice
(non-penetrating cryoprotectants such as polyvinylpyrrolidone (PVP) (Fig. need to be used at relatively high c . nee ag_am, it has been found that they
2.35) and hydroxyethyl starch). The different modes of action of these two nounced osmotic effects can conc(eFntration ( > 20%) at which quite pro-
occur ranks 1982· Wil d R
classes of cryoprotectant have been discussed by McGann ( 1978). 1982). However, their usefulness in so me s1tuat10ns
. . '.is beyond
son an t. obards (
Glycerol has been by far the most frequently used cryoprotectant for papers from Echlin ' Franks and Sk aer a b ove and also Schill ques t 1 1 see
10n
ultrastructural purposes. Its properties and cryoprotective capabilities have 1979) and, in particular the fact that th d er e a . 978,
been well studied (Luyet et al. 1958; Luyet and Kroener 1966) and, although brane (but see Barnard, 1980 h d e~b o not penetrate the cell mem-
, w o escn es endocytotic t k )
an increasingly long list of glycerol-induced artefacts is accumulating (Table considerable advantages if matena . 1 is
. to be subject . d t up a· e confers
2.12), for many purposes it remains the cryoprotectant of choice. Although Furthermore, these polymers pr ovi.de a very satisfactor . e o microanalysis.
· .
a 'penetrating' compound, the permeability of cell membranes to glycerol for cryosectioning (see , for exampl e, p·h i a k as k.i and Seveus "'f supportmg
1980). matnx
varies and prior fixation is sometimes required if the molecule is to get into
cells (e.g. Richter l 968a). As with most cryoprotectants, glycerol must be 2.3.3 Practical methods of sampling and pretreatment
used in relatively high concentrations ( > 20%) to provide good cryoprotec-
tion. This means that, for example, osmotic effects must be taken into ac-
dures suitable for suspensions of isolated c ll o ~stmgmsh between proce-
From a practical point of view it is convenient t d. . .
count when preparing processing solutions.
Dimethyl sulphoxide (DMSO) has also been widely studied (Rasmussen and bulk specimens such as animal and 1 e ts ~r ISolated cell components,
and MacKenzie 1968) and is well established as a cryoprotectant during the schedules is very wide indeed a d .t . . pan ~issues. The range of fixation
cryopreservation of cells in a viable state. However, it can have toxic effects all methods that would b n i_ is im~oss1ble, in this book, to refer to
e appropnate pnor to cry fi t.
on cell metabolism (Mathes and Hackenseller 1981 ), damages cell mem- referred to Glauert (1974) f o ixa ion. The reader is
branes and has been little used in ultrastructural work in comparison to gly- ~•• ods of fixation. or a more complete d iscussion
· · o f practical meth-

cerol. A useful study of its general toxicology was made by Wilson et al.
(1965). 23.3a Pretreatment of isolated cells an.d cell components
Other penetrating cryoprotectants have also been used, for example ethy"
lene glycol (Richter 1968a,b) and ethanol (Richter 1968a,b; Schiller an may take isolated cells to be those cell . .
Taugner 1980). However, for the majority of specimens for which a pen ave been cultured in this state. For s which are naturally free living
aryotic cells and so . the present purposes, these include
trating cryoprotectant is required, glycerol is the best substance to cho . me eucaryotlc cells Viru
usmg similar methods Su h . : ses can generally be dealt
The non-penetrating compounds, such as polyvinylpyrrolido~e (PVP
initially. :0 µl) in suspension or. as ~ pmellattena! his best frozen as a small sample
. ' e , wit out cryopr t t. ·
(Luyet and Rasmussen 1967; Luyet and Sager 1967; Rapatz and Luyet 19 freezmg methods (§2 4) D o ec wn usmg stan-
MacKenzie and Rasmussen 1972), hydroxyethyl starch (HES) (Allen a ystal damage occurs . "th. ue to the low average water content, little
Weatherbee 1979; Korber and Scheiwe 1977) and dextran,had been wells can cause distortio':s t~nc:~st procaryotic c_ells but extracellular ice
died in their own right in relation to cryopreservation of cell viability bef and disrupt highly h d d s. ~xtracellular ice crystals can also dis-
h 1972; Plattner et aly lr9a;;) shme, capsular material (Remsen and
their use in ultrastructural studies was first investigated. Allen and Weath . ' or cell appendages such as flagella or
bee (1980) compared the relative usefulness of HES and glycerol for t
Freezing 81
Robards and Sleytr Low temperature methods in biological electron microscopy
80

fimbriae (Sleytr 1978). These structures are often then tr~pped in the eutec- .. Depending on the incubation time, the temperature and the concentration
tic network surrounding the ice crystals. The average size of extracellular f of the glycerol solution, different rehydration or even germination states
ice crystals can be reduced by freezing dense (close-packed) cell pellets, or i may be attained (Remsen et al. 1968; Stocks and Hess 1970).
Most studies on fungal mycelia have been made using chemically fixed
by using a cryoprotectant. . I
Procaryotic cells can be prepared for freezing by followmg one of m~ny
and glycerinated specimens (Takeo et al. 1973; Takeo and Nishiura 1974).
. t.·
different protocols. The cells are centrifuged down from the growth medmm .c
Fix with buffered 2-4% glutaraldehyde at room temperature for a few hours
or taken directly from an agar surface. If a cryoprotectant is .requi:ed, then or at 4°C overnight. Vegetative fungal mycelia are best cut out and separ-
the cells are resuspended in 20-30% buffered glycerol solut10n w~th a pH ated from the agar surface in suitably sized pieces ( < 1.0 mm 3) for direct
and a final ionic strength, equivalent to that of the growth environment mounting and freezing. For mounting prior to freezing, it is necessary to
(Lickfeld 1973). Infiltration with glycerol is performed a~ room :em~era- wet the pieces to provide enough adhesion to the specimen support. This
ture at the growth temperature, or at 4°C. Recommended mfiltrat10n times is done by flooding the colony with water or growth medium, or by sus-
var; from a few minutes up to a few hours and the correct time can often pending the trimmed samples. Air inclusions are removed by briefly expos-
be judged by observing the complete reversal ofplasmolysis tha~ can be ~een ing the submerged mycelium to reduced atmospheric pressure in a desicca-
initially. In some instances it is even possible to grow cells on or ma medmm tor. Fungi grown as agitated submerged cultures generally maintain a pellet-
shaped colony structure during long incubation periods. Pellets of suitable
containing 15-20% glycerol (Moor 1970). . .
Dormant and resting spores offungi mostly have thick, impen_neable w~lls size ( < 1.0 mm 3) can be mounted and then frozen, or frozen directly, while
and, consequently, severe ultrastructural changes can occur.during chemi~al larger pellets can be trimmed down with a razor blade under a drop of
fixation procedures. These can be avoided, whenever possible, by selectmg growth medium on a dental wax plate.
the best method of fixation. The protoplasts of dormant spores ha_ve a low Algal cells and protozoa usually have too high a water content to allow
water content and are therefore best frozen without cryoprotect10n as a small ice crystals to be produced without cryoprotective measures when nor-
dense pellet (Sleytr et al. 1969; Van Gool et al. 1970; Stocks and He~s 19'.0; mal freezing methods are used, although spray-freezing and jet-freezing
Allen et al. 1971; Hess et al. 1972). To harvest spores, rinse the colomes w.ith have been applied successfully to a number of algal and protozoan cells
growth medium, water or, if previous experience has shown cryopr~tect10n (Plattner et al. 1972, 1973; Lefort-Tran et al. 1978; Schuler et al. 1978). To
to be necessary, with a 20-30% solution of glycerol and then spm .them minimise the risk of misinterpretation from artefact formation such direct
down. Spores are sometimes difficult to 'wet' and ~his may make sedimen- freezing methods should be tried before using pretreatments in~olving che-
tation difficult. This can often be overcome by addmg a few drops of deter- mical fixation and cryoprotection). If no fast freezing methods are available
gent (Tween, Teepo[) to the suspension before centrifugation (Sassen et al. or applicable, methods for pretreatment of algal cells and protozoa are simi-
lar to those already described for fungal material. Grow cells directly, or
1967). .
If spores are to be kept from rehydrating during harves~mg and pretre~t- incubate them, in a cryoprotectant solution of 10-30% glycerol. Alternati-
ment, they can be mixed as a dry powder with paraffin 011 and froz~n dir-< ely, fix the cells with glutaraldehyde (1-4%) prior to glycerination (Neus-
ectly. The paraffin oil technique can also be applied to other specime~~ 1 1970; Schwelitz et al. 1970; Werz and Keller 1970; Easterbrook 1971;
such as spores of mosses (Schultz and Lehmann 1974) and pollen grai erban et al. 1973; Bray et al. 1974; Pueschel 1977; Fineran 1978).
As mentioned earlier, the incubation time is a direct function of the diffu-
(Southworth and Branton 1971), which need to be kept dehydrated.
Although cryoprotection may not be strictly necessary, many work~ constants for glycerol through various compartments and boundaries
have nevertheless suspended spores in a solution of glycerol b~fore freezi can vary considerably from one cell type to another, especially in
(Stocks and Hess 1970; Brouchant and Demoulin 1971; Richmond a ed plant material. The rate of diffusion must be determined experimen-
· 197la b· Cole 1973 1975). Among other benefits, this leads to mu by studying cell morphology, by observing cell shrinkage or plasmoly-
P nng ' ' ' · · 15-3 ffects and by analysing the intracellular ice crystal size in relation to
smaller ice crystals in the extracellular solut10n. Imp~egnate with a 4-200
ration time. If the reaction and permeability of the specimen to glycerol
(v/v) aqueous solution of glycerol for between 30 mm and 24 hat
Freezing 83
82 Robards and Sleytr Low temperature methods in biological electron microscopy

with glutaraldehyde before infiltration with 1 .


is not known, it is safer to start with an appropriately chosen glutaraldehyde dures similar to those for the 11 d . g ycerol. Use fixat10n proce-
. ce s an trssues from wh"ch 1 th f .
fixation (see Glauert 1974) before glycerol infiltration. For example, using 1ved (Glauert 1974) or make up th e fiixatrve . m . the isolaf
. e ract10n
d. is der-
Micrasterias denticulata Staehelin and Kiermayer (1970) first fixed cells for cussed
. before, fixation and cryo wn
pro ec wn procedures sh ld 1 b dis-
t t. me mm. As
10 min at room temperature in 1% phosphate buffered glutaraldehyde and 1f, by comparison with ultra -1;.-ast freezmg
. method "t bou on y e used
then infiltrated them for 2 h with 30% glycerol in a 3% glutaraldehyde/phos- do not produce unacceptable artefacts (Zin s _s, 1 can e shown that they
phate buffer solution. Fixation with glutaraldehyde can be performed in the seldom realised that with g heim and Plattner 1976). It is
' non-cryoprotected sa 1 f
growth medium, a buffered solution, or in sea water (as described for the organelles, the ice crystal size 0 ft en d oes not mp es othemembranes
affect fi or cell
red alga Palmaria palmata; Pueshel 1977). After fixation, specimens are membrane fracture faces and etch ed sunaces .. If th me
t · 1· structure of
either transferred directly into a buffered 20-30% glycerol solution (Marti- in very dilute buffer or distilled wa t er, no solute/solvent
· e ma ena h is suspended.
nez-Paloma et al. 1976), or slowly infiltrated with 25-30% glycerol over a can occur
.
and
.
even the formation 0 f 1arge ice . crystals is n Pt ase separat10n
.1 .
period varying from a few hours up to two days (Wunderlich and Speth turbmg (Fmeran 1970a·' South wor th et a.1 1975; Willison 1975, 1976). y dis-
o necessan
1972; Sattler and Staehelin 1974; Kitajima and Thompson 1977).
With other isolated eucaryotic cells, such as tissue or monolayer cultures, 2.3.3b Pretreatment ofplant material
spermatozoa and blood cells, the same problems as those already discussed
concerning fixation and cryoprotection become apparent if no fast freezing Untreated, living plant material can onl be su .
methods can be used directly (Plattner 1971; Plattner et al. 1973) and a var- small ice crystals if the water t t .y 1 ccessfully frozen, with very
' con en is ow enough Th.
iety of pretreatment procedures has been reported in the literature (Koehler a few specimens such as dormant S · is only occurs in
1970; Kouri 1971; Reith and Oftebro 1971; Pfenninger and Bunge 1974; in cases of natu;al cold-harden1· sAporehs ( chul~z et al. 1973) or seeds and
ng. s s own with pi dl
Friend and Fawcett 1974; Flechon 1974). Again, diffusibility of glycerol and Sterling 1972) ' excellent cell u 1ar preservat10n
. is. obta·
ne nee des (Leonhard
h .
into unfixed cells can vary considerably among similar cell types. For exam- cold-hardened tissue is frozen . H owever, t h ere are consid me bl wr en. natrve
ple, sheep red blood cells require several hours for proper infiltration where- to examining most non-cryoprotected r . . era e 1m1tations
as those of rat or humans reach an equilibrium state after a few minutes In the absence of natural cryo t ~:mg ~l~nt trss~es (Fineran 1978).
(Koehler 1970, 1972). For isolated cells which are sensitive to osmotic ef- living tissue with glycerol or t pro ec IO~, it is possible to try to infiltrate
fects, 'complex' antifreeze media, containing hydrophilic substances with second method has been ~- grlo~ specimens in glycerol solutions. This
place Allium ce ap:~ icu a~ y us~ful for studying root tips. For
low osmotic activity, have been used successfully (Wecke 1968). . 'P lbs with theu bases in contact with 20% 1
If specimens are cooled down before chemical fixation, then it is possible m water and grow them in the dark at 25oc D o g y-
for phase separation to occur in the membrane lipids with concomitant de- 1.0 cm roots cut out 200 th. k or 4 days. From 0.5-
velopment of areas devoid of intramembraneous particles as seen by freeze- freezing (Moo~ 1964· B t µm d1c , 1 mm long longitudinal sections
, ran on an Moor 1964) s· ·1
fracture techniques (Tsien and Higgins 1974; Verkleij et al. 1975; Bayer et s of specimens grown in t . · 1m1 ar studies on root
con act with glycerol ha b
al. 1977). Thus if, for experimental purposes, pre-cooled samples are to be rth cote and Lewis (1968) and H ve een reported by
frozen, then structural comparisons with samples quenched from the norc d germinated seeds of pea ~eward and Northcote (1973). They
mal growth temperature are desirable to avoid possible misinterpretation o h 25% glycerol. grown or 1-10 days on cotton wool soaked

ineran (1978) examined roots of A . . .


membrane structure. v~na sat1va,
Isolated cell components can generally be frozen as a dense pellet or in sus
pension. With fast freezing methods such as 'sandwich' or spray-freezi
s. Seeds are first germinat d
ges from the coleorhiza et o: Trztzcum vulgare and Zea
:01stened filter paper until the radicle
e_gnated with 20% 1 , a wl ic st~ge t~ey are placed on filter paper
no cryoprotection of the specimen is necessary. Used with normal freezi o g ycero . Aerat10n is im t t b
methods, the ice crystal size can be minimised by adding between 5 and 30 rsed in glycerol often d por an , ecause roots
glycerol to the suspending solution. Osmotically sensitive cell organell aced on top of narrow t e~ener;:j To achieve optimal aeration, seeds
u es o I ter paper soaked in glycerol solution
such as chloroplasts, vacuoles or mitochondria, are also often best prefix
iii

thods in biological electron microscopy Freezing 85


84 Robards and Sleytr Low temperature me

.
bated in 20% glycerol for 5-7 days, advisable. This will help to distinguish the cryoprotectant-induced changes
(Fineran 1978). Sprouted seeds are mc~f bout a further 3 days (Fineran in cell structure. Structural studies on botanical material have also been
or for 3 days followed by 25% glycero . or a tact with glycerol for 7 days reported where the formation of intracellular ice has not limited observa-
1970b, 1972, 1~75). ~o.ot tips ~;::no~~:e~~ organelles, as compared ~ith tions. Examples of specimens which have been frozen in their native state
show an essentially similar stru f t re replication or thin section- are: phloem (Johnson 1973, 1978), cellulose fibrils (Cox and Juniper 1973),
h amined by freeze- rac u . 1
control roots, w en ex . d t hold for all plant matena tracheids (Puritch and Johnson 1973), water-conducting cells of moss
. n 1970c) This oes no d
ing metho d s (F mera . 1 . Glycerol is osmotically active an (Hebant and Johnson 1976), roots (Robards et al. 1981) and germinated
. b t d in glycerol so utions. h f and ungerminated cereal grains (Buttrose 1971, 1973; Swift and Buttrose
grown or mcu a e . . ll unbalanced solutions t ey o -
when unfixed tissues are placed m osmotica y 1972, 1973; Barlow et al. 1973).
ten plasmolyse. . cies dependent and tissues of For tissues which cannot tolerate glycerol, or where unacceptably coarse
1
The ability to metabolise glycero l is sp~ . -when exposed to glycerol. structural alterations occur, mild chemical fixation prior to infiltration with
higher plants can suffer irreversi~le p a~r:~~:~~ organelle morphology and glycerol is unavoidable. By comparison with other aldehydes tested, glutar-
Even when cells deplasmolyse, senous c ad H hli· 1970· Moor 1971). Con- aldehyde has been shown to be the most suitable fixative for both plant and
r (Moor an oec '
membrane structure can occu. . entrations of glycerol can pre- animal material (Moor 1966). It increases the permeability of the plasma-
. fil f with mcreasmg cone .
trolled, slow mi tra ion (R. ht a b) A simple electromc lemma and stabilises cell structure. A detailed description of the problems
"bl 1 olytic effects ic er 1968 , . involved in processing botanical material for freeze-fracture replication pro-
vent irreversi e p asm . f 1 1 has been described by Rotten-
1 admixture o g ycero h
apparatus for t h e s ow . d. nt makers would serve t e cedures is given by Fineran (1978), who also discusses other structural
. h (l 969) but simple density gra ie . changes introduced by glutaraldehyde fixation (such as the shrinkage of cell
burg and Ric ter . f fi d pieces of tissue with cryo-
ful infiltration o un ixe 5)
same purpose. success f . (Hall 1967· Pyliotis et al. 197 ' organelles). For some membranes, it has been shown that the general mor-
orted for lea tissue '
protectant has b een rep d 1970· Hereward and Northcote phology is better preserved in glutaraldehyde-fixed tissue than in material
d (Chafe and War rop ' 1 grown or incubated in glycerol (Fineran 1978).
collenchyma stran s . 1972) The method of choice strong y
1973) and lichen (Elhs and Brown · tly many different methods For fixation, glutaraldehyde is generally used in concentrations of0.2-6%
h f tissue and consequen , d
depends on t e type o . , otectants concentrations an in buffered solutions of appropriate ionic strength and osmolarity for
of retreatment involving different cry?pr , several hours. Alternatively, the fixative can be made up in glycerol sol-
P . h been used (Fmeran 1978). .
infiltration procedures ave . h 1 sul hoxide (DMSO) is less smtable as ution, which allows simultaneous fixation and infiltration (Fineran 1978).
Compared with glycerol, dimet y
.
?
t for living plant tissues or
isolated cells However, it
.
Fineran (1978) found no marked differences in the quality of preservation
a cryoprotective agen ful comparative cryoprotec- root cells between material fixed before glycerination and material
b ·t may serve as a use
is mentioned here ecause i fi d .t recommended in some papers. that had been fixed and glycerinated simultaneously. For root tips, a fixa-
b e the reader may m i
taut and, a1so, ecaus . ded to give adequate cryoprotec" }lon procedure of 3-6% glutaraldehyde in 0.025 M phosphate buffer for 1.5-
It is often :oxic at the ~oncentrations ~:i~ artefacts (Pyliotis et al. 1975): h was used. Glutaraldehyde fixation prior to infiltration with glycerol
tion and it can also mtroduce spe_c t it is necessary to decide wheth been used for a variety of botanical materials such as seeds (Barlow et
Depending on the ~atur~ of the ~x~~~:~~al effects induced by glycerol 1973; Swift and Buttrose 1973), thick-walled tissue (Robards and Parish
the potential physiological an d . neral that glycerol is fir 1), differentiating shoot cambium (Parish 1974), isolated protoplasts
DMSO can be tolerated. We recommen , tm ge fsp~cimen is under stud .illison and Cocking 1975), infected tissue (Aist 1974), and tissue cultures
t taut whenever a new ype 0
used as the cryopro ec . r ation of non-penetrating cryopr und and Smith 1974). Glutaraldehyde can be used successfully in con-
With the possible exception of th~ adpp icf t hardiness of tissue in the li. tions as low as 0.2% (Sjolund and Smith 1974). Fixed specimens are
. other way to m uce ros - .
tectants, t h ere is no . Thus if cryoprotected living tiss ly infiltrated with 20-25% glycerol but in some procedures up to 50%
by cryoprotection. ' ·
ing state oth er th an . f h samples with specime ol has been used (Steere 1971).
. d tructural companson o sue 1
is to be examme , a s . l ld hyde before infiltration with glycero. practice, the lowest concentration of glycerol capable of providing ade-
that have been fixed with g utara e
- 86 Robards and Sleytr Low temperature methods in biological electron microscopy
Freezing 87

cubes and preferably much less) can be fix . .


quate cryoprotection should be used. To avoid the build-up of concentra- been obtained (e.g by biops ) b . . ed immediately after they have
. y y rmmersmg them i 2 301 b
tion gradients around specimens and to increase the rate of penetration until aldehyde (Moor 1973· South th n - 10 uffered glutar-
' wor et a1 1975· St 0 r k.
equilibrium is reached, samples should be agitated in the medium (e.g. by Brown 1979) or in Karnovsk , ,. · ' ms I 1977; Willison and
' Y s para1ormaldeh d / 1
using a rotating tumbler). No general rule can be given about the optimal (Karnovsky 1965· Yee and Re ). Y .e g utaraldehyde fixative
· ' ve1 1978 . Alternat 1 · · ·
size of a sample for fixation and infiltration since this will depend strongly the tissue in vivo by perfus· (F ive y, rt rs possible to fix
wn orssmann et 1 1975
on the kind of tissue and the minimal size to which it can be trimmed with- Robards
. et al. 1981). This sec ond meth od has the ad a· t ; Bullock
h 1979;
out mechanical damage. The sample should not normally exceed a thickness tissues, better structural preser t. . . . van age t at, for many
. . va 10n is obtamed f fi f .
arnmal is still alive. Prelimina t t I rxa wn starts whrle the
of more than 2-3 mm at least in one direction and should preferably be ry rea ments such as r
much less than this. Similarly, although it is not possible to provide a pre- dehyde fixation are of great ·m t . ' samp mg and glutaral-
' 1 por ance m determ · · h
treatment schedule that will be appropriate for all specimens, the informa- ture is faithfully preserved . F ur th ermore the corre mmg t w . ether fine struc-
d.
tion provided above will allow the beginner to make rational choices in the strength and osmolarity ' as d et ermmed .' for cth.con rt10ns · . of pH, ionic
(Glauert 1974), must be established and used. m-sect10nmg methods
selection of all relevant parameters.
Transfer fixed samples into a buffered 20-300 .
for between 1.0 and 5 oh
.
filt . . % (v /v) solut10n of glycerol
. . 1n I ration rs generally be t d
2.3.3c Pretreatment of animal tissue o .
quent agitation to reduce the t.1me necessary to rea h s one Tb at
. 4 C with fre-
Since, with few exceptions, such as the stratum corneum of skin, the water vely, the material can be transferred ft fi . c. eqm I num. Alternati-
content of animal tissues is high and even the best cooling methods only solution of glycerol for 10-20 h 'ad er rxat10n, mto a buffered 10% (v/v)
· · an then for th ·
produce sufficiently small ice crystals in the peripheral 10-15 µm layer, cryo- a 25% glycerol solution. There ha t b ' e same time again, into
protective pretreatments are again essential. As for many plant specimens, lished to show whether such st . s ~ofil ee~ much systematic work pub-
. epw1se m I trat10n of ch . ll fi
most animal tissues must be fixed before infiltration with glycerol to avoid al improves structural preservat·wn b u t we recomm dem1ca ·t y ' rxed materi-
necrosis or other structural changes induced by the cryoprotective agent. cedure which is less likely to le a d t o artefacts. en I as a gentler' pro-
Even now that chemical fixation before infiltration with cryoprotective If artefacts associated with chemical fi .
agents such as glycerol or DMSO has become routine for animal tissue, such structural comparisons bet fi ixat10n are to be assessed, then
ween un rxed and fi d/
procedures must be seen as a compromise. Strictly speaking, they can only can be helpful When tiss h rxe cryoprotected samples
a 1 ~se y chemical fixation,
. . . ues ave not been st bT d b .
be considered 'safe', in terms of artefact formation if it is possible to obtain i.t ls preferable to infiltrate th .h
em wit gradually mer ·
a comparison in the form of a sample that has received minimal chemical o.f cryoprotectant This . 11 . easmg concentrations
~ap1 _Y .as possible to
. . . . rs usua y earned out as "dl .
treatment. Even using methods that give good freezing in a peripheral zone, mm1m1se cryoprotectant-induced art f . .
it is necessary to be aware that this well-frozen layer itself may have been e.•ial
. care must be given t .d. e acts. If hvmg tissue rs mfiltrated, spe-
o prov1 mg the proper ·
damaged during the excision and/or trimming processes. Thus, as long a For example Stolinski (1977) ,. d . environmental conditions.
. ioun it useful to b ff h
there are no methods available to give good freezing throughout pieces o th a tissue culture solution to s t . . u ~r t e cryoprotectant
e sample was frozen E . llus,. am tissue metabohsm and viability until
untreated tissue, the use of chemical fixation in combination with cryopro . spec1a y ior unfixed t . 1
provided for the opt1·m . fil . . ma ena ' no general rule can
tective agents has to be the compromise of choice. um m l trat10n time B th h
As with plant material, brief fixation with formaldehyde or glutarald ltration time and the 1 f" . . o t e shortest practicable
owest e iective cone t f f
hyde preserves the structure, renders the tissue permeable to glycerol ( be established from stud in .en ra 1011 o cryoprotectant
ski (1977) found ,. y gl morphological preservation. In practice
other cryoprotectants) and allows infiltration without obvious shrinka ' ior examp e that sm 11 b
ation periods of 1.0 h in 100 'v a cu es of rodent liver needed
and plasmolysis. glycerol i . %( /v) glycerol, followed by 1.0 h in 2501
If fixation is essential, it should be carried out as gently and as mildly n a tissue culture solution To . . . . lo
possible (Glauert 1974; Hayat 1981). Two main types of fixation proced ge by osmotic shock infilt f . mm1m1se excessive tissue
gradient. ' I ra wn can be also performed using a conti-
are possible. Firstly, small pieces of tissue (ideally not larger than 3.0
Freezing 89
Robards and Sleytr Low temperature methods in biological electron microscopy
88
Solenoid
Glycerol is generally superior to DMSO for infiltration ofliving tissue be- trigger

cause it is less toxic at higher concentrations. Since even continuous infilt-


ration of fresh animal tissue with the relatively non-toxic glycerol can result l ____::r==
in gross structural alterations (Moor 1971; Echlin et al. 1977; Skaer et al. Explosion chamber
Coll.action chamber containing
1977), it is usually 'safer' to fix the material chemically before cryoprotec- tissue fragments in liquid
propane at -1aa•c {85Kl
tant treatment. This will particularly be the method of choice if no compara-
tive studies on fixed and unfixed cryoprotected samples are to be carried
Switch t
out.
As well as penetrating cryoprotectants, non-penetrating compounds such t
as polyvinylpyrrolidone (PVP) and hydroxyethyl starch (HES) have been
applied to animal tissues (Franks 1977; Schiller et al. 1978, 1979), although
they can cause considerable cell shrinkage and consequent artefacts (Mazur Delay circuit Stim11lator

1970; Meryman 1971). They do, however, seem to have an important role Fig. 2.36. Schematic diagram of the ballistic sam !in t h .
to collect and freeze cardiac tissue The st· . pd g ec mque used by Monroe et al. (1968)
to play in the range of cryoprotectants available. They are used at concen- · 1mu1us-m ucmg cont f · h ·
be variably
. delayed to trigger the explosi· on, so d nvmg . rac wn
. . t h e bwpsy sa m rt e isolated
. .heart
. may
trations of between 20 and 30% made up in water or growth medium. As ventncle at any time during the contracf I 1 . mp mg projectile mto the
these solutions are osmotically active and as specimens to be treated with and entering the propane the needle gets!~~ cydc ef. hn the mterval between leaving the heart
· . ea o t e tissue which thus c I . 'di ·
non-penetrating cryoprotectants are usually unfixed, exposure of the speci- hqmd propane at -188oC (85 K) · (R ed rawn f rom Monroe et al.oo s rap1 y m the
1968.)
men to the cryoprotectant should be kept to a mimimum consistent with
ensuring that the whole specimen surface is thoroughly coated with the cryo- Cryochamber
protectant. At the concentrations used, the solutions are rather viscous and,
therefore, some form of mechanical 'stirring', or other agitation, may be
necessary.

2.3.3d Ballistic methods of sampling and freezing


Valve assembly Piston

From much of what has gone before, it will be clear that not only should
the cooling rate be maximised to ensure the 'best' freezing, but the method Fig. 2.37. A combined rapid sam r df · . t
Mergner. This design supersedes :p~:~i~~s ;:~~:~d~:~~~~:;~gn~d by S.H. Chang and W.G.
of obtaining the specimen prior to freezing will also have a strong bearing component parts are illustrated Th , ,. Y hang et al. 1980). The four
on the final quality of the preparation. The time interval between sampling •·. .through flexible metal tubing The tr:1 e crtyogun IS connected to a supply ofliquid nitrogen
'b · gger ac 1vates propulsio f th di ·
and freezing should be as short as possible. One means of achieving this for '.; e sampled· release of the tri· ° n e nee e mto the organ to
samples from bulk material is the use of 'ballistic' methods of sampling: that
'< d
.
iate
'
withdrawal. A supply
gger reverses the movem t f th ·
of cold nitro en en o . e piston and brings about imme-
zchamber so also cooling the needle a d g t ga.s cools the msulated copper walls of the cryo-
. h n pro ectmg the spec1me f ·
is, to 'shoot' a sampler into or through the tissue to be examined. Where s ig t positive pressure of th
1 . e gas prevents the entr 0 f
n rom environmental changes·
· · · '
e is mounted on a PTFE · t d . . Y m01st a!f mto the chamber. The
such methods have been employed, they have usually involved the rapidly I pis on an 1s designed t f · ·
. Cutting is aided by rotation of th di o op imise cuttmg and removal of the
sampled specimen being projected straight into the coolant, so that sam- lsion mechanism is powered b e nee eat 1200 rpm as it is projected forward. The
pling and freezing occur within an extremely short time interval and as par erature is monitored at the tip ~fd:;;'e n~1rogen gas at 60-WO psi (4.13-6.89 x 10s Pa). The
ber and within the cryochamber its If 0 opsy needle,. withm the copper mass of the cryo-
of the same process. As an extreme of such methods may be instanced t a special tool and is transferred te ~h n~ the specimen has been obtained, it is removed
hypodermic needle fired explosively by a modified 0.22 rifle that was us oned. (Iilustration red f o e c uck. of a cryoultramicrotome where it can be
M rawn rom an ongmal kmdly provided by S.H. Chang and W G
by Monroe et al. (1968) to acquire and freeze heart tissue (Fig. 2.3 ergner and reproduced with their permission). . .
Clearly, there are potential problems with such methods, not least those o
Freezing 91
ods in biological electron microscopy
Robards and Sleytr Low tempera t ure meth
90
(100 m s- 1) and then rapidly cooled to -135°C (138 K) in a cold halocar-
bon (Fig. 2.38). While the cooling rate of such a device is low, it does have
the advantage of sampling very quickly and of allowing minimal time
between excision and freezing.

2.4 Freezing methods

A summary of rapid freezing methods is provided in Table 2.14 and a 'flow-


diagram' for alternative preparation and freezing pathways is shown in Fig.
2.39. T here are many publications reviewing rapid freezing methods , among
which may be cited: Sitte (1979), Costello (1980), Elder et al. (1982), Escaig
(1982), Plattner and Bachmann (1982), Plattner (1984) and Sitte (1984).

2.4. 1 Freezing with liquid cryogens

The principles of freezing with liquid cryogens should now be clear but the
selection of the most appropriate practical method and its proper applica-
.~ r 'cryogun ' d escn.b ed by Hearse et al · (1981) to obta in. and
.
freeze
tion need to be carefully considered. Firstly, it must be stressed again that
Fig. 2.38. Multiple samp mg 'd db DJ Hearse and reproduced with perm1ss1on.)
heart tissue. (lllustrat1on prov1 e y .. rapid cooling rates(> 1x104 K s- 1) can only be obtained throughout extre-
mely small specimens ( « 100 µmin one dimension) or within the surface 10-
. k to the cells and local frictional heating caused
percuss10n (pressure) shoe d H it has been shown to be pos- 20 µm of larger ones. Consequently, if large ( > 0.1 mm in all dimensions)
f the sample ho\ er. owever, . . specimens are to be used, there is really very little point in applying ultra-
by the passage o II . f tissue and rapidly freeze it with m1mmal
sible to extract a very sma piece o rapid cooling methods because only the surface layer will cool quickly (as
damage. , , (F 2 37) which they was demonstrated in Fig. 2.34 and Table 2.11 in §2.3.2). Such material will
Chang et al. (1980) have described a cryogu~ ig. . , . . This need to be cryoprotected and cryogens (such as liquid nitrogen at its equili-
have used for cryofixation of renal spedc_imens pno~ ~ s~i~s:~o~~~:· from brium boiling point) that lead to severe film-boiling (§2.2.1) should be
, , 1 of 0 2-1 o mm iameter an · · avoided. Liquid baths of sub-cooled nitrogen or halocarbons 12 or 22
gun ~an cut samp es\ d The ~opper needle-type specimen holder, cooled
the tissue to be samp e . . '. t d towards the specimen; the sam- should prove quite satisfactory. In addition, attempts should be made to en-
to liquid nitrogen temperature, is proJebc e t t w1.th the cold surface and sure that the cooling is as consistent and reproducible as possible. This
d . d. tely frozen y con ac means that conditions should be recorded and standardised from run to
pie is cut out an _1mme ia d bod of the gun by a secondary spring.
is then retracted mto the coole _Y d cryofixation of tissue speci- run. One way of helping to achieve this is to use a cooling rate meter
. d . the authors report improve (Appendix 2), or some other device for monitoring cooling rates . This al-
Using t h is ev1ce, . I t' I data There remains scope
mens with improved reli~bility_ of m1c~o;n~ y I~~ mini~ise the possibility of lows cooling conditions to be optimised for a specific method and also pro-
for further experimentation with su~ ev1ces vides the possibility of checking that optimum rates are maintained from
post-morte~ ~rt(~~~clt; ~;;;o~e~r:e:;~~r different approach in constructing run to run .
As the next few sections demonstrate, liquid cryogens can be used in a
Hearse e a . . each 2 0 mm x 2.0 mm m sec-
a brass or stainless steel matnx of 70 cutters, d 34 . Jong that was first variety of ways. However, there are some basic principles that apply for all
f 4o x 4 0 mm) an mm ' applications. For example, the heat transfer characteristics of the cryogen
tion (or 40 cutters o . mm .' b led (heart) at 1os mm s- 1
explosively projected against the tissue to e samp

N

8'
"'....""
"'~
"'"'
"'"'
~....
~
;t
;:;:
~
"'~
TABLE 2.14
A comparison of rapid freezing methods
""':l
....

Disadvantages "'s..a
Applications ~
Description and advantages s·
Method
Complex, expensive "'s·""
Tissue blocks, suspensions, etc.
c
High pressure-freezing
Freezing at 2,000 bar (2.1 x l05kPa) Only method of obtaining well-frozen ""';:;·12..
to sub-cool water so that the critical large ( < 1.0 mm) blocks of uncryo- "'(;;"
cooling range is much reduced and protected hydrated tissue "a
1
cooling rates of only 102 K s- are
required for good freezing.
""':l
Commercially available (Apperidix 2)
Q'
a
.g""'
Moderately complex apparatus
Low viscosity solutions, suspensions, required, although this can made in
Spraying fine droplets of suspensions,
Spray-freezing etc. any normal workshop. '""'
emulsions, etc. into liquid cryogens Droplets ( ,s; 0.1 mm) or solid Only suitable for 'sprayable'
(e.g. propane). specimens ( ,s; 0.05 mm). Yields
Equipment commercially available specimens
extremely good freezing results

Thin, flat, specimens such as cultures, Complex and relatively expensive


saridwiehed between thin metal monolayers, etc. apparatus required.
plates, by impacting a jet of cold Thickness of specimen ~ 0.05 mm. Jet may dislodge specimens
cryogen (usually propane) on one, or Ideal for preparation of suitable
both, sides. Commercially available specimens for freeze-fracturing
(Appendix 2; §2.4.1 b)

Plunge-freezing Rapid immersion into liquid cryogen. Small tissue blocks. Unless plunge conditions are
The conditions of plunging must be This technique is very versatile. optimised, poor cooling rates may be
optimised. Some appropriate Simple and inexpensive. obtained. The surface may be rough
apparatus available (Appendix 2; Recommended as an initial and irregular and, therefore, the thin
§2.4. la) procedure and in most circumstances zone of well-frozen material may not
where the highest cooling rates are be easy to examine by freeze-
unobtainable (because the specimen fracturing, cryosectioning etc.
is too large). ~
~
~-
Cold block-freezing Rapid freezing by impacting a Gives highest heat transfer and yields Very complex and expensive
specimen onto the polished surface of excellent freezing in thin (10-20 µm) apparatus required for routine use.
a high thermal conductivity metal surface zone Specimen may be deformed by
block at low temperature (77 or 4 K). contant with block. Liquid helium is
Commercially available (Appendix 2; expensive
§2.4.2)


w
hods in biological electron microscopy Freezing 95
Robards and Sleytr Low tempera t ure met
94
(and hence the cooling rate) are strongly dependent on the temperature dif-
ference between the specimen and coolant (AI) and the likelihood of film
e ..
~ boiling is reduced if the temperature interval between the boiling point and
~ 'C
~"'
o9£0
-' Q. •
the temperature of use is as large as possible. Therefore, liquid cryogens
EE
.! should generally be used at their lowest liquid temperature. Thus, even a
difference of 10 K (e.g. 85-95 K) can have a significant effect on cooling
rates obtained using liquid propane (Costello 1980; Fig. 2.40). In addition,
it is always beneficial to present the specimen constantly with new coolant
at its lowest temperature during the freezing process, either by plunging the
specimen through the coolant or by 'squirting' the coolant over the speci-

·- "
c:-

-~... ~
~
Temperature (K)
Q) N
"z
u: -' 60 80 100 120 140

~
100

80

..,
l.:

"'0
I 60
Freon 22
~
QI

~
"'
.!: 40
0
0
u

"'
.E
ii
20 ~
E
1-1------.co
"'
-210 -190 -170
Temperature (°C)

0
c: 0
="~..c:
u
"..
~

".
~ Variation of cooling rate in relation to the temperature of the coolant. Curves for
Qi
.S! .!!?
If) -
c Gl .....
'O
4l
~
0
a. .
"'
0c:
:;:
;;
!
c.
ane, Freon 22 and liquid nitrogen are shown. A bare copper/Constantan thermocouple
a 70 µm diameter bead and 25 µm diameter wires was plunged 15 mm into the coolant
~-~ ;! E E
·;:
II> c 0 0 0
:i :J 0 <I: 0:: at a rate of approximately 0.48 m s- 1• (Redrawn from Costello 1980.)
·-
Cl)
:IE
"'
Freezing 97
Robards and Sleytr Low temperature methods in biological electron microscopy
96
Air admittance valve
men. Other considerations will, of course, also apply but the following sec-
tions describe practical methods of achieving rapid cooling rates using cryo-
genic liquids that have proved satisfactory. Thus many potential coolants
are eliminated on the grounds of their physical characteristics (such as too
Polystyrene container
high a melting point, rapidly increasing viscosity as the freezing point is
approached, too hazardous to use, or even too expensive). Furthermore,
- Rotary pump
other coolants will also be ruled out because the complications associated
with their use are not justified by the results obtained (liquid helium II ap-
Liquid nitrogen
pears to fall into this category although it also appears to be a poor coolant;
§2.2.3a) or because they produce undesirable side effects (such as halogen
contamination of samples by halocarbon cryogens prior to microanalysis) Fig. ·2.41. A simple system for prepari·n g 1.1qm.d mtrogen
. 'slu h' b I ·
well-msulated (e.g. expanded polystyrene or po 1yurethane) s t"or su · -cooed mtrogen. A
a1~er
in relation to the particular experimental requirements. With such thoughts I .
vacuum chamber. The chamber is evacuated throu h a . con is p aced m a small
in mind, we can first look at the simplest and most widely used meth'od of diameter). The liquid nitrogen firstly b . . g relatively wide-bore tube (15-25 mm
. 01 1s vigorously and then f A. .
cooling: that of immersing the specimen directly in the liquid cryogen. qmckly and the nitrogen remains sub-cooled . f or a cons1.d erable penod,
. reezes. If 1s readmitted
depending on the ther-
ma1properties of the container.

2.4.la Immersion method


liquid to 'explode' so thro . r "d .
The principles governing the selection of a suitable technique for rapid cool- chamber. This can'b "dwdmbg Iqm. mtrogen over the interior of the
e avoI e Y placmg hollow drinkin t ' ·
ing by immersion should now be apparent (§2.2.5). Liquid nitrogen at its bump' granules in the container. Alternative! . g s raws or antl-
equilibrium boiling point is not a good coolant, but sub-cooled nitrogen, be partially solidified and then reliquefied b y, th~ mt~ogen can repeatedly
some halocarbons and cryogens such as propane or ethane may all be used. ing the whole pool to have sufficient ti~e ~ovaryI~g t e vacuum, so allow-
Current evidence suggests that, if the very fastest rates are to be obtained, and freeze as an entity As a furth . reac the same temperature
· er precaut10n it · d ·d
then propane is the coolant of choice. We shall look at propane more closely a perforated polystyrene lid over th t . ' IS a goo I ea to place
later. However, we will first consider cooling methods where the require- ing. When the nitrogen has soli~ificioedn at1nher
'
:o
trlap adny excessive splash-
e msu ate conta1· b
ment is that moderately sized specimens (perhaps cryoprotected pieces of removed from the vacuum chamber. ner can e
tissue of about 1.0 mm 3) should be well-frozen under controlled conditions. i I
As soon
starts as air
to warm u isI leaked
d" into the v acuu~ chamber, the nitrogen at 63 K
In these circumstances, the very fastest cooling rates are not required and
the use of either sub-cooled liquid nitrogen or one of the halocarbons con- ('slushy' nitroge:): ~~e1;~~:rua~~'/.o a mixture ofliquid and solid nitrogen
. . . e a Is normally used for rapid r S h
densed into a bath cooled by liquid nitrogen is recommended. Methods for a mixture is not in a state of stable e uilibr· . coo mg. uc
cooling in liquid cryogens are reviewed, amongst others, by Umrath (1974b, adients and th q mm, there are strong temperature
<
ing sub-cooled ere may C
liquid not be any so rd .
I mtrogen present, the pool simply
1975), Costello (1980) and Elder et al. (1982).
Sub-cooled liquid nitrogen. Liquid nitrogen is pumped down to the tripl in cooling rates bet. onsequently, there can be quite significant varia-
ween runs when using this th 0 d Q .
point of nitrogen (about 13.5 kPa, 1.01x10 2 Torr) and solidifies at a tern r considerations the . . me · mte apart from
perature of about 63 K. To obtain this transition, pour about 100-200 m tallisation is tak' pbre~nce of solid mtrogen means that the heat of
en up e1ore the whole Ib . . .
of liquid nitrogen into a well-insulated container (such as an expande ccepted without wa . . poo ecomes hqmd (i.e. heat
thermod . rn:mg t~e pool until the solid has all gone) and thus
polystyrene cup) and evacuate this within a chamber (a desiccator can
t h
s u~ or o~ t e sub.-cooled liquid
ynam1c considerat10ns for the I h
used) pumped by a rotary pump through a relatively wide-bore tube (pret are rather different. If the b-
ably ~25 mm in diameter; Fig. 2.41). Nitrogen 'ice' can form on the surf ated cup then it ·n _su_ cooled mtrogen is contamed in a well-
, w1 remam m the sub-cooled state for some period
of the liquid, thus impeding further gas escape and causing the trapp
99
98 Robards and Sleytr Low temperature methods in biological electron microscopy

of time (a few minutes). However, as the temperature drifts upwards, the


likelihood of significant film boiling during cooling becomes increasingly
great and therefore the coolant should be used as soon as possible after re-
moval from the vacuum apparatus.
A rather more convenient method of maintaining a liquid pool of sub-
cooled nitrogen has been described by Umrath (1974a, 1975) in which an
outer chamber of liquid nitrogen, which can be evacuated, surrounds an Pressurised
haiocarbon
inner chamber which remains at atmospheric pressure (Fig. 2.42). Using this cylinder

device, nitrogen can be maintained sub-cooled for prolonged periods, thus


simplifying the consecutive freezing of a number of specimens. In all in-
stances where sub-cooled nitrogen is used, it is advisable to check the tem-
perature of the pool, either by noting the presence of solid nitrogen (63 K)
or by using an electronic thermometer which has been calibrated for use at
such low temperatures.
Halocarbons. The next most used group of liquid cryogens is the halocar- Magnetic stirrer

bons, of which halocarbons 12 (CC1 2F 2) and 22 (CHC1F 2) are the most


popular. These gases become liquids at about 230-240 Kand solids at about
114 K (Table 2.6; §2.2.3a) and they are therefore usually condensed into Fig. 2.43.
. Preparation of halocarbons ior~ use as cryogens Th .
t e cy mder and allowed to flow throu h fi ( . · e pressunsed gas is ducted from
the .metal cryogen container which is c;ol:d ~e t::~~~;imately 1.0-2.0 mm diameter) tube into
h I
containers cooled by liquid nitrogen. The gases are available from suppliers
con~e~:~ ~~~ ~;~uid
(Appendix 2) in large cylinders or in small pressurised cans for laboratory m1tially, be rather slow so that there i t" / h . oundmg hqmd nitrogen. Gas flow must
Once a small _'pool' of liquid has to condense on the cold metal walls'.
use. The simplest way of obtaining the cooled liquid is to attach a piece of be immersed m the liquid and furthe : rate_ may be mcreased as the nozzle can
b d r gas w1 11 condense d1rectl Wh .
tubing to the container outlet and then to lead this tubing into the bottom een con ensed, the gas supply is turned ff d h . Y· en sufficient cryogen has
of a container partially immersed in liquid nitrogen (Fig. 2.43). The con- to cool until it is close to its freezing po~nt a~d\~en~zzk is ren:10ved. The cryogen is allowed
or some other means to ensure that thermal d. qmd IS st1rred usmg a magnetic stirrer
d· · gra 1ents are · · · d
use m this way will freeze solid when cooled with li ui _mm1m1se ".Most cryogens that are
Thin stainless steel cylinder ?y.adiustmg (reducing) the level of the surroundin lq d mtrogen. This may be avoided either
JS msufficient to maintain the tern t b 1 g iqmd mtrogen so that the heat withdrawal
, . pera ure e ow the fre · · .
(a cartndge
ld .
heater' (see Appendix 2) . . bl . ezmg pomt or by mserting a small
is smta e) mto the p 00 l 0 f
: I
is used to avoid excessive condensat" f . cryogen. An insulated
1 °
wn mmsture mto the cryogen. I

·ner should ideally be made of a hi h-con . . .


- Rotary pump
. Gas is admitted until sufficient I~ uid ductrv1ty mate~1al such as cop-
e care that liquid does not f . q h has condensed m the container.
ay be blown out by increasireeze m t e end of the supply pipe, otherwise
erature of the hal b ng ~ressure. Subsequent cooling reduces the
using the cryog;:ara ho~ huntth1l, ultimately, it freezes. Immediately pri-
Fig. 2.42. Apparatus for the continuous maintenance of sub-cooled liquid nitrogen (as d ' ig - ermal cond f ·t
e.r) is thrust into the s I'd h I uc IVI y metal rod (brass or
scribed by Umrath 1974a). The outer vessel is evacuated, so sub-cooling the liquid nitro 0 1 a ocarbon th ·
lted pool. As this p I b . ' us causmg the formation of
This cools the liquid nitrogen in the inner container by conduction through the copper W . . oo egms to refreeze and . th I .
Thus the inner vessel contains liquid nitrogen which is continuously maintained at low tem 'It Is ready for use. A "th b . IS us c ose to its freezing
ature (close to -210°C; 59 K) and can be used without the need for constantly breaking is not without practica~ WI blsu -coo_le~ mtrogen, cooling in halocar-
vacuum as required for the apparatus illustrated in Fig. 2.41. pro ems. It rs important that there should be
Freezing 101
100 Robards and Sleytr Low temperature methods in biological electron microscopy

a large enough liquid pool to allow an adequate depth of plunge of the spe-
cimen, but it is also important that all the liquid is very close to its freezing
point. Using this method, there are bound to be thermal gradients and,
again, it is difficult to use the same liquid pool to cool a number of speci- Metal cryogen
container
mens consecutively. Another consequence of plunging specimens into a pool
that is re-freezing is that the specimen can become trapped in the solid cryo- Insulated container

gen. For these reasons it is useful if a system of maintaining the halocarbon


at constant temperature is devised so that the excessive cooling by liquid
nitrogen is offset by additional heat input. This can be achieved either by ~--J'-1-=SH~_j_ Magnetic
flea
placing a small heater in the cryogen or by adjusting the contact area of the
cryogen container with the surrounding liquid nitrogen. If this is done and
some sort of stirring or agitation method is used to minimise temperature
gradients within the pool (Fig. 2.43), the results will be far better and more Magnetic stirrer
reproducible. Some workers have achieved good results by very rapidly rais-
ing the container of cold halocarbon up to the specimen (e.g. Somlyo and
Silcox 1979; Wendt-Gallitelli 1984). It is good practice to use halocarbons
.
Fig. 2.44. . Liquefaction of propane "o1 ' r use as a cryogen Th h . .
in a fume cupboard to avoid the build-up of high levels of vapour which trate m Fig. 2.43 for halocarbons but b .
d
: e met od IS similar to that illus-
°.
~eta! p~~v1~e
- 190oc.· 83 . '
K) is lower, it is necessar to ecause the
. meltmg .
. . porn t f propane (approximate!
may be toxic (Taylor and Harris 1970). usually by leading it through a coil of add1t10nal. coolmg for the incoming ga{,
Propane. Propane (or other cryogens with similar properties, such as eth- gen. Propane usually stays liquid for quite~ !~~ r) pipe that IS immersed .in the liquid nitro-
ane) can be condensed in exactly the same manner as the halocarbons but, (especially if 'commercial' rather tha g penod when cooled with liquid nitrogen
r1qm'd state by the addition' of 20- 0"/n .pure, propane. IS US ed) an d it. can be maintained i th
because of its lower freezing point, the gas flow should be slower. It is useful · . . . . 3 /o 1sopentane (Fig 2 28· §2 2 S) n e
msu atmg lid 1s essential (as in Fig. 2 43 ) C Id · · ' · · · Although not shown an
to partially cool the incoming propane gas by passing it through a copper 1
longed periods due to the danger of.e , . .o propane should not be left standing for p' ro-
xcess1ve amounts of Ii uid . .
coil immersed in liquid nitrogen (Fig. 2.44). It cannot be stressed too strongly thus forming an explos1·ve m· t ( q oxygen condensmg mto it
IX ure see text). '
that propane is potentially an extremely hazardous chemical. On the other
hand, this should not cause its considerable virtues as a coolant to be nitrogen gas over the top of the r .
ignored. Under proper conditions, there is no reason why propane should n µie propane is close to its usable t p opane contamer. Similarly, when
• emperature ( ~85 K) 't
be used both safely and routinely. The major hazard lies in the flammabili so low and it is so far removed f . fl - I s vapour pressure
of propane and the potentially explosive mixture that it may form wi e hazards are minimal Th . rom I.ts ash point ( 104oc; 169 K) that
liquid oxygen which, with a boiling point of 90 K, may condense from t ne is to be disposed .of: ::ut~i~r ~:n:~ for concern arises when the pro-
air into other liquids at lower temperatures. There are reports that ult rrned up; the vapour pressure will ris: an~et::opane ~ust ~ecessarily be
sonic devices may provide sufficient energy to cause liquid propane our experience, small quantities of Ii uid flash-pomt will be reached.
explode and, therefore, they should not be used adjacent to such coola rm up and evaporate in a f bq propane may be left safely to
ume-cup oard pr .d d h
baths. Clearly, simple precautions, such as avoiding naked flames, usi any other adjacent electrical e . ov1 e t at the fan motors
the liquid may be ca . d qdmpment are spark-protected. Alternati-
spark-protected switches, using the minimum quantities required a rne out oors and d' d
minimising the time for the pool of liquid propane to 'stand around', ground well away from buildin . ispose of by pouring onto
all help to minimise any risk. During the condensation phase there sho ns should be taken b th gs, vehicles or any other object. Pre-
be little risk of propane gas escaping provided that the flow rate from opane as he pours it it i e ope.rator to see that he stands up-wind of
supply is kept low. In addition, the relatively warm incoming gas causes pletely in liquid nitr~ge: ~:~s1~lle t~ st~re the propane by immersing
a owmg it to freeze prior to storage in
surrounding liquid nitrogen to boil and thus to form a protecting 'blan
thods in biological electron microscopy Freezing 103
102 Robards and Sleytr Low temperature me

. . . i erated vessel. However, this process should the Chapters relevant to the specific techniques. It is sufficient to comment
a suitable hqmd mtrogen-refr g f 1. .d oxygen condensation into the here that the supports should have the minimum possible mass and should
d to the danger o iqm d
not be repeate d ue . Th blem ofliquid oxygen con ensa- be of high thermal diffusivity. Conversely, the means of actually holding
. ecutive cycles e pro
propane dunng cons . 11 .d d that the propane con- onto the specimen, together with its support, during freezing should allow
. l e is probably sma provi e . l f
tion dunng norma us "bl d that the cryogen is not e t the least possible heat conduction into the cooling specimen. In general,
d as much as possi e an .
tainer is kept covere . d The constant flow of mtrogen metal holders of low thermal conductivity and small cross-sectional area
·vely long peno s.
standing around f or excessi t . r also impedes the access of oxy- should be used. For example, very fine stainless steel tweezers are good for
h f the propane con ame
r .d W can see t h ere f ore, that propane is a perfectly rea-
gas over the mout o this purpose. Copper has probably been used more than any other metal
for the specimen supports themselves and there are no obvious reasons for
gen gas to the iqm . e . d' d .t has some practical advantages apart
sonable cryogen to use and, m eed' i .b d above (§2.2.5). For example, its preferring gold or silver to copper (Table 2.10; §2.2.6)
. . .t s a coolant escn e 1
from its supenon y a . . 1. .d in liquid nitrogen for ong Handley et al. (1981) drew attention to the critical importance of the
. · that it remams iqm .
low meltmg pomt means . . problem that arises with the thickness of any metal specimen support across which heat needs to be with-
eriods (thus alleviating the specimen-dtdr~~pmgf p to 3001 isopentane, the drawn during rapid cooling. As the effective time interval between cooling
P h by the a it10n o u lo
halocarbons). Furt ermo~e, db l that of liquid nitrogen although the outer and inner surfaces of such a support is proportional to the square
melting point of propane is r~duce e ?w ffected (Jehl et al. 1981; Fig. of its thickness, it is clear that it should be as thin as possible. Unfortuna-
the coolant properties are said to remam una tely, metals with good heat transfer characteristics (e.g. gold and silver) are
2.28 in §2.2.5). . . h . II is also a liquid coolant, its not very strong. Consequently, it is often not practically possible to use
· II Although hqmd e1mm ·b d
Liquid helium · . f th t of the cryogens descn e them as sheets less than 25-50 µm thick. Handley et al. used titanium which,
. · · ite different rom a . ,
use and apphcat10n is qu . . .t n temperature or above. Fernan- while having thermal properties inferior to copper or silver (Table 2.10), has
above, which are all used at h_qu: n~ roge duction of liquid helium II by better tensile strength (Ti= 860 MP a; Cu= 211 MP a; Ag= 130 MP a) and
dez-Moran (1960) has describe .t e.dprhol. m I to about 70 µm (9.3 Pa), was usable as foil envelopes only 4.0 µm thick. The time taken for heat to
· h ssure on hqm em f
rapidly lowenng t e pre f 0 8 K Because of the low heat o ·. be transferred across a thin metal layer was calculated (Carslaw and Jaeger
. . 1 t to a temperature o . . d d
which is e~mva e~ . . relatively large quantities must be use an 1959) from:
vapourisat10n of hqmd helmm, d d by a liquid nitrogen-filled.
. el must be surroun e . d 0.36 d 2 pep
the helium Dewar vess the helium II must be maintamed un er (2.14)
k
chamber. Furthermore, becau~e d . vacuum lock. This method h
· must be mserte via a . .
vacuum, the specimen . (b t Bullivant 1960) because it is n
· d · recent times u see · · ·
not be wide1y use m . nsive Reference to it is 1 =time interval for 100 K temperature decrease
t (§ 2 2 3a) and is very expe ·
a very good coo 1an . .. . ft quoted and it is importa =thickness of foil
1 d d here because its possible virtues are o .en . density
cu e d t nd the actual situat10n. .
that the reader should un ers a S f only methods for obtai =specific heat
· · liquid cryogens o ar,
Methods of immerszon zn b d . ·bed However equally imp =thermal conductivity.
s have een escn · ' d
ing liquid pools o f cryogen . . plunged into the pool an
whereby the specimen is M t
tant are t h e means f d after the freezing process. os s ig. 2.45 illustrates the relationship between the thickness of copper or
. ns both be ore an M h
handling of specime f m of support or carrier. uc ·um sheets and the time taken for a 100 K temperature change to pass
e mounted on some or f
cimens nee d t o b h t .al once it has been rozen. one side of a sheet to the other. Using figures for thermal conductivity
. t be done to t e ma en
depend on wh at is o . ecimen in a freeze-fracture ecific heat at 25°C (298 K), Handley et al. (1981) found that it would
example, the requirements for mounltmg ~ s~tome or for freeze-substitu . ake 0.62 µs for a 100 K temperature change across a 4.0 µm thick tita-
th se for a cryou tram1cr .
are different from o . f ecimen supports will be given layer, equalling a rate of 1.6 x 10 8 K s~ 1 , which is well above the criti-
or freeze-drying. Recommendat10ns or sp
Freezing 105
Robards and Sleytr Low temperature methods in biological electron microscopy
104 if mounted, the support should be of h.igh thermal diffusivity
.
mass; and low

--
--- --- 6
liquid bath (Ryan and Purse (1984) h: pre~cool m co~d g~s above the
the specimen should not be allowed t .
100

,, ,,
---
,,,.,,,,,.,,,.Titanium
10

(v)
the specimen should be pl
ble. .
d . ve discussed this pomt) and
unge mto the coolant as rapidly as possi-
I
,, ,, 1 o7
I
10 I
I
I The ·first three points have already b een d"IScussed but th 1
I require further consideration. e ast two aspects
u; 1.0 I
I
I
10
8
·:..:.
3-
GI
E
j:.
I
I
I
I
I

10
9
-'"
GI

g:

0.1

1010
0.01

Thickness (/"ml
Spring _ _Ja>-Mic:::JP'J
Fig. 2.45. The time taken to conduct heat across copper or titanium sheets. This graph shows
Close fitting piston ----1.AL:...._-""'=:..J/I
curves obtained by using Eq. (2.14) to calculate how long it takes for a 100 K temperature
change to pass from one side of a metal sheet to the other. The better thermal conductivity
of copper gives significantly shorter times although titanium can be used in thinner sheets.

cal cooling rate necessary for the formation of sufficiently small ice crystals
and, therefore, not limiting to the ultimate specimen cooling rate that could
be achieved by their methods. From Fig. 2.45 it can also be seen that th
time taken to cool 100 K across a 50 µm titanium sheet would be about l.
ms (1.03 x 105 K s- 1) whereas across a copper sheet of equivalent thickness, Air damper
needle valve
the time would only be 7.6 µs (1.3 x 107 K s- 1): two orders of magnitud ·
faster. Fig. 2.45 is thus useful in providing an estimate of the effect oft
thickness of the specimen support on the highest cooling rate that can
obtained within the specimen itself.
The major requirements during the sampling and freezing stages ar
(ignoring the coolant characteristics):
.46. Mechanical dev1·ce r " . . Thermocouple support
. . ior mJectmg' . . .. .
(i) the specimen should be damaged as little as possible, either by ni ector is 'damped' by usmg
. a piston
. specimens
to force air oumtof hqmd .cryogens . Th e sprmg-pow-
.
needle valve. (Redrawn from GI t o the cylmder through an adjustable
chanical means or by desiccation; over and Garv1tch 1974.)
(ii) the specimen should be as small as possible;
l

Freezing 107
Robards and Sleytr Low temperature methods in biological electron microscopy
106
the coolant is kept at a low temperature and th . . .
When the specimen has been obtained and mounted on a support (if ap- cold gas above the liquid surface. ere is mm1mal depth of dense,
propriate) it should be frozen as quickly as possible. Over any pool of very The method of inserting the specimen into the c .
cold liquid there will be cold air which will pre-cool the specimen and may comparatively recently been gr· h . oolant ts one that has only
ven muc attent10n H ·
actually freeze it at very slow cooling rates if precautions are not taken to seem clear that within reason h. h · owever, it does now
avoid this. Such precautions take two forms: firstly to minimise the layer tello 1980; Ha;dley et al 1981'. Rtgberdratesdof entry are advantageous (Cos-
. . ' o ar s an Crosby 1983) E 1 .
of cold air above the bath, and secondly to move the specimen into the bath mvolved 'dipping' the specimen mo . t the coolant bath Th" · ar Y techmques
as rapidly as possible. In general, it is preferable to maintain the liquid level recommended: firstly , the entry r a t e ts
. b ound to b · th ts method1 is not
as close to the top of the container as possible and, except at the moment ondly such hand-held methods pro d uce very mcons1st
. _e ra t ert s ow,
A and .sec-
of freezing, to keep the container covered with an insulated lid. In this way work should be undertaken using some form f. . en ra es. . ny senous
anism that allows predetermined ent 1 .~ mJector an~/ or sturer mech-
ry ve oc1ttes to be obtamed reliably and

Brass tubing
20mm
t--i

,.!'!]..__--~--Specimen head
Filling holes
Plywood disc
Polystyrene
container and lid

Gas escape holE>

Brass container
I I
I I
I I Liquid nitrogen
lJ~--- Specimen support

Fig. 2.47. A liquid damping system for specimen injectors. This system, as opposed to
air compression mechanism illustrated in Fig. 2.46, uses glycerol in a 'dash-pot' to damp 2.48 . sprmg-ass1sted
. ..
ens. The cryogen usedmechanism
here i for
f in"ecti
J
.
ng specimens at up to IO m s-1 into liquid
progress of the specimen rod into the cryogen. The specimen support and support arm are
liquid; the descending speci~eca:~ ully cooled so that its surface is as cold as the bulk
ried on a plate which is integral with the damper rod. When the specimen falls. the rat n us meets the coolant at its lowest temperature (Red
movement is retarded as the piston moves through the glycerol. (Redrawn from Costello a rawn f rom Handley et al. 1981.) · · -
Corless 1978.)
-

Freezing 109
108 Robards and Sleytr Low temperature methods in biological electron microscopy

reproducibly. Many such devices have been described and some are
illustrated in Figs. 2.46 to 2.50. Luyet and Gonzales (1951) used a 'guillo-
tine' for rapid specimen insertion. Rebhun (1972) describes a small har-
poon-like device that uses a rubber band to fire the specimen into the cool-
ant. Glover and Garvitch (1974) built a mechanical injector (Fig. 2.46) that
1
shot the specimen into the coolant at an average velocity of 2 m s- and
simultaneously triggered the oscilloscope as it fired. Costello and Corless
(1978) constructed another guillotine (based in part on the work of Heuser
et al. (1976) for copper block freezing; Fig. 2.62; §2.4.2) that gave unassisted
1
entry rates of 0.39 Ill s- 1 and variable rates of 0.49-0.74 m s- when ener- Support
brackot
gised by rubber bands; it was damped using a glycerol-filled dash-pot (Fig.
1
2.47). Handley et al. (1981) produced rates up to 10 m s- with a spring-
actuated device (Fig. 2.48); Echlin and Moreton (1976) described a free-fall Dewar
injector giving an entry speed of 3 Ill s- 1 and Bald and Robards (1978; see
:Ji<~+-1--- Aluminium
also Bald 1978a) also used a gravity-driven device giving a plunge velocity cylinder

~i.g.
e .sp~c1men IS held between very fine
2.50. Solenoid-operated rapid immersion device Th . .
pmcers' (tweezers) to minimise thermal contact Volt.
v1des a variation in immersion velocity betwe 0 4 ;~eOvan~t10n of the electromagnet pro-
ation (200 m s-2 maximum) The st k . ~n · an · m s with almost constant acceler-
2
1

. ro e is mm and the specimen is immersed 15 mm into


the coolant. (Redrawn from Escaig 1982.)

't:~ m~ -1 indepen~e~tly of springs, rubber bands, or other auxiliary sys-


. 1st
.s. d umerous
· · ·s1m11ar
· devices have been described · A s1·mp Ie spnng- .
e
F' 2 Ll9 mJector givmg rates
. between about 2 ·0 and 8 0 -1. ·11
· m s 1s 1 ustrated
ig.. . ' . and an electnc solenoid-driven plunger designed by Escaig is
o;.~ m Fig. 2.50. fto'n elegant device for rapid plunging has been described
•. 1 e (1984) and is commercially available (Appendix 2) Th .
mbl h Id , , · e specimen
fl Y_ s ou be held so that the lowest possible pathway for heat
s:r l~to the speci~en ~s .p:ovided: fine stainless steel forceps or wire
F' ffic21ent mechamcal ng1d1ty with relatively low thermal conductivity
Fig. 2.49. Simple spring-assisted mechanism for plunge-cooling specimens. Use of diffe ~ and .2.52 for exampeso
igs. . .50 I f specimen
. holders). Summarising the
1
spring strengths allows different entry velocities up to about 8.0 m s- • (Reproduced, with ab! e mtormat10n ' 1·t is
mission, from Robards and Crosby 1983.)
· poss1'ble to construct a list of conditions that will
Freezing 111
Robards and Sleytr Low temperature methods in biological electron microscopy
110

TABLE 2.15
· Some criteria for efficient cooling in liquid cryogens

Specimens should be as small as possible


(i)

Specimen supports should be as small as possible, of low mass and high thermal
(ii)
diffusivity

For most purposes propane is the best liquid cryogen


(iii)

The coolant should be at its lowest liquid temperature


(iv)

The coolant pool should be stirred to avoid temperature gradients


(v)
The coolant vessel should be filled to the top (to minimise pre-cooling in the vapour
(vi)
phase)

The entry rate of the specimen into the coolant should be high(> 1.0 ms-')
(vii)

There should be a sufficient depth of liquid to allow the specimen to keep moving
(viii)
at the initial entry velocity, at least over the early part of the critical coolmg range
(effectively 273-173 K)

The specimen should be the first part of the injection assembly to contact the coolant
(ix) Fig. 2.51. Propane jet-freezing device designed by Muller et al. (1980). The purpose of the
apparatus is to provide an opposed pair of high velocity jets of cold liquid propane, between
The specimen should be supported and 'held' using low mass, low thermal
(x) which the specimen (see Fig. 2.52) can be thrust. This apparatus is commercially available
conductivity contacts (Appendix 2). (Reproduced, with permission, from Muller et al. 1980.)

help to produce the most efficient cooling method available using the meth n. The principle of jet-freezing is simple but a specially constructed appar-
s is necessary for its application (Fig. 2.51) and is commercially available
od of plunging into liquid cryogens (Table 2.15).
ppendix
The propane is cooled down to working temperature (which should be
2.4.lb Propanejets 93 K) in a vessel cooled by liquid nitrogen (Fig. 2.51). At the appropriate
In the previous section, the importance of moving the specime~ rela:i~e t
ment, the specimen (which is usually sandwiched between very thin cop-
lates; 2.52) is inserted between the jet nozzles. This is done using
the cryogen was stressed. One of the important reasons for domg this is
ensure that the temperature of the coolant effectively remains the sam~ d
cial specimen holder. In the commercial equipment, this movement
ing the whole cooling process. If this is not done, then the ?oolan~ adJaC~
s pressurisation of the propane container which propels the cold
to the specimen warms-up and may even boil, thus pr.oducmg ~n msulatl
propane onto either side of the specimen. The propane must be at
gas jacket. An alternative approach is to keep th~ specm~en ,statl? but to P
a temperature as possible prior to each run. This means that there
be some slight delay while the propane re-cools after consecutive runs.
ject the coolant over it: this is the method of 'Jet-freezmg which was fi
tmosphere between the jets is bound to be cold and it is therefore
introduced by Moor et al. (1976) and subsequently described in greater
·al that specimen is brought extremely quickly between the jets and
tail by Muller et al. (1980). These authors again chose propane as the er
Freezing 113
Robards and Sleytr Low temperature methods in biological electron microscopy
112
Insulating material
Specimen carrier
Specimen - Nitrogen gas
Specimen sandwiched
between grids
!I yopper

~ -=- Liquid propane

/
/ /
/ /
/ ///
/
/
/ Fig. 2.53. Simple single-sided jet-freezing. Cold liquid propane is projected against one side
of the specimen, the other side being made from a thermally insulating material to reduce heat
input during freezing. (Redrawn from Pscheid et al. 1981.)

workshop to construct suitable equipment. There is no reason why this


should not be accomplished successfully, but if it is tried it is imperative that
due consideration is given to both the cryogenic and mechanical require-
<rn.ents of the apparatus. The points mentioned above must be met and fea-
. .tures such as insulation around the jet orifices, exact alignment of jets and
Fig. 2.52. Specimen holders for use with a propane jet-freezing device. A pair of o:din
electron microscope grids 'sandwich' the specimen. In the system shown m th~ upper diagr yery quick response times to trigger the pressurisation must be optimised.
the grids themselves are mounted between a pair of shaped copper supports, with central h Pscheid et al. (1981) and Knoll et al. (1982) have used a rather simpler
so that the whole assembly can be transferred to a freeze-etching apparatus .. Othe.r arra pane than the one described above to project a single stream of pro-
ments may also be used as, for example, shown in the lower diagram. The two iet orifices
be exactly aligned horizontally opposite each other. The specimen. holders are, m both the .. e against one side of a specimen (Fig. 2.53). This equipment has the ad-
terns illustrated, approximately 3.0 mm across. (Upper diagram kmdly provided by M. M ntage that precise alignment and synchronisation of two jets is avoided
and reproduced with his permission. Lower diagram redrawn from Robards and Severs 198 d the actual 'gun' is much easier to construct (Fig. 2.53). It is therefore
re easily applicable for use in laboratories where, either from choice or
ility to purchase expensive equipment, the two-sided jet apparatus is
that the propane stream is triggered immediately. If this is not done, th
available. The performance of this system appears good: cooling rates
is a real risk of slow pre-cooling. The commercial equipment enables th
. 18,000 K s- 1 have been recorded, which compare well with those
criteria to be satisfied easily and routinely. Finally, when the jet has sto~
med from two-sided jets. The explanation for this is partly that, in prac-
the very small, thermally sensitive, specimen will start to warm up qmc
' recise synchronisation of double-sided jets is probably not possible
and it is therefore important to transfer it into a liquid nitrogen storage c
1ll addition, the thermally insulating backing to the specimen in the sin-
tainer as quickly as possible. . system minimises heat input from this side while heat is being rapidly
While the commercially available propane jet-freezing apparatus pr~
awn from the other.
all the requirements for this technique, it is also possible for a well-eqmp
Freezing 115
114 Robards and Sleytr Low temperature methods in biological electron microscopy

d LED Temperature controlled


j brass block
Alternatives

Compressed air

o'T"~ I 1 0 '-Propane jet


~
I l .. .......
, , Butyl

riU
nozzles

DI / liquid
Position A
I 0 nitrogen

L__+J""_r
Teflon r o d [/ _ _ _ _ _ _ -f/-_.J
light sensor Specimen in clamps

c=J-q-l\!========~--.~Propane
0 .t

D.
JI!! COPP"'
block
Position B V6
c~
-190°C(83K) - - - - - - -85°C(188K) - - - - - - -196°C(77K)

Fig. 2.55. The sprar-freezing technique. Propane is condensed into a cold copper or brass
Fig. 2.54. System for maintaining specimens at controHed, known temperature immediately block. at about . -190 C (83
. K). Usmg
. an a!f brush or a '1'et' sprayer , the l'iqm'd specimen · (sus-
prior to freezing. The arrangement illustrated here 1s designed for use with a propane Jet device pens10n, emuls10n, etc.) is sprayed mto the coolant The temperature of th e l'1qm'd propane 1s
• • • ·
0

but could easily be adapted for other applications. When the specimen 1s thrust forward, the raised to about - 85 C and 1s then evaporated by connection to a rotary vacuum pump. The
light from the light emitting diode (LED) falls onto the light sensor and thus tnggers the Jet frozen droplets are then mixed mto a 'slurry' with butyl benzene or some alternative substance
of propane. (Redrawn from Van Venetie et al. 1981.) havmg a meltmg
, po . . (§4 ·2· l •· Table 4 · 1) · So me o f th'1s s1urry 1s
. mt around - 90 to -. 100°C · trans-
ferred to tne specimen holder, after which 1t 1s re-cooled to liquid nitrogen temperature. This
techmque has mamly been. applied to freeze-etching (§6.2) but could equally well be used in
Van Venetie et al. (1981) made a simple modification for the Muller e o~her low t~mperature applicat10ns. A_ spray-freezing system is commercially available (Appen-

al. (1980)-type propane jet by arranging a specimen st~ge tha~ would ~ain
dix 2). (Adapted from vanous 1llustrat10ns published by the originators of this technique: Bach-
mann and Schmitt 1971.)
tain temperatures around the specimen up to 330 K im~ed~ately _pnor t
rapid cooling (Fig. 2.54). This allowed them to _study hqmd-so_hd phas
transitions for some phospholipids. Although designed for use with a pro
same process is by the use of an atomiser. This technique was pioneered by
pane jet apparatus, a similar system could equally well be used with oth
hmann, Plattner and their colleagues (Bachmann and Schmitt 1971 ·
rapid cooling methods. . . hmann and Schmitt-Fumian 1973; Plattner et al. 1973; Plattner and
While excellent results have been achieved using propane Jet-freezm
hmann 1982), primarily as a preliminary to freeze-fracture methods
ough there is no reason why it should be so restricted. The freeze-frac~
there is little reason to believe that optimised plunging into liquid propa
(as described in §2.4. la) cannot produce equally acceptable re~ults: T
asp~cts of the method are dealt with more fully in Chapter 6 and only
apparatus required is both simpler and cheaper and those startmg m
eezmg technique will be discussed here.
field of cryobiology are advised to experiment with such sy~tems be~
e apparatus is essentially rather simple (Fig. 2.55) and, while it can be
deciding whether they need to resort to more sophisticated devices.
vely easily constructed in the average workshop, a fully equipped form
. ercially available (Appendix 2). It comprises an artist's air brush
2 .4 .1 c Spray-freezing IS u~ed to spray a fine 'mist' of microdroplets (typically about 10 µm
er) mto a small container of liquid propane at about 83 K. The cool-
It has been repeatedly stressed that, to obtain the highest cooling velocif
are extremely high and, by small modifications to the size of the
the specimen size should be minimised. With emulsions and hom?gen
r orifice, it is possible to spray cell fractions, micro-organisms and
suspensions of small particles, it is possible to freeze very small microd cells.
lets (10-50 µm diameter) which will be representative of the bulk sa.
originators of the spray-freezing method used a Grapho air brush
The most effective way of producing such droplets and freezing them i
Freezing 117
116 Robards and Sleytr Low temperature methods in biological electron microscopy

8
Pressure (Pa x 1 o- >
type I at pressures between 0.5 and 1.5 atmospheres (50-150 kPa); an Aero-
1.0 1.2 1.4 1.6 1.8 2.0
graph 'Super 63' Model A-504 retouching brush operated at pressures
100
between 40 and 60 kPa has also been found suitable for atomising liquids
(Lang et al. 1976) and other combinations would undoubtedly also be
50
acceptable (Appendix 2). Ververgaert et al. (1973) modified the specimen
loading arrangements so that volumes as low as 0.05 ml could be used.
There must be enough suspending liquid to 'wrap' each isolated cell in a 20

spray droplet and so the final ratio of specimen to suspending medium is ~


(ij
> 10
important. This is usually about 1: l (medium to pellet). ·;: 1.0s
The small propane container is covered with a plate, in which is a hole :; '
CJ)
5 \ ''
slightly smaller than the top of the container: this prevents the ice, which \
'
forms from the air during spraying, falling into the cup (Fig. 2.55). Subse-
\
\
\
''
2.0s
quently, the propane is evaporated, using the vacuum of a rotary pump, so 2
4.0s
\

leaving the rapidly frozen microdroplets in the cooled block and available
0
for further processing. 1000 1200 1400 1600 1800 2000
The distance between the mouth of the spray gun and the coolant should Pressure (bar)

be 50--100 mm and the sample (typically 0.2-0.3 ml) is sprayed at intervalsc Fig. 2.56. The survival of Euglena gracilis sub. ected .
(Redrawn from Rieh?e and :
0~~~~~~~;; ; different periods of time.
0
of about 1.0 s to minimise warm-up of the coolant. Plattner et al. (197
used the spray-freezing technique successfully to cool relatively large is
lated cells: this was achieved by using low pressure (0.5 bar; 50 kPa) an (~968a,b) who found that it was possible to rod
for Paramecium (30-40 µm diameter, 120--150 µm long), by using a 50 µ diameter) ice crystals when freezin 50 I p uce ~ery small (5-10 nm
electron microscope aperture positioned in the mouth of the spray gun. . . .·.·.tion temperature fell by about 30 Kg %dg ycerol solut10ns. The crystallisa-
•··· an the speed at h' h h ·
A small modification of the spray-freezing technique allows 'jet-freezin ····.formed decreased by a factor of 103 . w IC t e ice crystals
whereby a continuous jet (rather than individual droplets) of suspension One of the major problems in the d . .
propelled into the coolant (Fig. 2.55). A simpler 'gun' is required and t st biological cells are very ra 'di ~1gn of a smtable apparatus is that
jet is formed by using a pressure of 1.0--2.5 bar (100--250 kPa) to propel t ures (Moor and Hoechli 197~~;. ~maged by exposure to such high
suspension through an electron microscope aperture of between 10 and ply the pressure for only a ·1'1· ig. d.56) and, therefore, it is necessary
m1 1secon or so beD th .
µm. It is necessary to move the jet about over the surface of the cool ences. Riehle (1968a b· Riehl dH . ore e coolmg process
as it is used so that undue warming of the coolant is minimised. tus which was subs~q~ently ~aondifie~echl! 1973) described a suitable
980) to study the influence of hi h and has be~n used by Moor et
tissue. Although this a ~ pressure freezmg on mammalian
2 .4 .1 d High pressure freezing . pparatus is commerciall ·1
s expensive; it is also unlikely t b . h' y avaI able (Appendix
As described in §2. l .2, when water is subjected to a pressure of about 2 s to construct a copy of tho ew b!I.t m the capability of most labor-
e pu Ished vers· Th · .
bar (2.1 x 108 Pa), its freezing point is lowered to about - 22°C (251
4
K). eless potentially so important that 't d . I~n. . e techmque is
critical cooling rate for vitrification is then reduced to 2 x 10 Kr ides only curre t "' I s escnpt10n IS warranted here.
n or 1orseeable p 'bT
water and to about l x 102 K s- 1 for biological specimens. Thus, if it is diameter) ice crystals at depth f ossI I tty of .pr~ducing small
sible to construct an apparatus that allows cells to be subjected to this original version of th so up to 0.5 mm w1thm a specimen.
. e apparatus used a ve th· 1
pressure and then rapidly cooled, a very significant improvement in fre specimen holder while th I . ry m-wa led brass tube
e ater vers10n described by Moor et al.
should be obtained. This principle was first elaborated in detail by R
l

Freezing 119
118 Robards and Sleytr Low temperature methods in biological electron microscopy

HL
tot PC

IPA

0 HP
DV
HL

Fig. 2.57. Diagram of a high pressure-freezing apparatus. DV, dewar flask; GB, gas balloon~
HC, high pressure cylinder; HL, high pressure line; HP, hydraulic pump; IN, position of the
injection chamber (see Fig. 2.58); LC, low pressure cylinder; LL, low pressure line; PA, p LN2 fromt HL
sure accumulator; PC, freezing chamber containing a specimen sandwich (Fig. 2.58); PP, pr
sure piston; PV, pressure valve. Small circles indicate oil and dots liquid nitrogen. The er Fig. 2.58. Top, profile of a specimen sandwich for t . .
section (CS) at the upper right is of the high pressure chamber showing the narrow apertu 2.57). B, droplet of buffer solution· T t. bl k he_ high pressure freezmg device (Fig
. ' ' issue oc Below mJ· f h b ·
(NE) for gas escape. (Reproduced, with permission, from Moor et al. 1980.) o 1sopropyl alcohol (IPA) This ch b -.. . · . ' ec ion c am er (C) for the uptake
f . h · am er is mserted mto the high ·
e 1g pressure chamber (PC in Fig ) V d pressure !me at its entry into
· · ht.!Ono r·isopropyl alcohol which · 2·d57 ·1· ,an
th
m3ec . . Vz are pre ssure vaI ves for regulating the
prece es iqmd mtrogen in o d t 0 d 1
ing. (Reproduced with permi· . f r er e ay the onset of cool-
(1980) used a more conventional freeze-fracture specimen sandwich (Fi ' ssion, rom Moor et al. 1980.)
2.57 and 2.58). In this system, liquid nitrogen is used as both the 8coola
and the pressurising agent so that a pressure of 2100 bar (2.1 x l 0 kPa)
maintained while 100 ml of liquid nitrogen is forced into the specim Liquid paraffin method
chamber through two 0.3 mm diameter apertures within 0.5 s. Temperat
fall (to 173 K) and pressure rise (to 2.1 x 108 kPa) occur sim ultaneo uchheim (1972 1977) and B hh .
within 30 ms unless precautions are taken to delay cooling until pressuri ethod of rapidly freezin sm:~I ~m and ~elsch (1978) have described a
tion is complete. This is achieved by injecting a little warm isopropyl alco g them with liquid paraffi I ( 30 µm diameter) globules by emulsify-
as a pressurising medium immediately prior to the liquid nitrogen. Th rted. The technique in~~jv ce c~y~tals of less than 50 nm diameter were
about 8 0 ml f . es m1xmg about 0.5 ml of the tissue or liquid
about 2.0 ml of isopropyl alcohol is forced into the specimen chamber . o viscous paraffin containin 501 1
the pressure develops, after which the high pressure liquid nitrogen com -3% of a liquid poly I 1 g lo g ycerol mono-oleate
g ycero ester, as an emulsifying agent Th' . '
es the cooling process. This systems allows the full pressure to be achi en h omogenised in a normal 1 b . Is mixture
mni-Mixer see Ap d' 2 : oratory apparatus (e.g. Ultra-Turrax
and maintained prior to cooling but for a sufficiently short time so ' pen ix ior man f t
according to the authors, damage from pressurisation or contact wit ity and duration of the horn . .u ac urers and suppliers). The
globules are 5-10 . d~gemsat10n should be sufficient so that
alcohol is avoided. It is clear that, provided the technical complexitie µm m iameter so th· h
be overcome and cells can be shown to withstand high pressure for ed using a light · ' me mg t at can easily be
microscope. Buchheim used small quanitities of the
necessary time, this is a technique worthy of further study.
Freezing 121
Robards and Sleytr Low temperature methods in biological electron microscopy
120

homogenate on standard freeze-fracture holders pri~r to freezing in Freon 2.4.2 Freezing on cold surfaces
22 but other, better, methods of cooling could obviously als? be used. ~t
is suggested that the reason for the production of very small ice crystals is The improved hea_t transfer characteristics between hydrated biological spe-
that the differential thermal contraction of the two phases leads to the pro- cimens and solid surfaces of high thermal conductivity and diffusivity have
duction of increased pressure within the small aqueous droplets and he~ce been used by some workers in attempting to raise cooling rates (§2.2.6). If
gain from the lowering of the freezing point so brought about (as for high cold blocks are to be used, then it is sensible to chose the best combination
of block materials and temperatures. For maximum heat transfer, ultra-
pressure freezing; §2.4. ld).
pure copper cooled with liquid helium is clearly good (Table 2.10) but syn-
thetic sapphire has even better thermal properties and, as pointed out by
2.4.lf Droplet method Meisner and Hagins (1978), could well be used in place of copper. (It might
The principle of trying to reduce the specimen size so that high cooling rat~s be added that there are no immediately obvious reasons for using gold or
can be achieved has been approached in a novel way by Menold ~n_d his silver blocks in preference to copper; the thermal properties of copper are
colleagues (Menold et al. 1974, 1976; Menold and Luttge 1979!. Thi~ ~s the better than those of silver and gold at the relevant temperatures (Table 2.10)
so-called 'droplet method' in which a droplet of a ~ispersion is positioned but it is essential that the copper is ultra-pure and that its surface is absolu-
between the ends of two wires and is then frozen (Fig. 2.59). _such a syst~m tely clean and untarnished, otherwise the beneficial properties will be greatly
again allows maximum cooling rates to be achieved in relation to the size reduced.) For metals, the heat transfer characteristics improve dramatically
with falling temperature. Thus it is generally preferable, for obtaining the
of the specimen. . .
Menold et al. (1976) describe a proprietary device that allows a pai~ of fastest possible cooling rates, to use such cold blocks closer to the tempera-
droplets to be positioned on the ends of wires so that they can be rapidly ture of liquid helium than of liquid nitrogen. The work of Bald (1983) sug-
frozen. The wires should just be touching (Fig. 2.59) and should be of l?w gests that the optimum temperature for rapid cooling on a cold copper
thermal conductivity so that the specimen is not warmed by condu_ct10~ block is 19 K (§2.2.6).
once it has been frozen. It is recommended by the auth~rs that freezmg is Before deciding to use such a method, some thought should be given to
carried out in liquid Freon 13 at approximately 93 K b~t it may be assumed the balance between advantages and disadvantages. High cooling velocities
that liquid propane would improve the cooling rate still further. Although can undoubtedly be achieved but, as in all other cooling methods, the bene-
fit will only be apparent within the first 10-20 µm depth from the surface
this technique has primarily been used as a preliminary to freeze-fracture
of hydrated biological specimens (Table 2.11; §2.3.2). Furthermore, the heat
methods, there is no reason why the cooling procedure should not be used
extraction is from one side only and all the heat of the specimen (and its
prior to other subsequent steps. support) must move across this one boundary. For ideal cooling, it is essen-
tial that the wet, deformable specimen is brought extremely rapidly into very
Fractured drop
Drop in a coolant ?~ood, close contact with the surface of the cold block. This must inevitably
Drop between
two wires
· · an that some degree of surface deformation is induced. The fact that the
pidly cooled surface is essentially flat can be most useful if the frozen spe-
en is to be sectioned on a cryoultramicrotome because it is then possible
cut sections within the plane of the well-frozen layer. All such consider-
ons must be taken into account before deciding on the cooling method
e used. However, if it is concluded that cold block cooling is the techni-
Fig. 2.59. Droplet freezing method. A small droplet is suspended between the ends ~ t
0
to be followed, then a number of practical methods have been de-
closely adjacent wires. High cooling rates are obtained because the ~roplet ca~ be ver~ r bed.
and has comparatively little metal 'supporting' matenal. This techmque is usually use P
to freeze-fracturing (see §6.3.3a). (Redrawn from Menold et al. 1976.) he first practical use of cold block freezing was made by Eriinko (1954)
Freezing 123
122 Robards and Sleytr Low temperature methods in biological electron microscopy

and this was followed by experiments by Van Harreveld and Crowell (1964) ~~~ ~~~ s:;~m~n (in the ~rigin~l description, a frog muscle) is attached to
. p unger .device which moves rapidly towards the cold block
and Van Harreveld et al. (1965), who give detailed practical instructions on
so thrustmg the specimen against the co Id sur1ace
"' Van Ha Id 1,
the operation of such a method using sub-cooled liquid nitrogen to cool a (1974) Idused a rate ofO.l m s-'with a 20 d . . rreve et a.
silver block to about 67 K. Christensen (1971) used a similar system for . mm eep well immediately above
t he co surface, this well being filled with th e co Id h e1·mm gas (Fig.
. 2.60).
freezing pancreas and liver prior to frozen thin sectioning. Van Harreveld
et al. (1974) later described an improved version of their device which over-
came some earlier difficulties (Fig. 2.60). In this apparatus, the highly Muscle
polished silver surface is protected from condensation of moisture and gases Nerve

by a flow of cold helium gas. The specimen passes through this gas layer I

sufficiently rapidly so that pre-cooling does not affect physiological process- ~~Liver
I
~.i____ Filter paper
I

Specimen rod supports


~Plastic ring

Tongue
Well
0-""•'"'"• ' "
Tube
Rod
Threaded
Specimen brass cuff
aoomm
rod
I
~c'""'"
Silver
50mm rod Iron
ring

a
slot
b
l
Helium gas tube
Piston
Support plate
.~xpenments Detailed
<Fig. 2.61. view of specimen mountin
of Heuser et al. ( 1979). (For the ur g
arran em
g. .en ts for the cold block freezing
;isshown 'upside-down' (In pract1'c th .P poses ofth1s 1llustrat10n, the freezing assembly
> ) • · e e specimen would face d d ) .
IS seen on the freezing head ready t b f T . ownwar s. The specimen (mus-
impact with the copper blo~k and to ~. rozen. he tissue hes on a piece ofliver to cushion
Barrel . o raise its surface above th 1 · ·
as ip of filter paper glued to an al . . d' h e P ast1c nng. The liver sits
·1 ummmm 1sc s aped t fit th
ng machine. The filter paper absorbs some . o ,1 on .e cold stage of the freeze-
muscle. When the Ringer's l t' f of the Rmger s solut10n that clings to the liver
· . so u 10n reezes it bo d th r
mm disc. The plastic ring preve t th n s e iver and muscle tightly to the
Needle knife in the freeze-etching unit ~ sth' e muscle from being flattened beyond the reach
valve c
iately prior to impact through t.hn 1 IS particular example the nerve can be stimulated
· . ea ummmm wifes wh· h h d .
Fig. 2.60. Apparatus used by Van Harreveld et al. (1974) for the rapid freezing of tissue e time of impact the free . h d 1 IC are attac e to a stimulator
. ' zmg ea te escapes a · t th · ·
a cold block. Diagram (a) shows the general arrangement of the damped plunger device. on nng hits the electromagnet d' gams e spnng mounted inside it until
tails of the cold block are shown in (b): the polished silver surface is cooled to liquid nitro 'ng head from bouncing away f sur~~un mg the copper block. The magnet prevents the
temperature and purged with helium gas to avoid contamination: the glass protection P thin layer of foam rubber t ro;:i e copper block after impact. The iron ring is covered
o cus wn its impact with the magnet. (Redrawn fro H
is removed immediately prior to cooling. The general layout of the Dewar vessel and cold bl et al. 19 _) m euser
79
is shown in (c). (Redrawn from Van Harreveld et al. 1974.)
124 Robards and Sleytr Low temperature methods in biological electron microscopy
Freezing
125

~-...__----:;f:'{;:=::;;;:;
(:_____, o STARTER
Such a depth was considered to be the minimum compatible with protecting
the surface from contamination while not leading to unacceptable specimen
pre-cooling.
Boyne (1979) used a modification of the Van Harreveld apparatus to

Vi\~~ I
SOl..Ef()lD
freeze thin isolated stacks of the electric organ of a fish. He paid consider-
able attention to one of the major practical problems of cold block devices:
PWNGER ON
Ir··"···
•. . . . /... . . .,.,.,. .,._,.,.
ROLLER BEARINGS-...__
l TRIGGER FOR
STIMULATOR
optimising the force with which the specimen is thrust against the cold
block. Excessive force deforms the specimen and, if no precautions are

'-r-T---0p-=--
taken to avoid it, the freezing specimen may bounce away from the cold sur-
0 face, so giving inconsistent cooling. To overcome these difficulties, it is usu-
---.-r al to mount the specimen on a small piece of energy-absorbing material,
such as foam rubber and/or even a thin slice of liver tissue (Fig. 2.61); this
/' """::·
' ' protects the specimen from being pushed too forcefully against the cold
block. The problem of bounce has been overcome in the Heuser device (Fig.
2.62) using a 'magnetic catch' that locks-on and prevents the specimen
assembly from retracting once it has hit the cold block. Alternatively, Boyne
(1979) eliminated bounce by using a hydraulic-pneumatic damping system
on his plunger (see also Phillips and Boyne 1984).
Authors subsequent to Van Harreveld and his colleagues have described
various modifications to the orginal cold block apparatus. Some (for exam-
ple, Dempsey and Bullivant 1976a; Heath 1984), use extremely simple meth-
ods while others (e.g. Heuser et al. 1976, 1979) have described more sophisti-

Specimen

Teflon
specimen support

Wooden 15mm
diameter dowel

A simple cold block freezing method. The copper block is cooled in liquid nitrogen
c:::====.....:....------:-::---:---:-=.
f th
:::-:d;:.
ld block freezing device es1gne
d b Heuser et
Y
hich, just prior to freezing, it is lifted so that its surface is just clear. The specimen is
Fig. 2.62. General arrangement·~ bl e(lo endix 2). (Figure kindly supplied and reprodu nnly thrust against the highly polished surface of the cold copper block. (Redrawn from
(1979) and now commercially ava1 a e pp an original illustration in Dempsey and Bullivant 1976a.)
by courtesy of J. Heuser.)
Freezing 127
Robards and Sleytr Low temperature methods in biological electron microscopy
126

·vant (l 976a) technique must certainly be the sim- Heuser et al. (1976) (also Heuser et al. 1979), Escaig et al. (1977) (also
The Dempsey an d Bulll d Escaig 1982) and Sitte et al. (1977) are among those who have described
. 2 63) but the results demonstrate compare
plest cold block meth o d (F ig. · · · more elaborate cold block freezing devices. The system devised by Heuser
well with those achieved using more elaborate methods: Theh spec1ren i~
. . . u with tweezers and quickly pushed agamst t e sur ace o and his colleagues has produced some particularly elegant results and is now
simply p1c~ed dp bl k that has J·ust been lifted above the surface of commercially available (see Appendix 2). It utilises a guided free-fall slider
a cold, pohshe copper oc . . I mechanism whereby the specimen is thrust against the copper block which
. . . . . h. h it has been cooled. Even with such s1mp e appara-
hqmd mtrogen m w 1c Id is cooled with liquid helium. A shutter opens as the specimen falls, thus
the benefits of the higher rate of heat transfer_ across_ the co . copper
tus, 1· d However the depth of tissue with small ice crys- overcoming problems of condensation on the block face as well as of pre-
surface at 77 Kare rea JSe · ' . cooling in a cold gas layer above the specimen. In the particular application
. . 11·m1.ted by the properties of the specimen to about 12 µm, a
ta1s is agam . . · of Heuser et al. (1979) the cold block device was used to freeze neuromuscu-
distance comparable to that obtained with hqmd cryogens.
lar junctions of frog muscle where stimulation had been triggered by the
closing of a microswitch as the specimen descended towards the cold sur-
Spring 1 - - - :
face. Thus, it was possible to control very precisely the time interval between
stimulation and rapid freezing, a unique achievement.
Electromagnet Rubber foam
The device of Escaig et al. (1977) (Fig. 2.64) also uses a copper block
A6~===30Z--Filter paper
?--:::44----Specimen
cooled with liquid helium. In the latest form of this apparatus (Escaig 1982),
Electromagnetic plunger--1-1;*'1-+- the specimen is projected against the block by means of an electromagnet
and there is a shutter to protect the block from contamination.
I ,
The system of Sitte et al. (1977) provides a simple and neat method gf I

using a copper cold block at liquid nitrogen temperature. In this device, the
specimen is allowed to fall against the cold block. The authors point to the
particular advantages of being able to obtain large areas of relatively flat,
well-frozen tissue for cryoultramicrotomy.
High purity
copper block -+-==-t- Any of the commercially available cold block devices should give satisfac-
ry results (suppliers are listed in Appendix 2) although the theoretical as-
cts of this technique are still being evaluated (Bald 1983).

t---4
2omm Comparison offreezing methods

. . db E ai (1982) In this system a high the rapid freezing methods described above have been shown to produce
Fig. 2.64. Cold block freezing devi~; t~~~c~1~e wh!ch :cre:ulated ~upply of liquid helium
copper block is fixed ms1de a c.hamb d gth ample holder which is fixed to the tip
!lent results, in terms of small ice crystal size, in some specimens. There
· n is mounte on e s e been arguments about the relative merits of the different methods but
!ates (A to B). T h e specime . . 2 4 I The vacuum chamber is pumped out t
electromagnetic plunger (see Fig. 2.510 m§ .. a). t d to the device. Liquid helium then information available tends to support the view that small specimens
,p d th, transfer me is connec e . h d
than 1.3 x 10-" a an ~ . When the required temperature is reac e '
the device down to 6.0 K m less than 2.0 m1~.the stem insert, so allowing cold helium gas be well-frozen by any of the methods while large ones will, with any
ration of the electropneumat1c system open . ached spring (2) pulls the shutter od, only be well-frozen to a maximum depth of about 20 µm. This is
Wh tmosphenc pressure is re •
the vacuum cham b er. en a . h. then stops the electropneumatic system damental limitation imposed by poor heat transfer through hydrated
at the end of its travel, operates the swfl1tch. Tf his!. as) and also activates the electro
· . d tops the ow o e mm g
closes the stem msert an s .h h r block) When the electromagne
t cal specimens. Cryoprotection (§2.3.2a), or high pressure freezing
(which brings the specimen into contact wit t ehcohppsethen st;red in liquid nitrogen. (Ill d), offe': the only possibilities for decreasing ice crystal size in large
. d · (1) frees the specimen w ic i .
to be activate , spnng . t t. kindly provided by J. Esca1g.) ens. The choice of freezing technique therefore depends more on the
tion redrawn from an i11 us ra 10n
Freezing 129
128 Robards and Sleytr Low temperature methods in biological electron microscopy

Lumen
nature of the specimen and what is to be done with it once it has been fro-
zen, than on the efficiency of the technique itself (Table 2.14; §2.4).
Final seal
'W
In the absence of any other criteria, it is best to start freezing tissues using
rapid plunging methods. These can be established in any laboratory using
simple and inexpensive apparatus. For many purposes they are be totally
adequate. Spray-freezing is excellent for 'liquid' specimens and gives an
extremely short time delay between 'sampling' and freezing whereas, in all
Tissue
compartment 8 D 0,
the other methods, specimen sampling and mounting can be rather pro-
,_. :D.
I
T d I
A
tracted. The droplet and liquid paraffin methods are also suitable for liquid I t:l q
'd'
I I I I I
Seals
specimens. Propane jets seem to offer advantages as much in terms of speci- ,_,
I I I
I
I
I
I
I
I
I
I
L,
I

I,_,
I I
men handling (thin metal sandwiches) as in absolute cooling rates: this tech- Weighted ,_, ,_,
I I
I
I
I
I
compartment
nique is thus particularly appropriate as a preparation for freeze-fracture
methods. High pressure freezing should provide a very significant improve-
ment in freezing quality but apparatus remains to be made generally avail- 1--i t-------l >---I
10mm 10mm 10mm
able. Although cold blocks offer theoretical advantages over most of the
a b c
other methods and some dramatic applications have been demonstrated
plastic laminate (Appendix 2) which can be heat sealedm:e-~ature_s.
(e.g. Heuser et al. 1979), their general usefulness has been less well docu- Fig. 2.65. Storage containers for specimens at low te
1
(a) .A sachet made from a
mented (e.g. Table 2.11; §2.3.2). It hardly seems worthwhile using cold ment_under liquid nitrogen. (Redrawn from A leton 19 : 7 e mamtammg the tissue compart-
blocks unless at liquid helium temperature (Bald 1983) or unless the benefit alummmrn where a hole in the lid can be ,. dpp d' _ .) (b) Simple contamer made from
, · . m exe to align with st
A d1p-stick'-type storage container for use with riqm.d mtrogen
. Dewarorage compartments.
vessels. (c)
(Redrawn from
of having a relative flat frozen layer (as for cryoultramicrotomy) outweighs
Krah et al. 1973.)
the additional complexity. Alternatively, the requirement for a precise time
course of freezing in relation to cellular activity may indicate the usefulness
of cold block freezing. Thus, the individual microscopist must ultimately
make his own choice of method but the most helpful advice is probably, f I~ many ins_tances, the frozen specimen will be used immediately for
urt er processmg. ~t may well have been frozen on a specimen support that
initially, to 'keep it simple'! ~an be mounted directly onto a cryoultramicrotome or a freeze-fracture
apparatus. However, it is often useful to be able to store a specimen"
at some f u t ure t'ime. Al t h ough the principles of storage are simple itior use
is im-
2 .5 Subsequent processing and storage of specimens
firt~nt t~at the technical aspects are fully considered so that no d~leterious
ec s ~a1 e place at this stage. For most purposes storage under liquid ni-
When the specimen has been frozen, it must be removed from the apparatus
g~n IS ooth saf~st and easiest. A number of suitable containers for stor1'ng
and then either used directly for further processing or stored for use at som
c1mens und er liqm'd mtrogen
· have been recommended (Fig. 2.65). Apple-
future time. As stressed previously, once the specimen has been frozen ·
(1967) .used a heat-sealed plastic/aluminium foil bag for cryoultramicro-
must always be handled in such a way that unnecessary temperature i
i:e~p~c1mens while a typical container for freeze-etch specimens is de-
creases do not occur. In particular, unless the protocol of the technique d
y Krah et al. (1973) and Williamson (1984) has also described a
mands it, the temperature should not be allowed to rise above about 90
u1) system
d . for stor·mg specimens
· · hqmd
m · . nitrogen. Rigler and Patton
the temperature above which it may be anticipated that some ice crys
4 estchnbed another 'cryocell' storage container. The important fea-
growth may occur. Thus, it is important that pre-cooled tools are used a
that transfers are accomplished as quickly as possible. These precautions a rareld at the . .specimen
· sh ou1d be transferred into the container while
co cond1t10ns (preferably under liquid nitrogen), that the container
particularly relevant when using very small, thermally sensitive specimens,,
130 Robards and Sleytr Low temperature methods in biological electron microscopy Freezing 131

should itself be protected from warming while being transferred into the roscopists (and micro analysts) ha .
that 'look good' rather than w"th vthe, m gheneral, preferred to work with cells
storage Dewar vessel and that, similarly, precautions should be taken to ose t at can be dem t t d .
. ons ra e to ~urv1ve
I

avoid heating the specimen when it is removed from the nitrogen and trans- the freezing process. This should act as a ca .
of existing techniques and perh ut10n agamst the unheedmg use
ferred to the next stage of whatever process is being used. It is also impor-
what we are trying to achieve i~;, ca~se u~ to pa~se to think about exactly
tant that the containers should be indelibly and unambiguously labelled.
Some felt-tip markers are stable under liquid nitrogen; engraving on metal
use it in one of the methods descri:=~~~~hte :olslpec~menChbefore
owmg
1' apters.
going on to
is safe and permanent for containers that are frequently used; and some
adhesive tapes are satisfactory for low temperature use.
It is possible to purchase ultra-low temperature deep freezes that main- References
tain temperatures of approximately 170 K (Appendix 2). These may be safe
Adrian, M., J. Dubochet, J. Lepault and A M D II
for storing some specimens but they should be used with caution. Their low- viruses, Nature, Lond. 308, . · c owe (1984), Cryo-electron microscopy of
32
est temperature is close to the expected temperature for ice recrystallisation Aist,
. c J.R.d (I 974), A freeze-etch st u d Y o f membranes m . pla d" h ·
m1ecte cabbage root hairs, Can. J. Bot. 52 smo !Op ora-mfected and non-
and it is advised that they should only be considered 'safe' in this respect 1441
Allen, E.D. and . L. Weatherbee (1979) ' UI trastructure
' . of red II f .
if rigorous tests on the particular specimens have been carried out. stare h , J . Microscopy 117, 381. ce s rozen with hydroxyethyl
Allen, E.D.
c and
. . L. Weatherbee (1980) , Ul trastructural appear f d
cero l 1or chmcal use, Cryobiology l7, . ance o re cells frozen with gly-
448
Allen, J.V., W.M. Hess and D.J. Weber (1971) Ult . .
2.6 Cooling and viability after thawing tia caries teliospores, Mycologia 63, 144. ' rastructural mvestigations of dormant Til/e-
Ananthanarayanan, V.S. and C.L. Hew 0977 A . .
ty, Nature, Lond. 268, 560. ), synthetic polypeptide with antifreeze activi-
The subject of cooling rates in relation to structure and viability has already
Appleton, T.C. (1967), Storage of frozen materials f . .
been discussed ih §2.3. l but we should not leave the critical topic of freezing pound autoradiography, J. Microscopy 87 489 or cryostat sect10mng and soluble-com-
biological specimens without again briefly pausing to contemplate the inter- Aranc1a, G., F. Rosati Valente and p Troval' ..19
on freeze-fractured Escherichia colt J M. usc1 ( 79), Effects of glutaraldehyde and glycerol
relationships between cooling, structural preservation and viability. This .A · v , · 1croscopy 118 161
rpaci, · (1966), Conduction heat transfer (Add· ' ·
topic has been well reviewed by a number of authors but the paper by Far- A.urich, F. and Th. Foster (1984) T tson-Wesley Publ. Co., London).
. . , emperature measurement d . .
rant et al. (1977) is particularly germane. These authors point to the proba- mma usrng a fluorescence label. Cryo-Letters 5, 231. unng rapid cooling in pl-volu-
bility of inducing toxic changes or structural alterations by using cryopro- pachmann, L. and W.W. Schmitt (1971) Im roved .
Proc. natl. Acad. Sci. U.S.A. 68, 2149.' p cryofixat10n applicable to freeze-etching,
tectants at high enough concentrations to prevent ice crystal artefacts. Such Bachmann ' L · and w·w· Sch m1tt-Fumian
. (1973) s .
cryoprotectants are usually only protective at low rates of cooling and may Freeze-etching: techniques and applications E ' pray-freezmg and freeze- etching, in:
~1crosc. Elect., Paris), p.73. ' .L. Benedetti and P. Favard, eds. (Soc. Fr.
be positively harmful at higher rates. Lower concentrations of cryoprotec-
, W.B (!975), A proposed method for s ec1f .
tants ( < 10%) may reduce ice crystal artefacts but they are unlikely to pro- rapid cooldown of plant specimens J p Bymg the temperature history of cells during
tect cells in a functional sense unless cooling rates are low. Thus, cells which , W.B (1978 ) . ' ·exp. ot. 26, 103.
. a , A multi-purpose cryostat for freezing biological materials , C ryogemcs
. 18
are to survive upon thawing may be frozen by cooling very slowly or by
' W.B. (l978b), The thermal history of cells d . '
using a two-step method; however, these cells will also be severely shrunken;. oc. 7th. Int. Cryogenic Eng Conf L d unng the freezing of biological materials
, WB ( .. · ., on on, p.441. · ,
The microscopist is presented with a choice: to cool cells as rapidly as possi:: ·. . 1979), A cnt1cal appraisal of the ex . .
olog1cal materials, Proc. 15th Int C R i~nmental techmques used in the freezing of
ble, tolerating the presence of some intracellular ice which will almost cer , W.B. (1983) Opt· . . h · .ongr. e ng., Vemce 3, 225.
, 11. , 1m1smg t e coolmg bl oc k ior
c th e qmck-freeze
. method J M"
tainly mean that the cells would not be viable after thawing, or to cool eel 1w ··=~
slowly, causing them to become shrunken but knowing that such cells m .. B..Cl 984), The relative efficiency of er o . . .
~logical samples, J. Microscopy 134, 26[. gemc flmds used m the rapid quench cooling
well show good viability on thawing. No general guidance can be given.
b .. B..and A.B. Crowley (1979) On definin .
choice must made for each experiment carried out in relation to the ty . iological materials, J. Microsc;py 117, 395.g thermal h1s~ory of cells during the freezing
and quality of information required. It should, however, be noted that mi
Freezing 133
Robards and Sleytr Low temperatw.e me thods in biological electron microscopy
132

) Modifications to the paper "On defining the thermal Burton, E.F. and W.F. Oliver (1935), Structure of ice at low temperatures, Proc. R. Soc. Lond.
Bald, W.B. and A.B: Crowley (1_981 fb. I gical materials", J. Microscopy 122, 331. Ser. A 153, 166.
history of cells dunng the freezmg o w o er Re . Prog. Phys. 45, 1381. . Buttrose, M.S. (1971), Ultrastructure of barley aleurone cells as shown by freeze-etching,
Bald W.B. and J. Fraser (1982), Cryogemc surg fy, thprapid freezing of biological specimens Planta 96, 13.
, R bards ( 1978) A device or e
Bald, W.B. and A.W. o 'ducible conditions, J. Microscopy 112, 3. . Buttrose, M.S. (1973), Rapid water uptake and structural changes in imbibing seed tissues,
under precisely controlled and repro d putation of thermojunct10n response Protoplasma 77, 111.
l968) Measurement an com
Balko, B. and R.L. Be.rger ( ' · t Instrum. 39, 498. Carslaw, H.S. and J.C. Jaeger (1959), Conduction of heat in solids (Oxford University Press).
times in the submilhsecond range, Rev .. sc~e~f.freezing damage, J. Cell Biol. 57, 729. Chafe, S.C. and A.B. Wardrop (1970), Microfibril orientation in plant cell walls, Planta 92,
Bank H. and P. Mazur (1973), Visuahzat10 d M V k (1973) The nature of the starch 13.
' D H Simmonds an · es '
Barlow, K.K., M.S . Buttrose, · · Cereal Chem. 50, 443. Chandler, D.E. (1979), Quick freezing avoids specimen preparation artifacts in membrane
protein interface m wheat endosperm, f h. h molecular weight cryoprotectants dextran fusion studies, in: Freeze-fracture: methods, artefacts and interpretations, J.E. Rash and
Barnard, T. (1980), Ultrastructural re!~~t~r~w~!dipose tissue, J. Microscopy 120, 93: . C.S. Hudson, eds. (Raven Press, New York), p.81.
and polyvinyl pyrrohdone on hve . ff .ng tolerance in an antarctic midge, Chang, S.H., W.J. Mergner, R.E. Pendergrass, R.E. Bulger, I.K. Berezesky and B.F. Trump
and J.S. Edwards (1979), Mechamsms o reezi
Baust , JG
. · . . I 4 1. . (1980), A rapid method of cryofixation of tissues in situ for ultracryomicrotomy, J. Histo-
Belgica antarctica, Phys101. Entomo . S2) The h siology and biochemistry oflow tempera- chem. Cytochem. 28 47.
Baust, J.G., R.E. Lee and R.AdKi;;e?:erre~trial ~rihropods: a bibliography, Cryo-Letters 3, Christensen, A.K. (1971), Frozen thin sections of fresh tissue for electron microscopy, with
ture tolerance m msects an o a description of pancreas and liver, J. Cell Biol. 51, 772.
191. H ( ) Effects of lipid phase transition on freeze- Clark, J., P. Echlin, R. Moreton, A. Saubermann and P. Taylor (1976), Thin film thermocou-
M D 1 ck and E ouser 1977 ,
Bayer, M .E. , · 0 a . ." r J B cteriol. 129, 1563. ples for use in scanning electron microscopy, Scanning Electron Microscopy 1976, 1, 83.
cleaved envelope of Escherichia co l, . ~Tt f the amorphous state in the system water- Cole, G.T. (1973), A correlation between rodlet orientation and conidiogenesis in Hyphomy-
and A. Kaufmann (1979), Sta ii y o cetes, Can. J. Bot. 51, 2413.
Bout ron, P · · I 16 83 ·
glycerol-ethylene glycol, Cryob10 ogy , . bl for quick-freezing tissues for electron m1c- Cole, G.T. (1975), The thallic mode of conidiogenesis in the Fungi imperfecti, Can. J. Bot.
Boyne, A.F. (1979)'.A gentle, bl ou:~e-fre~;~;e:y ~lectrocyte stacks, J. Neurosci. Methods J, 53, 2983.
roscopy: application to iso ate orpe Costello, M. J. (1980), Ultra-rapid freezing of biological samples, Scanning Electron Micros-
353. d H M (1964) Fine structure in freeze-etched Allium cepa L. roots. J. copy l 980, 2, 361.
Branton, D. an · oor ' Costello, M.J. and J.M. Corless (1978), The direct measurement of temperature changes within
Ultrastruct. Res. 11, 401. d EB Wagenaar (!974), Ultrastructure of Chla- freeze-fracture specimens during rapid quenching in liquid coolants, J. Microscopy I 12, 17.
Bray D.F., K. Nakamura, J.W. Costerton an ·h: ll wall plasmalemma and chloroplast Costello, M.J., R. Fetter and M. Hochli (1982), Simple procedures for evaluating the cryofixa-
m;domonas eugametos as revealed by freeze-etc mg ce , tion of biological samples, J. Microscopy 125, 125.
membrane, J. Ultrastruct. Res. 97, 125Ultrastructure de le paroi des basidiospores de Lyco- Cowley, C.W., W.J. Timson and J.A. Sawdye (1961), Ultra rapid cooling techniques in the
Brouchant, R. and V. Demoulm (1971), a celle de quelques autres spores de freezing of biological materials, Biodynamica 8, 317.
perdon et de Scleroderma, Gasteromysetes comparee . Cox, G. and B. Juniper (1973), Electron microscopy of cellulose in entire tissue, J. Microscopy
champignons, Protoplasma 72, 179. C !ete vitrification in pure liquid water and dilute 97, 343.
.. II r p and E. Mayer (1980), omp 'De Quervain, M.R. (1975), Crystallization of water, a review, in: Freeze-drying and advanced
Brugge e ' · . d 288 569.
aqueous solutions, Nature, Lon ·. ' .. . Losungen Naturwissenschaften 59, 12L food technology, S.A. Goldblith, L. Rey and W. W. Rothmayr, eds. (Academic Press, Lon-
Buchheim, W. (1972), Zur Gefrierfixi~rung was:~;:: cryoprote~tants, Naturwissenschaften 59, don, New York) p.3.
Buchheim, W. (1977), Cryofixat10n o tissue wi empsey, G.P. and S. Bullivant (1976a), Copper block method for freezing non cryoprotected
121. . U Welsch (1978), Freeze-etching of unglycerinated tissue dispersions by tissue to produce ice crystal free regions for electron microscopy. I. Evaluation using freeze-
Buchheim, W. and : . h . e J Microscopy Ill, 339. :'substitution, J. Microscopy 106, 251.
application of the 01! emulsion tee nd1qhu. '_fiJ thermocouples for accurate measurement · mpsey, G.P. and S. Bullivant (1976b), Copper block method for freezing non cryoprotected
. L H (1963) Vacuum-depos1te t m 1 m tissue to produce ice crystal free regions for electron microscopy. II. Evaluation using freeze-
Bu IIis, · · ' J · t Instrum. 40, 592. ·
substrate surface temperature, . scien . t' f mouse pancreas prepared by the Fema fracture with a cryoultramicrotome, J. Microscopy 106, 261.
h · · of thm sec ions o 639
Bullivant, S. (1960), T e stammg b t't t' method J Biophys. Biochem. Cytol. 8, · psey, G.P., S. Bullivant and W.B. Watkins (1973), Endothelial cell membranes: polarity
dez-Morim Helium II freeze-su_ s '. u ion o;ti~ techniques, Lab. Invest. 14, 440. fparticles as seen by freeze-fracturing, Science 179, 190.
Bullivant, S. (1965), Freeze-substitutiofnfand.supptechniqgues in: Some biological techniques well, L.G. and A.P. Rinfret (1960), Low temperature forms of ice as studied by x-ray dif-
0) p ent status o reezmg '
Bullivant, S. (1 97 ' res d (Academic Press, New York), P· 10 1· . taction, Nature, Lond. 188, 1144.
electron microscopy, D.F. Parsons[, e..ht . ctions Proc. 9th Int. Congr. Electron Mier bochet, J. (1984), Electron microscopy of vitrified specimens, Proc. 8th Eur. Conf. Electron
Bullivant, S. (1978), The structure o tlg JUn , 'croscopy, Budapest 2, 1379.
copy, Toronto 3, 65 9 . . d. m for ultrastructural and enzymatic . s~ chet, J. and A.W. McDowall (1981), Vitrification of pure water for electron microscopy,
Bullock, G.R. (1979), Preservation of myodcar m h DJ. Hearse and J. de Leiris, eds. (W 'croscopy 124, RP 3.
. Enzymes m
m: . card'ology·
i · diagnosis an researc , · het, J., J. Lt;iault, R. Freeman, J.A. Berryman, and J.-C. Homo (1982), Electron mic-
Interscience, Chichester), P· 461. opy of frozen water and aqueous solutions, J. Microscopy 128, 219.
Freezing 135
Low temperature methods in biological electron microscopy
134 Robards and Sleytr

71) Ultrastructure of Crithidiafasciculata - a freeze-etching study, Can. Franks, F. (1982), Apparent osmotic activities of water soluble polymers used as cryoprotec-
Easterb roo k , KB· · (19 ' tants, Cryo-Letters 3, 115.
J. Microbiol. 17, 277. t techniques for scanning electron micros- Franks, F. and H.le B. Skaer (1976), Aqueous glasses as matrices in freeze-fracture electron
E chlin ' P. and .R. Moreton (1976),
.
Low tempera ure
1976 753
microscopy, Nature, Lond. 262, 323.
copy, Sclmmng Electbron Micro(~~~~y) Prep,arati.on of biological specimens for X-ray microa- Franks, F., M.H. Asquith, C.C. Hammond, H.le B. Skaer and P. Echlin (1977), Polymeric
Echlin, p. and A.J. Sau ermann , cryoprotectants in the preservation of biological ultrastructure. I. Low temperature states
nalysis, Scanning Electron Micros.copy l 9F77, I ,k62 l. d M H Asquith (1977), Polymeric cryo- of aqueous solutions of hydrophilic polymers, J. Microscopy 110, 223.
. p H B Sk er B O C Gardmer F. ran s an · · . J M.
Echlm, ., · · a ' · · ·. f b. j . ·I ltrastructure. II. Physiological effects, . ic- Frederik, P.M. and W.M. Busing (1981), Strong evidence against section thawing whilst cut-
protectants in the preservat10n o 10 og1ca u ting on the cryo-ultratome, J. Microscopy 122, 217.
roscopy 110, 239. J d. JN Chapman and W.H. Bidd!ecombe (1982), Opti- Friend, D.S. and D.W. Fawcett (1974), Membrane differentiations in freeze-fractured mamma-
Elder, H.Y., C.C. Gray, A.G. ar mhe: .
mum conditions for the cryoquenc mg o sma
f
11 tissue blocks in liquid coolants, J. Micros- lian sperm, J. Cell Biol. 63, 641.
Fujikawa, S. (1981), The effect of various cooling rates on the membrane ultrastructure of fro-
copy 126, 45. B (1972) Freeze-etch ultrastructure of Parmelia caparata (L) Ach., zen human erythrocytes and its relation to the extent of haemolysis after thawing, J. Cell
Ellis, E. A. and R. M. rown ' Sci. 49, 369.
Trans. Amer. Microsc. Soc. 9 l, 4 J 1. · eta Anat. 22, 331. Furch!, L.T. and R.E. Scott (1974), Influence of cell cycle and cell movement on the distribu-
.. k 0 (1954) Quenchmg of tissue for freeze-drymg, A .
E ran d 6 K J Micros- tion of intra-membrane particles in contact-inhibited and transformed cells, Exp. Cell Res.
o, · ' h' h c ·1·t t
. J . (! 982) , New instruments w ic iaci i a e rap id freezmg
· at 83 K an ' ·
E scaig, 88,311.
copy 126, 221. d G N' !· (1977) Congelation rapide de tissus biologiques. Mesure Furcht, L.T. and R.E. Scott (1975), Modulation of the distribution of plasma membrane intra-
· J ., G · Geraud an
E sca1g, '. ico as
d ' f par thermocouple en couch e mmce,· C . R. membranous particles in contact-inhibited and transformed cells, Biochim. Biophys. Acta
des temperatures et des vitesses e conge1a ion 401, 213.
Acad. Sci. Paris 284, 2289. . d K Cl rke (1977), Structural and function- Ghosh, B.K. and N. N anninga (! 976), Polymorphism of the mesosome in Bacillus licheniformis
Farrant, J., C.A. Walter, H. Lee, G.J. Morns, an . .1. a 17 (749c and 749): influence of chemical fixation monitored by freeze-etching, J. Ultrastruct.
. · If · t chniques J Microscopy 111, ·
al aspects of b10log1ca reezmg e h d' · d . . !es of fixation by freeze-substitution. Res. 56, 107.
Feder, N. and R.L. Sidman (1958), Met o s an prmcip Gilleland, H.E., J.D. Stinnett, I.L. Roth and R.G. Eagon (1973), Freeze-etch study of Pseudo-
593 monas aeruginosa: localization within the cell wall of an ethylenediaminetetraacetate extrac-
J. Biophys. Biochem. Cytol. 4, · t reparation techniques for electron micros-
, d M · H (1960) Low tempera ure P y A d table component, J. Bacteriol. 113, 417.
Fernan ez-. oran, · . ' 'd f · with liquid helium II, Ann. N. . ca ·
copy of biological specimens based on rap1 reezmg Glauert, A.M. (1974), Fixation, dehydration and embedding of biological specimens, in: Prac-
Sci. 85, 689. . f th f rm of vacuoles in thin sections and freeze-etch repli- tical methods in electron microscopy, Vol. 3, A.M. Glauert, ed. (North-Holland, Amster-
Fineran, B.A. (1970a), Evaluat10n o e o dam).
457
cas of root tips, Protoplasma 7o, f · . t tments on the freeze-etching of root tips, .Glover, A.J. and Z.S. Garvitch (1974), The freezing rate of freeze-etch specimens for electron
Fineran, B.A. (1970b), The effects o vanous pre- rea microscopy, Cryobiology 11, 248.
J. Microscopy 92, 85. f I ·nclusions in root tips Protoplasma 71, I. C. de and W. Leene (1979), The influence of cryoprotectants, temperature, divalent
. BA (1971) Ultrastructure o vacuo ar i ' .. f cations and serum proteins on the structure of the plasma membrane in rabbit peripheral
Fmeran, · · ' f t ·n root tips after various condit10ns o pre-
. BA ( 1972) Fracture faces o tonop 1as i blood lymphocytes, Eur. J. Cell Biol. 19, 19.
F meran, · · ' . M. 96 333
treatment prior to freeze-etchmg. J. icrosc~pyl , ·c r~ticulum and vacuoles in frozen root 'Gulik Krzwicki, T. and M.J. Costello (1978), Use of low temperature X-ray diffraction to eva-
Fineran, B.A. (1973a), Assoc1at10n between en op asmi luate freezing methods in freeze-fracture electron microscopy, J. Microscopy 112, 103.
tips, J. Ultrastruct. Res. 43' _7 5.. f th Golgi apparatus in frozen etched root tips, Cyto- pta, B.L. (1979), Electron microprobe X-ra.y analysis of frozen hydrated sections with new
Fineran, B.A. (1973b), Orgamzat10n o e information on fluid transporting epithelia, in: Microbeam analysis in biology, C.P. Lechene
biology 8, 175. h. d the ultrastructural morphology of root tip cells, Phy-· d R.R. Warner, eds. (Academic Press, London) p. 375.
Fineran, B.A. (1975), Freeze-etc mg an , J.A. (1966), The measurement of temperature (Chapman and Hall, London).
tomorphology 25, 398. . . . copy and cytochemistry of plant cells,· D.M. (!967), Wax microchannels in the epidermis of white clover, Science 158, 505.
. B A (1978) Freeze-etchmg, m: E 1ectron micros ley, D.A., J.T. Alexander and S. Chien (1981), Design and use ofa simple device for rapid
F meran, · · ' d ) 279
J ·L · Hall ' ed. (Elsevier/North-Holland,
.
Amster am 'p.
f bb"t
. . . 19 59
matozoa J Microscopie , · ch freezing of biological samples, J. Microscopy 121, 273.
Fl ech on, J ·E · (1974) ' Freeze-fracturing o . ra f i· sper . ' · . .
(C b idge Umvers1ty Press .
) •.· en, 0.G. (1971), Cryogenic fundamentals, (Academic Press, New York).
Fletcher, N.F. (1970), The chemical phdysiVcsMo icRe amh ~ Aebi U. Tschope and B. Meng D.L. and E.D. Hay (1977), Freeze-fracture studies of the developing cell surfaces. II.
th M K G Lickfiel . . eusc ' . ' . . l
Fooke-Ach terra ' ., . . . ' h I reus mesosomes for transmiss10n e ec 'cle-free membrane blisters on glutaraldehyde fixed corneal fibroblasts are artefacts, J.
(1974) Close-to-life preservat10n of Stap y ococcus au Biol. 78, 756.
tron n::icroscopy, J. Ultrastruct. Res_. 49,~7(~975) Gap junctions in the Hemotrichorial p .D. and D.L. Hasty (1979), Extrusion of particle-free membrane blisters during glutaral-
Forsmann, W.G., J. Metz and D. Hemnc ' e fixation, in: Freeze-fracture: methods, artefacts and interpretations, J.E. Rash and
centa of the rat, J. Ultrastruct. Res. 53, 374.
. . f · d yofixatwn J
. M.
1croscopy
Ill 3
, · ·Hudson, eds. (~.;:aven Press, New York), p. 59.
Franks, F. (1977), B10logical reezm~~al c:nd hysioio~ical effects of low temperatures a
Franks, F. (1980), Physical, bioche I bl I ~ers Scanning Electron Microscopy 1980,
. freezing: modification by water so u e po Y '
349.
II'
Freezing 137
Robards and Sleytr Low temperature methods in biological electron microscopy
136
Kopstad, G. a.nd A. Elgsaeter (1982), Theoretical anal s· . .
Hayat, M.A. (1981), Fixation for electron microscopy (Academic Press, New York). impact freezmg and liquid-jet freezing off h y r_s of specimen coolmg rate during
Hearse, D.J., D.M. Yellon, D.A. Chappell, R.K.H. Wyse and G.R. Ball (1981), A high velocity Korber, C and M.W. Scheiwe (1977) Er·c treezfe-hetcd specimens, B10phys. J. 40, 163.
· · ' iec s o Y roxyethyl sta h d.
impact device for obtaining multiple, contiguous, myocardial biopsies, J. Mol. Cell. Cardiol. diagram and rts influence on the free · re on so mm-water phase
. zmg process of cells Cryob· I 14 70
Koun, J. (1971),. Ultrastructure of BHK 21 ceIIs as revealed
, by IOfreogy t, h. 5. .
13, 197. met o d s, J. Mrcroscopie JJ, 331. eze-e c mg and fixation
Heath, LB. (1984), A simple and inexpensive liquid helium 'slam freezing' device, J. Micros- h
copy 135, 75. Krah, S., L.A. Staehelin and E. Nettersheim 0973) A
Hebant, C. and R.P.C. Johnson (1976), Ultrastructural features of freeze-etched water con- etch specimens, J. Microscopy , . ' new type of storage container for freeze-
99 349
ducting cells in Polytrichum, Cytobiologie 13, 354. La;;4LB. (1925), Freezing points of glycerol and its aqueous solutions, Ind. Eng. Chem. 17,
Heide, H.-G. and E. Zeitler (1984), Water in cryomicroscopy, Proc. 8th Eur. Reg. Conf. Elec-
tron Microscopy, Budapest 2, 1388. Lang, R.D.A., P. Crosby and A.W. Robards (1976) An . . .
Hereward, F.V. and D.H. Northcote (1973), Freeze-fracture planes of the plasmalemma of preparmg specimens for freeze-etchr.ng J M. , mexpensrve spray-freezmg unit for
, . rcroscopy, 108, IOI.
some higher plants revealed by freeze-etch, J. Cell Sci. 13, 621. Lange, N.A. (1961), Handbook of chemistry, 10th edition (M G ff!! N
Hess, W.M., J.L. Bushnell and D.J. Weber (1972), Surface structures and unidentified organel- Lefort-Tran,
. M., T. Gulik ' H · Plattner, J . Berssen
. an d W Weis ' c raw 1 , 1 fl
(1978) ew York).
les of Lycoperdon perlatum Pers. basidiospores, Can. J. Micro biol. 18, 270. at10n procedures on organisation and part.t. 1 10n o f mtramemb
. · sner , ·n 1 uence of cryofix-
Heuser, J.E. and T.S. Reese (1976), Freeze substitution applied to the study of quick frozen Congr. Electron Microscopy Toronto rane partrc es, Proc. 9th Int.
2 146
synapses, J. Cell Biol. 70, 357a. Le!bo, S.P., J.J. McGrath and E.G. Cravalho {i978) M. . . .
Heuser, J.E., T.S. Reese and D.M. Landis (1976), Preservation of synaptic structure by rapid rce formation in unfertilized mouse ov f . , rcroscoprc observat10n of mtracellular
Leonhard, R. and C. Sterling (1972) Fa as a ~~ct~on of cooling rate, Cryobiology 15, 257.
freezing, Cold Spring Harbor Syrop. Quant. Biol. 40, 17. Res. 39, 85. , , reeze-e c e surfaces m potato starch, J. Ultrastruct.
Heuser, J.E., T.S. Reese, M.J. Dennis, Y. Jan, L. Jan and L. Evans (1979), Synaptic vesicle
exocytosis captured by quick freezing and correlated with quanta! transmitter release, J. Cell Lickfeld, K. G. (1968), Der frostgeriitzte B kt . k . .
tiiirstruktur, Z. Zellforsch. 88, . a enen ern. Em Bertrag zur Kliirung seiner Ter-
Biol. 81, 275. 560
Heverly, J.R. (1949), Supercooling and crystallisation, Trans. Amer. Geophys. Union 30, 205. Lickfeld,. K.G. (1973) ' Die gefriera··tzung von Ba k tenen
. . Meth 0 d
rn- I
Higgins, M.L., H.C. Tsien and L. Daneo-Moore (1976), Organisation of mesosomes in fixed nenm1kroskopie, G. Schimmel and W Vo ell ed '. · ensamm ung der Elektro-
Lonsdale K. (1958) Th t . · g ' s. (Wiss. Verlags. Ges., Stuttgart) p. 1.
and unfixed cells, J. Bacterial. 127, 1519.
James, R. and D. Branton (1971), Correlation between saturation of membrane fatty acids and
Lusena CV 1960, I es ructure ofrce, Proc. R. Soc. Land. Ser A. 247, 424.
N. Y'. A.cad( Sci. ~5,c~fr_opagat10n m glycerol solutions at temperatures below -40oc, Ann.
presence of membrane fracture faces after osmium fixation, Biochim. Biophys. Acta 233,
504. Luyet, B.J. (1960), On various phase transitions occurr· . .
Jehl, B., R. Bauer, A. D6rge and R. Rick (1981), The use of propane/isopentane mixtures for •.. tures, Ann. N. Y. Acad. Sci. 85, 549. mg m aqueous solut10ns at low tempera-
rapid freezing of biological specimens, J. Microscopy 123, 307. ;J;u~et,_ B.r(1961), A method for increasing the coolin rat . . . . .
Johnson, R.P.C. (1973), Filaments but no membranous transcellular strands in sieve pores i hqmd mtrogen or in other boiling bath B. d .g e m refngerat10n by 1mmers10n in
s. 10 ynamrca 8 331
freeze-etched translocating phloem, Nature, Land. 244, 464. uyet, B.J. (1966), Anatomy of the freezin roe · '· · .
Johnson, R.P.C. (1978), The microscopy of P-protein filaments in freeze- etched sieve pores\ Merymman, ed. (Academic Press, New~ ~rk), e;~ ;~f hysrcal systems, m: Cryobiology, H.T.
yet, B.J. (1970), Physical changes occurrin · f . .
Planta 143, 191.
Jones, G.J. (1984), On estimating freezing times during tissue rapid freezing, J. Microscop ing, in: The frozen cell GE W Wol t h ~ m rozen solut10ns durmg rewarming and melt-
London), p. 27. ' . . . sen o me and M. O'Connor eds. (J. and A. Churchill,
136, 349.
Kachar, B., J.A. Serrano and P. Pinto da Silva (1980), Particle displacement in epithelial yet,191.
5, B.J. and F . Gonzales (1951) , R ecord'mg ultra rapid changes in temperature, Refrig. Eng.
membranes of rat prostate and pancreas induced by routine low temperature fixation,
Biol. Int. Reports 4, 347. ryet, B.J. and C. Kroener (1960) Th h' h . .
Kappen, M.M. (1979), Protective agents with respect of freezing tolerance of leaves. Investi Meeting Amer Biophys oc Ph, ·1 de I rhg' est obtamable coohng velocities. Proc. 4th Ann
s'

BJ · · ., ra ep ra,L7. · .
tions with the halophyte Halimione portulacoides, Oecologia 38, 303. .. and C. Kroener (1966), The temperature of th " I .. , .
Karnovsky, M.J. (1965), A formaldehyde-glutaraldehyde fixative of high osmolality for ns of glycerol and ethylene glycol B. d . 1 e g ass trans1t10n , m aqueous sol-
B · 10 ynamrca O 33
.J. and D.H. Rasmussen (1967) Stud b d.f , ..
of instability in rapidly cool d ' ( y ~ ferentral thermal analysis of the tempera-
in electron microscopy, J. Cell Biol. 27, 137A. 1
Kitajima, Y. and G.A. Thompson (1977), Tetrahymena strives to maintain fluidity inte 1 wns o polyvinylpyrrolidone, Biodynamica JO 137
.I. and DH R e sou
tionships of all its membranes constant, J. Cell Biol. 72, 744. · · asmussen (1968) Stud b d.fD · , ·
Knoll, G., G. Oebel, and H. Plattner (1982), A simple sandwich-cryogen-jet procedure of instability of rapidly co d ' . y Y 1 erentral thermal analysis of the tempera-
.
iodynamica JO, 167. o1e so 1ut10ns of glycerol 'eth YIene g1ycol, sucrose and glu-
high cooling rates for cryofixation of biological materials in the native state, Proto B
111,161. B.J. and D. Sager (1967) On thee ·
Koehler, J.A. (1970), Freeze etching study of rabbit spermatozoa with particular referenc ly cooled solutions ofpolyvi l xrsltdence of.two temperature ranges of instability in
BJ c K ny pyrro 1 one, Brodynamica JO 133
head structures, J. Ultrastruct. Res. 33, 598. .. , . roener and G. Rapatz (1958) D . , .
Koehler, J.A. (1972), Freeze etching technique, in: Principles and techniques of electron er mixtur-'s, Biodynamica 8, 73. ' etect10n of heat of recrystallisation in glycerol-
roscopy, Vol. 2., M.A. Hayat, ed. (Van Nostrand Reinhold, New York), p. 53.
138 Robards and Sleytr Low temperature methods in biological electron microscopy Freezing 139

MacKenzie, A.P. (1969), Apparatus for the partial freezing of liquid nitrogen for the rapid Mishima, L., L.D. Calvert and E. Whalley (1984) 'M If . ,
method of making amorphous solids N t L' e mg ice I at 77K and 10 kBar a new
cooling of cells and tissues, Biodynamica JO, 341. Monroe R G w J G bl ' a ure, ond. 310, 393. ·
MacKenzie, A.P. and D.H. Rasmussen (1972), Interactions in the water-PVP system at low , . . ., . . am e, C.G. La Farge, R Gambo CL
S. Bulhvant (1968), Myocardial ultrast t .. a, . . Morgan, A. Rosenthal and
temperatures, in: Water structure at the water-polymer interface, H.H.G. Jellinek, ed. (Ple- · J rue ure m systole and d' t 0 I · . .
t10n, . Ultrastruct. Res. 22, 22. ias e usmg balhstlc cryofixa-
num Publishing Corporation, New York), p. 146.
Marchesse-Ragona, S.P. (1984), Ethanol, an efficient coolant for rapid freezing of biological Moor, H. (1964), Die Gefrierfixation Iebender Zellen und ih .
m1kroskop1e, Z. Zellforsch. 6 , . re Anwendung m der Elektronen-
2 546
material, J. Microscopy 134, 169.
Martinez-Paloma, A., P. Pinto Da Silva and B. Chavez (1976), Membrane structure of Enta- Moor, H. (1970), Die Gefrieratztechnik in· M th d
G. Schimmel and W. Vogell eds (W.' . h ef .o ensammlung der Elektronenmikroskopie,
Moor, H. (1971), Recent pro~ress .in t~:s~~~e a tl~c~~ Verlagsanstallt, Stuttgart), p. I.
moeba histolytica - fine structure of freeze-fracture membranes, J. Ultrastruct. Res. 54, 148.
Marto, P.J., J.A. Moulson and M.D. Maymard (1968), Nucleate pool boiling of nitrogen with 261, 121. ze-e c mg techmque, Phil. Trans. R. Soc. Ser B
different surface conditions, J. Heat Transfer 90, 437.
Mathes, G. and H.A. Hackenseller (1981), Correlations between purity of dimethyl sulfoxide Mo.or, H. (1973), Etching and related problems in· . .
t10ns, E.L. Benedetti and P. Favard d (S, . Freeze.-etchmg, .techmques and applica-
and survival after freezing and thawing, Cryo-Letters 2, 389. Paris), p. 21. ' e s. oc. Franc;:aise de Microscopie Electronique,
Maul, G.G. (1979), Temperature dependent changes in intramembrane particle distribution,
7th Int. Con~r. Electron Microscopy, Grenoble 1, :45pressure freezmg on living cells, Proc.
in: Freeze-fracture: methods, artefacts and interpretations, J.E. Rash and C.S. Hudson, eds. Moor, H. and M. Hoechli (1970), The influence of hi h .
(Raven Press, New York) p. 37.
Moor, H., J. Kistler and M. Muller (\ 976) F . . .
Moor, H., G ..Bellin, C. Sandri and K. Ake;tee~~~~m ;::~pane Jet, Exp~rientia 32, 805.
Mayer, E. and P. Briiggeller (1982), Vitrification of pure liquid water by high pressure jet freez-
ing, Nature, Lond. 298, 715.
Mazur P. (1966), Physical and chemical basis of injury in single-celled micro-organisms sub-
~n mammahan nerve tissue, Cell Tissue Res. So9
~01 mfluence of high pressure freezing
jected to freezing and thawing. In: Cryobiology, H.T. Meryman, ed. (Academic Press, Lon- Muller, M., N. Meister and H. Moor (\ 980) F '. : .
freeze-fracturing, Mikroskopie 36, . ' reezmg ma propane Jet and its application in
129
don). Nannmga, . N. (1973) ' Freeze-fractur1·
. ng o f m1croorgamsms
. . Ph · I d .
Mazur, P. (1970), Cryobiology- the freezing of biological systems, Science 168, 939. BaC1llus subtilis in: Freeze-etchi.ng t h . · ysica an chemical fixation of
Mazur, P. (1977), The role of intracellular freezing in the death of cells cooled at supraoptimal ' , ec mques and appr f E
Favard, eds. (Soc. Franc;:aise de Microscop· EI t . ica wns, · .L. Benedetti and P.
· ie ec romque Paris) 151
rates, Cryobiology 14, 251.
Mazur, P., S.P. Leibo, J. Farrant, E.H.Y. Chu, M.G. Hanna and L.H. Smith (1970), Interacc Nas h ' T · (1966), Chemical constitution and h · 1 '. 'P· ·
living cells against damage due to free . p ysdicahpropert_ies of compounds able to protect
zmg an t awmg m· c
tions of cooling rate, warming rate and protective additive on the survival of frozen mamma-
lian cells, in: The frozen cell, G.E.W. Wolstenholme and M. O'Connor, eds. (Churchill, Lon-
~d . (Academic Press, New York), p. 179.
. b' 1
' · ryo 10 ogy, H.T. Meryman,
Ne~, T. (1973), Growth of ice crystals in frozen s ecimens .
don), p. 69. Ne1, T. (1977), Freezing patterns and ost-th p . , J. Microscopy 99, 227.
Nei, T .. (l 978), Structure and functionpof froz:: ~~~~'.~al o~yeast cells, Cryobiology 14, 711.
McGann, L.E. (1978), Differing actions of penetrating and nonpenetrating cryoprotecti
agents, Cryobiology 15, 382. ,• J. Microscopy 112, 197. · reezmg patterns and post-thaw survival,
Mcintyre, J.A., N.B. Gilula and M.J. Kamovsky (1974), Cryoprotectant induced redistribu
tion ofintramembranous particles in mouse lymphocytes, J. Cell Biol. 60, 192. J'lermut, M.V. and
J. Microscopy 102,B.J.
29.Ward (1974) ' Effect of fixatives
. on fracture plane in red blood cells,
McMillan, J.A. and S.C. Los (1965), Vitreous ice: irreversible transformations during warm
up, Nature, Lond. 206, 806.
eushul, · study of the red a Iga p orphyrzd1um,
1231. M. (1970), A freeze-etching .. Amer. J. Bot. 57,
Meisner, J. and W.A. Hagins (1978), Fast freezing of thin tissues by thermal conduction in
sapphire crystals at 77 Kelvin, Biophys. J. 21, l 49a. rmeyer, W. and H. Moor (1976) Effect of I
etch study, Proc. 6th Europ C ,El g .ycerol on structure of membranes: a freeze-
Menold, R. and G. Liittge (1979), Freeze-etching of dispersions without contamination oft rthcote, D.H. and D.R. Le~i:(nlg9r6. 8) eFctron Microscopy, Jerusalem 2, 108.
1~ he cells of pea root tips, J. Cell Sci. ,3 , reeze-etched
fracture faces, Microscopica Acta 81, 317.Menold, R., B. Liittge and W. Kaiser (1974), Z · th surface 0 f b
Einfrieren von Suspensionen fiir elektronenmikroskopische Untersuchungen. Coll 199 . s mem ranes and organelles
' ?.R. (1974), Seasonal variation in the membr .
Polymer Sci. 252, 530. mbial cells demonstrated by freeze et h' C b' ane. structure of differentiating shoot
Menold, R., G. Liittge, and W. Kaiser (1976), Freeze-fracturing: a new method for investi GR - c mg, yto 10log1e 9 131
tion of dispersions by electron microscopy, Adv. Colloid Interface Sci. 5, 28 \. , ... (1975), Changes of particle frequenc in freeze , .
ixatlon, J. Microscopy 104, 245. . y -etched erythrocyte membranes af-
Meryman, H.T. (1966), Review of biological freezing, in: Cryobiology, H.T. Meryman,
ger, K.H. and R.P. Bunge (1974) Freeze fractu . f
(Academic Press, New York), p. l. res. A study of developing plasma , b - J nng o nerve growth cones and young
Meryman, H. T. (1971), Cryoprotective agents, Cryobiology 8, 173. mem rane, · Cell Biol 63 180
Meryman, H.T. (1974), Freezing injury and its prevention in living cells, Ann. Rev. Biop ps, T .E. and A.F. Boyne (1984) ' L.iqm·d mtrogen-based
· · 'k f · ·
qu· .
unce- ree delivery of choline · . IC reezmg: expenences with
h. J, f9. rg:ic nerve termmals to a metal surface, J. Electron Microsc.
Bioeng. 3, 341.
Meyer, H.W. and H. Winkelmann (1970), Die Darstellung von Lipiden bei der gefrierat
ration und ihre Beziehung zur Gefrieranalyse biologischer Membranen, Exp. Pathol. 4,
Miller, K.R. (1979), Artefacts associated with deep etching technique, in: Freeze-fra
aski, K. and L. Seveus (1980), iii h
y of plants. I. PVP infusion offixe~
da Silva p (19?2) T .
toI . . ..
[;1enc protective add1t1ves in cryoultramicro-
Pan issues. Cryo-Letters 1 494
methods, artefacts and interpretations, J.E. Rash and C.S. Hudson, eds. (Raven Press, ' · , rans1at1onal mobility f b · ' ·
ocyte ghosts. pH dependent rev .bl o mem rane mtercalated particles of human
, ers1 e aggregat10n, J. Cell Biol. 53, 777.
York), p. 31.
Freezing 141
Robards and Sleytr Low temperature methods in biological electron microscopy
140
Rebhun, L.L (1972), Freeze-substitution and freeze-drying, in: Principles and techniques of
Pinto da Silva, P. and B. Kachar (1980), Quick freezing versus chemical fixation: capture and
electron microscopy, M.A. Hayat, ed. (Van Nostrand Reinhold, New York), p. 3.
identification of membrane fusion intermediates, Cell Biol. Int. Reports 4, 625.
Reith, A. and R. Oftebro (1971), Structure of HeLa cells after treatment with glycerol as
Pinto da Silva, P. and R. G. Miller (1975), Membrane particles on fracture faces of frozen mye-
revealed by freeze-etching and electron microscopic methods, Exp. Cell Res. 66, 385.
lin, Proc. Natl. Acad. Sci. U.S.A. 72, 4046. . .
Plattner, H. (1970), A study on the interpretation of freeze-etched ammal tissues and cell orga- Remsen, C.C. and S.W. Watson (1972), Freeze-etching of bacteria, Int. Rev. Cytol. 33, 253.
nelles, Mikroskopie 26, 233. . . . . Remsen, C.C., S.W. Watson, J.B. Waterbury and H.G. Triiper (1968), Fine structure of
Plattner, H. (1971), Bull spermatozoa - a reinvestigation by freeze-etchmg usmg widely differ- Ectothiorhodospira mobilis Pelsh, J. Bacteriol. 95, 2374.
ent cryofixation procedures, J. Submicrosc. Cytol. 3,19 . Rey, L.R. (1960), Thermal analysis of eutectics in freezing solutions, Ann. N. Y. Acad. Sci.
Plattner, H. (1984), Ultrarapid freezing and freeze-etchmg of cell monolayers and suspens10ns, 510.
Proc. 8th Europ. Reg. Conf. Electron Microscopy, Budapes: 3, 1.731.. Richmond, D. V. and R. J. Pring (197la), Fine structure of Botrytis fabae Sardina conidia
Plattner, H. and L. Bachmann (1982), Cryofixation: a tool m b10log1cal ultrastructural re- Ann. Bot. 35, 175. '
search, Int. Rev. Cytol. 79, 237. . Richmond, D. V. and R. J. Pring (1971 b), Fine structure of germinating Botrytisfabae Sardina
Plattner, H., W.M. Fischer, W.W. Schmitt and L. Bachmann (1972), Freeze-etchmg of cells
conidia, Ann. Bot. 35, 493.
without cryoprotectants, J. Cell Biol. 53, 116. . .
Plattner, H., W.W. Schmitt-Fumian and L. Bachmann (1973), Cryofixat10n of smgle cells by Richter, H. (1968a), Die Reaktion hochpermeabler Pflanzenzellen auf die Gefrierschutzstoffe
spray freezing, in: Freeze-etchirig, techniques and applications, E.L. Benedetti and P. (Glyzerin, Athylenglykol, Dimethylsulfoxid), Protoplasma 65, 155.
Favard, eds. (Soc. Frarn;:aise de Microscopie Electronique, Paris), p. 81. Richter, H. (1968b), Low-temperature resistance of Campanula-cells treated with glycerol, Pro-
Pryde, J.A. and G.0. Jones (1964), Properties of vitreous water, Nature, Lond. 170, 685. toplasma 66, 63.
Pscheid, P., C. Schudt and H. Plattner (1981), Cryofixation of monolayer cell cultures for Richter, H. and U. Sleytr (1970), Gefrieratzung des Assimilations Parenchyms von Asparagus
freeze-fracturing without chemical pretreatments, J. Microscopy 121, 149. sprengeri Regel, Mikroskopie 26, 329.
Pueschel, c.M. (1977), A freeze-etch study of the ultrastructure of red algal pit plugs, Protop- Riehle, U. (1968a), Uber die Vitrifizierung verdunnter wassriger Losungen, Doctoral thesis
(Juris Druck and Verlag, Zurich).
lasma 91, 15.
Puritch, G.S. and R.P.C. Johnson (1973), The structure of freeze-etched bordered-pit mem-. Riehle, U. (1968b), Fast freezing of organic specimens for electron microscopy, Chem. Ing.
Tech. 40, 213.
branes of Abies grandis, Wood Sci. Technol. 7, 256. · .•
. t. AN DJ Goodchild and L.H. Grimme (1975), The regreening of nitrogen deficient Riehle, U. and M. Hoechli (1973), The theory and technique of high pressure freezing, in:
y
P 110 IS, . . , · · . . .
Chlorellafusca. II. Structural changes during synchronous regreenmg, Arch. M1crob101. 103,. Freeze-etching, techniques and applications, E.L. Benedetti and P. Favard, eds. (Soc. Fran-
~aise Microscopie Electronique, Paris), p. 31.
259.
Rapatz, G.L. and B.J. Luyet (1968), Combined effects of freezing rates and various protectiv~· Rig~er, M.W.. and J.S. Patton (1984), The cryocell. A simple inexpensive cryogenic storage de-

agents on the preservation of human erythrocytes, Cryobiology 4, 215. vice for m1croscop1c specimens, J. Microscopy, 134, 335.
enz and B.J. Luyet (1966), Anatomy of the freezing process in biologi Rinderer, L. and F. Haenseler (1959), Heat transfer in superfluid helium, in: Proc. 10th Congr.
J M
R apa t z, G .L. , L ..
materials, in: Cryobiology, H.T. Meryman, ed. (Academic Press, New York), p. 139. Refrig. (Pergamon, London)J, 243.
Rasmussen, D.H. (1982), Ice formation in aqueous systems, J. Microscopy 128, 167. obards, A.W. (1980), A microprocessor controlled rapid cooling rate meter Cryo-Letters J
384. , ,
Rasmussen, D.H. and B.J. Luyet ( 1969), Complementary study of some non-equilibrium P
transitions in frozen solutions of glycerol, ethylene glycol, glucose and sucrose, B10dyna obards, A.W. and P. Crosby (1979), A comprehensive freezing, fracturing and coating system
for low temperature SEM, Scanning Electron Microscopy 1979, 2, 325.
JO, 319. .
Rasmussen, D.H. and B.J. Luyet (1970), Contribution to the establishment of the temperatu 'obards, A.W .. and P. Crosby (1983), Optimisation of plunge freezing: linear relationship
concentration curves of homogeneous nucleation in solutions of some cryoprotect1ve age etween coolmg rate and entry velocity into liquid propane, Cryo-Letters 4, 23.
~s, A.W. and G.R. Parish (1971), Preparation and mechanical requirements for freeze-
Biodynamica 11, 33.
Rasmussen, D.H. and A.P. MacKenzie (1968), The glass transition in amorphous water. g thick walled plant tissue, J. Microscopy 93, 61.
plication of the measurements to problems arising in cryobiology, J. Phys. Chem. 75,. 97 s, A.W. and N.i. Severs (1982), A comparison between cooling rates achieved using
Rasmussen, D.H. and A.P. MacKenzie (1968), Phase diagram for the system water/dim ~pane Jet device and plunging into liquid propane, Cryo-Letters 2, 135.
s, A.W.'. U.B. Sleytr, G.R. Bullock, P.D. Sibbons and K.H. Quine (1980), Thin plate
sulphoxide. Nature, Lond. 220, 1315. . . . .
Rasmussen, D .H. and A.P. Mackenzie (1972), Effects of solutes on 1ce-solut10n mterfac1al vage device for Leybold- Heraeus Bioetch 2005 freeze-etching unit, Proc. 7th Europ.
energy: calculation from measured homogeneous nucleation temperatures, m: Water Conf. Electron Microscopy, The Hague 2, 740.
ture at the water-polymer interface, H.H.G. Jellinek, ed. (Plenum Press, New York), P· s, A.W., G.R. Bullock, M.A. Goodall and P.D. Sibbons (1981), Computer assisted ana-
Raviola, E., D.A. Goodenough and G. Raviola (1980), Structure of rapidly frozen gap offreeze-fractured membranes following exposure to different temperatures, in: Effects
w temperatures on biological membranes, G.J. Morris and A. Clarke, eds. (Academic
tions, J. Cell Biol. 87, 273. , London), p. 219.
Robards and Sleytr Low temperature methods in biological electron microscopy Freezing 143
142

5) Course of temperature variation in an object during freeze-etching pro- Silvester, N.R., S. Marchese-Ragona and D.N. Johnston (1982), The relative efficiency of vari-
R osen k ranz, J . (197 , ous fluids in the rapid freezing of protozoa, J. Microscopy 128, 175.
cedure, Arzneim Forsch. 25, 454. . Sitte, H. (1979), Cryofixation of biological material without pretreatment, Mikroskopie 35, 14.
.chter (1969) Automatische Glyzerinbehandlung pflanzhcher Dauer-
Rotten b erg, W . an d H . Rl , Sitte, H. (1984), Equipment for cryofixation, cryoultramicrotomy and cryosubstitution in bio-
gewebszellen fur die Gefrieratzung, Mikroskopi.e 25, 313.'
Ryan, K.R. and D.H. Purse (1984), Rapid freezmg: specimen supports and cold gas layers. medical TEM-routine, Zeiss Magazine for Electron Microscopists 3, 25.
Sitte, H., H. Fell, W: Hobli, H. Kleber and K. Neumann (1977), Fast freezing device, J. Mic-
J. Microscopy 136, RPS. . roscopy 111, 35.
J Roli R Myklebust and H. Engedal (1977), Preservation of shock-frozen
1 TS
Saeters d a, .. , . , . . . Sjiiland, R.D. and D.A. Smith (1974), Freeze-fracture studies of photosynthetic deficient
. 1 t' shown by cryoultramicrotomy and freeze-fracture studies, J. M1cros-
myocard ia issue as supergranal chloroplasts in tissue cultures containing virus-like particles, J. Cell Biol. 60,
copy 111, 297. . ..
Sassen, MM
c C Rems_ en and W.M. Hess (1967), Fine structure of Pemczllzum megas-
. .A ., . .
285.
Skaer, H.le B. (1982), Chemical cryoprotection for structural studies, J. Microscopy 125, 137.
porum conidiospores, Protoplasma 64, 75. . .
· B H . d p Satir (1979) Partitioning of intramembrane particles dunng freeze-fracture Skaer, H.le B., F. Franks, M.H. Asquith and P. Echlin (1977), Polymeric cryoprotectants in
Satlf . . an . , the preservation of biological ultrastructure. III. Morphological aspects, J. Microscopy 110,
' d · · Freeze fracture· methods artefacts and interpretations, J.E. Rash and C.S.
proce ures, m. - . ,
257.
Hudson, eds. (Raven Press, New York), p. 43. . . .
Sattler, C.A. and L.A. Staehelin (1974), Ciliary membrane differentiations m Tetrehymena pyr- Skaer, H.le B., F. Franks and P. Echlin (1978), Nonpenetrating polymeric cryofixatives for
u!trastructural and analytical studies of biological tissues, Cryobiology 15, 589.
iformis, J. Cell Biol. 62, 473. Skaer, H.le B., F. Franks and P. Echlin (1979), Freeze-fracture studies of ice recrystallization
Schenk, H. (1959), Heat transfer engineering (Prentice-Hall, London). . .
Schiller, A. and R. Taugner (1980), Freeze-fracturing and deep etchmg with the_ volatile cryo- in tissues quenched in aqueous solutions ofpolyvinylpyrrolidone, Cryo-Letters 1, 61.
protectant ethanol reveals true membrane surfaces of kidney structures, Cell Tissue Res. 210, Sleytr, U.B. (1974), Freeze-fracturing at liquid helium temperature for freeze-etching, Proc. 8th
Int. Congr. Electron Microscopy, Canberra 2, 30.
57. b . d Sleytr, U .B. (1978) Regular arrays of macromolecules on bacterial cell walls - structure
Schiller, A., R. Taugner and E. Rix (1978), Die Anwendung kryoprotectiver Su stanzen m er
chemistry, assembly and function, Int. Rev. Cytol. 53, 1. '
Gefrierbruchtechnik, Mikroskopie 34, 19. . .
'll A u Sonnhof and R. Taugner (1979), Tissue function-compatible cryoprotectants ;S\eytr, U.B., H. Adam and H. Klaushofer (1969), Die Feinstruktur der Konidien von Aspergil-
Sch 1 er, ., . lus niger, V. Tiegh., dargestellt mit Hilfe der Gefrieriitztechnik, Mikroskopie 25, 320.
in cryoultramicrotomy and freeze fracturing, Mikroskopie 35, 23. . .
r, U.B., H. Groesz and W. Umrath (1981), Interpretation of morphological data obtained
Schmalbruch, H. (1980), Delayed fixation alters pattern of intramembrane particles m roam·
by freeze-fracturing at very low temperatures, Acta Histochemica 23, Suppl., S29.
malian muscle fibres, J. Ultrastruct. Res. 70, 15. . .
H Plattner W. Aberer and H. Winkler (1978), Particle segregat10n m chromaffi mlyo, A.V. and J. Silcox (1979), Cryoultramicrotomy for electron probe analysis, in: Micro-
Sch u1er, G ., . , beam analysis in biology, C.P. Lechene and R.R. Warner, eds. (Academic Press, New York)
granule membranes by forced physical contact, Biochim. Biophys. Acta 513, 244.
Schultz, D. and H. Lehmann (1974), Use of freeze-etching technique to demonstrate the ultr . 535.
tructure of dormant cells, Proc. 8th. Int. Congr. Electron Microscopy, Canberra 2, 712. thworth, D. and D. Branton (1971), Freeze-etched pollen walls of Artemesia pycnocephala
Neidhart E. Pemer J. Jaenicke and J. Sommer (1973), Verbesserte Darst. d Lilium humboldt, J. Cell Sci. 9, 193. .
S h ltZ, D ., H . V .
CU , ,
orth, D., K. Fisher and D. Branton (1975), Principles of freeze-fracturing and etching,
Jung der Feinstruktur ruhender Pflanzenzellen, Protoplasma 78, 41. .
Schwabe, K.G. and L. Terracio (1980), Ultrastructural and thermocouple evaluat10n ofr echniques in biochemical and biophysical morphology, D. Glick and R.M. Androsen-
, eds. (Wiley Interscience, New York, London), 2, 247.
freezing techniques, Cryobiology 17, 571.
Schwelitz, F.D., W.R. Evans, H.H. Mollenhauer and R.A. Dilley (1970), The fine struct !in, L.A. (1973), Analysis and critical evaluation of the information contained in freeze-
of the pellicle of Euglena gracilis as revealed by freeze-etching, Protoplasma 69, 341.. micrographs, in: Freeze-etching, techniques and applications, E.L. Benedetti and P.
Seveus, L. (1978), Preparation of biological material for X-ray microanalysis of diffus ard, eds. (Soc. Franc;:aise de Microscopie Electronique, Paris), p. 223.
solutes. J. Rapid freezing of biological tissue in nitrogen slush and preparat10n of ultra : , L.A. (1979), A simple guide to the evaluation of the quality of a freeze-fracture rep-
m: Freeze-fracture: methods, artefacts and interpretations, J.E. Rash and C.S. Hudson,
frozen sections in the absence of trough liquid, J. Microscopy 112, 269. .
Shelton, E. and W.E. Mowczko (1979), Scanning electron microscop.y of membrane ~I
(Raven Press, New York), p. 11.
produced by glutaraldehyde fixation and stabilised by post fixation m osmrnm _tetroxi . , L.A. and W.S. Bertaud (1971), Temperature and contamination dependent freeze-
Freeze-fracture: methods, artefacts and interpretations, J.E. Rash and C.S. Hudson mages of frozen water and glycerol solutions, J. Ultrastruct. Res. 37, 146.
n, L.A. and 0. Kiermayer (1970), Membrane differentiation in the Golgi complex of
(Raven Press, New York), p. 67. .
Shepard, M.L., C.S. Goldstein and F.H. Cocks (1976), The HiO-NaCl-glycerol phased ~terias denticulata ~reb. Visualised by freeze-etching, J. Cell Sci. 7, 787 .
.L. (1971), Retent10n of three dimensional contours by replicas of freeze-fracture spe-
and its application to cryobiology, Cryobiology 13, 9.
, Proc. 29th Ann. Meeting EMSA, p.242.
Robards and Sleytr Low temperature methods in biological electron microscopy Freezing 145
144

1954) Caution in the use of liquid propane for freezing biological specimens, Van Venetie, R., W.J. Hage, J.G. Bluemink and A.J. Verkleij (1981), Propane jet freezing -
Steph enson, J .L . ( ,
Nature, Lond. 174, 235. . . . . . a valid ultra-rapid freezing method for the preservation of temperature dependent lipid phas-
1956) Ice crystal growth during the rapid freezmg of tissues, J. b10phys. b10- es, J. Microscopy 123, 287.
Steph enson, J .L . ( ,
chem. Cytol. 2, 45. . . . ·d f · Vargaftik, N.B. (1975), Tables on the thermophysical properties of liquids and gases in normal
1960) Ice crystal formation in biological matenals durmg rap1 reezmg, and dissociated states. 2nd. Edn. (Hempisphere, Washington, London).
Step h enson, J .L . ( ,
Ann. N. Y. Acad. Sci. 85, 535. . . Venrooij, G.E.P.M., A.M.H.J. Aertsen, W.M.A. Hax, P.H.J.T. Ververgaert, J.J. Verhoeven
Ultrastructure of dormant and germinated bas1d10spores and H.A. Van der Vost (1975), Freeze-etching: freezing velocity and crystal size at different
Stocks, D.L. and WM . . H ess (1970) ,
of a species of Psilocybe, Mycolgia 62, 176. . . locations in samples, Cryobiology 12, 46.
. k. c (1977) Freeze-fracture replication in b10log1cal research development, current Verkleij, A.J., P.H. Ververgaert, R.A. Prins and L.M.G. van Golde (1975), Lipid phase transi-
Sto1ins 1, . ,
practice and future prospects, Micron 8, 87. . . . . tions of the strictly anaerobic bacteria Vei/lonella parvula and Anaerovibrio lipolytica, J. Bac-
·rt JG d Ms . . Buttrose (1972) , Freeze-etch studies ofprotem bodies m wheat scutellum, teriol. 124, 1522.
Sw1 , . . an
J. Ultrastruct. Res. 40, 378. . Verkleij, A.J. and P.H. Ververgaert (1975), Architecture of biological and artificial membranes
Swift, J.G. and M.S. Buttrose (1973), Protein bodies, lipid layers and amyloplasts m freeze- as visualised by freeze-etching, Ann. Rev. Phys. Chem. 26, 101.
etched pea cotyledons, Planta 109, 61. Verkleij, A.J. and P.H. Ververgaert (1978), Freeze-fracture morphology of biological mem-
Takeo, K. and M. Nishiura (1974), Ultrastructure of polymorphic mucor as observed by means branes, Biochim. Biophys. Acta 515, 303.
of freeze-etching, Arch. Microbial. 98, 175. Ververgaert, P.H.J.T., A.J. Verkleij, J.J. Verhoeven and P.F. Elbers (1973), Spray-freezing of
·ra (1973) Fine structure of Cryptococcus neoformans grown liposomes. Biochim. Biophys. Acta 311, 651.
Takeo, K., I . U esa k a an d K . Uehl ,
in vitro as observed by freeze-etching, J. Bact. 113, 1442. Weast, R., ed. (1971 and other editions), CRC handbook of chemistry and physics, (Chemical
· (1970) Cardiac toxicity of aerosol propellants, J. Amer. Med. Rubber Company, Ohio).
Taylor, G.J. an d W .S . H arns ,
Wecke, J. (1968), Vergleichende elektronenmikroskopische Untersuchungen an Membran-
Ass. 214, 81. f p d
74) Freeze-etching of outer membranes o seu omonas strukturen von latent lebenden Organismen verschiedener systematischer Stellung zur Frage
Thornley, M.J. an d U .B. Sleyt r (19 ,
and Acinetobacter, Arch. Microbial. 100, 149. der Unit-Membran-Konzeptes, Inauguraldissertation, (Freie Universitat, Berlin).
. y s et al (1970) Thermophysical properties of matter, Vols.J-13 (!FI/Plenum, Wendt-Gallitelli, M.F. (1984), Importance of freezing tissue in defined functional states for
T OU1OUk 1an, . . '
New York, Washington). . . . . reproducible X-ray microanalysis in heart muscle, Proc. 8th Europ. Reg. Conf. Electron
· Hc d ML Higgins (1974) Effect of temperature on d1stnbut10n of membrane part!, Microscopy, Budapest 3, 1677.
T s1en, . . an . . , . 118 725
cles in Streptococcusfaecalis as seen by the freeze-fracture technique, J. Bactenol. . , · Werz, G. and G. Keller (1970), Die Struktur des Golgi-Apparates bei gefriergeatzen Duna-
Umrath, w. (1974a), Cooling bath for rapid freezing in electron microscopy, J. Microscop liella-Zellen, Protoplasma 69, 351.
.Westwater, J.W. (1959), Boiling heat transfer, American AEEE Scientist 47, 427.
JOI, 103. ·k k · B ·t 1
Umrath, w. (1974b), Gefrierschock in Kuhlbadern fur die Elektronenm1 ros op1e, ei r. e e · hite, G.K. (1979), Experimental techniques in low temperature physics (Clarendon Press,
tronenmikroskop. Direktabb. Oferfl. 7, 103. . Oxford).
Umrath, w. (1975), Rapid freezing in open cooling baths, Arnze1m Forsch. 25, 450. .illiams, L. and S. Hodson (1978), Quench cooling and ice crystal formation in biological tis-
Umrath, w. (1977), Prinzipien der Kryopraparationsmethoden - Gefnertrocknung u sues, Cryobiology 15, 323.
Gefrieratzung, Mikroskopie 33, 11. . . · iamson, P.A. (1984), Storage of small samples under liquid nitrogen, J. Microscopy 134,
. M E G Cravalho and R.L. Levin (1979), Volumetric changes m yeast cells dun 25.
U sh 1yama, ., . .
freezing at constant cooling rates, J. Membr. Biol. 6, 112. . ison, J.H.M. (1975), Plant cell wall microfibril deposition revealed by freeze-fractured plas-
J H Luft (1979) Effects of glutaraldehyde fixation on the structure oft! alemma not treated with glycerol, Planta 126, 93.
Van D eurs, B. an . .d ,
junctions: a quantitative freeze-fracture analysis, J. Ultrastruct. Res. 68, 160. . n, J.H.M. (1976), An examination of the relationship between freeze-fractured plasma-
eyer and R. Lambert (1970), The fine structure of frozen etched Fusar ma and cell wall micro fibrils, Protoplasma 88, 187.
Van G oo, 1 A .P. , J . M
conidiospores, J. Microscopie 9, 653. . . n, J.H.M. and R.M. Brown (1979), Pretreatment artefacts in plant cells, in: Freeze-frac-
Van Harreveld, A. and J. Crowell (1964), Electron microscopy after rapid freezmg on a : methods, artefacts and interpretations, J.E. Rash and C.S. Hudson, eds. (Raven Press,
' surface and substitution fixation, Anat. Rec.149, 381. . . e York), p. 51.
Van Harreveld, A. and J. Trubatch (1979), Progress of fusion during rapid freezmg for el on, J.H.M. and E.C. Cocking (1975), Microfibril synthesis at the surface of isolated To-
microscopy, J. Microscopy 115, 243. . co mesophyll protoplasts, a freeze-etch study, Protoplasma 84, 147.
Crowell and S.K. Malhotra (1965), A study of extracellular space in n, J.E., D.E. Brown and E.K. Timmens (1965), A toxicologic study of dimethyl sulfox-
Van H arreveld , A ., J .
tral nervous tissue by freeze-substitution. J. Cell Biol. 25, 117. . Toxicology appl. Pharmacol. 7, 104.
Van Harreveld, A., J. Trubatch and J. Steiner (1974), Rapid freezing and electron micros , A.J. and A.W. Robards (1982), Some experiences in the use of a polymeric cryoprotec-
for the arrest of physiological processes, J. Microscopy 100, 189. in the freezing of plant tissue, J. Microscopy 125, 287.
11
',/,

,11
Robards and Sleytr Low temperature methods in biological electron microscopy
146

. h F
Wund er11c , . an d V . Speth (1972) , Membranes in Tetrahymena. I. The cortical pattern, J.
Chapter 3
Ultrastruct. Res. 41, 258. .
Y annas, I · ( 1968) ' Vitrification temperature of water, Science 160,
·
298.
f ·
.
erate liver J
Yee, A.G. and J. Revel (1978), Loss and reappearance of gap JUnc 10ns rn regen . ' ·
Cell Biol. 78, 554. · t
· KE d J A Husby (1982) Antifreeze effect of thermal hysteresis agen s pro-
Zachanassen, · · an · · '
tects highly supercooled insects, Nature, Lond. 298, 865.
Zerban, H., M. Weh ner and G . W erz (
1973) Uber die Feinstruktur des Zellkerns von Acetabu-
, Direct viewing and analysis
!aria nach Gefrieratzung, Planta 114, 239. . d
· · t d" d b coolrng rate measurements an
Zierold, K. (1980), Cryofixation of tissue specimens s u 1e Y of cold specimens
scanning electron microscopy, Microscopica Acta 83, 25. . . . .
. h · Hp and H Plattner (1976) Electron microscopy methods rn b10logy, rn. Methods
Zrngs e1m, · · · ' . . N y k) 1
· memb rane b"101ogy, Vol · 7, ED
rn · · Korn ' ed. (Plenum Pubhshrng Company, ew or p ..

3.1 Principles and problems

The most obvious and direct way of dealing with a biological specimen once
it has been frozen is to look at it while still in this hydrated frozen state.
This, however, is not without its attendant problems. While bulk, frozen
specimens are now readily and routinely observed in the scanning electron
microscope (SEM), both the preparation and examination of thin frozen
specimens required for tr:;msmission electron microscopy (TEM) still pre-
sent very substantial difficulties.
It is technically very demanding to produce thin specimens from bulk fro-
zen material and, moreover, when such samples are irradiated in an electron
'croscope, they can be all too easily irreversibly damaged by the effects
heating, by radiolysis of ice or by other direct or indirect results of elec-
n bombardment. A major reason for putting cold or frozen specimens
o electron microscopes is to make use of the potentially improved reten-
n of diffusible components in their in vivo positions so that different mic-
nalytical techniques can be applied. It is not the purpose of this book
deal with microanalytical methods, which are fully discussed elsewhere
this series (Chandler 1977). However, the preparation of specimens for
use must obviously take into account the requirements of the final ob-
tional and microanalytical processes. Thus, while we will not discuss
oanalysis per se in any detail, the implication that many of the prepara-
techniques will be used with this goal in mind must necessarily condi-
our assessment of the experimental procedures used as we move from
ative biological structure towards the hostile environment of an elec-
microscope column.

147
Robards and Sleytr Low temperature methods in biological electron microscopy Direct viewing and analysis of cold specimens 149
148

units of e- nm - 2 where 6.25 e- nm- 2 is equivalent to a dose of 1 coulomb


per square metre (C m- 2). Misell (1978) provides a useful discussion of elec-
tron dose in relation to radiation damage. Taylor (1978) has commented
that the 'safe' dose for high resolution studies on unstained biological speci-
mens is less than about 100 e- nm- 2 (and, for some specimens, less than
10 e- nm- 2) while Cosslett has stated that most biological specimens will
be damaged with an input energy of between 1 and 10 C m- 2 (approxima-
tely 6 to 60 e- nm - 2). When it is realised that the normal electron dose
simply to record an image (i.e. ignoring the time taken for finding the right
position on the specimen, focusing, etc.) at a magnification of 4000 times
is about 20 e- nm - 2 (Glaeser and Taylor 1978), the magnitude of the prob-
lem is clearly seen. There has been considerable debate about the actual de-
gree of protection from radiation damage conferred by the use of low tem-
peratures. In part this has arisen because of the difficulties inherent in ensur-
ing that the actual specimen in an electron microscope is held at a specific,
very low temperature (e.g. 4.2 K). Knapeck and Dubochet (1980) suggested
that radiation damage was reduced by a factor of 30 to 300 when the speci-
. · d · stals of L-valine at different temperatures in relation to elec-
Fig. 3.1. Rad iat10n amage m cry . . . . . . _ men was observed at 4.2 Kin a microscope using a liquid helium super-con-
'ndicates the maximum resolut10n remammg m the d1ffract10n pat
tron dose. Th e va1ue o f R l . h h' h t d'ff f ducting lens. However, subsequent reports (e.g. Lamvik et al. 1983) have
t . R is the inverse of the average spatial frequency correspondmg to t e ig es 1 r~c ~on
oe;;~rs visible in the two axial directions of the pattern. Curve (a) represents data rec~rd e .a~ rated the protective value of low temperatures much more conservatively
with the L-valine at room temperature while (b) was recor e wit
a beam vo It age of 60 kV I I ~ with an improvement factor of between four and six. Discussions of this
f 65 kV at liquid helium temperature. The lower temperature c ear Y con ers
a b earn vo It age o f · ·1 · t matter are contained in Talmon (1982a,b) and Dubochet et al. (1982) (see
· h · This result is typical of a number o s1m1 ar expenmen s <
some protection on t e specimen. .
although earlier work had suggested greater protection of the specimen at low temperatures also Knapeck et al. 1984). We shall return to this problem in §3.4.1 and
than indicated here. (Abridged and redrawn from Lamv1k et al. 1983.)
·. §3.4.2 but suffice it to say that, for some high resolution purposes, even a
·gain of a factor of two would be sufficient to warrant the use of low temper-
ure stages and, therefore, the practical methods of handling and viewing
Although most of this book is concerned with the production, and subse- y specimen at low temperature fall within the ambit of this book.
quent utilisation, of hydrated frozen specimens, there are also go?d reason While the direct insertion of a cold specimen onto the low temperature
for using cold stages in electron microscopes so that other specimens (no tage of an electron microscope is conceptually one of the simplest ways of
necessarily hydrated) can be viewed and analysed a~ low temperatures. r. ting frozen specimen, even this is not without its difficulties. The speci-
particular, it has been shown that radiation damage is reduced when spe must be kept cold while it is moved from its place of cooling into the
mens are maintained at low temperature instead of at room temperatu roscope and, furthermore, it must be protected from the accumulation
( g Cosslett 1978· Glaeser and Taylor 1978; Taylor 1978; Dubochet et. contamination by condensing gases. In addition, few specimens can be
1e9 8],. Talmon 198la,b; Fig. 3.1), although the actual quantitative redu~t1 wed immediately without additional processing. This Chapter describes
rem~ins a matter of debate (see below). A number of a~thors h.ave pom different means of preparing specimens for observation and analysis at
to the extremely low electron dose that will bring a~out Ir~evers1ble cha temperatures, methods for the protection of specimens during transfer
in biological specimens at 'normal' temperatures m a ~mcr~scope cold , finally, cold stages for both scanning and transmission electron micro-
(e.g. Kellenberger 1982). The radiation dose to the specimen is measu:: · es.
the number of electrons (n 0 ) incident on one square nanometre. no th
151
Direct viewing and analysis of cold specimens

Robards and Sleytr Low temperature methods in biological electron microscopy 3.2 Instrumental requirements for low temperature electron
150
microscopy
3.1.1 Types of specimen
This is primarily a book of practical methods in electron microscopy. How-
The types of specimen viewed at low temperatures may be broadly divided ever, it has ~lso been part of the authors' philosophy in this, still emergent,
field of low temperatures that the background and underlying principles of
into four main categories (Fig. 3.2):
frozen-hydrated - with the possibility for partial or complete freeze- the use of specific techniques and procedures should be fully understood so
(1) that sensible and correct decisions can be made in selecting appropriate
drying at some stage during the preparation pathway; .
frozen-fractured - a simple modification of (l) that allows mternal methods. This approach becomes especially important when discussing
(2) equipment. Thus, while it is possible to provide a detailed statement on how
surfaces to be exposed; . . to use a cryoultramicrotome or how best to freeze a particular type of speci-
frozen-section - thin pieces of frozen-hydrated or frozen-dned tissue
(3) men, the same approach cannot be employed when dealing with, for exam-
that can be viewed by transmission imaging, and
frozen thin-layer - which are not necessarily hydrated but which are ple, stages in electron microscopes. For freeze-sectioning (see Chapter
(4) 4) a relatively high degree of operator knowledge and training is necessary,
observed at low temperature to minimise radiation damage.
Viewing whereas with cold stages, provided that the right equipment is available, the
and operator has little scope for individual variation and merely has to follow
a~S
the specific instructions of the designer. In this section we shall, therefore,
be more concerned with the basic principles underlying the design of the
1 a Bulk specimens stages as illustrated with a few selected examples; see Butler and Hale (1981)
for a full discussion of cold stages. It is certainly not the intention of this
SEM
book to provide the information necessary for the reader to produce his
Fracturing Freeze- coating
.. wn cold stage or transfer device! This is a specialist area which relatively
-Pre-
treat-
Freezing
2

1
b Bulk
drying
0 individuals will have either the aptitude, facilities or inclination to enter.
few
purpose of what follows is therefore to provide the basic information
ment specimens
d fundamentals that will apply to any system and then to look at one or
,~i-i- 3 STEM o examples of how these have been put into practice.
sectioning I or
Specimen
I
TEM
Microscopes

he principle requirements for cold stages in electron microscopes are that


y should easily maintain the specimen at the lowest required tempera-
e, that they should be mechanically and thermally stable and that the spe-
. 2 Diagrammatic illustration of preparation possibilities for the direct observatio1 en should be protected from contamination by molecules condensing
ig. · · II · be frozen direct
the residual gas in the vacuum system. If microanalysis is to be carried
~
3
Fbiological specimens using \ow temperature methods. A specimens may. f he e

sta~~rov~:i
after pretreatment (depending on the nature of the specimen and the requirements o S then the cold stage should also be constructed so that minimum back-
~EM
iment). (la) Bulk specimens, after coating, may be viewed directly on the cold
Alternatively (lb) they can be frozen-dried and then transferred to the iall fro nd radiation is generated.
ambient temperature. (2) Bulk specimens can also be fractured open, perhaps part. the S e technology required for the construction and use of cold stages to
ried (freeze-etched) to reveal further structure,
· · and
d then tcoated
b before so
epared viewing
that tihney ar e ct specimens from the effects of radiation damage is, to quite a large
d(3) Specimens for viewing in the transmiss10n mo e mus e pr. . ibl foll t, the same as that required for purposes of observation and analysis,
enough for electrons to traverse. This can be done by cryosect10nmg poss y in aft
freeze-drying and/or coating. (4) Thin layers. are usually prep.ared directly and, aga
possibility of coating, are viewed directly.
152 Robards and Sleytr Low temperature methods in biological electron microscopy Direct viewing and analysis of cold specimens
153

although there may be substantial differences in detail. Therefore, with the Temperature- sensing
element
present exception ofliquid helium-cooled stages, it is possible to discuss the Specimen table
technological requirements generally, irrespective of the purposes that the
specific cold stage will ultimately serve. It should, however, be noted that,
whereas for high resolution molecular studies the prime requirement is for
low temperature and high spatial resolution, for X-ray microanalysis the
emphasis is on good X-ray performance (thus precluding massive sample b
holders) and high temperature stability so that specimen drift is minimised -Flexible copper connection
over the relatively long counting period. For these reasons low temperature into liquid nitrogen vessel

stages for X-ray use need to be more delicate than those for high resolution a
structural studies.
The actual temperature of the cold stage will depend primarily on
whether liquid nitrogen or liquid helium is the cryogenic source. For SEMs,
~xpansion orifice
'\(
liquid nitrogen (or adiabatic expansion of nitrogen - the Joule-Thomson ef-
fect) has been adequate for most purposes but, for high resolution TEMs
or scanning transmission electron microscopes (STEMs), helium-cooled l~r
l~I
stages have advantages in terms of increased protection from radiation
damage. Liquid helium-cooled stages are very specialised pieces of equip- c
ment and they will not be considered further here. The majority of cold stag-
Nitrogen gas
es have used nitrogen as a coolant and operate down to a lowest tempera-
Fig: 3.3. Some possibilities for cooling specimen stages in scanning electron microscopes and
ture of about 63 K (the temperature of sub-cooled nitrogen). It is now s1m1lar eqmpment. (a) D!fect supply of piped liquid or gaseous nitrogen. This method can pro-
generally accepted that cold stages for biological work should be designed duce temperatures close to -196°C (77 K) A t ·
. . . · emperature-sensmg element (thermocou le
to operate at as low a temperature as possible (even if this is not always thermistor or plat1~um resistance element) is usually incorporated so that the temperature ~a~
be accurately momtored and also so that, using a heater, the temperature of the stage can
required). Some early SEM cold stages, including commercial designs, with. '. ~~ually be controUed. (b) A flexible metallic connection can be used as an alternative to pipes.
a minimum working temperature of only about 140 K are no longer consi~ e metal chosen 1s almost always copper because of its high thermal conductivity. The stage
dered adequate: any cold stage should be capable of being cooled to c~nnot be cooled to such a_ low temperature as with directly piped nitrogen but there may be
a v~ntages of ease of mampulat10n of the stage and lack of mechanical disturbance. (c) The
<110 K. . ~-Thomson refngerator makes use of the principle of the cooling effect produced by the
SEM and TEM cold stages differ considerably in their construction. SEM la atic expans10n of a compressed gas. This system can also produce temperatures very close

stages tend to be quite large and capable of maintaining a relatively bulky to liquid nitrogen temperature.

specimen at low temperature, whereas TEM stages need 'less' cooling bu~
much better thermal and mechanical stability and precision. The parameter tween the specimen and the coolant. To provide the best conduction of
determining the lowest operating temperature for the specimen is the rati t from_ the specimen ~t ~s necessary to have a conduction pathway of large
between the rate of heat input to the specimen and heat withdrawal fr ~ss-sect~o.nal area, mm1mum length, and high thermal diffusivity, and
it. Heat input to the specimen mainly arises from beam heating (which ~·a mm1mui_u number of mechanic.al junctions. Thus, a most efficient
normally be reduced to an acceptable level; see §3.4.1) and from h
radiated from surrounding surfaces. It is usually possible to minimise t
:ng system is
0
:ohave sub-cooled mtrogen passing through copper tub-
. :he underside of a copper specimen support. It is then possible to
source of heat by using well-positioned cold shrouds. The rate of heat wi :~~m the speci~en tempe~ature close to that of the coolant (Fig. 3.3).
drawal depends upon the temperature of the cooling source, its temperat er, mechamcal constramts on specimen movement, and other fac-
difference from the specimen, and the magnitude of the thermal resistanq s, such as the possibility of vibration arising from flowing liquid, often
154 Robards and Sleytr Low temperature methods in biological electron microscopy Direct viewing and analysis of cold specimens 155

preclude direct liquid cooling. Cold nitrogen gas can also be used altho~gh Specimen

this again necessitates the use of piping between the coolant (a contamer
of vapourising liquid nitrogen) and the specimen stage. For these reasons,
many designers of cold stages (particularly for SEMs) have pr~ferred to use
flexible, solid metallic conduction pathways between the specimen and the Cold stage
coolant. These provide improved flexibility of stage movement and reduce
problems of vibration. If such solid connection_s ~re use~, ~~en it is necessary -----'-Temperature
to maximise the cross-sectional area while retammg flex1b11Ity. Stranded fine sensor

copper thread is ideal and preferable to braided c~ppe~ (R?bar~s and Fig. 3.4. Heat flow through a specimen and cold stage. Heat is withdrawn through the speci-
Crosby 1979). It is obviously !1ecessary to have mechamcalJunct10_ns_ m sue~ men towards the coolmg source. It is important to position the temperature-sensing element
a 'solid' pathway but they should be kept to a minimum because 1t 1s possi- as close as possible to the specimen itself. Any mechanical junctions in the thermal pathway
will mtroduce a loss of cooling capacity.
ble to 'lose' up to 1O K cooling capacity across each junction (Taylor and
Burgess 1977). The flexible copper wire must be very tightly clamped at its
ends and solder should only be used to a minimum extent as a space filler As the specimen will be very cold, it will act as an efficient cold trap for
(its thermal conductivity is much worse than that of pure copper)._ Apart condensable molecules in the vacuum system (especially water), against
from solid or piped conduction pathways using liquid or gaseous mtrogen which it should be protected. The usual method of achieving this is to posi-
as the cooling source, some systems have used the Joule-Thomson ~ef~igera­ tion one or more cold plates (anticontaminators) close to the specimen and,
tor effect (e.g. Pawley and Norton 1978). This operates on the pnnc1ple of /preferably, making a large solid angle with it (see §5.2.1). This plate should
cooling through the adiabatic expansion of a gas (nitrogen) under pressure be at least 10-20 K colder than the specimen itself. From this it will be
and can provide temperatures very close to that of liquid nitrogen. apparent that, i.f the specimen stage is operated at the lowest possible tem-
While the specimen stage may be capable of good thermal performance, erature, then the probability of contamination is greatly increased since the
it is, of course, the temperature of the specimen itself that is cri~ical. It is nticontaminator is unlikely to be cooler than the specimen and therefore
therefore most important that thermal contact between the specimen and ··'11 have a much reduced effect. The basic principles of a simple EM cold
the cooled specimen stage is as good as possible. This means that mechan~\ age are illustrated in Fig. 3.4.
cal connections should all be very close ('tight') and high thermal conduct!~;
vity metals (preferably copper) should be used. If all the criteria menti?neq .2 Transfer devices
here are taken into account, then it should be possible to keep a specimen
very close to the coolant temperature (i.e. within a few Kelvin for liquid/ga~ specimen is to be viewed at low temperature in an EM, then it is usually
piped systems and for the Joule-Thomson refrigerator; within 10-20 K f? led externally before being put into the microscope. This means that a
solid conducting pathways). For many reasons it will not always be destr cold specimen has to be moved from its original cooling site, through
able to keep the specimen at the lowest possible temperature on the co · us preparative procedures, into the EM through the air (except in a
stage in the microscope. In consequence it is usually advisable to ha~e ively few instruments where the specimen is cooled in the EM itself or
heater in.the cold stage so that, by balancing heat input with heat extractlo chamber attached directly to it). The specimen must be protected
it is possible to maintain any desired temperature. Clearly, there must a st both excessive warming and contamination with condensing
be a temperature monitoring device (thermocouple or resistance element) urs during this transfer. It is usual, therefore, to protect the specimen
the stage so that the actual temperature can be determined a~d so t~ati g such transfer stages by use of a 'transfer device'. Together, the
required, the temperature can be automatically controlled u~mg a smta er device and the specimen mount should allow the specimen to be
electronic device to balance heating and cooling and thus achieve the act maintained at low temperature and protected from contamination.
equirement for low temperature can be met.either by provision of a
temperature demanded.
156 Robards and Sleytr Low temperature methods in biological electron microscopy Direct viewing and analysis of cold specimens 157

constant cooling source or oy relying on the thermal capacity of a pre-


Specimen
viously cooled block. Contamination is avoided either by providing a good Cold Hot
vacuum environment (with low partial pressure of condensable molecules)
y
and/or by shrouding the specimen with a cold, protective surface. The cold
shroud may either be cooled continuously or rely on thermal capacity. If
the specimen is well shrouded with low temperature surfaces and is sup-
Ho i .

~
Cold

ported on low thermal conductivity mounts, then its own heat gain will be Specimen
support
very much reduced. A variety of transfer devices has been constructed (see
§3.3.7 and §3.4.7), some with continuous evacuation and liquid nitrogen a b
cooling, others relying on cold shielding alone. The physical demands are
Heat flow to
clear, and it is simply a matter of ensuring that they are met within the geo- cold source
metric demands of the microscope and accessory units being used. Fig. 3.5. Thermal consequences of subliming ice f .
specimen (a) or by radiant heat to the the s . ron;, specimens either by warming the whole
pec1men sur1ace (b) (a) The free t h'
warms t h e whole specimen block to the etchin tern · ze-e c mg approach
3.2.3 Other ancillary equipment furthest from the heating source remains th g Id perature. The specimen surface, being
. ' as e co est site along the te t d'
(b) If a ra d1ant heater is used it applies h t t th f mpera ure gra 1ent.
'll b · ' ea o e sur ace the remaind f th ·
So far, we have discussed the freezing of specimens and their transfer onto ~ 1st
st1 emg cooled. Thus ' in this situation , th e sur1ace . h e warmest
' er o d'e specimen
part of th E'
the cold stage of an electron microscope. In addition, there are other interc met hod of etching can. be .used but. the th e gra 1ent.
erma1 consequences should be taken into conside 1ther
at10n m relat10n to the requirements of th e specimen.
· r-
mediate processes that may have to be accommodated, notably fracturing
of bulk specimens, sublimation of ice and coating of specimens with an elec~
trically conducting layer. Once again, the precise methods and equipment are situations
. where
. . the. very purpose of th e expenment· · to observe the
1s
will vary but the principles are clear. The requirement for maintenance of hav10ur
bl' · of subhmmg ice or to 'etch' to a very precise level . I n sue h cases
low temperature with low contamination is a paramount consideration. 1mat10_n must be performed within the microscope but, in other circum~
The frozen specimen, if it is a 'bulk' object, may be crudely fracture ances'. rs better to protect the microscope column against contamination.
under liquid nitrogen (§6.3.3c), cleaved or sectioned in a special chamber ( CoatmgI of· a low temperature specimen ' with gold ' plat1·n um or car b on
on a cryoultramicrotome), or manipulated while under observation in examp e, is a pr~cess that has occasioned misgivings in relation to prob-
SEM. s~~c1men he~tmg. However, it is now clear that a variety of thin film
Ice can be sublimed by raising the temperature of the whole block, eith os1t10n
h h techmques
· can be used without leading to und ue d amage
in a special chamber or on the SEM cold stage, or by surface heating usi ui ~atmg or other effects. Evaporative methods were the first to be
a radiant heater. Heating of the whole specimen is the method used ( ch_lm 1975, 197_8b, 1981; Pawley and Norton 1978; Echlin et al. 1980)
freeze-etching (§6.5) and is capable of close control of bulk temperat . remam necessar~ if carbon is to be deposited (although ion beam de o-
However, it can be slow for large specimens and raises the whole block n appear~ practicable; Franks et al. 1980). More recently it has b~en
approximately the same temperature. It also means that the surface will ~ ~h:t d~ode sputter-coating with gold can also be accomplished with
the coldest part of the specimen (Fig. 3.5). If a radiant heater is used, t ~a eatmg (Robards et al. 1981). Using estimates of the .
ble he t I d . maximum
it may be .difficult to obtain such a precise control over sublimation ~ oa to specimens when sputter-coated with gold using the
(although Talmon (1980) has provided some useful calculations rel ra illustrated in. F ig. 3· 11 (§ 3·3.4), it
ment · was concluded that the surface
heater current to rate of ice removal), but surface ice (hoar frost) c u ture h not mcrease by more than 6 K (2 K w-1 energy mput).
h th could .
removed easily and quickly; only the surface is warmed appreciably, th g ey ~ve not, as_ yet, been routinely applied to low temperature
mainder of the specimen staying cold. Sublimation of ice is preferably< ens, Pe~mng sputtenng (Jakopic et al. 1978; Peters 1980) and ion-
ried out in an ancillary chamber rather than in the EM itself, although t depos1t10n (Franks et al. 1980) would undoubtedly also provide
158 Robards and Sleytr Low temperature methods in biological electron microscopy Direct viewing and analysis 0~rcold specimens
. 159

acceptable layers without damage. In fact, it may well be found that these 3.3.I Frozen-hydrated bulk specimens
methods will provide fine grain films that will allow higher resolution studi-
es on low temperature specimens. The subject of grain structure in relation Conc~ptually t~e simplest of all low temperature techniques for EM is th
to deposition conditions during sputter-coating and ion-beam coating has ?f~~i;g a specimen, freezing it, and then looking at it while it is maintain:~
been investigated by Echlin et al. (1982) and Kemmenoe and Bullock (1983) 1ll e rozen state on the cold stage of a SEM Wh . bl .
and, while again not directly concerned with low temperature specimens, is available, such a method now provides . .I edn sm:a e eqmpment
· · a simp e an rapid means of pre
some useful relevant conclusions arise. In general it is preferable to use the parmg specimens with minimal damage For ex I .t. . .
delicate structures such as small flowe . h ampfe, _i. is possible to take
lowest possible voltage and current consistent with depositing an adequate · · rs, mus room rmtmg bodies or insect
film within a reasonable period in order to minimise the heat arising from larvae (to give JU~t ~ few typical examples) and to have them on the cold
the dispersion of energy within the specimen. stage of a SEM withm a_few minutes of starting operations. Quite often the
Evaporative and sputtering techniques have their own relative advantages surfaces that are to be viewed will be covered wit . h a film of water that will
and disadvantages. Evaporation is carried out in a high vacuum, hence faci- have to ~e removed (by_ sublimation of ice) before the true surface can be
litating contamination protection, but avoidance of 'shadowing effects' may seen. With most _scannmg electron microscope specimens it is normall
be necessary. Sputtering is simpler and may give a more even coating but, necessary
· · · · d" to provide a conducting film over the fima I frozen surf;ace beforey
as it is carried out at low vacuum (0.1 Torr, 13.3 Pa), it is important to en- it is Irra iated by the electron beam . F or b u lk specimens
. the preparation
sure protection from contamination. Despite the relative merits and demer-
its, it is a fact that both methods have been operated successfully on low Slow cooling Transfer to coating
temperature specimens. Evaporation of carbon can provide problems be-
'Bulk' specimen
apparatus A
cause the energy input needs to be so much higher than for metals, but both on stub

~\
carbon arc evaporation and resistance heating of carbon fibre 'string'
(Vesely and Woodisse 1982; Peters 1984; Smits et al. 1984) have been Coating

employed successfully. The problems of heat damage during specimen coat;- \)/

~~
ing obviously become more formidable for specimens of low thermal ca
city, such as those for observation in the transmission mode, than for b
SEM specimens. Nevertheless, it appears that, provided the physical prin
Examine on
ples of specimen heating are appreciated and the practical application cold stage
Rapid cooling in
appropriately selected, significant heat damage can be avoided. microscope
'Small' specimen

<Oi
on grid

3 .3 Frozen bulk specimens JL


in hollow rivet

By bulk specimens we generally mean those that are too thick to be


~
versed by electrons. Such specimens will normally be viewed in an SEM on surface of
~ metal pin'
may be required for either observation and/or analysis. There has
number of useful papers in this area, among which the most general incl 3.6. Alternative methods of re arin f .
on microscopy If larges ec· p p g rozen specimens for low temperature scanning
Nei (1974), Nei et al. (1973, 1974), Fuchs and Lindemann (1975), e whole speci~en can b~ fr imens are us~d,_ the cooling rate is of little significance. There-
(1975), Marshall (1975, 1977, 1980), Echlin and Moreton (1976), Ee n). However if high c 1·ozen on a re at1vely large stub (for example, using sub-cooled
· ( '
rt such as a 'pin' g ·d
00 mg rates are reqmred the
. ) . ' ·
specimen · fi f
JS Irst rozen on a small
(1978a), Fuchs et al. (1978b), Pawley and Norton (1978), Robards ' n or nvet before bemg attached, under liquid nitrogen, to a larger
Crosby (1979) and Zeirold (1979). stub.
160 Robards and Sleytr Low temperature methods in biological electron microscopy Direct viewing and analysis of cold specimens 161

route is thus clear. Unless the specimen is very small ( < 0.1 mm in at least controllable and that it may be effected in either the microscope column or
one dimension) there will be little point in trying to freeze it very rapidly. an accessory chamber. On the other hand, it can be rather slow because the
Furthermore, the actual resolution that is attained in the SEM should be temperature of the whole bulk specimen on the stub must be raised substan-
considered in relation to the probable ice crystal size developed in the speci- tially (Fig. 3.5) ..Radiant heating is less controllable, although empirical de-
men. In most circumstances it is perfectly adequate to cool the relatively tenninations can lead to reproducible results and Talmon (1980) has made
large specimens in sub-cooled nitrogen or in a halocarbon b~th (§2.4. l~) some calculations relating radiant energy to the amount of sublimation
and, indeed, the specimen may be mounted on the S~M st~b its~lf for this achieved, and is not carried out in the microscope. It is relatively fast as the
purpose. If the specimen is much smaller or, perhaps, if a microbial suspen- bulk temperature of the specimen is effectively unchanged. An advantage
sion or cell fraction is to be examined, then it may be thought preferable of sublimation in the microscope is that the surface can be watched as the
to use a rapid cooling method. The specimen should then be frozen on a ice disappears. However, as a general rule the sublimation of excessive
support of low thermal mass that can subsequently be mechanically at- amounts of additional contamination into electron-optical columns is not
tached to the cold SEM stub (Fig. 3.6). recommended and, if this is done, it is essential that the microscope should
Once the frozen specimen is on the stub, it needs to be protected against have effective cold traps close to the source of contamination. This discus-
two main damaging conditions: firstly it should not be warmed unintention- sion relates to hydrated frozen specimens; that is, those containing frozen
ally above about 160 K (which would cause the inadvertant ~ubl~mation of water. The physical behaviour of ice at different temperatures under
ice, see §5.2.1); secondly, because it is at low temperature, it will act as a .vacuum is now well understood (§5.2.1; §6.5) and the operational paramet-
good trap for condensing molecules (e.g. water and hydrocarbon oils) that ers for the design of equipment may thus be determined. However, low tem-
will contaminate the surface. The means of avoiding these pitfalls are clear: perature techniques, including low temperature SEM, may often be called
the specimen should be maintained (by continuous cooling or by being on for the processing and observation of non-aqueous systems. For such
mounted on a support of adequate thermal capacity) at a temperature below·.•. ····specimens, the behaviour of cold material under vacuum may be quite dif-
160 K · and it should be protected against contamination either by surround- ferent from that of ice. For example, it may be that a component of the spe-
ing it ~ith even colder surfaces and/or by evacuating the vessel in which it 0imen is an oil of low melting point and high vapour pressure, in which case
is contained so that the partial pressures of condensible gases are drastically en very low operational temperatures may be inadequate to prevent subli-
reduced. Different experimental systems have used different combinatio ation into the vacuum. The only remedy would be to use a system with
of these two solutions (§5.2. l ). ficient cooling capacity (possibly using liquid helium) to reduce the
If a cryogen with a melting point warmer than the temperature of the col our pressure of the relevant oil below the point at which it would sub-
stage is used (e.g. a halocarbon), then this may stay as a solid residual con e into the vacuum system. The theoretical calculations would be as for
minant on the specimen surface. This may then either be removed by wa ter/ice, but substituting the appropriate parameters for the oil.
ing the specimen under controlled conditions so that sublimati.on of the so nee the required, possibly etched, surface has been revealed, it is usually
id cryogen occurs, or the problem may be circumvented by usmg a cryog essary to coat it with a conducting layer prior to viewing. Frozen-hyd-
ii
oflow melting point (e.g. propane). . d specimens usually show serious charging if not so coated. This subject 'I
If the sublimation of water (or contaminating cryogen) from the spec1m been fully discussed (Nei et al. 1973; Brombach 1975; Marshall 1975;
surface is required, then some means of doing this under controlled con lin, 1978b; Fuchs et al. 1978b). Brombach (1975) and Marshall (1975)
tions is necessary. Two possibilities exist: firstly, the whole temperature described the phenomenon of surface charging, which was evaluated
the specimen may be raised to > 170 K so that freeze-drying in vacuu.m fully in the paper of Fuchs et al. (1978a). While it is to be expected
start to take place (exactly as in freeze-etching; §5.2.1 and §6.5!. This ure ice (with a DC conductivity of approximately 1-10 nohms m-1;
in principle, be carried out either on the cold stage of an EM or m an ~c s 1974) will show substantial charging, the presence of electrolytes in
sory chamber. Alternatively, surface ice may be sublimed away usi biological samples increases the conductivity by some orders of mag-
radiant heater. The 'freeze-etching' approach has the advantage that r e and thus reduces the charging problem considerably. When a speci-
Robards and Sleytr Low temperature methods in biological electron microscopy Direct viewing and analysis of cold specimens 163
162

men is freeze-dried, its conductivity is reduced and it charges more .readily sputtering systems working at higher pressures than resistance evaporators
than the equivalent fully hydrated material, and consequently chargmg can have also been used in the total absence of contamination (Robards and
be seen to increase as etching progresses. In general, it has been found that Crosby 1979; Beckett et al. 1982). (In fact, to our knowledge, K. Oates was
quite high accelerating potentials can be used (up to 30 keV) so long as the the first microscopist to demonstrate (in about 1977) that the sputter-coat-
beam current is kept very low (Echlin 1975; Woods and Ledbetter 1976; ing technique could be satisfactorily applied to low temperature specimens;
Echlin 1978a,b,c; Echlin 1981). Fuchs et al. (1978a) showed t~at an ~d~­ personal communication.) Other methods of coating at low temperatures,
quate conducting film deposited on the specimen surface effectt~ely elimi- such as ion-beam sputtering, can certainly be used although they have not
nates surface charging but does not necessarily reduce space chargmg. ~~ese yet been applied routinely. For most purposes gold, gold alloy or platinum
authors used electrodes to collect electron currents at different pos1t10ns layers have proved satisfactory for observational work, while carbon, beryl-
around model specimens in an EM and came to the conclusion that charg- lium or chromium have been favoured for microanalytical determinations.
ing need not be a problem so long as the electron range (mean depth of.pen- If the system uses a sputter-coater, then the carbon may have to be evapor- I'
etration) is greater than about half the specimen thickness and the specimen ated in a poor vacuum ( > 5 x 10- 4 Torr; 6.67 x 10- 2 Pa). This is not neces-
is mounted on an effective rear 'electrode' (earth contact). They also noted sarily a disadvantage because there are more collisions of the carbon atoms
that the irradiation of bulk samples with a narrow electron beam leads to with residual gas molecules, leading to a less directional coating (see Willi-
a space charge field close to the surface that reduces the effective depth for son and Rowe 1980 for these techniques). Using carbon fibre 'string' as the
analysis (excitation volume) and, furthermore, that char~e storage de~elops evaporation source, it is simple to produce a carbon layer on a frozen speci-
i i

very quickly, even in coated specimens. Quite ~learly, th1c~ b~l~ specm~ens men in a vacuum of 10- 3 to 10- 4 Torr (1.33 x 10- 1 to 1.33 x 10- 2 Pa).
will normally need to be coated with a conductmg layer to mh1b1t ch~rgmg. Having been coated, the specimen now simply has to be inserted into the
A number of possibilities exists for accomplishing this, includ~ng resistance microscope and onto the cold stage: it must again be borne in mind that
heating, diode and triode sputter-coating, ion-be~m sputtenng, electron· the same stringent conditions of protection from contamination during
beam coating and Penning sputtering. These techmques have all been well transfer must be applied.
reviewed in the SEM literature (Echlin 1978a,c; Echlin et al. 1980) but only
the first two have been used to any significant extent in low temperature 3.3.2 Frozen-fractured bulk specimens
applications. .
To evaporate metals, they must be heated above their boilin~ pomt .<s elatively few specimens are such that the surface of interest is directly
also the previous discussion in §3.2.3). Consequently, evaporative coatm vealed in a SEM without further treatment. More usually, it is an internal
necessarily involve the generation of high temperatures (Table 6.2; §6.4.1 face of which views are required and it is then necessary to cut or fracture
locally around the source. Nevertheless, evaporation of gold, gold. allo specimen before coating it and looking at it in the SEM. Where possible
chromium and carbon onto frozen surfaces without unduly deletenous. · best that the frozen specimen should be cut or fractured because largely
fects has been demonstrated in a number of laboratories. Sputter-coat1 n fracture planes through the frozen material will then be produced. If
has the advantages of requiring only a low vacuum (0.1 Torr, 13.33 Pa) a. surface is revealed by exposing it when the tissue or sample is unfrozen,
also gives a more even coating over static, uneven surfaces than does res.I n deformation can occur and, if the specimen is wet, some etching will
tance evaporation. It has been shown both theoretically and in pra~tI necessary to show the real surface. Furthermore, debris will often conta-
(Robards and Crosby, 1979; Robards et al. 1981; §~.3.6) not only that d10 ate this exposed surface.
sputter-coating is acceptable for providing co~ductmg films on froze~ s o produce a fracture through the frozen specimen, many possibilities
mens but also that the heat input to the specimen can be very lo': md . (Fig. 3. 7) and each worker has to co~ider what is most suitable for
An advantage of resistance heating is that it is usually carried out .1~ ~ h articular material in hand. Whether the specimen is 'cut' or 'cleaved'
vacuum ( < 1 x 10-s Torr; 1.33 x 10-3 Pa) and the theoretical poss1b1h~y temperature, we may expect that a brittle fracture will result (§6.3.1).
contamination of the specimen is thus reduced. However, properly design may, of course, be specific reasons for fracturing at warmer tempera-
164 Robards and Sleytr Low temperature methods in biological electron microscopy Direct viewing and analysis of cold specimens 165

ma.te~ial such ~s polyvinylpyrrolidone (PVP), 'Tissue-TeK' (Appendix 2) or


a similar medmm. The two criteria of good thermal contact with the stub
and good mechanical stability are critical to the attainment of satisfactory
results .
. When the sp:cimen is frozen on a large block or stub, it must be appre-
a ciated that, while the specimen itself may cool to the temperature of the
cryo~en. quite rapidly, the stub can take very much longer to achieve an
eqmh~nu.m t~mper.ature. Thus, it must be ensured that the specimen/stub
co~bmat10n is left m the coolant for a sufficiently long time for this equilib-
ration to take
. place. This can easily mean as much as a minute or so in the
coolant. It is best to test the time taken for each different type of stub to
cool down by carrying out some test runs with a thermocouple positioned
at the centre of the supporting block.
Specimen
support With the frozen specimen on its stub or carrier, we can go on to the next
b
stage. This may well involve inserting the carrier into a transfer device (see
Fig. 3.7. Possibilities for producing fractures through frozen SEM specimens. (a) Using some
sort of knife system (cooled microtome, scalpel, razor blade) a fracture can be produced ,§3.3.7). For the moment, however, we shall concern ourselves with methods
through the specimen while it is maintained cold. Alternatively (b ), a simple fracture can be ~or revealing faces for observation and analysis, i.e. fracturing and/or etch-
induced, for example by breaking apart pairs of rivets that have been frozen together or by mg.
cross-fracturing a tube containing the specimen.

tures (for example, to look at temperature-related changes in an emulsion) Fracturing


but even then it is usually required that the specimen is subsequently main-
tained at the lowest practical temperature consistent with the absence of Many frozen specimens have surfaces that are not of direct interest to the
contamination. Once the fractured surface has been exposed, it may be observer, who requires some internal plane. It is possible either to fracture
etched by subliming ice away so that biological structures are revealed, and e specimen (usually under liquid nitrogen) before mounting onto a speci-
the specimen is then ready for coating and viewing. en stub, or to carefully 'section' the specimen using a cryoultramicrotome
.that a relatively smooth, frozen surface is produced. The first method re-
3.3.3 Preparation lresn? further c?mment, since it is simple and can easily be adapted to
pa~tlcular reqmrements of di~ferent specimens (Fig. 3.7; §3.3.2). Cryo-
The methods of preliminary preparation and freezing of specimens ha ram1crotomy is described in the next Chapter and is used in the same way
been described in Chapter 2. It remains to be decided whether the speci prepa~e smooth surfaces for low temperature SEM as to prepare thin fro-
should be frozen on a relatively large specimen block (stub) or on am spec1mens. In either of these two methods the final specimen ends up
smaller support which can subsequently be mounted onto a larger car .Jhe stub to be put on the SEM cold stage.
These are choices for the individual user to make, bearing in mind the hen a special accessory chamber, either remote from or attached to the
mate objective that is to be achieved. In a typical system (Fig. 3.8), thew~ (a 'de.dicated system'), is used, the frozen specimen is fractured in the
specimen, on its stub, is cooled in sub-cooled nitrogen. Whenever poss1 .~er with .a needl:, knife o~ microtome. The system shown in Fig. 3.8
it is preferable that the stub has good thermal conductivity and a high . es a vanety of simple devices that allow the specimen to be relatively
capacity: copper is the usual choice. It is important that the specimen ly :ractured. This chamber is part of an integrated system that enables
as good contact with the metallic stub, or holder, as possible. Before co. ecimen. to be moved, in a transfer device, from the freezing chamber
it should be 'buttressed' with the tissue solution or, if appropriate, w1 fracturmg/coating chamber and subsequently to the microscope itself.

Direct viewing and analysis of cold specimens 167
Low temperature me/hods in biological eleclron microscopy
166 Robards and Sleytr

Trana fer Gate Knife


device valve mechanism
Sputter
head

Specimen Specimen
a tub table

Fig. 3.8. The EMscopc SP2000 Sputler-Cryo System (see Appendix 2) for low temperature
~anning electron microscopy (see also Fig. 3. 11). (a) Illustration of whole Sputter-Cryo System
with control unit on the left and the preparation unit on the right. (b) Components of the sys-
tem. (c) Schematic diagram o f the system. (Illustrations by courtesy of EM scope laboratories,
Ltd.)
Cold stage
Transf er
for SEM Similar systems have been built-up using equipment already available in the
SEM devic e
laboratory. F or example, a number of workers (Saubermann and Echlin
interface
1975; Fuchs a nd Lindemann 1975) have used a transfer device to move the
frozen specimen onto the cold stage of a vacuum or freeze-etch apparatus
where fracturing and coating can easily be accomplished (Fig. 3.9). Some
dedicated systems have been constructed with rather more precise (but
nevertheless simple) microtomes. An example is illustra ted in Fig. 3.10, in
which the cham ber is actually attached to the microscope itself. In some spe-
cial circumstances it may be useful to observe the specimen in the micro-
scope while it is being fractured. Again, this causes no special problems be-
yond the need to provide suitable vacuum lead-throughs to the micromani-
pulators or other fracturing devices (Hayes and Koch 1975; Koch 1975).
Whatever means is used to fracture the specimen, it is important that the
fracturing tool is as cold as, if not colder than, the specimen itself. If a
b
168 Robards ond Sleytr Low temperature methods in biological electron microscopy
Direct viewing and analysis ofcold specimens
169
mjcrotome is used, then it is usual to provide a positive means of cooling
the blade (see Figs. 3.9 and 3.10) but if simple needle or scalpel blades are
employed then they can be cooled by 'parking' them in close contact with
some convenient cold surface in the system (e.g. in the apparatus in Figs.
3.8 and 3.11 it is possible to cool the needles by placing them in specially
machined recesses in the cold table). It should also be remembered that any
material that is fractured or otherwise broken away from the main body of
the cold specimen will probably be in poorer thermal contact with the cold
sink and w'ill therefore warm up, with the consequent possibility of water
sublimation and condensation onto the adjacent, colder, specimen surface

Side view Front view Specimen


change

Ji~~
Razor blade
Isolation valve

Air lock valv e -<>=--=~~.

n=--+1--Joule-
Thomeon
r.etrigerator

C ryofr acture s tage

Fig. 3.9. A method of using the Balzers BAE 120 vacuum evaporator (see A ppendix 2) as Fig 3. JO. (a)TheAMR IOOOA · ·
a fracturing and coating apparatus for preparing specimens for low temperature scanning clcc· Ptrature preparation unit attach:~a~~~ng clectr~n microscope with the. Bio-Chamber low tem-
Iron microscopy. These illustrations sho w the cold stage, cooled microtome knife and evapora· 810-Chamber show· th h. h ( Appendix 2 ). (b) Diagrammatic representation of the
tive coating arrangements. A large cold trap minimises specimen contamination. (Relabelled tcm. and the joule-~~~ c ig /acuum system, evaporative coating system, the 'shuttle' sys-
and reproduced, with permission, from Sa ubermann a nd Echlin 1975.) rnson re ngerator in the SEM. (Fig. 3.I Ob redrawn from Pawley and
Norton 1978.)
Direct viewing and analysis ofcold specimens 171
Low temperature methods in biological electron microscopy
170 Robards and Sleytr
it should not also be achieved through careful control of temperature under
a dry inert gas (e.g. nitrogen) at atmospheric pressure. The theory of ice sub-
limation under vacuum is fully discussed in Chapter 5, from which it will
be seen that etching is effected, and controlled, by manipulating the temper-
Manipulator
ature of the specimen (ice) surface under vacuum. F or bulk low temperature
specimens, ice sublimation can therefore be brought about by:

(i) raising the bulk temperature of the specimen above about 170 K while
Gate valve under high vacuum (strictly, in a vacuum with a low partial pressure
of water; see §5.2. la) on a temperature-controlled stage such as in the
accessory chambers shown in Figs. 3.8 and 3.9;
(it) raising the bulk temperature of the specimen on its stage in the mic-
roscope; or,
(iii) etching the surface with a radiant heater as provided in the apparatus
shown in Figs. 3.8 and 3.10.

Deeper etching is usually necessary fo r low temperature SEM than for


Specimen stub
freeze-etching, although this depends on what is required to be revealed and
. . hamber of the Sputter-Crye system designed by Robards
Fig. 3. 11. The fractun~g/coatmg ~ o osed to the equipment illustrated in Fig. 3.10, is must be determined empirically for most specimens. (Data provided in Ta-
and Crosby ( 1979). This syste°:; ~smJPt If There is thus a gate valve so that the specimen ble 5.2 in §5.2. la a llow a preliminary estimate of etching times and tempera-
1
designed to operate away fro~ t e se k. The specimen is positioned on a temperature-
. d . d moved via a vacuum-1oc · . tures to be made). After etching by warming the whole bulk specimen, the
can be mserte an re d f ctured or otherwise manipulated. It 1s then
. h e it can be c1eave • ra . block should be recooled so that etching ceases and the surface is stable.
controlled cold stage w er . b tt r coated with a metal (gold or platinum)
h Id 1. so that 1t can e spu e -
moved along t e co s age orated carbon (not illustrated). The same principles as those When fracturing and/or etching have taken place, the specimen is ready for
or, altemat1vely, coated with evap SP 2000 stem (Fig. 3.8). (Redrawn from Robards
described here are used Ill the EMscope sy viewing but mus t usually be firs t coated with a conducting film . If etching
a nd Crosby 1979.) has been effected in the m icroscope, then the specimen will have to be
retrieved so that it can be coated. Some operators like to watch the specimen
(this subject is more fully discussed in §6.5. 1). Care should be taken to try as ice sublimes from its surface in the microscope. Such specimens will, of
to avoid this problem, either by attempting to move the fractured fra~ents course, charge badly as the surface will not be coated or, if a coating had
as far away from the specimen surface as p?ssible and/~r by prov1~mg .as already been deposited, this will become disrupted by the etching process.
much cold-trapping surface around the specimen as possible (e.g. as m Fig. If this method is used (and contamination of the microscope column in this
3.11). · f ·t mayeither way is not recommended), then the specimen (on its stage) should be rapidly
Having successfully produced the required specimen sur ace, I re-cooled once the desired degree of etching has been achieved, before it is
be coated and viewed directly or it m ay be ' etched' so that t.h~ surface tt~~~~ withdrawn for coating.
graphy is accentuated (by lowering the_ice.level) or so that, if 1t has no
fractured at all, any obscuring surface ice is removed. 1.3.6 Coating

3.3.5 Sublimation As described previously (§3.2.3), coating is usually achieved either by resis-
. . .1 , ·ed out under tance heating or by sputtering methods. These two approaches are repre-
Sublimation of ice from bulk specimens is almost a ways earn h- ect·1ons sented in the apparatus in Figs. 3.10 and 3.11. During coating, the specimen
- - · 1 hy as for t m s
vacuum although there is no reason m pnnc1p e w ,
Robards and Sleytr Low temperature methods in biological electron microscopy Direct viewing and analysis of cold specimens 173
l72

a the heat source (radiant heat load falls off in relation to the inverse square
333 of the distance), that the input power for evaporation or sputtering should
be kept as low as possible, and that it is better to minimise heat input by
323 coating for a long time at low deposition rates rather than for a short time
at high rates (Fig. 3.12; see §6.4.1 b for further discussion of this topic). If
g evaporative techniques are used, then some means of ensuring that the spe-
-------b
~
• 313

"~
--- --- cimen is evenly coated should be sought (such as moving it during coating).
It should be mentioned that a disadvantage of coating with carbon for low
,,.. ,,.. .... temperature work is that the thin carbon film often cracks, so leading to
~ 303
...• //
./!.-··-----·-··-
··-··-··-··-··-··-··-··- ··-··-·· -··- c
impaired themial and electrical conductivity.
/
293
3.3.7 Transfer

60 80 100 120 To accomplish the different steps necessary in preparing a specimen for low .
20 40
Time (s)
temperature SEM, the specimen must be moved to different work positions.
F
. · d ·
3 12 Typical heat mg curves unng spu
lier-coating. (a) Specimen heating during sputter·
. · A
This is done either within a dedicated accessory chamber attached to the
tg .. · · . kV (b) Specimen heating during sputter-coating using 20 m at SEM (Fig. 3.10), or by using a portable transfer device. The first method
1 2
coating usmg 40. mA at '. · . tter-coating using an experimental quadrupole head
1.0 kV. (c) Specimen heatmg dunngdspu ' b db Robards et al (1981). This is similar to the has the advantage of simplicity and, perhaps, speed in operation. The
t' ' th 25 mA at o 44 kV as escn e Y · d h h
opera mg w1 . . co e SP2000 Sputter-Crye System (Fig. 3.8) an s ows t at second can be more versatile and does not require the accessory chamber
system incorporated mto the EMs P. . . si'gnilicant problem as lo ng as the operating
. h t'
specimen ea mg
during sputter-coatmg IS an 1Il
d h
f
t' head is properly designed. (Drawn rorn ata
d to be attached to the SEM. There do not appear to be overwhelming advan-
parameters are chosen correctly an t e coa mg d
presented m Robar s et a ·
l 1981 )
·
tages or disadvantages to either method. If the cold specimen is to be moved
external to the vacuum of the microscope column, then a protecting transfer
h Id be maintained at the lowest possible temperature cons.istent with the device is essential. A very simple type of device is illustrated in Fig. 3.13
~a;~city of the system to protect it from contamination. It will ~e seen that (Saubermann and Echlin 1975). This merely relies on the cold-trapping
the Bio-Chamber (Fig. 3.10; Pawley and Norton 1978) uses a high vacuum capacity of a relatively high thermal capacity tube into which the specimen
whereas the Sputter-Cryo (Fig. 3.8; Robards and Crosby 1979) works at low is withdrawn. Provided that the tube is reasonably long, most condensable
owever the critical point is whether or not there are any condens· molecules will be trapped on the walls of the tube before they can arrive
vacuum. H , . h s tter Cryo
able molecules in the residual vacuum and, in .this respect,_ t. e pu - ened at the specimen. A rather more elaborate system is shown in Fig. 3.1 4 in
·des a very low contamination environment because tt is rarely op which the specimen is retracted into a cold-shrouded ' garage' and is also
prov1 . . d · t tly purged
to the atmosphere (only for occasional serv1cmg) an is c~ns an. .h contained within an evacuated chamber. The transfer device carries its own
'th dry argon gas. Consequently, while the total pressure is relat1v~\y htg .
wi · I I H O) 1s extre-
the partial pressure of condensable molecules (particu ar y 2 . . I Brass tube Low thermal Spec imen
mely low. Furthermore, the cold shrouds in both systems serve as ~d~~~":. \ conductivity rod holder

~
anticontamination surfaces. As well as these types of purpose-_des1~ th ior
other authors have successfully interfaced transfer devices w1 ', ed
tem , · an be transierr
example freeze-etching units so that the frozen specimen c h nd
into tha~ apparatus to be coated (Saubermann and Echlin 1975; ~uc ~~ow
Fig 3.13. A simple transfer device for cold SEM specimens. The tube is first cooled with
quid nitrogen so that, when the specimen is retracted into it, potentially contamina ting
Lindemann, 1975). The salient points to be remembered for co.attn~: frorn ~pour molecules (such as water) are trapped on the cold metal inner wall. (After a design de-
temperature are that the specimen should be at a reasonable d1stan scribed by Saubermann and Echlin 1975.)
Robards and Sleytr Low temperature methods in biological electron microscopy Direct viewing and analysis of cold specimens 175
174

fig. 3.1 5 A 'self-shrouding' stub as used in the EMscope SP2000 Sputter-Cryo system . (Dia-
gram redrawn and reproduced with pennission from EMscope Laboratories Ltd.)

coated chamber (Fig. 3.8) but with the specimen stub carrying its own high
a thermal capacity, automatically closing cold shroud. These examples repre-
sent typical approaches and constructions; there are many others described
in the literature (e.g. Aldrian and Windisch 1984). What is important is that,
once the specimen has been frozen, its temperature should never rise so
much that ice sublimes (unintentionally) or so that unwarranted ice recrys-

tO
Shuttle temper ature
sensing device
tallisation takes place. A good working rule is that the surface of the speci-
men should not inadvertently rise above 140 K. It is thus necessary to be
.---.J.,........-tcOD<;;r~c_J~'°I T"'m" '""'·•U• able to monitor the complete temperature history of a specimen from freez-
ing to viewing and to see that no untoward temperature excursions have oc-
Transfer ro d curred. Contamination during transfer is minimised by the provision of cold
shrouds and by ensuring tha t the vacuum vessels are always back-filled with
a dry inert gas (usually argon or ni trogen) rather than with air, and also
b vacuum p umping line . h by evacuating 'dead volumes' between adjacent vacuum valves before these
F 3 14 A cooled transfer device for frozen SEM specimens. (a) Photograph showin~ \' are opened to allow continuity between evacuated compartments.
1
ig. · · ·holder gate valve and Dewar vessel. (b) Diagram illustrati ng the components od '
specimen , · · t tly coole 1''
transfer device. When enclosed in the evacuated device, the specimen is cons an rtesv ol
temperature is monitored at all times. (Photograph supplied and reproduced by ~ou d ;~11h
13.8 Viewing and analysis
Cambridge Scientific Instruments, Ltd. (Appendix 2) .. Diagram redrawn and repro uce
permission from Cambridge Sc1ent1fic Instruments Ltd.) The basic requirement of the EM is that it has a cold stage and a cold anti-
liquid nitrogen supply and Fhus maintains a very low temperatu re through~ contaminator plate. The cold stage should be temperature-controlled and
out the change from one work position to another. Yet anot?er approa~. capable of going to as low a temperature as possible (certainly < JOO K).
is embodied in the apparatus illustrated in Fig. 3.15, which agam has an e~a In operation, the anticontaminator sho uld preferably be kept 20-30 K
s in biological electron microscopy Direct viewing and analysis ofcold specimens 177
Low temperature me 'hod
176 Robards and Sleytr
colder than the specimen itself. For SEM work, the cold stage can be kept
rather simple and, provided that the 'fit' of the specimen stub into the stage
is good and secure, there should not be a problem in temperature attainment
and control. Three forms of cold stage are illustrated in Fig. 3.3 in §3.2. l;
one uses a stranded copper cooling pathway (as in the system illustrated in
Figs. 3.8 and 3. I 1), another uses a piped system (Fig. 3.16), and the third
type uses the Joule-Thomson refrigerator principle (Fig. 3.10). Other com-
mercial forms of heating/cooling stages are also available and may be sui-
table for low temperature applications, although some of these are more
complex (and hence more expensive) than is required and may not have the
ultimate low temperature capability required for hydrated specimens. It is
important that the cold stage is mounted on thermally insulating (e.g. ny-
lon) spacers to separate it from the main body of the SEM stage. However,
some provision for maintaining an electrically conducting pathway from the
specimen to instrument earth must be made. A fine wire of low heat transfer
capacity but high electrical conductance is adequate.
If microanalysis is contemplated, then the normal instrumental modifica-
a tions to minimise background counts should be performed (see Chandler
1977). Surfaces close to the specimen which are likely to generate spurious
X-rays should be coated with carbon or fabricated from beryllium. It should
be remembered, however, that carbon is not a good heat conductor and any
carbon shrouding should take this into account. For example, it would be
Stainless steel inner Dewar preferable to coat the top of a copper specimen stub with a thin film or
·wafer' of carbon (or, even better if available, beryllium) rather than to
vacuum 'jacket'
fabricate the whole stub from carbon as might be done for analysis at
Outer vacuum Dewar
LNz gravity feed
ambient temperatures.

ItfAJ.1.~~& ·::·~·:~· ::~~:~i~'"


to cold stage
after initial
pressurisation
3.4 Frozen thin specimens

•Exhaust' TL-'..:!~'""""""tt;;;;=n=~~\~;;;;; Inter! ace PTF~ Surface signals are the predominant ones retrieved from bulk specimens
to air
LL------~ Flexible to stage insulation (§3.3). At low temperatures, such samples have almost always been studied
cryogenic mSEMs. However, when it comes to thin specimens, we have the option
tubing
of low temperature conventional TEM, or STEM in an instrument that is
b Hexland liquid nitrogen cold stage either basically a TEM or an SEM. If an SEM cold stage is adapted for
· · d ti uid nitrogen (see Appcndi•' transmission work, then the problems are not too severe as there is generally
Fig. 3. J6. The Hexland cold sta~e for an SE~l~:~~:i~~of t~ cold stage, showing the d1 plenty of space for necessary components and the engineering is relatively
(a) The Hexland cold stage. (b) D1agrammat1c h t· e to accept the specimen mounts. (Phot raightforward and similar to that required for bulk specimens (indeed, the
liquid nitrogen piping and the dovetail slot m t f e;~gTechnology - H exland Ltd. Diagram
graph supplied and reproduced .b y lcourtelsyd ~y EM Technology - Hexland Ltd.) llrnc stage can often be adapted for either mode of operation). When it
drawn from an ongma supp ie
178 Robards and Sley tr Low temperature methods in biological electron microscopy Direc/ viewing and analysis of cold specimens
179

comes to thin specimens for observation in true TEM or STEM instru- unn~e~sary to provi~e a conductive coating for thin frozen sections rior
ments, then the problems are all magnified by an order of magnitude or so. to v1ewmg and analysis. p
The specimen is much smaller, has a lower thermal capacity, and its temper- Talman and his colleagues (Talman and Thomas 1977a,b, 1978, 1979;
ature is therefore much more liable to be changed, especially during Talmon et.al. 1979; T~lmo~ l982a) have considered in detail the effects of
transfer. The specimen mount must also necessarily be smaller and this bea~ he~tmg and rad10lys1s of ice. U sing a model for a 'modera tely thick
means that it is common to mount the specimen into a holder which is itself specimen (a red blood cell mounted on a metalised polymer film) Talm
transferred into the microscope (Fig. 3.6; §3.3.1 ). and Thom~s (I977b) det~rmined the theoretical temperature ris: undero:
range of different operat1.ng conditions. It was calculated that a STEM
3.4.1 Frozen thin sections probe current of < 10 nA m a 20 nm diameter beam Id · h ·
wou raise t e speci-
men temperature by a~out 5 K , the temperature increase being proportional
The preparation of frozen thin sections is described in the next Chapter and, to t~e probe current 1f all other variables were maintained constant. The
for the purposes of the present discussion, it is sufficient to assume that we maximum surface temperature rise of a specimen on the cold stage of a SEM
have produced a section which is essentially a fragment of biological tissue depends on beam current, incident beam energy, ratio of sample area to
'embedded' in a matrix of ice. Reference to tables of sublimation rates (Fig. scanned area, and the temp~ratu~e and thermal diffusivity of the sample.
5.3 and Table 5.2 in §5.2. l a) show how very quickly such a section would The temperature .change vanes d!fectly in relation to beam current and is
completely freeze-dry at temperatures much above 160 K, and so the main- '". versely proportional to specimen thermal conducti'vi'ty s
. . ome representa-
tenance of a very low temperature during all processing and observational tive data .p rovided by Talman and Thomas (1977b) are provided in Table
stages is of paramount concern. This may be more difficult than is immedia- 3..1 and
. Fig. 3.17. . . It can easily be seen (Fig· 3. J 7) that water Ioss by subl'1ma-
tely obvious. Whereas a bulk frozen specimen can be mounted on a large, llon 1s very sens11Ive to surface temperature, and the beam conditions should
high thermal capacity, high thermal conductivity metal block, the thin sec- therefore be selected to minimise temperature rise (see a lso §5.2. I and §6.5).
tions must be mounted on a grid (or similar supporting structure) where Talm~n and Thomas went on (1978) to calculate heating temperature
much of the frozen material will not be in contact with an efficient heat profiles m
. . moderately thick SEM/STEM spec1'mens on a co Jd stage. F or
transfer element (e.g. a copper grid bar) and may, in fact, be supported on thm .specimens, one temperature maximum in the profile was found near the
a plastic film with poor heat transfer characteristics. There is always a rather specimen su~face and this maximum shifted away from the surface as the
poor heat conduction pathway through the ice of the specimen to the near- tempera.ture mcreased. For thicker specimens (those approaching the maxi-
est cold sink. A second criterion for the effective use of frozen thin sections mum thickness for electron penetration), two maxima were found: one close
is, therefore, that the sections should be mounted in the best possible ther- to the su.rface, t~e other close to the sample/substrate interface. These re-
mal contact with the ultima te cooling source, this mounting being cons1s· :ults are mterest~ng and potentially important, particularly in relation to X-
tent, of course, with the requirements of the particular viewing/analytical ~ay .m 1 croanalys1~ where (as opposed to using the secondary electron mode
methods to be used. If these two main criteria are fulfilled (lowest operatio- ofr image formation) the temperature at the main site of X-ray emission is
0 some consequence.
nal temperature and best possible thermal contact with the substrate), then
the stages of coating the section with a conducting film (if necessary) and bleThe c~n.clusion of this and other work on beam heating is that it is possi-
transferring it into the microscope can be accomplished without damage lo to mimm1se temperature increase in the specimen by operating with th
The environment within the microscope is, however, far more hostile. In a i:~~'!ractical. beam current and using well-designed cold stages to ensur:
high vacuum the specimen is subjected to beam intensities greater than. 300 OCcur m~t poss1~~e sample temperature. However, radiation damage may
C m - 2 which may lead both to direct heating and also to the radiolys1s of
cross-'1In
· kistly ansmg from ionisation processes that lead to bond breakage
ng mas I df . ,
ice. As shown in the discussion in §3.3. l , however, problems of charging in Pfocc ' s oss an ree radical formation (Talman I 982a) These
thin sections are less than for bulk frozen specimens (Echlin 1975, 1978b, .I-. sses proceed independently of tempera ture a lthough it has been.clearly
~onstrated that th ,rr. . .
1981 ; Fuchs et al. 1978a). This means that many workers have found 1 e e11ects of radiation damage are reduced if the
Direct viewing and analysis of cold specimens 181
hods in biological electron microscopy
Robards and Sleytr Low temperature met
180

TABLE 3.1
Maximum surface temperature rise of thin and thick sections

~T ~T
Conditions
Beam voltage Thickness (Jul/scan) (spot mode)
(K) (K)
Beam current Spot size
(kV)
< I I
80 pA 20nm
25 IOOnm <I 8
6.0nA 0.1 µm
25 JOO nm <I <1
80pA 20nm Fig. 3. 18. The general molecular structure ofpolyimides. (Redrawn from Talmon et al. 1979 .)
25 2.0µm < I 4
6.0 nA 0.1 µm
25 2.0 µm <I <I
1.0nA 20nm
100 2.0µm <I 15 specimen temperature is lowered. The explanation of this appears to reside
50 nA 0.1 µm
100 2.0µm in the reduced mobility of the products of radiation damage lessening the
chances of subsequent deleterious interactions (Cosslett 1978). Under the
From Talmon and Thomas (1977b).
conditions within an electron microscope, ice can be readily radiolysed by
an impinging electron beam . .This effect may well be far more important
than direct heating of the specimen and becomes increasingly pronounced
when high electron doses are required as, for example, for high magnifica-
t10ns or for microanalysis (Talmon et al. 1979; Talmon 1982a,b). Talmon
0.1 et al. (1979) showed that free radicals produced from the radiolysis of ice

-... (OH", H" and H0 2°radicals) significantly etched Formvar films used as sub-
strates fo r transmission microscopy. They concluded that more resistant,
.sE 10
-2
hydrophilic polymers, such as the polyimides (Fig. 3.18), are preferable and

.!!
demonstrated a reduction in etching when a 20-40 nm thick polyimide film
0 was used in place of Formvar. Carbon films and carbon films coated with
• 10
-3
Si0 1 were also resistant to etching by the radiolytic products of ice.
~
.,. Finally, in relation to the radiolysis of ice, it should be noted that not only
...
£ may the specimen be etched by the free radicals and ionised molecular frag-
ments, but there is also the mass loss of the ice itself to be taken into ac-
•c 10-4
.>/.
count. This loss increases in proportion to the beam density and is therefore
~" al a maximum when magnifications are high, when (in a scanning mode)
E
"E 10 -5 a small area is scanned, and when microanalysis is being carried out. An
·;;
::E
.. example provided from theoretical calculations by Talmon et al. (1979) is
illustrative. Using a beam current of O. l nA at 100 keV (10 nC) to scan an
area of 0.1 x 0.1 µm on a 1.0 µm thick frozen-hyd rated section for JOO s,
the beam heating would be quite negligible but 1.5 x 10- 14 g of ice would
be lost through radiolysis. However, assuming that the specimen has the
120 130 140 150 160 170 density of water, it would only have a mass of 1.0 x 10- 14 g ! Although the
Specimen temp (K)
etching effect may have been exaggerated in this example, it can be seen that
d ecimen as a function of temperature. \\a
Fig. 3.17. Thickness loss of a ~rozen hydrate sp f m Talmon and Thomas l 977b.} most of the water would be lost, and this greatly limits the application of
loss through only one face ts assumed. (Redrawn ro
Direct viewing and analysis of cold specimens 183
182 Robards and Sleytr Low temperature methods in hiological electron microscopy

quantitative analysis. Although in the example provided 100 keV electrons the frozen-hydrated to the freeze-dried state.
were used, lower voltage electrons (as in SEMs) would lose more energy to (vii) If the specimen is deliberately freeze-dried in the microscope there
the specimen and the anticipated mass loss per coulomb of impinging elec- should be a disappearance of the oxygen peak from the resid~al gas
spectrum.
trons is likely to be greater in an SEM than in a STEM. As stated earlier,
even though the radiolysis itself is temperature-independent, the subsequent .
diffusion of the reaction products is not and it is therefore still beneficial · A further possibility is the 'electron transmission method', which allows
mass-thickness determination by monitoring either the transmitted electron
to maintain the specimen at the lowest possible temperature.
The above comments concerning possible sublimation or radiolysis of ice'.-. beam (Halloran et al. 1978) or the zero energy-loss electrons (Hosoi et al 'I
1981; Leapm an et al. 1984) ·
lead on to a major matter of concern in all low temperature work with thin_:
specimens: how can we tell that the specimen is still frozen-hydrated an > Clearly, it is not possible to apply all these criteria to every observation
:I

not freeze-dried? For bulk specimens this is a relatively insignificant ques --·-an frqzen-hydrated specimens. Nevertheless, it is of the utmost importance !

tion, but for thin specimens it is a paramount consideration. In a detaile to. know whether the specimen is fully frozen-hydrated, partially freeze- I

discussion of this problem, Saubermann and Echlin (1975) outlined the fol dned or completely freeze-dried. Only by giving this matter detailed consid-
I
':'.eration~ in conjunction with a precise knowledge of the relevant temperatur-
lowing possible methods of detecting whether ice has been lost:
' total pressures and partial pressures of surrounding gases, is it possible I

I 1

By monitoring the specimen temperature and the surrounding to have any confidence in the interpretation of data, whether structural or
(i)
pressure and partial pressure for water during preparation, exarnin · roanal~tical. Now that the requirements for preparing and viewing thin
tion and warm-up. In this way it is possible to detect changes in zen sections are better understood, some remarkable results are starting , I

temperature and partial pressure of water vapour that would indi emerge (e.g. Chang et al. 1983).
that sublimation is occurring. When the, presumptively hydrated, s
cimen is warmed-up there should be a corresponding increase in .2 Frozen thin layers
partial pressure of water and the total pressure of the residual ga
the vacuum system. (It should be borne in mind, however, tha. :View of the very clear correlation between the effects of radiation damage
small change in partial pressure of water brought about by wa temperature (Cosslett 1978; Glaeser and Taylor 1978), those who wish
a frozen thin section would, in practice, be almost or completely ork at the highest resolntion on biological specimens will necessarily
.- to u.se low temperature stages. Indeed, the ultimate aim is to work at
tectable.) .
The determination of section mass by measuring 'continuum' r_-_ hel1:1m temperature although this may necessitate complex and costly
(ii)
tion during and after analysis will reveal no significant change · e~L, possibly 1nvolv1ng superconducting lenses (Fernandez-Moran

specimen has remained fully hydrated. Heide and Urban 1972; Hiirl 1974; Dietrich 1978; Butler and Hale
If the thin specimen is in an instrument equipped with both Lefranc et al. 1982). It is not the intention to discuss here the reasons
(iii)
(STEM) and SEM imaging, then it is possible to observe d' aking observations on specimens in the form of thin layers at low tem-
changes between the frozen-hydrated and freeze-dried states. es: these are fully discussed elsewhere (e.g. Dubochet et al. 1982).
Similarly, structural changes are seen between frozen-hydra! . rt to say that thin layers (of the order of 100 nm thick) can be frozen
(iv)
ently rapidly so that good structural preservation is achieved. This is
freeze-dried specimens.
Sections shrink during warm-up if they are still frozen-hydrate. ularly true when observing protein crystals, for example, where the
(v)
ontent may be only about 50%. As with thin frozen sections, it is
Two further methods were proposed by Dorge et al. (1975): thaUhe specimen is maintained in the fully frozen-hydrated state
(vi) A comparison of peak: background ratios of the elements bei rt rs mtended to be partially or completely freeze-dried). Specimens
lysed should show a dramatic increase as the specimen chan refore either be produced directly at the correct thickness for obser-
184 Robards and Sleytr Low temperature methods in biological electron microscopy Direct viewing and analysis of cold specimens 185

Subllmatlon of ice
ties, mesh size, whether it will interfere with X-ray analysis, and other cri-
\ I teria. The specimen on the grid is often sandwiched between plastic films
\, \ ,' ,I which serve both to support it and also to reduce water loss by sublimation.
, ,'
'' \ I
, I
' Thin specimens are usually of one of two types: frozen suspensions, liquids

~t~i%~~: )~
and the hke, or frozen thin sections. Thin sections will be dealt with more
film ____,,@ f fully in the next Chapter although the technology for processing the sections
once they have been obtained is essentially similar to that for thin layers.
Specimen support Thus, we shall look at this topic first.
· 3.19.
Fig.
Controlled sublimation of ice to reveal the specimen while maintaining it in a
frozen-hydrated state. Thin layers
vation or they must be reduced to working thickness by controlled sublima- /
tion of ice (Fig. 3.19). Furthermore, once the required thin hydrated layer!> Thin layers may take the form of liquids in which components such as virus
has been produced, it must be treated most carefully to ensure that _no addi"o' ,particles, membraneous fragments, subcellular fractions, and protein crys-
tional water loss occurs. This protection can be achieved by combmmg the•• >.tals may be suspended or with which other liquids may form an emulsion.
right conditions of temperature in relation to vacuum and by physically pr()--! n other words, it is possible to produce a layer of such materials so that
tecting the specimen on both sides by thin films of plastic and/or carb can be frozen very rapidly indeed and then observed in the transmission
(§3.4.4). Frozen thin specimens have also been observed on a specially c~ ode without further treatment (apart, perhaps, from some sublimation of
structed cold stage in the high voltage electron microscope (Fotmo and G ). It is usual to freeze such specimens directly on their supporting coated
dings 1984). · . . •ct. As the specimen is so small, freezing can be extremely good. For exam-
The production of frozen-hydrated thin layers of specimens such as vir , Dubochet and McDowall (1981) and Dubochet et al. (1982) have shown
es, protein crystals and cell membranes does not involve any ch~mi~al tr at if copper grids, coated with a thin, hydrophilic carbon film, are allowed
ment and specimens are obtained from which very high resolution tnfo fall through a mist of microdroplets from a nebulizer into liquid ethane
tion can be retrieved. There have been some remarkable demonstrations' liquid propane (Fig. 3.20), water layers < 1.0 µm thick are truly vitrified
the applications of such techniques (Taylor 1978; Lepault et al. 1 judged from electron diffraction). This is a recent development that will
Adrian et al. 1984; Dubochet 1984; Lepault 1984) and, while the techni oubtedly be more widely explored using specimens other than simple
demands of specimen preparation and microscope cold stage construct: ·ds alone. 1i

are complex, they are, nevertheless, now quite well understood. With:! ylor and Glaeser (1976), Taylor (1978) and Talmon et al. (1979) are
i
exception of the preparation steps, subsequent treatment .of th1_n froz~n g those who have provided details for the satisfactory production of 111

ers is essentially the same as for frozen-hydrated thin sections, 1nclud1n . n thin layers. The third paper is particularly useful in describing the
1\

requirement for adequate protection during transfer stages. Finally, nt practical details. The liquid sample is trapped between two thin
worth commenting that, although it is imperative that specimens are : . supported by copper grids. For reasons already given (§3.4.1 ), Talmon I
I
cold if high resolution ( <0.34 nm cited by Taylor 1978) is to be achr very much prefer polyimide films about 20-40 nm thick, whereas Tay-
d Glaeser (1976) have used SiO-coated carbon films for their work. 1:
few of the present cold stages currently available for TEMs have s
good mechanical stability as those for use at normal ambient tempe on and Rowe (1980) give details for preparing silicon monoxide films.
(Taylor 1978). ides have good heat, radiation and chemical resistance and are much I
in these respects than, for example, Formvar. Thin films of polyimide
3.4.3 Preparation produced by dipping a glass microscope slide length-wise and verti-
Thin specimens usually end up on some form of grid, with or without" to a 0.75% (w/v) solution of the prepolymer (the poly(amide-acid)),
porting film. As usual, the grid is selected on the basis of its thermal for 10 min at 90°C and then curing at 300°C for 3.5 h to obtain the
Robards and Sleytr Low tempera I ure m"lhods
c
in biological electron microscopy Direct viewing and analysis of cold specimens 187
186

to allow liquid to get onto the outside surfaces of the grids or the specimen
will be too thick when frozen. To avoid excessive premature drying through
evaporation (especiaJly if particularly volatile components are present) this
E mounting process should take place in a glove box in which there is a satur-
Suspension E
ated atmosphere of the relevant liquid. The sandwich is then rapidly frozen
by one of the methods described in §2.4.1. Talman et al. (I 979) plunged their
specimens into sub-cooled nitrogen but propane would give higher cooling
rates, particularly if mixed with isopentane so that it does not form a solid
layer when the specimen is transferred into liquid nitrogen. Although the
method described here is simply one specific technique, it provides accept-
! able solutions to all the fundamental problems that need to be overcome
• .· in preparing this sort of specimen. Other support films could be used (such
"'"' as hydrophilic carbon films) but the principles and constraints wiJI be the
"• same. The next stage is to pick the specimen up from the liquid nitrogen
/ . . . f -freezing apparatus as used by Dubochet et iif which it wiJI have been placed after freezing and mount it for transfer
Fig 3 20 Schematic 1llustrat1on o sprfayh t . d droplets have a volume less than 3 the next processing operation.
(t98i).to. produce vitrified water. 50% o t ea omtdse 'data frequency of about 1 drop
. d n the 200 mesh coate gn .
µm3 and these are d epost1e 0 . . df . . about 0 12 5 and the rate of entry I
T h . b tween depos1t1on an reez1ng ts . 1 1982)
per square. e time e . _ 1 (Redrawn from Dubochet ct a. ·
the cryogen (ethane) is about 2. 0 ms .
100 mesh grid

fi 1 1 .mi de The coated glass slides are dipped into 12% hydro flu Liquid layer Polyimide
ma po y~lO s .so that the films can easily be floated off onto a water
films

;:~~. ~~c~ films are very fragile and are put onto grids usmg t;e ~paa~~ac
.b db HaJI (1966). This involves makmg a gnd spa e Y .
od desert e Y . mm to a wire handle whic
. f 200 h foil about 50 mm x 50 , Preparation ofa thin layer for freezing. (Redrawn from Talmon et al. 1979.)
a piece o mes d , d su ort the mesh itself. The grids are place
also bent to surroun an
the floating film and then gen y pie
.
~i ked up by raising the spade from b
d ·ih 'Skybond 705' (Monsa
d .d Talman et al. have use ei er . .
the,;:~~~(C~~~-~eigy) as their original material for casting . ;.otm1~; Cryosections

~:ee Appendix 2 for list of su~~lier:rw:~~~eisa~~e~o~~:t~~~~ea 3~~ m( ctions, once mounted firmly onto a grid (§4.3.2c) can be treated in
solntions of the prepolymer, e1 er .d the final solution into
of n-methyl pyrrolidone and xylene to prov1 e ' lly the same way as thin films. Some workers prefer to place an addi-
Iastic film over the frozen section. This can be accomplished using
the microscope slides are dipped. . f EM 400 mesh film-covere
p' method or by using a second, 'matched', coated grid. Although
To prepare a.thin layer of:1~:da :eel of~lter paper: a microdro yimide technique of Talman et al. (1979) clearly has merits, many
is placed, film side uppennos , h fil d then a JOO mesh, coat
. d 1. .d . laced on t e I m an s still use Formvar or CoJJodion films for this application. Produc-
the reqmre iqm is p d the droplet (taking care, if. .these is straightforward and described in detail in Chapter 3 of Willi-
. .. d film side downwar s, over . ·ct
IS pos1t1one ' .d ) This squeezes some hqu1 Rowe (1980). Nylon films have also been used (Saubermann and
ary, to align the bars of the two gn; . ·ct (Fig 3 21) It is import . 975).
leaving a thin film between the coate gn s . . .
Direct viewing and analysis of cold specimens 189
Rohard~ and Sleytr Low temperature methods in biological electron microscopy
188
J.4,6 Coating

In many instances coating is not required for thin specimens, since the prob-
lem of charging is not the same as for bulk specimens where the vast major-
ity of the impinging electrons are trapped by the specimen, However, depo-
sition of a conducting layer, such as aluminium or carbon, may be required,
and then the precautions already described in §33,6 are necessary, The
problems of obtaining a uniform coating are less acute (the specimen is flat-
ter) but the probability of heat damage is significantly increased, The speci- i!
men rr1ust be maintained at its lowest possible temperature while on the cold
stage, Once again, it should be recalled that it is much better to use low
Ii
deposition rates (low energy input) for longer periods rather than high rates '
!
for shorter times (see Fig, 3, 12),
! II
Transfer
i 11
specimens are to be observed in transmission mode in an SEM, then the 1I

ansfer devices described in §33, 7 will, with small modifications, be suit-


e. However, the situation becomes more complicated if a true transmis-

Baffle plate

LN 2 reservoir

Protecting
sleeve

Co!d stage transfer module described by Perlov et al. (1983) for the JEOL JEM
OOCX electron microscope (see also Fig. 3.28). (Redrawn from Perlov et al. 1983.)

. Cryotransfer unit PW6361 for the Philips 400 TEM (see Appendix 2). This is part
_plete system (see also Fig. 3.25) for the transfer and viewing of frozen thin specimens.
allows the frozen specimen to be brought, at liquid nitrogen temperature, into the
1 where it can be loaded into a cooled specimen holder in a dry nitrogen atmos-
e specimen holder is then transferred into the microscope under the protection of
echanical shielding system (Fig. 3.25). (Reproduced by courtesy of Philips Electron
Optics.)
Direct viewing and analysis of cold specimens 189
Low temperature methods in biological electron microscopy
188 Robards and Sleytr
J. 4.6 Coating

In many instances coating is not required for thin specimens, since the prob-
lem of charging is not the same as for bulk specimens where the vast major-
ity of the impinging electrons are trapped by the specimen. However, depo-
sition of a conducting layer, such as aluminium or carbon, may be required,
and then the precautions already described in §3.3.6 are necessary. The
problems of obtaining a uniform coating are less acute (the specimen is flat-
ter) but the probability of heat damage is significantly increased. The speci-
men must be maintained at its lowest possible temperature while on the cold
stage. Once again, it should be recalled that it is much better to use low
deposition rates (low energy input) for longer periods rather than high rates
for shorter times (see Fig. 3. 12).
Loading _ _....~­
block
3.4.7 Transfer

Cable If specimens are to be observed in transmission mode in an SEM, then the


release
transfer devices described in §3 .3.7 will, with small modifications, be suit-
able. However, the situation becomes more complicated if a true transmis-
Dewar
vessel
a

Protecting
sleeve

Fig. 3.23. Cold stage transfer module described by Perlov et al. (1983) for the JEOL JEM
IOOCX electron microscope (see also Fig. 3.28). (Redrawn from Perlov et al. 1983.)

Fig. 3.22. Cryotransfer unit PW6361 for the Philips 400 TEM (see Appendix 2). This is part
of a complete system (see also Fig. 3.25) for the transfer and viewing of frozen thin specimens.
This system allows the frozen specimen to be brought, at liquid nitrogen temperature, into the
lransfer vessel where it can be loaded into a cooled specimen holder in a dry nitrogen atmos-
phere. The specimen holder is then transferred into the microscope under the protection of
1 double mechanical shielding system (Fig. 3.25). (Reproduced by courtesy of Philips Electron

Optics.)
190 Robards and Sleytr Low temperature methods in biological electron microscopy
Direct viewing and analysis of cold specimens
191
Cut outs for Xray detector

Specimen
Tip of carrier Thumb
h Ider......_

Spec i men

,..,,;":±:~~Ji:::ifi~I
Vitreous silica for

Beryllium
Specimen tip couplin g

Fig. 3.24. Double-til t specimen tip for the Phi li ps EM400 cold stage as manu fa ct ured by
Gatan Inc. (Redrawn and reproduced by courtesy of Ga tan Inc. ; see Appendi x 2. )

sion instrument is used; not least because the low thermal capacity of the
specimen means that extra precautions must be taken to ensure that und ue
warming does not occur during transfer. One practical problem has been
that, while a number of manufacturers have provided cold stages for tra ns- b
mission microscopes, few have also supplied suitable transfer devices. The
situation is now gradually changing, with some well-designed systems, such

Cooling hose
Shield
J

Flushing

Fig. 3.25. (a-c) Transmission holder PW6599/00 for the Philips Cryosystem (Appendix 1)
The specimen can be loaded into this holder using the injector system ill ustra ted in Fig. 3.11
This side-entry holder for the EM 400 series of Philips microscopes allo ws for viewin g tempera-
tures down to - l 95°C and is cooled by the circulation of cold, dry nitrogen gas. Tempera1urt
Heeter,
monito ring is incorporated and there is a double mechanical shroud to avoi d contamination temperat ure measurement
during transfer. (Reproduced by courtesy of Philips Electron Optics.) C •nd con t rol connections
Direct viewing and analysis of cold specimens 193
192 Ro bar ds an d SIey t r L ow temperature methods in biological electron microscopy

as that available for the Philips 400 (Fig. 3.22; Hax and Lichtenegger 1982;
see also Tatlock et al. 1984) and for Zeiss TEMs (Hezel I 984), but many
workers have found it necessary to construct their own accessory to fulfil
this function. In addition to the necessity of maintaining low specimen tem-
perature during transfer, the restricted space available in a TEM means that
the mechanical design has to be neater and, furthermore , the higher instru-
mental resolution makes additional demands on stage stability. A detailed
description of a transfer device for TEM work has been provided by Hutchi-
PTFE spacers
son et al. (1978a,b), while another example of a transfer device for TEM
work is illustrated in Fig. 3.23.

.'' 3.4.8 Viewing and analysis


''

A number of manufacturers now provide liquid nitrogen-cooled stages for


transmission instruments (e.g. Figs. 3.24 and 3.25) and others have been de-
scribed in the literature (see Butler and Hale 1981). For transmission view-
ing in an SEM, the problems are relatively easily overcome, merely requir-
ing a suitable transmission holder attached to a slightly modified SEM cold
r-----1 stage (e.g. Fig. 3.26). As such stages are almost always required for use in
20 mm
conjunction with microanalysis, precautions must be taken to provide a low
a X-ray background environment, and a beam current meter (Faraday cage)
and space fo r standards (see Chandler 1977) are also required.
The higher resolution of a true TEM or STEM instrument requires a ma-
jor step forward in the construction of a suitable cold stage (see Butler and
Hale 1981). Not only must the temperature be maintained low enough but
the mechanical stability (in relation to varying temperature) must also be
of a high order. It is usually necessary for stages to be allowed to equilibrate
at a particular temperature for a period of time so that they are stable before
r-----1
20 mm observation begins. To produce such stages requires specialised knowledge
and precision engineering for successful construction, and such details fall
outside the scope of this book. However, such stages are now being pro-
duced commercially, together with appropriate transfer devices, so that a
·~··'.~::.~~'"" '" """" complete low temperature pathway from freezing the specimen to observa-
tion and analysis in the microscope can be pursued.
The Philips integrated low temperature transfer device and cold stage for
the EM 400 provides an excellent example of a commercially available sys-
· I t on microscopy to pro-
tem capable of routine use for observational studies and/or analytical ones
Fig 3 26 Modification of a low temperature stage for scanning e elc rd old stage (b) Plan (Figs. 3.22 and 3.25). Recently, a number of other stages for work at high
· · · ( )G l arrangement of Dewar vesse an c ·
vide for transmission use. a enera (Rd f m Taylor and Burgess 1977.) resolution have also been described. Hayward and Glaeser (1980) published
and vertical section of cold stage. e rawn ro
Robards and Sleytr Low temperature methods in biological electron microscopy Direct viewing and analysis of cold specimens
194 195

Spring clips for seating Copper strip with


cartridge electrical insulation
on the lower side
3mm copper plate
(bolts to dewar rod)

Solder

40 leaves of
0.025mm copper Silver strip Heating Connector Wire to I
with insulation filament tube power
on both sides centre wire supply I

I
support buttons
I
0 Retainer spring
~ Upper spacer
I I
~ Trap grid Specimen
assembly ~ Specimen I

~ Specimen assembly
@ Lower spacer

Nema G-10 insulator


~ Trap grid

Stainless steel body

Jig
1-----1
20mm

Fig. 3.27. A high resolution cold stage for the JEOL lOOB and IOOC transmission elect!'
microscopes. (Redrawn from Hayward and Glaeser 1980.) ,1i,
3.28. (a) The modified tip section of the JEOL r .
as described by Perlov et al (1983) d d . coo ~ng holder with a heating bridge in- 111
· an use m conjunct" ·th
(Red rawn from Perlov et al. 19 83 ·) (b) Th e specimen
. wn WI a transfer device (Fig. , I
assembl d · ··
a design for a high resolution cold stage compatible with a JEOL mi of holder illustrated in (a) (R d f Y an its Jig for the type
· e rawn rom Talman et al. 1979.)
11

scope (Fig. 3.27). This liquid nitrogen-cooled stage only achieves a temp
I
ture of 183 K but provides a resolution of at least 0.6 nm as demonstra ,1
on purple membranes of bacteria. Perlov et al. (1983) have also prod
a transfer module and a simple cooling holder for use with a JEOL Cu-Ni gas pipes
lOOC TEM. This allows frost-free transfer during which the temperatu - Thermocouples
.............
the sample does not exceed 120 K. The specimen may be held at a tern .
ture between 100 and 450 K during observation while retaining a resol
of 1.5 nm (Fig. 3.28). Nicholson et al. (1982) have designed and built a
that is cooled to 83 K using nitrogen gas, the minimum specimen tern r--R-e-a+r-S_S
_ __J 11
ture achievable being 108 K. This stage has a drift rate of 0.2-0.3 Rear isolator Out
'O' ring Gas flow
and, being designed for microanalytical work, contributes less than
Brehmsstrahlung (X-ray 'white' radiation or 'continuum' radiation) . Low temperature specimen holder desi . .
microscope by Nicholson et al. (I 9~~e)d(~r m1croanalyti~al work in conjunc-
.11 JEOL
tages between 9.5 and 14.5 keV (Fig. 3.29). . edrawn from Nicholson et al. 1982.)
Robards and Sleytr Low temperature methods in biological electron microscopy Direct viewing and analysis of cold specimens 197
196

Now that both SEM and TEM cold stages and accessory devices are Dubochet,
- . J. and A.W. McDowall (! 981 ) ' v·1tn.fi1catlon
. of pure wat D 1 .
J. Microscopy 124, RP3. er ore ectron microscopy,
becoming more widely available it is inevitable that applications will be
Dubochet, J., J. Lepault ' R . Freeman ' 1.A . Bernman
. and J -C Homo (1982) El .
broadened as more users have the opportunity to use low temperature tech- copy of frozen water and aqueous s0 1 f J . · · , ectron micros-
. . u 10ns, ·Microscopy 128, 219.
niques to investigate their particular material. However, in concluding this Echlm, P. (1975), Sputter coatmg techniques for scannin 1 t .
tron Microscopy 1975, 1, 217. g e ec ron microscopy, Scanning Elec-
Chapter it is, perhaps, as well to stress that the final choice of system and
operating specifications ultimately rests with the individual user who Echlin, :· (1978a), Low temperature biological scannin ele t . .
techmques in biological electron microsco K g c ron microscopy, m: Advanced
should, therefore, have a clear understanding of what is to be achieved and 89. py, 1. . Koehler, ed. (Sprmger-Verlag, Berlin), p.
what operational conditions are necessary to facilitate the particular study. Echlin,
· p. · ), Coating techniques for scanning elect ron microscopy
s (1978b . and x ra ·
Thus, to provide an example, although a commercial cold stage may be 1ys1s, cannmg Electron Microscopy 1978 1 109 - Y m1croana-
available that will cool to 183 K, this will be totally unsuitable for frozen- Echlin, P. (1978c), Low temperature scannin~ ele;tron microscopy:
112, 47. . .
a review, J. Microscopy
hydrated specimens (they would rapidly freeze-dry). On the other hand,
copy P.
Echlin, (1981), Recent advances insp ec·imen coatmg
· techniques, Scanning Electron Micros-
such a stage may be very suitable for reducing radiation damage in dry spe- 1981, 1, 79.
cimens viewed at low temperatures. The ultimate decision lies with the user! Echlin, P. and R.B. Moreton (1976) ' Lo w t emperature techmques. for sc · 1
copy, Scanning Electron Microscopy 1976 anmng e ectron micros-
1 753
. A.N. Broers and W. Gee (1980) ' r'mprove
Echlin, P., , .d resolut10n
· of sputt d
Scannmg Electron Microscopy 1980, l, 163 . er-coate metal films,
References Echlin, P., B. Chapman ' L. Stoter ' W · Gee an d A . Burgess (1982) L I
Scanning Electron Microscopy , l, . ' ow vo tage sputter coating,
1982 29
Adrian, M., J. Dubochet, J. Lepault and A.W. McDowall (1984), Cryo-electron microscopy
Fernandez-Moran, H. (1966), High resolution electro . .
of viruses, Nature, Lond. 308, 32. es at liquid helium temperatures Proc ti A d n '.1ucroscopy with superconducting lens-
Aldrian, A.F. and G.A. Windisch (1984), A versatile cryo-transfer system for SEM, Proc. 8th . . . ' · na · ca · Sc1. U.S.A. 56, 801
otino, M. and T.H. G1ddmgs, Jr. (1984), Ultrastructural ima . .
Eur. Reg. Conf. Electron Microscopy, Budapest 1, 99. ing and cryo-HVEM, Proc. Sth Eur Re C gmg of whole cells by fast freez-
Beckett, A., R. Porter and N.D. Read (1982), Low temperature scanning electron microscopy ks F CS Cl d . g. onf. Electron Microscopy, Budapest 3 1823
, ., . . ay an G.W. Peace (1980) Ion beam thin fil d .. , .
of fungal material, J. Microscopy 125, 193. tron Microscopy 1980, 1, . ' m epos1t10n, Scanning Elec-
155
Brombach, J.D. (1975), Electron beam X-ray microanalysis of frozen biological bulk sped•
chs, W. and B. Lindemann (1975) Electron beam X .
mens below 130 K. II. The electrical charging of the sample in quantitative analysis, J. ulk specimens below 130 K 't . -ray microanalysis of frozen biological
· 1· 1ns rumentat10n and spec· .
roscopie Biol. Cell 22, 233. iol. Cell. 22, 227. imen preparat10n, J. Microscopie
Butler, E.P. and K.F. Hale (1981), Dynamic experiments in the electron microscope, in: Pr
hs, W., J.D. Brombach and W. Trosch (1978 ) Ch . .
cal methods in electron microscopy, Vol. 9, A.M. Glauert, ed. (Elsevier/North-Ho! .Microscopy 112, 63. a' argmg effect m electron irradiated ice,
Amsterdam). hs, W., B. Lindemann, J.D. Brombach and W. Trosch (I 978 .
Chandler, J.A. (1977), X-ray analysis in the electron microscope, in: Practical methods in el en preparation for electron bea X . . b), Instrumentation and speci-
tron microscopy, Vol. 5, A.M. Glauert, ed. (Elsevier/North-Holland, Amsterdam). . m -ray m1croanalys1s of froze h d db .
. Microscopy 112, 75. n Y rate ulk specimens,
Chang, J.J., A.W. McDowall, J. Lepault, R. Freeman, C.A. Walter and J. Dubochet (1
Freezing, sectioning and observation artefacts of frozen hydrated sections for electron opy of biological specimens at l
r, R.M and K.A. Taylor (1978) Radiation d
t
·
amage relative to transmission electron mic-
roscopy, J. Microscopy 132, 109. .E. (1966) Int d . ow emperature: a review, J. Microscopy 112, 127.
Cosslett, V.E. (1978), Radiation damage in the high resolution microscopy of biological 4. ' ro uct10n to electron microscopy, 2nd Edn. (McGraw-Hill, New York),

rials: a review, J. Microscopy 113, 113. 'B.P., R.G. Kirk and AR Spurr (1 978) Q . .
Dietrich, I. (1978), Superconducting lenses, Proc. 9th Int. Congr. Electron Microscopy, logical thin sections· the u.se. f STEM D ' uantltatlve electron probe microanalysis
oscopy 3, 175. . o or measurement of local mass thickness, Ultra-
onto 3, 173.
Dorge, A., R. Rick, K. Gehring, J. Mason and K. Thurau (1975), Preparation and applica
T.L and G. Koch (1975), Some problem . d .
of freeze-dried sections in the microprobe analysis of biological soft tissue, J. Micros in the scanning electron . s associate with low temperature manipula-
d SB dR microscope, Scanning Electron Microscopy 1975 1 35
Biol. Cell, 22, 205. ' · ·an M Glaeser (1980) ff h ' ' ·
Dubochet, J. (1984), Electron microscopy of vitrified specimens, Proc. 8th Eur. Reg. on microo0op~s ~It . ' ig resolution cold stage for the Jeol !OOB and !OOC
u~ ' ram1croscopy 5, 3.
Electron Microscopy, Budapest 2, 1379.
Robards and Sleytr Low temperature methods in biological electron microscopy Direct viewing and analysis of cold specimens 199
198

Hax, W.M.A. and S. Lichtenegger (1982), Transfer, observation and analysis of frozen hyd- mens, Microscopica Acta 79, 2 54 .
Marshall,
. A.T. (1980) ' Quantitative X-ray m1croanalys1s
. . of frozen hyd t d b lk b'
rated specimens, J. Microscopy 126, 275. specimens, Scanning Electron Microsco 1980 2 ra e u 10logical
Heide, H.G. and K. Urban (1972), A novel specimen stage permitting high resolution electron 335
Misell,
. D.L. (1978), Image enhanceme t pyd . ' ' .·
n an mterpretat10n in· p t' I h .
microscopy at low temperatures, J. Phys. E: scient. Instrum. 5, 803. microscopy, Vol. 7 AM GI . ' · rac ica met ods m electron
Hezel, U.B. (1984), A high-performance cryotransfer system and cryostage for transmission darn). , . . auert, ed. (Elsev1er/North-Holland Biomedical Press, Amster-
electron microscopy, Zeiss Magazine for Electron Microscopists 3, 23.
Nei'. T. (1_974), Cryotechniques, in: Principles and techni ue . .
Hobbs, P.V. (1974), Ice physics (Clarendon Press, Oxford). B10log1cal applications Vol MA H q s of scanmng electron microscopy.
Hori, E.M. (1974), A versatile liquid helium cooling stage for the transmission electron micros- 113. , 1
. ' . . ayat, ed. (Van Nostrand Reinhold, New York), p.
cope, Proc. 8th Int. Congr. Electron Microscopy, Canberra 1, I.
Hosoi, J., T. Oikawa, M. Inoue, Y. Kokubo and K. Hama (1981), Measurement of partial Nei, T., H. Y otsumoto ' y · Hasegawa a n d y · N agasawa (1973) Dire t b ·
specimens with a scanning electro . ' c o servat10ns of frozen "I

specific thickness (net thickness) of critical-point-dried cultured fibroblasts by energy analy- N . T H y t n m1croscope, J. Electron Microscopy 22 185
e1, ., . o sumoto, Y. Hasegawa and M H , . '!I
sis, Ultramicroscopy 7, 147. electron microscopy J Electron M' · asegawa (1 974), Freeze-etching in scanning 1

Hutchinson, T.E., W.A.P. Nicholson, B.W. Robertson, R.P. Ferrier and W.H. Biddlecombe . ' · 1croscopy 23, 137. 1i

Nicholson, W.A.P., W.H. Biddlecombe and H.Y. Elder 1982 ''


(1977), A new low temperature specimen stage and transfer system for the JEOL JEM IOOC,
temperature specimen stage for biol . . . (_ ), Low X-ray background low ''I
Developments in electron microscopy, D.L. Misell, ed., Ins. Phys. Conf. Ser. 36, 49. ogica1m1croanalys1s m the t · · l 111
cope, J. Microscopy 126, 307. ransm1ss10n e ectron micros-
Hutchinson, T.E., D.E. Johnson and A.P. MacKenzie (1978), Instrumentation for direct obser-
vation of frozen hydrated specimens in the electron microscope, Ultramicroscopy 3, 315. Perlov, G., Y. Talmon and AH · · Falls (! 983 ) • A n Improved
. transfer mod l d ·
perature control for a simple comm . 1 I' u e an vanable tem-
Jakopic, E., A. Brunegger, R. Essl and G. Windisch (1978), A sputter source for electron mic- Peters K.-R. ('1980) p . .ercia coo mg holder, Ultramicroscopy 11, 183.
roscopic preparation, Proc. 9th Int. Congr. Electron Microscopy, Toronto 1, 150. ' , ennmg sputtermg of ultra thin metal fil1 ,. .
microscopy, Scanning Electron Mi ms ior high resolution electron
Kellenberger, E. (1982), Radiation damage in biological materials in perspective with other croscopy 198 o, 1, 113.
limitations, Proc. 10th Int. Congr. Electron Microscopy, Hamburg 1, 33. eters, K.-R. (1984), Precise and reproducible d epos1t10n .. of thin and ult th· b
Kemmenoe, B.H. and G.R. Bullock (1983), Structure analysis of sputter-coated and ion-beam bY fl ash evaporation of carbon yarn in h.igh vacuum, J. Microscopy . ra m car on films
133 17
sputter-coated films: a comparative study, J. Microscopy.132, 153. ey, J.B. and J.T. Norton (1978), A chamber attached to the sc . ' . .
for fracturing and coating frozen biol . . l . annmg electron microscope
Knapeck, E. and J. Dubochet (1980), Beam damage to organic material is considerably .b ogwa1 samp es J Microscopy 112 169
o ards, A.W. and P. Crosby (1979) ' A compreh ens1ve . ' f.reezmg
. fracturing, d· .
reduced in cryo-electron microscopy, J. mo!. Biol. 141, 147.
Knapeck, E., H. Formanek, G. Lefranc and I. Dietrich (1984), New aspects for the interpreta~· Iior Iow temperature scanning electro n microscopy
.
Scann
' an coatmg system
325. ' ·
1 ng ect ron Microscopy
El . 1979, 2,
tion of cryoprotection illustrated on L-valine, Proc. 8th Eur. Reg. Conf. Electron Micros-'.
bards, A.W., A.J. Wilson and P. Crosby (1981) S eci . .
copy, Budapest 2, 1395. f· Microscopy 124, 143. ' p men heatmg dunng sputter coating,
Koch, G.R. (1975), Preparation and examination of specimens at low temperatures, in: Prin
pies and techniques of scanning electron microscopy. Biological Applications, Vol. 4, M: nnann, A.J. and P. Echlin (1975) The re ar . . .
rated tissue sections by scanning t' . p _P atwn, examma!Jon and analysis of frozen
Hayat, ed. (Van Nostrand Reinhold, New York), p. l. ransm1ss1on electron micr dX .
, J. Microscopy 105, 155. oscopy an -ray m1croana-
Lamvik, M.K., D.A. Kopf and J.D. Robertson (1983), Radiation damage in L-valine at liq ·
helium temperature, Nature, Land. 301, 332. H.Th.J., A.L.H. Stols and A.M. Stadhouders (19 .
zen hydrated cryo-sections Proc 8th E R 84), A carbon fibre evaporator to coat
Leapman, R.D., C.E. Fiori and C.R. Swyt (1984), Mass thickness determination by elect 1687. ' . ur. eg. Conf. Electron Microscopy, Budapest
energy Joss for.quantitative X-ray microanalysis in biology, J. Microscopy 133, 239.
Lefranc, G., E. Knapek and I. Dietrich (1982), Construction principles for stable super on, Y. ( 1980), Rate of sublimation of ice b r d. . . .
. Meeting EMSA S F . y a ia!Jve heatmg m freeze-etching, Proc. 38th
ducting lens systems as required for obtaining high resolution images, Proc. I 0th Int. Co , an ranc1sco, p. 618.
Y. (l 982a), Thermal and radiation da
Electron Microscopy, Hamburg 1, 391. yn,125, 227.
f
mage to rozen hydrated specimens, J. Micros-
Lepault, J. (1984), Frozen suspensions, Proc. 8th Eur. Reg. Conf. Electron Microscopy,
n, Y. (l 982b), Frozen hydrated s e ·
pest 2, 1413. burg I, 25. p c1mens, Proc. 10th Int. Congr. Electron Microscopy,
Lepault, J., F.P. Booy and J. Dubochet (1983), Frozen biological suspensions, J. Micro
Y. and EL Thomas (1977 ) T
129, 89. hydrated. s;ecimens in colda ,t emperature rise and sublimation of water from thin
Marshall, A.T. (1975), Electron probe X-ray microanalysis in: Principles and techniq s age microscopy ' Sea nnmg
· El ectron Microscopy
. 1977,
scanning electron microscopy, Vol. 2, M. A. Hayat, ed. (Van Nostrand Reinhold,
'Y. and E.L. Thomas (1977b) B ·
York), p. 103. SEM/STFM J M' ' earn heatmg of a moderately thick cold stage specimen
Marshall, A.T. (1977), Electron probe X-ray microanalysis of frozen hydrated biological L ' icroscopy 111, 151.
·
Robards and Sleytr Low temperature methods in biological electron microscopy
200

T I y and E.L. Thomas (1978), Electron b earn heating temperature profiles in modera- Chapter 4
a mon, . M/SEM s ecimens J Microscopy 113, 69.
tely thick cold stage STE P ' t. ·crothermometry - a technique for the
d E L Thomas (1979), Open sys em m1 S .
Talmon, Y . an · · . . th electron microscope, J. mater. c1. 14,
measurement of local specimen temperature m e

1697. . . d EL Th mas (l 979), Cold-stage microscopy system


Talmon, Y., H.T. Davis, L.E. Scnven an . . o
r ·d R cient Instrum. 50, 698.
for fast frozen 1p1 s, ev. s · d p J R hton (1984) A vacuum cryo-transfer
G J T J Hurd R.E.W. Banfield an
Tatlock, · ., · ·
,.
'
PhT TEM Proc 8th Eur eg. on ·
R
· · us '
C f Electron Microscopy, Budapest 1, 97. Freeze-sectioning
holder ior a 11ps ' · . ." ff hydrated crystalline biological speci-
T ayIor, KA . . ( 1978) ' Structure detennmat1ons o rozen, ,
115
mens, J. Microscopy 112, · f tron probe microanalyser, J. Micros-
Taylor, P.G. and A. Burgess (1977), Cold stage ore1ec

copy 111, 51. M GI (1976) Electron microscopy of frozen hydrated biological spe-
Taylor, K.A. and R. . aeser ,
cimens, J. Ultrastruct. Res. 55, 448. b f with carbon fibre filament, Proc. R. mic-
Vesely, D. and G. Woodisse (1982), Car on coa mg Principles and problems
rose. Soc. 17, 137. . h d wing and freeze-etching techniques, in:
M d A J Rowe (1980) Rep1ica, s a o
Willison, J. H · ·an · · . ' V 1 8 AM Glauert ed. (Elsevier/North-Hol- In principle, the idea of cutting thin frozen sections from biological tissue
Practical methods in electron m1croscopy, o. , . . , .
simple. In practice it turns out to be rather more difficult. Since the ear-
land, Amsterdam). ll ll at uncoated cryofractured surfaces as ,
Woods, P.S. and M.C. Ledbetter (1976), Ce organe es . 'est attempts to cut thin frozen sections (Fernandez-Moran 1952; Bernhard
. . I t microscope, J. Cell Sci. 21, 47. ..
viewed with the scannmg e ec ron . d X-ray microanalysis of cryofractured. 965; Bernhard and Leduc 1967; Appleton 1969, 1972a,b,c; Christensen
. Id K (1979) Scanning electron microscopy an
Z1ero , · ' · k bb Oberfl 12
cells and tissues, Beitr. Elektronenmikroskop. D1re ta . . .
9, 1971; Dollhopf et al. 1972), the rate of publication of papers on this
~ect has increased dramatically. This interest reflects the realisation that
en thin sections potentially offer great rewards in terms of ultrastruc-
1preservation and maintenance of the in vivo distribution and concen-
ion of solutes. However, thin frozen sections are difficult to obtain, diffi-
to maintain fully hydrated and uncontaminated, and difficult to view
out problems arising from radiation damage and other sources. Were
t for the very real and unique opportunities available, one wonders
her such a technique would be pursued at all. Various terms have been
in the literature to describe this technique: we shall call it 'cryoultrami-
y', following the practice of Reid (1974) in an earlier volume of this
. This is a very rapidly changing field and, therefore, only the most re-
evelopments will be described: except where their importance and/or
have been retained, equipment and methods which are now (even af-
ly a relatively few years) of historical interest only will not be dealt

e theory of cryosectioning is, as yet, by no means totally understood.


plest, should be possible to differentiate between a cutting and a
ing process, in the knowledge that hydrated biological tissues are
to contain substances that will show cleavage/cutting phenomena that

201
Freeze-sectioning 203
202 Robards and Sleytr Low temperature methods in biological electron microscopy

are strongly affected by temperature. Saubermann (1980) has discussed such Shear zone

matters in detail. He cites the melt-freeze theory of Thornburg and Mengers


(1957), who suggested that the local pressure exerted by the advancing knife .i Stress
caused melting of the ice remaining in the tissue as a result of melting poiriti / lines
-- Fracture plane
depression. However, this concept was based on the, possibly erroneous, as· Compression
sumption that the physical properties of frozen biological tissue blocks are:
b
similar to those of frozen pure water. It was estimated by Thornburg a
Mengers that the energy produced during sectioning would be about 2 x roi·
ergs (2 x 10- 1 J) although subsequent direct measurements by SaubermaJ
et al. (1977) indicated that the energy of cutting 0.5 µm thick cryosectio
4
ranged between about 1300 ergs (1.3 x l0- J) at 243 K (-30°C) to 5
ergs (5 x 10- 4 J) at 193 K ( -80"C). It was calculated that this could prov·
a melting zone of between 0.4 and 1.6 nm. A number of authors have tur
their attention to this problem because it could clearly be quite cruciah
melting takes place during sectioning (or at any other time), then sol c d
redistribution ma')' be possible and the whole authenticity of the techni .1. (a) Schematic . diagram of metal m·ac·h·in1ng
. model of . .
_an comp_ rcss1on in the block along th d' . cryosect1on1ng. 'The knife causes
for the localisation of soluble components is thrown open to doubt. . d · e ircchon shown b th h
o e -n1fe is defined as the angl b . Y e cavy arrow. The rake
concensus appears to be that melting either does not occur or, if it.<\ f th k
surface of the knife. The cutt!n
· e etween a hne
I .
d'
perpen icular to the cut surface and
it is to such a very limited degree as to be insignificant, and is therefor gange1sthesum 0 fth k ·r.
·The general distribution of stre f k . e ni c angle (ct) and the clearance
) ·- ' ss rom a ntfe ed e · h
a limitation to the use of the method (Hodson and Marshall 1972; Ap press1 ve forces are relieved b f
Y racture or rupt
g is s own by the stress lines (b)
1 ·
1974; Saubermann et al. 1977; Frederik and Busing 1981; Frederik of least resistance. This may b b t ure a ong the fracture plane which is
e e ween crystal pl b
cture planes as frequently obse d . f anes utmost probably follows the
1982a; Karp et al. 1982). · f rve in reeze-fractu d ·
ve orces occurs through plastic d f . . re material. Son1c relief of the
As an alternative theory, Saubermann (1980) has drawn the analog ' ·
on depends greatly on the ductility f th .
e ormat1on 1n the sh
e~r zone. The amount of plastic
.
metal machining, suggesting that the main difference is simply that, w ds t . .
o occur 1n the d1rection show b th
0 e tissue. Plastic flo f h
w, rom t c compressive
· b · n Y e arrows (c) Ad' ·
in cryosectioning it is the 'chip' that one wants to retain, in metal ma e JS nltle, non-compressible dd ·· 1scont1nuous chip is formed
' an oes not under l ·
it is the chip that is removed. Using this comparison, the knife is co d, stress from the knife is relieved b f , go p ashc deformation in the shear
I d Y racture along th t
to be forced into the frozen block, causing compression and stre s an 2). Slippage along fractures in th d' . e wo planes shown by the
hi ( e irechon sho b h
P se_qu_ence to overlap. Such sect· wn Y t e heavy arrow
knife edge. The force is eventually released by rupture along plane.s . · ions would not b f 1f
us c tp is formed when plastic d f . e use u or X-ray analysis. (d)
resistance with additional relief through plastic deformation which .h e ormatton can 0 ,, ·
ue is sufficiently ductile to P"O t I . ccur in the shear zone. In this
'moepashcflowo th k · .
greatly on the ductility of the specimen. If the block is brittle, it is s parted by the knife cause the ch' t ver e n1fe surface. The bending
·-;:. Ip o curl. (Redrawn fro1n Saubennann 1980.)
that the compressive force causes rupture along the shear angle pri
movement of the chip over the knife, so leading to a discontinuou .!her locally or in general th .
opposed to the continuous chip (section) that is produced from le. ntaining very large ice er' stal: ice crystals ~re small. Conversely,
more ductile material (Fig. 4.1 ). As the state of ductility /brittleness yosection. The explanati y c are usually dtfficult, if not imposs-
·1 on ,or many of th n· .
specimens depends both on the temperature and on the nature of · fat ure ' even using the same tissue
.
· th
e con 1ct1ng -reports of
men itself including, particularly, the size of the ice crystals (see S part, from differences . . , ts at they undoubtedly stem
· In ice crystal · · · '
Robards 1977; see also §6.3.1), the importance of selecting the a Fig. 2.34 in §2.3.2). size withm the blocks (see
preparation/sectioning conditions is self-evident. One of the l)l clear that the temperature at which sections are cut will greatly
contributions to ease of cryosectioning is to produce a speci
204 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-sectioning 205

affect the nature of the cutting/fracturing process. Saubermann (1979) tle specimen. The chip can be further constrained from bending as it moves
showed that continuous chips from mouse liver could be obtained with across the knife by use of an anti-roll plate (§4.3.2a), or by the ingenious
lower 'work' at 243 K ( - 30°C) compared with the greater work needed to vacuum method of Appleton (1974) described later (§4.3.2b). Saubermann
produce discontinuous chips at 193 K ( -80°C). Such findings point to the (1980) states that an advantage of metal knives for low temperature section-
possibility of improving sectioning by cutting at warmer temperatures (e.g. ing is that they have smaller knife angles than glass or diamond knives and
243 K; -30'C), although it can also be argued that ice recrystallisation is can thus provide a lower clearance angle and high rake angle. There is the
more likely to occur and any melting zone would be extended. Howevet;- ~/-i further advantage that metal knives have better thermal stability than glass
the work needed at the higher (warmer) temperature is less and the likeli-.; ones.
hood of inducing thermal damage is therefore also less ! Partly for the rea- The rate of heat input to the specimen is directly proportional to the cut-
sons given above, Saubermann et al. (198la,b) advocated sectioning_ at tern-;;·'.; ~ting speed used. If heat input is too rapid the specimen face will warm up.
peratures closer to 240 K (-33°C) than 200 K (-73°C); a suggest10n that_. _:·It is therefore necessary to create an acceptable balance between cutting
has met with strong opposition from some other workers. For example, 'tr \~':speed, block face temperature and section thickness. Given that the section
has been suggested that Saubermann did not see ice recrystallisation be'" ';thickness is usually determined by the requirements of the operator, it is sen-
cause the crystals in their specimens were already so large that addition~};; sible, if c~nsiste~t results are to be obtained, to maintain a constant cutting
growth would be extremely slow; only very small crystals_ (such as might ~.e: eed while varymg the block face temperature, which is determined by the
produced at the surface of a large block) would grow rapidly at the warme_! lance between the thermal properties of the specimen together with the
temperatures (Seveus 1980). More work remains to. be done before mperature of the specimen holder and the surrounding atmosphere, to
theoretical basis of variations in the physical charactensllcs of the spec1 -. e optimum sectioning properties. However, the hardness of the specimen
at different temperatures in relation to both sectioning and ice transfo varies with te~perature and cutting speed may have to be varied to help
tion phenomena are properly understood. Nevertheless, enough is kn pensate for tlus temperature-dependent variation in hardness.
and agreed to allow us to go some way towards producing and sectio
frozen specimens under optimised conditions. ,
The critical temperature during the cryosectioning process is clearly_
at the block face itself. This, in turn, is largely dependent on the temper
of the surrounding atmosphere. Saubermann (1980) stresses the impor ill be clear that the basic requirement for frozen thin sectioning is an
of achieving a stable thermal equilibrium between the block face an. icrotome that will operate with the specimen area at a very low tem-
knife and that consequently there is little point in separately coohn ure. 'Cryostats' (a poor term for conventional microtomes in cold
speci:Uen and the knife. However, most cryomicrotomists find it prefi s) have long been available for preparing sections with thicknesses
to have full control over both knife and specimen cooling. Such con n to about 10 µm and are in widespread routine use for light microsco-
ations have important implications for the design of cryoultramicro investigations. The more rigorous requirements of thin frozen sections
Similarly, it bas been found that thin ( < 100 nm) sections need to meant that, over a decade or so, cryoultrarnicrotomes have evolved
at lower temperatures than !bickers ones (Seveus and Kmdel 1974; Sa slightly modified conventional ultramicrotomes to specially designed
mann 1980). Thus it is apparent that any cryoultramicrotome should a gmeered_ apparatus. As the specimen has to be protected at all stages,
have a means of obtaining a good range of specimen temperature youltram1crotome cannot be considered alone: it must be 'interfaced'
so that the operator can experiment with the best conditions for the P th the initial freezing process and with subsequent procedures such
lar specimen in hand. imen coating and viewing. This usually necessitates some sort of
Consideration should be given to the knife angle and rake angle ( r device (see §3.2.2). An important aspect that should also not be
4.1) when sectioning (Reid 1974). Saubennann (1980) recommend , ked_ is that the success or failure of many difficult techniques, and
knife angle and a high rake angle to minimise the tendency to bend t tramicrotomy is certainly no exception, can be determined by the
206 Robards and Sleytr Low ten1perature methods in hiological electron 1nicroscopy Freeze-sectioning 207

accessory tools that are available. It is necessary to have a range of forceps, is not generally possible to clamp frozen specimens directly because ice is
clamps, grid supports and so on that have been selected and/or modified too brit~le. Therefore, material is usually either frozen on a very small sup-
to work at the specific temperatures chosen. Frozen thin sections are often port which can then be mounted onto the microtome arm, or the specimen
allowed to freeze-dry prior to subsequent handling and some provision must froz.en alone and subsequently stuck to a chuck, or other support, using
also be made for this process to be carried out under controlled conditions, special low temperature 'glue' (§4.3.lb; §4.3.2c; Table 4.1).
Cryoultramicrotomes are of two basically different types: either the whole
4.2.l Basic requirements of cryoultramicrotomes '..µltramicrotome is placed in a cold box (the conventional 'cryostat' ap-
_;proach), or JUSt the specimen/knife area of the ultramicrotome is enclosed
The cryoultramicrotome should have good temperature control of the speci'.;c• jso that this region can be kept at the appropriate selected temperature. Both
men and knife (either independently or by cooling the whole surroundin&>' ~orms are commercially available and both the instruments and the methods
atmosphere, see §4.2.2) to temperatures considerably lower than about 17-0f2 ~fop.crating them have often been modified over the past few years. The
K (-100°C); the specimen area should be protected from water vapour, an escnpt10ns provided here predominatly refer only to the latest versions.
hence condensation; and the usual facilities of an ultramicrotome (adj hichever form of cryoultramicrotome is used, it should be possible to
table cutting speed, knife angle, section thickness, etc.) should be availa , n~fer the fr_ozen sections to a suitable area for freeze-drying, jf this is
and operable down to the lowest achievable temperature of the equipm mred, and it should also be possible to take up the frozen sections and
The manner in which the specimen is mounted is important, because it transfer them into other ancillary equipment, without allowing them
often required that the smallest possible specimen is frozen (to provide hi er to freeze-dry or to become contaminated.
est cooling rates) whereas thermal and mechanical stability during secti
ing demand that the specimen must be mounted firmly onto some form Cryoultramicrotomes
substantial support having good thermal conductivity and heat capaci
two types of cryoultramicrotome (cryostat or specific specimen area
TABLE 4.1 ling) have their own particular merits and demerits. The advantage of
Some low temperature 'glues' and other substances used during cryoultramicrotomy
-·-·------···----·--·--·-----·-
cryostat approach is that it provides a large working volume at a stable
Boiling point Melting pointa Vapour pressureb · temperature around the specimen. It is easy to keep the many tools
. and chemicals in the large cold box. However, such equipment is ofte~
Liquid
°C (K) (K) cc
Torr (Pa) (K)
uncomfortable to use. The operator must usually stand bent-over
Toluene 110.6 (384) -95(178) 755 (1.0 x IO')
his/her hands thrust down into the cold chamber, to mou~t the speci'.
C6HsCH3 and t.o make adjustments. Furthermore, change in specimen tempera-
183.3 (456) -- 88 (I 85) (parl!cularly warming) can take an extremely long time because the
n-Butyl benzene
CoH 5(CH2)3Cl-I3
mass of the microtome must be brought to the required temperature.
/_arly cryosectioning attachments for specimen area cooling on conven-
Tsopentane 27.9 (301) -159.9(113) 1 ultrarnicrotomes were also often less than satisfactory in many re-
(CH 3) 2CI!CII2Cl-!3 ' but more recent versions appear to have overcome many of the pre-
98.4 (372) -90.6 (183) problems, such as the maintenance of temperature stability. More-
n-Heptanc
CH 3(CI12)sCH3
the ultramicrotomes can be used in the conventional sitting position:
:portant consideration during the painstaking processes involved.
192 (465) 18.0 (292) at systems tend to use more liquid nitrogen than the attachments
DMSO
CI-! 30SCH 3 have smaller volumes to cool. In principle, in either cryostat or
- - · -·
ent, cooling can be achieved by using cold nitrogen gas ducted con-
a Temperatures given for 760 mm pressure.
n Vapour pressures given together with relevant temperature.
208 Robards and Sleytr L ow temperature methods in biological electron microscopy
Freeze-seclioning
209
tinuously or intermittently into the chamber, or by evaporators in the case
of cryostats. Specimen mount and knife holder can be warmed indepen.
dently, using heaters, so that their bulk temperature can be varied from that
of the su rrounding atmosphere, and the temperature of incoming nitrogen
gas can be adjusted by passing the cold gas over some form of heater. Vari-
ous combinations of these cooling methods have been used in different de-
signs of cryoultramicrotome (see below). The critical considerations are that
the specimen block face and the knife should achieve the desired tempera-
tures and that these temperatures should be very stable. Early studies often
stressed the importance of differences between apparent specimen and knife
temperatures. More recently the trend has been towards using the same tem-
perature fo r both. This simplifies control systems because it means that only
the temperature of the surro unding atmosphere needs to be accurately and
stably controlled. When cooling is effected by enclosing just the specimen/
knife region within a small chamber, there is the additional advantage that
all the mechanical parts of the microtome remain at ambient temperature
This allows a low consumption of coolant and better operatio n of the fine
controls of the microtome. However, the chamber must be continuously
cooled with a flow of dry, cold nitrogen gas so that no humid air can reach
the cooled parts. Also, there must be no opening below the top of the
chamber through which the cold nitrogen gas can 'overflow', since th11
would leave the upper area unprotected by dry gas and, therefore, at ris~
from ' frosting-up' by condensing water vapour. Two alternative s olution ~
to these problems have been found (§4.2.2b).
It has already been noted that a versatile arrangement must be available
for mounting the specimen. Thermal and mechanical stability should be
combined with good thermal conducti vity once the specimen has been
mounted. Saubermann (1980) has emphasised the advantages of metal km\ ·
es when used a t low temperatures, partly due to their better thermal condul·
tivity a nd pa rtl y to thei r small knife angle. He notes that razor blades, which
can have a knife angle of <25°, can be used to cut sections down to 0.'
µm .

4.2.2a Cryostat ultramicrotomes

Typical of this approach is the Slee Type TUL cryoultramicro tomc (F1
4.2; see Appendix 2 for supplier). Details of using such a device have be(ll
given by Appleto n ( 19 74, 1978) an d G upta (1979). The top-opening co
box contains a Porter-Blum ultramicrotome; this box can be cooled do,.
210 Robards and Sleylr Low tempera/are me/hods in biological electron microscopy Freeze-sectioning
211
to about 220 K ( - 53°C) using an electrically driven compressor system and which liquid nitrogen is fed into an insulated cha
is further cooled by liquid nitrogen which is injected into the work chamber and the specimen The speci·men . mber a round both the knife
· is supported on a 'b ·d , h
under electronic control. This provides a vibration-free and reliable system the wall of the cryochamber this . f n ge t at hangs over
. ' is one o the two most c
capable of operating down to 170 K ( - l03°C). It is necessary for the ultra. keeping the specimen and knife cold h.I h ommon ways of
. w i e t e rest of the · t ·
microtome to be specially prepared to operate satisfactorily at low tempera- ambien t temperature (Fig 4 3) Th 1 t . micro ome 1s at
. · · · e a est version (the FS IOOO) · d .
tures (something not required for cryoattachments to ultrarnicrotomes). for operat10n with the Sorvall MT 5000 . - is es1gned
1
Section thicknesses between 25 and 500 nm can be set and cut at a range prises a freezing chamber a modi~ d u ~ram1crotome (Fig. 4.4). It com-
' ie specimen mount a lo t
of speeds (1- 12 mm s - 1) using this electronically controlled, motor-driven controller, which senses and mainta· th • w emperature
ms e preset temperat f b h
microtome. Specimen feed is obtained through a mechanical lever system specimen and the knife, and an insulated Ii uid . ur~ o ot the
It is possible to heat the knife, on its 'universal' stage, so that it is wanner The stage of the FS- 1000 · h . q mtrogen delivery system.
is a t ermally insulated al . .
than the ambient temperature within the cabinet. Unique to this system is taining a kn ife holder which accepts 1 d' ummmm chamber con-
the vacuum probe device described by Appleton (1974, 1978) which allows (Fig. 4.5). , The specimen holder acce gt ass, i~mond or ra.zor blade knives
ribbo ns of frozen sections to be gently drawn out from the knife edge (Fig. 4.5). P s a variety of specimen 'pins' (Fig.
4. 12 in §4.3.2b). Within the main work chamber is a separate, small com- The con troller initiates the delivery of th
. . e correct amount of liq ·d ·
partment for storage and/or freeze-drying of sections. This has its own tem- gen to mamtam the preset temperatur . th h u1 mtro-
. es m e c amber A pha
perature control arrangements, allowing temperatures as low as 120 K in the supply Ime ensures that only r .d . · . se separator
iqm gas is allowed mto the chamber.
( - 153°C) to be used for storage or temperatures between 250 and 150 K
( - 23 and - l 23°C) to be maintained for freeze-drying. A special set of han-
Christensen type
dling tools is provided with the equipment. Dollh opf type

' Bridge'
4.2.2b Ultramicrotome cryoattachments
Specimen arm

This approach to cryoultramicrotomy has been more popular than th,


cryostat system over the past few years. In part, this is because a number
of (ultramicrotome) manufacturers have produced commercially availabk
'kits' which enable a conventional ultramicrotome to be used at low temper
atures. Interestingly, and fortunately, the three major systems currentl)
available (from Dupont-Sorvall, LKB Produkter AB and Reichert-Jung
Appendix 2) embody a number of different design features which allow use:
ful comparisons to be made. It is very significant, and a reflection of tht
still evolving state of the art, that numerous papers have been pubhshl'd , 43 c
. .
.
ompanson of two methods of therma l! . .
which describe modifications to commercially available apparatus. Thu microtome from the main body of th . y msulatmg t he cold specimen in a cryoul-
969) e equipment The 'Ch · ,
while the basic equipmen t can now be bought, the newcomer to cryouftra uses a 'bridge' pro 'd' · nstensen type (Christensen 1967
lpcc: • v1 mg a poor heat cond uction th ,
microtomy must still be willing, not only to experiment to fi nd the best pre rmcn arm is thermally isolated f h . .pa way, so that the main part of the
n._mple) tbe Reichert-Jung FC4 cry~omtt e specimen itself. This type is represented in (for
aration methods for the particular biological material in hand, but also l ,.,. alt · . sys em and the Sorva ll FS l 000 · ·
ematrve Dollhopf type' (D llh f . - cryosectJomng system .
spend some time in adapting the equipment to facilitate the different pro- I). specimen arm that enters t ho op h e t al. 1972) has a continuo us, low thermal conduc-
llCllcd . e cryoc a mber through h I .
cesses that are to be carried out. with a thin, flexible plastic film whi h h a o e m the rear wall. This hole
The Dupont-Sorvall equipment (Hagler et al. 1980; Barnard 1982; Brdd LKB Cryokit 14800 employs this meth ~ ~~~allows movement of the microtome arm.
combe et al. 1982) is a derivative of the Christensen (1967, 197 1) system lion kindly provided by H Sitt 0 u' (. IS fi~ure redrawn from an original illustra-
. e, m versllat des Saarla ndcs, F.R.G.)
Low temperature methods in biological electron microscopy Freeze-sectioning 213
212 Robards and Sleytr

Spec imen chuck holder-----+-

Adapter locking s crew

Specimen pin
locking scr ew

~ ~SoooOmoo , ;, " " " " "

/ /,....._/_/
~
~ ~ ~---Specimen p ins

Diamond Glass Razor b lade

Specimen heatin g eleme n t


Knife holders
fig. 4.S. Specimen and knife h andling facilities for the Sorvall FS 1000 cryoseclioning system.
\ >ariely of specimen pins can be used lo mount specimens prior lo freezing. The low heat
Specimen cooling member ,Jpacity pins can be attached to the pin adapters, under liquid nitrogen, and then quickly trans-
Specimen crred into the cryochamber and inserted into the cooled specimen chuck holder. The FSIOOO
Knife heating element temperature
llows the use of diamond , glass or metal (razor blade) knives by using different holders. (Re-
~k-..,,~.:::,..~"""'~:--- sensor
Knife temperature rawn from an o riginal kindly provided by Biomedical Products Division, Dupont (U .K.)
sensor
Ltd.)
Knife cooling
member
Entry ofmo1st, surround ing air is completely eliminated through a constan t
now of nitrogen gas generated by a hea ting element in the bottom of the
hamber.
Accessories supplied with the Sorvall system include: a grid holder/freeze-
-Hr......f''------,r--- Section collector I l) tng chamber; different sized specimen pins and pin holders; a gnd
freeze -drying
c hambe r
Chamber
heating (a) Sorvall FS IOOO Cryoscclioning system. The attachment of a cold-box to the Sor-
element MTSOOO ultramicrotome, together with liq uid nitrogen supply and control unit, is illus-
ted here. (b) Diagrammatic illustration of the cryochamber of the FSIOOO as seen from
Grid c(~ce also Fig. 4.5). (Fig. 4.4a kind ly supplied by Biomedical Products D ivision, Dupont
hand l ing K.) Ltd. and reproduced with permission. F ig. 4.4b redrawn from an original diagram pro-
b device
vided by D upont U.K. Ltd .)
f
214 Robardi a11d Sleytr Low temperature methods in biological electron microscopy
Freeze-sectioning
215
handling device; knife holders; film support rings; a cryochamber cover with
three doors a nd various other cryotools. This system allows two ranges of
chamber temperature to be selected: - 120 to - I 60°C (l 53-113 K) or - 20
to -80°C (253- 193 K).
The LKB Cryokit 14800 attachment, together with the LKB 14860 cryo-
too ls, represents one of the best known cryoultramicrotomy systems. Nu-
merous authors have published results from this equipment, including
Seveus (1979, 1980), who has been concerned with its commercial develop-
ment. The cryochamber encloses the whole specimen/knife area and gives
sufficient space for other ancillary items to be included (Fig. 4.6). Two
liquid nitrogen reservoirs are provided (one serving the specimen head, the
other the knife stage) and thus the knife and specimen stage temperatures
can be independently and automatically controlled by varying the coolant
level. The specimen head can achieve temperatures down to JOO K
( - 173°C) while the knife cools to about 120 K ( - 153°C). There is automa-
tic control between these temperatures and temperatures approaching 273
K (0°C). The specimen holder is rapidly interchangeable, so several holders
can be pre-cooled and inserted into the chamber. A rotatable specimen
holder allows trimming to be carried out. Either diamond or glass knives
can be used, spare knives also being housed within the chamber . Using a
liquid nitrogen 'booster' it is possible to evaporate some liquid nitrogen
from within the cryochamber, thus providing the anticondensation advan-
tages of purging with dry nitrogen gas.
The Reichert-Jung Cryochamber FC 4 (Sitte et al. 1979) was specificall}
developed for use with the Reichert OmU4 Ultracut ultramicrotome (Sitte
et al. 1979, 1980; Sitte 1982, 1984). It represents an extremely adva nced ap-
proach to cryoultramicrotomy. The walls of the 'cryochamber' a re ofalum1·
nium, the side walls being cooled directly by about 1.0 litre of liquid nilro·
gen stored in an adjacent ' twin-tank' (Fig. 4.7). This lank, together with the
cooled chamber, is insulated by surrounding styrofoam which is itself
enclosed within an a luminium jacket that is continuously held at 293 K
(20°C). More than 75% of the nitrogen gas escaping from the supply tank
is warmed up to room temperature and hence causes minimal disturbanL't
d
of ultramicrotome function. A small ( < 25%) amount of the cold gas 1
guided by aluminium plates to the bottom of the chamber and provides Fog 4.6. LK B 14800 Cryokit cryosectionin . .
continuous, smooth and laminar flow of gas from the bottom to the top 0 llached lo lhc U ltrotome V (b) S . . g system. (a) General view of LKB 14800 Cryokit
her 'pins' on which specime~s a re ~ec1men holder and locking ring together with one of lhe
the open chamber. The specimen is fixed on a bridge which is co nnected t
llltn holder under liq uid nitrogen be~ozcnt. (c) ;he specunen. on its pin, is secured lo the speci-
the specimen arm of the Ultracut and enters the chamber through the Opell d· . •Ore rans.er to them·. 1 .
gn Jn a honzontal plane, close lo the knife ed .1cro ome-, (cl) The grid carrier holds
top (Fig. 4.7b). All parts of the specimen bridge exposed to the roo ~I) be manoeuvred o nto lhe grid ( ) Th ge, while sect1on111g. T he sections can thus
Ider Ill !he cryochamber is shown ~h- ·1~ general layout of the specimen , knife and grid
in
JS ' ustrallon. (Illustrations by courtesy of LKB-P -
dukter AB.) ro
Freeze-sectioning 21 7
Robards and Sleytr Low temperature methods in biological electron microscopy
216
atmosphere are separately heated to about 293 K (20°C), so preventing
frosting of these parts or any other thermal effect. The specimen holder,
knife holder and preparation plate each contain a temperature sensor and
a heating element. The preparation plate is used for flattening specimens
and/or for freeze-drying. Since the atmosphere within the chamber is cooled
down by the cold aluminium walls, together with the cold nitrogen gas flow
from bottom (77 K; -196°C) to top (123 K; - 150°C), specimen, knife and
preparation plate reach temperatures of about 100 K ( - l 73°C). By partial-
ly filling the chamber with liquid nitrogen, a further reduction in tempera-
ture (or more rapid initial cool-down) can be achieved. The heaters can be
used to warm the three separate components rapidly to up to 390 K
(+ J l 7°C), so eliminating films of water condensation that would otherwise
occur during more gradual warming to lower temperatures.
Any temperature of the three main components (knife, specimen and
preparation plate) between 100 K (-173°C) and 273 K (0°C) can be accu-
rately maintained(± 5°C) over prolonged periods and the chamber temper-
ature can be held to a constant value ± l °C at any given level. Since the
walls of the chamber are, under any operational conditions, colder than the
otber components, any condensation will occur on the walls, so leaving the
important areas free of contamination.
Specimen
bridge The FC 4 chamber is easily and quickly mounted onto the Ultracut knife
support. There is an entry system so that sections can be transferred out of
the cryochamber and onto other equipment. Cryogen consumption is low
(3.0 I N2 h - 1) even at the lowest temperatures and the twin cryogen tanks
are automatically refilled from the main Dewar.
The standard knife holder system allows the subsequent use of two pre-
cooled glass knives or one glass and one special diamond knife, as well as
LN connection
2
betw een twin tanks one stainless steel knife for trimming. Adaptors allow the use of special dia-

hg. 4.7. Reichert-Jung Ultracut ultramicrotome with FC4 cryoattachment. (a) Photograph
llustrating general layout of the ultramicrotome, FC4 cryochamber, su p porting table and
hqu1d nitrogen D ewar vessel.(b) Diagrammatic illust ration of the main components of the FC4
'l)Oattachment. Liquid nitrogen is supplied , under pressure from the D ewar vessel, to the twin
Heater
cartridge
tanks surrounding the specimen/knife a rea of the ultramicrotome. Ra pid cooling can be
btamcd by 'overfilling' the tanks so that some liquid nitrogen spills into the cryocbambcr.
b Once working temperature has been achieved, it can be mainta ined constant within precise lim-
h (5ee text). (Fig. 4.7a reproduced by courtesy of Reiche rt-J ung (U .K.) Ltd. Fig. 4. 7b redra wn
rom an original illustration provided by H. Sille, Universitiit des Saarlandes, F.R.G. and
reproduced by courtesy of H. Sitte and Reichert-Jung.)
218 Robards and Sleytr Low tempera/ure me /hods in biological e/eciron microscopy Freeze-seclioning 219

. an le can be adjusted between 0 and 90 whil_e trend has been much more towards minimal use of pretreatments and any
mood kmves .. The_ clearance f hg knife around the vertical axis is also poss1-
other 'wet' processes which would detract from the effectiveness of such
± 100 eucentnc adjustment o t e . . g signals exists to prevent system
b foptical or acoustic warnm techniques. It may be decided that the best course of action is to fix the tis-
ble. A n um
. er o to avoid
. loss o f cooIan t , e tc . While the details provided here
sue, in which case one of the conventional recipes can be used (Glauert 1974;
malfunct10ns or
· d. t the roam . fieatures o f the FC 4 cryoattachment, it can
Hayat 1981). Because cryoultramicrotomy is often used in an attempt to re-
only serve to m ica e . . t ment that can be used with rela-
be seen that this gives a very versat1 1ems ru tain otherwise diffusible substances within the specimen, it has sometimes
been suggested that vapour fixation (e.g. with formaldehyde fumes; Bern-
tive ease over prolonged periodds. "bed bove each have their own virtues
h It amicrotomes escn a . ha rd and Viron 1971) is preferable to wet fixation prio r to cryoultramicro-
T e cryou r
.fi fi atures· all have b een use d successfully for the preparation of tomy. T he individual must assess the requirements of the particular p repa-
and spec1 ic e . . . H the success or failure of cryoultra-
frozen-hydrated thm sections. owever, ration method and chose accordingly. After fixation, some cryoprotectant
microtomy depends very greatly on: treatment may be used. Indeed, fixation may often be necessitated because
cryoprotection is required (§2.3.2a). Glycerol is the most commonly used
(i) the nature of the specimen itself, . . cryoprotectant (§ 2.3.2a) but infiltration with sucrose (20% w/ v) gives ad-
(ii) the manner in which it is prepared for sect1onmg, . . van tages in terms of better 'sectionability'.
(iii) . 1 parameters that are selected fo r the sect1onmg process,
the phys1ca
4.3.lb Mounting specimens for freezing
and ·
the skill of, and/o r investment of time by th e o perator
. , which must
(iv)
be considerably more than for conventional ultram1crotomy. Two main considerations apply in chosing how best to mount the specimen
for freezing. Firstly, the specimen needs to be mounted firmly onto the mic-
It is these factors that will be d iscussed below. rotome specimen suppor t and, secondly, the opportunity may be taken (par-
ticularly if a penetrating cryoprotectant was not used during the earlier pro-
cessing stages) to confer some cryoprotection by 'embedding' or 'encapsu-
4.3 Sectioning lating' the specimen in a suitable medium. Appleton (1978) used small brass
or copper specimen holden in the end of which a small, rough-surfaced cu p
4.3. 1 Specimen preparation for sectioning had been machined (Fig. 4.~ ; see also Fig. 4.5). Tissues often adhere to the

. re aration prior to ultra-thin frozen sec·


The basic methods of specimen _P dp . Cha ter 2 and it is important that
. . have already been descnbe m p Specimen
t10nmg d. f these Ice - - -
the reader should have a good understan mg o .

4.3.la Pretreatment
Freezing
. . 0\
fi the specimen or to pretreat it a . in
The choice of whether or not to. ix d , d n the nature of th~
.h h menter and epen s o
other way, resides wit ~ e expen s of er oultramicrotomy, foa· Fig 4.8. Methods of mounting specimens for sectioning. Simple means of attaching a speci-
information that is reqmred. In the early day . y se of tro ugh liquids m(n l o a metal support prior to freez ing. The water or surrounding medium (for mounting
tion, cryoprotection, encapsulation (of the spe~t~~;h ~ie recognition of the llkdia see Appendix 2) provides good adhesion of the tissue block to the support, o nce frozen.
and post-staining were all commonly advocate in t~e localisation of diffu· Thr same princip le can be used for mounting specim ens onto other supports (e.g. see a lso Figs.
particular importance of c~yo~ltram1crotomy d t chemical methods. the '5 and 4.6). If frozen specimens are to be attached to supports, then low temperature glues
sible elements and the apphcatton of h1sto- an cy o may be necessary (§4.2.1 ; T a ble 4.1). (Red rawn fro m an original diagram in Roland 1978.)
thods in biological electron microscopy Freeze-sectioning 22 1
Robards and Sleytr Low tempera lure me
220
. b t a thin application of a suitable bons because they can contaminate the specimen (possibly leading to spuri-
bth . inherent m01sture u ous peaks during microanalysis) and may also remain as liquid surface lay-
holder Y eir . A ndix 2) or simply a 6% solution of methyl
medium such as T1ssue-Tek ( ppe · d by Appleton any ers over the rest of the specimen during sectioning. Nevertheless, Gupta
, 979) help although as cautione ,
cellulose (Seveus 1 can . h t'.ng medium should be regarded considers that the virtues of Freon 13 outweigh the disadvantages so long
ll . ontact with t e moun i as excess cryogen is shaken from the tissue blocks after freezing. If the speci-
tissue actua Y m c . . t al or analytical information. The
with suspicion ~hen mte~pretm(Pg :~~~~ydroxyethyl starch (HES) as both men is of any significant size (>0.l mm in all dimensions), and particularly
virtues ofpolyvmylpyrrol!done b reported by Echlin et al. (1977, if it is mounted on a specimen support, then it is best simply to freeze it
t d mountant have een by plunging into sub-cooled nitrogen (§2.4. la). Only if the specimens are
cryoprotectan an ortin medium confers cryopro-
1982). A 25% solution acts as af~ood sutpapls anJ importa~tly in the current l'ery small is it worth using the ultra-rapid cooling methods described in
. . ·ng growth o ice crys , , .
tect1on m suppressi B'ddl mbe et al 1982). Thus, ttssues §2.4.
· \l (see also I eco . Sitte et al. (1979, 1980) have given the matter of freezing prior to cryo-
context, sections we f to be more amenable to sectioning
. t ·x of PVP o ten prove ul tramicrotomy special attention and conclude that the vast improvement
embedded ma ma n h t ent Other 'encapsulating' media
. f without sue trea m . in 'sectionability' obtained with very small ice crystals is a goal well worth
than tissues rozen . %bovine serum albumin subsequently c~oss-
that have been used mclud~ 1 o 200 elatin which may need to be mlil- pursuing. For this reason they have advocated the use of carefully con-
linked with glutaraldebyde, ~nd % g · t 37oc followed by 1 week trolled cold block-freezing (§2.4.2), rather than plunge-cooling, prior to
'd rable penod (e g 20 mm a cryosectioning. This not only yields very small ice crystals within the surface
trated for a cons1 e . '. depends on the nature of the speci-
al 4oq. The de~ision f~r ~~;~~r :~~1: sensible first attempt would be to zone but also provides a relatively flat surface that can be orientated parallel
men and what ts require. ' . 601 methyl cellulose and some others to the knife edge, so making it possible to cut quite large areas of section
mount some specimens simply usmg / o . through well-frozen cells. If the tissue surface is very carefully orientated as
using PVP or HES as an embedding matnx. parallel as possible to the surface of the polished metal block (metal 'mir-
ror') prior to freezing, then deformation is reduced to a minimum (Fig. 4.9).
4.3.lc Freezing specimens Cold block freezing also leads to dimensional changes which are made
. leton (l978), in which quite large (1.0 worse by incomplete relief of shear stress caused by dry sectioning. This
The method of freez1.ng. used ~ Ap;. I metal holder are cooled by plunging technique is relatively complex and the beginner would do well to experi-
mm1) tissue blocks within a s~ stan i~ 1 t provide ultra-rapid cooling ment with simple plunge-cooling methods until he or she understands well
I d · t en will certam Y no enough what is involved before embarking on cold block-freezing.
into sub-coo e m rog , be btat'ning a suitable large, unda-
h t be drawn tween o It may be advantageous to freeze the specimen alone (i.e. without a sup-
rates. The balance as o . . bi h enough cooling rate to suppress
maged, tissue bloc~ and obtammg :ate~ above, cryoprotectives will, there· port) so that very high cooling rates can be obtained, such as when spray-
formation of large ice crystals. Ass . more fully discussed JO freezing cells. In these circumstances the frozen cells will need to be 'stuck'
The reqmrements are to the specimen support (while, of course, maintaining low temperature).
fore, often be necessary. d 'b d the well-established method of
§2.3.2a. Seveus (1979, 1980) bas escrl1 ~l ' . s' She bas used 'solid' ni· This requirement has led to the search for suitable low temperature 'glues',
. th nds of smal s1 ver pm · . e substances with a melting point lower than the temperature at which spe-
freezing specimens on e e b t 6-8 um of the specimen
208oC) and states that t e ou er ,,. .
trogen at 65 K ( - . 0 again the absolute cooling rate cimens will be mounted (typically lower than about 80°C; 193 K), with a
is essentially free of observable;c~. t' neeature ~f the surface, together with gnificantly higher boiling point and with a relatively low vapour pressure
is unlikely to be high and the un. u a m~ln have taken place at the surface at the temperatures to be -used (Table 4.1; §4.2.1). The use of all such low
the mechanical damage that ':ill u~ua y 'bl to take complete advantage temperature glues must be viewed with particular caution when rnicroanaly-
itself means that it is difficult, if not impossi. e, G ta (1979) bas used llS is subsequently to be carried out because they may contribute either to
' . · the superficial zone. up ,.
of the very small ice crysta1s m . , . , or 'microcbucks mto specific elemental peaks (as is the case with the halocarbons) and/or to the
·ng specimens on pms
rather similar method s, d roppi f. d not to use hatocar continuum radiation. In practice, substances with melting points around
(Fr eon 13) Many workers have pre erre
halocarb on 13 ·
Low temperature methods in biological electron microsrnpy Freeze-sectioning 223
222 Robards and Sleytr

_ 100°C (170 K) have proved useful. For example, n-heptane (Steinbrecht


and Zierold 1982, 1984) has a melting point of -90.6°C (182 K) and a boil-
ing point of +98.4°C (372 K) together with a vapour pressure of 4 x 10- 3
Insulating Torr (5.3 3 x 10 - 1 Pa) at -100°C (compared with I x J0- 5 Torr (l.33 x 10-3
holder
Pa) for ice a t the same temperature). Thus, specimens can be 'mounted' with
the 'glue' a t temperatures warmer than - 90°C and will become stuck to the
support as the temperature is lowered. Furthermore, the vapour pressure
_which is higher than that of water - allows freeze-drying of sections with-
out risk of solvent extraction of the specimen. Steinbrecht and Zierold
reported that n-heptane sections satisfactorily at temperatures between
- 11 5 and - J 30°C (158 and 143 K). Some other low temperature glues are
listed in Table 4.1 in §4.2.1.

4.3.Jd M ounting specimens on the ultramicrotome

Shaft (low weight)


When the specimen has been frozen on its holder, it is transferred to the
chuck of the cryoultramicrotome. This should present no special problems
I Precise alignment of
although, of course, it is always essential that the specimen is not allowed
lo warm up inadvertently once it has been frozen. The simplest method is
I injection shaft
usually to tra nsfer the specimen, protected within a small insulated con-
tainer of liquid nitrogen, into the cold chamber of the ultramicrotome from
I ~ here it can be picked up, using pre-cooled tools, and inserted into the
chuck. Because ice is so brittle, it is not normally possible to clamp frozen
specimens directl y. However, if there is a need to fill a space between a fro-
Cold metal
1cn specimen, or a specimen holder, and its chuck, then thin, soft, indium
block film (Appendix 2) is very suitable for ensuring good thermal contact.
d Although some workers have preferred not to trim frozen specimens,
c
most appear to believe this to be necessary, particularly if ribbons of sec-
. . . d Id block-freezing in preparation for cryt>-<'C
Fig. 4.9. Comparison of plunge-frcezm.g : oc~ns only have a superficial layer of 5-lll µ
tions are to be produced. Because of the need to maintain a low specimen
tioning. (a) Tissue blocks frozen m hqu1d y·tgl This is the region from which cryosccll•' temperature, trimming is usually carried out wi th the specimen mounted on
thickness that is well-frozen with small ice crys a sf. h bl k 1·s likely to be irregular. er)•'"' the microtome arm. Appleton (1978) recommends trimming in much the
. b se the edge o t e oc
are cut most easily. However, ecau f poorly frozen material (sec §2.3 I I same way as for conventional ultramicrotomy (see Reid 1974). The leading
. d/ ·ontain large areas o - ·
tions will be mcomplcte a n .or c 1 bl k then a relatively large, Oat area ol " nd !railing block edges are made parallel to the knife edge, while the lateral
. · f
If the spccnnen ts rozen a . gamst a cold meta
. ·r t xacrly parallel to the sur actiJ"
oc ' r ..
. . d th t cryosect10ns even I no e edges are trimmed to produce an approximately square face. Trimming can
ll ice crystals m them see ~ 2 ·4 · 2 ·
froze n tissue 1s obtame so a ' . ( ~ ) (c) To be able to pn
.h be accomplished by rota ting the knife holder and then 'cutting' a pyramid
contain large area s wit very sma that the direction of ;pc<t
large Oat areas o f well-frozen tissue, it is necessary toh ens.ure(d) only a very small reg1<•n wnh approximately 30° sides. It is, of course, true that trimming often re-
' h . Id block surface ot erw1se H
travel is exactly normal to t e ~o . ori i~al illustration kindly provided b> \es much of the best frozen parts of the specimen. If cold block-freezing
well-frozen tissue results. (Redrawn from an g . . )
and reproduced with perm1ss1on. s been used, then trimming should be confined to the edges of the block
Freeze-sectioning 225
Low temperature methods in biological electron microscopy
224 Robard~ and Sleytr
abrade opposing faces until the required configuration is obtained (Biddle-
o that the well-frozen face is left intact. In some microtomes (e;g..th~ R~i­ combe et al. 1982). An important point that can easily be overlooked is that
s . . 79 1980) a special steel alloy kmfe w1 th
chert-Jung FC 4, Sttte et a·1 19 ' '. · (F. 4 10) An alternative the specimen should be allowed to come to thermal equilibration with its
n elli tical edge is used specifically for tnmmmg ig. . . holder/chuck before sectioning is commenced. This may take a considerable
~eth:ci of shaping the block is to use a pre-cooled metal file and gently to time (30 min-LO h) but, if not allowed to take place, sectioning is erratic
and not under proper control.

4.3.2 Dry sectioning

4.3.2a Cutting the sec/ions

Whereas the general principles of sectioning apply irrespective of the type


of cryoultramicrotome being used, there are inevitably substantial differ-
ences in details of operation from one instrument to another. Therefore,
only general procedures will be described here, and the reader will need to
refer to the relevant manual of the microtome being used for details of parti-
cular operational conditions.
For most work involving thin frozen sections, temperatures in the region
of the specimen/knife of the order of - 80°C (193 K) and lower are required.
For the LKB 14800 Cryokit the following standardised sectioning condi-
tions are suggested (Seveus 1980):
(1) specimen temperature -1 40°C (133 K);
(ii) chamber temperature between - 100 and - 120°C (173 and 153 K);
(iii) knife temperature between -100 and -120°C (173 and 153 K);
(iv) knife angle 40°, clearance angle 4°; and
(11) sectioning speed as low as possible, usually 1-2 mm s- 1.

Only trial and error will determine precisely the best conditions for any par-
ticular specimen. Under such conditions the sections appear transparent,
glossy and hard, resembling cellophane.
Steel knives are preferred for cutting sections of frozen material down to
about 0.5 µm thickness, after which glass or diamond knives are necessary.
Special diamond knives are available for cryosectioning from Diatome (see
Appendix 2), while Reichert Jung AG (see Appendix 2) make knives of har-
b dened stainless steel which are suitable for low temperature work.
It is difficult to cut frozen sections to a precise predetermined thickness.
. FC4-Cr osystem. (a) GeneraI arran11emcnl
- '''
Fig. 4.10. Knife holder for the Reichert-Jung , oc~amber (b) Three knives are mounted Although interference colours can often (if not always) be seen, they do not
the knife holder in relation to the spec1m_en in the cry d k ·r r . dry sectioning; in the centrt provide the good indication of thickness that is available during conven-
. (A d x 2) diamon m1e 1or 1
o n the left, a special Di_atome ppen i . h t a normal glass knife. (Reproduced by cou . ional ultramicrotomy. Also the colours do not indicate actual thicknesses
an eliptical metal trimrrung edge; and. o n the ng ' )
tesy of Reichert-Jung.
thuds in biological electron microscopy Freeze-sectio11ing 227
226 Robards and S leytr Low tempera ture me

d ·ng wet sectioning of resin-embedded blocks (1 975) devised a simple system for stretching sections and preven ting them
equivalent to those seen ur1 . . from rolling by feeding them as they are cut between the adjacent edges of
because of differences in refractive inlde,x. those that are cut in the 'thicker' two carefully aligned glass knives (Fig. 4.11 b ). Schiller and T augner (1979)
. tions and part1cu ar Y .
F rozen t h m sec ' I they are cut. T his has actua lly coated the cut surface of a frozen tissue block with metal (platinum)
ften show a tendency to cur as
range ( > 0.2 µm) ' 0 f t. roll plates (Fig. 4.11) o r other and carbon prior to cutting each section. They found that this made the
f ons for the use o an I-
led to many sugges I . ha ening. The simplest type of mechanism handl ing of t he sectio ns easier, the advantages including:
devices to preclude this from PP . cryotomy and consists of a glass
1 (t) enhancement of contrast without the need for staining;
is similar to that used for convenht!Onka ·c dge (Fig. 4.1 l d). Hellstrom and
. .t. d very close tot e m1e e (ii) stabilisation of the sections for subsequent freeze-d rying, histo-
covershp pos1 10ne f and use of such a plate. Dorge
Sjostrom (1974) describe the con! struc It~noll plate which was manoeuvred chemistry, TEM observation, or analysis witho~t further coating;
d a small g ass an 1-r (iii) protection of the sections against artefacts during autoradiograph y;
et al. (1978) a lso use h . µm away from the cutting edge,
. · pulator so t at 1t was 5.0 and
with a m1cromam h hi h the sections could pass. Boll et al
so forming a small gap throug w c (iv) allowing t he production of useful surfaces for bulk specimen micro-
an alysis.

T a ugner sy stem
Anti-r oll plate Frozen sections can be extremely difficult to handle because they often
become electrostatically charged, in which case they may well be subject to
strong repulsive forces when attempts a re made to place them onto grids.
Spe c i men

"-<r=.
Various methods have been tried to overcome this problem (e.g. Nicholson
1978). Fine metal needles are better than hairs for handling the specimen
Sections Knife 1 ~
(Biddlecombe et al. 1982). A n tistatic 'pistols' (Appendix 2) have sometimes
~~ Knife 2/
proved to be helpful, and Sitte has successfully incorporated a Simco-
,mglespike (Appendix 2) into the chamber of his cryoultramicrotome to re-
c
a
b duce problems arising from static electricity. Seveus (1980) refers to coating
Anti-roll device 1he knife with an evaporated carbon film that has electrical continuity with
1he frame of the microtome. This is said to be useful although the thickness
of the carbon layer appears to be critical.

4.3.2b Section collection

Once the sections have been cut, they are collected, transferred to a grid (or
ther support) and made to adhere firmly to this 'carrier' .
Appleton (1973) introduced the method of vacuum collection of ribbons
f sections at temperatu res of about -60 to - 70°C (193 to 203 K ) (Fig.
ti ey tend to cu 4 12) Although this accessory is standard on the Slee cryoultram icrotome,
. . (a) When cryosections are cut d ry. 1 . . '"
. 4 11 C ryosecuon antt-roll systems. . "bl . k them up easily as nbl><
F ig. . . k. it difficult or 1mposs1 e to pie h c number of workers have adapted it to be used on other systems. It is im-
away from the knife edge, so ma mg . b d vised Some are illustrated er
portant that the vacuum 'pull' is both even and reproducible. The leading
Different systems f~r p rcvent'.ng~~is4~~~.1~~)h;~:
.;::gn:r' sy~tem
(see BoB et aL 19:~~1~•1" edge of the first section to be cut is gently 'nudged', using a mounted hair
alternative method ts shown m F g. k . t direct the ribbon o f sections as it d
the slit between two closely adjacent glass mvcs\\o l· te when attached to a support arm' a cold micro-needle, towards the m outh of the vacuum tube. As succes-
· k . excellent anti-ro P '1
(c) A glass covershp ma es an (d) (F' 4 I Id redrawn from Rick ct al. 1982.)
as that shown m ig. ·
228 Robard~ and Sleytr Low temperature methods in biological electron microscopy
Freeze-sectioning
229
sive sections are cut, the vacuum tube is gradually and gently retracted fro m of using a droplet of freezing sucrose (T k .
the knife, thus drawing out the ribbon. When sufficient sections have been small syringe hair or platinum I do uyasu 1973; Fig. 4.13). Using a
. ' oop, a roplet (0 5-1 o 1) f
cut, they are lowered onto a support positioned immediately below the rib- (or approximately 2 o M) sue . . . · · mm o saturated
. rose is inserted mt0 th Id .
bon. the sectioning chamber As the d . e co environment of
. · rop 1et cools 1t beco ·
The vacuum collection technique cannot be used at temperatures much which stage it is touched against th . . mes very viscous, at
. e sect10n(s) which dh ·
below - 70°C (203 K), and so at lower temperatures the sections must be be hfted away from the knifie d Th. ' a ere to 1t and can
e ge. l S procedure ·
collected singly. T hey can be picked up using a mounted hair (e.g. an eye- to be carried out successfully· ,·f th d . requ1res some practice
· e rop1ets are st111 r ·d
lash) or thin metal needle and transferred to the grid which is usually posit- onto the knife edge! When the dr t . iqui they may freeze
. op1e is removed from th
ioned, for convenience, as close to the knife edge as possible. A method of 1t melts a nd the sections expand· th d . e cryochamber
• e rop1et with s f · '
picking up dry sections that has been fo und useful by many workers is that a coated grid which is floated sect" d ec wns IS then put onto
' wns ownwards, on water to remove the

Shell of PVC tape


'
' -Knife edge
'
a •''
~.
.
10°

W
----~
)
c

Fig. 4. 12. The vacuum method of producing a ribbon of d ry cryosections as devised h1


Appleton (1973). (a) A weak vacuum (using a 'fish-tank' pump or similar device) is applied
through the flattened orifice of a serum needle which is positioned very close to the knife edge
A Formvar-coated grid is supported just below the knife edge using a 'shelf of plastic tape
(b) As the first section is cut it is very gently 'induced', if necessary using a mounted eye·fa5h
to enter into the opening of the serum needle. The vacuum should be just sufficient to 'hold
the section without drawing it further into the needle. (c) As sectioning proceeds, the needle
is moved, using a micro-manipula tor, step-by-step away from the knife edge. so mainta1mn#
a continuous, extended ribbon of sections. When enough sections have been cut, the micr<'-
manipulator is used to lower the needle, and hence the sections, onto the wailing grid. (d) A
highly polished, cold , copper rod is then used to press the sections into close contact w1lh tilt'
grid , after which the mounted sections are either freeze-dried or used directly for frozcn-h)d
rated observation or analysis. (e) (opposite) lllustration showing cryosectioning in prog!C"
using the vacuum collection method. (Diagrams redrawn from Appleton (1973) and Fig. 4.1
reproduced with the permission of T .C. Appleton, North-Holland Publishing Compan) and
Slee Medical Equipment Ltd.)
230 Robards and Sley tr low temperalure methodsin
. biological electron microscopy
Free=e-sectioning
231
charge
of orsections.
frozen o ther means; see Wi11ison a nd Rowe 1980) prior to the collection

~
sections

Cryochamber
Frozen sections do not adhere weU to mounting films or grids and must
usually be firmly pressed against their supporting layer. This must o bviously
be done with the sectio ns maintained at Jow temperature. The most com-
mon method is to press the sections against the support using the highly
polished end of a cold meta l (copper) or plastic (nylon) rod (Christensen
transfer of secti ons to g r id 1971) This is done either by hand or using a simple system to ensure that
the pressure applied is consistent and even. F or example, Seveus (1980)
';•···· .... (ff.."·· recommends leaving the sections under a chilled weight in a press for a peri-

••~• "' ~::,:"•~IL"=s==::•_<4~~~=~rid 9 od of l h prior to further processing. It is often useful to 'press' the sections
between two coated grids (e.g. Biddlecombe et aJ. 1982) so that, when separ-
section s ated, the sections are well a ttached to one of the grids. Tved t et al. (1984)

fi "'" '""•" .
d ry cryoscctions as devised hy Tok-
have prod uced a device (Fig. 4. 14) for thi s purpose. Other devices for press-
ing sections have also been described, such as the one illustrated in Fig. 4. 15.
Freezing sucrose droplet method of p1cbng. ~~t is about lo freeze , is brought 111111
. . An alternative method of ensuring good contact between sections and
.
F ig. 4. 13. A dro le t of about 2.0 M sucrose. JUSt a. e lifted from the knife edge. The their substrate is to float them on a drop of cold ( - 100°C; 173 K) isopen-
uyasu ( 1973).
t· ct with the c ryose p ctions ' which adhere
M 'd a llowed to t h aw so that the sections Oat1en
lo it a nd can b tane on their supporting films. This serves the two-fold purpose of removing
con " I
frozen drop et is . the n transferred to an E gn free
,. hed • from Suero sc by noa ting the grid. ) scc1111n folds from the sections and helping to provide good contact with the sub-
o nto t h e gr id ' and the sectionsd 'arc
side downwards, o n is 1 t' then
lled wav.as
ter (see text )· (Redrawn fro m Roland 1978. ~trate when the temperature is red uced again (Saubermann et al. 1977).
11-Butyl benzene (Saubennann and Echlin 1975) or toluene (Somlyo et al.
Suc . f om unfixed spec1me . ns become dispersed dur-b 1982) can be used as low temperature 'glues' to 'stick' sections to their sup-
rose Sometimes sect10ns r . f'ects This may be overcome ) porting film. These glues have already been discussed in §4.3. lc (see also Ta-
ing this· process due to sur f a ce tens10nI etion subsequently was h'mg thc
1' •
and hie 4.1 in §4.2. l for other su bstances having similar properties).
adding 0.5-2% gelatin to t~e sucrose s~r~tions of sucrose (Tokuyasu and
soctions thrnugh demasmg col~~~.
While this method does invol". ':'
Sin er 1976; see also Tokuyas~
con~act "::~llcction)
. vides a useful intermediate pat
of liquids with the """'.ons( and othe< ' wet' metho<b
way between totall_y d_ry sectio ning an
which use trough !tqmds (§4.3.3).

4 3 2c Section mounting
Knife
· · f sections, man Screw
. h commonest supports for rozen etal col
While coated gnds ace t e d T hese usually comp,iso a small m mpk.
:~~~a~~:~l~:~i~ suspenddedp(soe:,t~~r;~:nu
-
other systemhs.chh
1 a :e film can be @
lar across w . d h pe depen s u
b
. 28 . §3 4 8). The precise size an s a . It is particularly 1111-
F1g. 3. m . . h articular nucroscope. I w dil- c Bolt 1
arrangements o n the s_tage oft e ~ould be made hydrophilic (by go
portant that supporting films s 4 14. 'Press· device fo r cryosections as devised by Tvedt et a l. (1984). (Redrawn after
Tvedt et al. 1984.)
232 Robards and Sleytr Low temperature m ethods in biological electron microscopy Freeze-sectioning
233

A second plastic (e.g. Formvar) film can be placed over the sections once
they have been mounted onto their support to ensure greater stability. This
does not appear to interfere significantly with any subsequent freeze-drying.
Once the sections have been made to adhere firmly to the support, they may
either be freeze-dried (§ 5.5.2d) or transferred directly onto the cold stage
of a microscope for transmission viewing and/or analysis.
Cryosections are often freeze-dried before viewing and/or analysis at
am bient temperature (Frederik et al. 1984). The drying process is conve-
niently carried out on a temperatu re-controlled platform in the microtome
cold chamber. The theory of freeze-drying is discussed in Chapter 5 and its
application to cryosections in §5.5.2d. Zingsheim (I 984) has made a study
of sublimation rates of ice in a cryoultramicrotome and has concluded that
the danger of even partial (unintentional) dehydration of 0.5-1.0 µm thick
sections is less serious than commonly assumed.
It has generally been believed that frozen-hydrated thin sections show
very little intrinsic contrast in transmission electron microscopes. However,
the recent work of Dubochet and his colleagues (Dubochet et al. 1982;
Chang et al. 1983; McDowall et al. 1983; Dubochet and McDowall 1984)
has indicated that quite remarkable contrast can be observed within such
sections. Much remains to be done before we are fully able to appreciate
the mechanisms of contrast formation in viewing frozen-hydrated sections.

4.3.3 Wet sectioning

Any liquid coming into contact with frozen sections must be regarded with
suspicion because of the probable induction of artefacts, including the pro-
duction of contaminati ng elements prior to microanalysis and, most impor-
tantly, the probable loss or relocation of diffusible components of the speci-
men. Such considerations have turned many workers against using any
hqu1ds during cryosectioning. Microanalytical or cytochemical studies are
now carried out almost exclusively using 'dry' sections throughout. Further-
more, wet sectioning can usually only be done at relatively high temperatur-
es ( > -60°C; 213 K). Despite these restrictions, the morphological appear-
ance of sections cut using trough liquids can be much better than that of
dry sections and the sectioning process itself is easier. It is for the individual
. . 1 rovided with the LKB 14800 Cryol11 Yiorker to make a choice after careful consideration of all the requirements
F . 4 15 'Pressing' cryosec!lons using the too s p . . bly. (b} The' volved.
ig. . . 'ct with sections to the press a sscm
The grid carrier is used to transport the gn . . Id' T n (PTFE) and bras; pf('
h ·d T his system has go • c on
specimen press is placed onto t e gn . . . . f LKB-Produkter. AB.)
Dimethylsulphoxide (DMSO) is the most commonly used trough liquid
surfaces provided. (Ill ustrations by courtesy o or cutting frozen wet sections (Bernhard 1971) but Sjostrom and Thornell
234 Robards and Sley tr Low temperature methods in biological electron microscopy
Freeze-sectioning
235
(1975) concluded that this led to the extraction of elements. DMSO is also 4.4 Applications offreeze-sectioning
known to be toxic to some cellular systems and can inhibit some enzymes
(e.g. acid phosphatases). Nevertheless, DMSO is a convenient choice for f rozen thin sections can serve a th .
s e startmg point fo b
someone contemplating a first attempt at cryoultramicrotomy using a tant morp h o logical and analytical a r . r a num er of impor-
trough liquid. A 60% solution ofDMSO freezes at about -60°C (213 K). nale is that rapid freezing will h PP i~at1ons (Table 4.2). The basic ratio-
Alternatives to DMSO are available. Cyclohexane can be used at temper- sou! ble compound distributio .ave retamed both ultrastructuraJ detail and
. h n m a more natural stat th
atures below - I 30°C as a trough liquid (Hodson and Marshall 1970, 1972). uon met ods. In interpreting th e an other prepara-
~ .. e resu 1ts from fro h' .
Ethylene glycol, methylene glycol, halocarbons, isopentane and even liquid there1ore, of critical importanc th t h zen t m sect10ns it is
. h
dunng t e preparation procedu e a t e changes that h ,
nitrogen have all been used as tro ugh liquids with varying degrees of suc- h . may ave occurred
cess. However, in general, their use at temperatures below the probable re- noted and understood. re, sue as ice crystal formation, are both
crystallisati on point for ice (i.e. lower than about -80°C; 193 K ) is prob-
lematical and not recommended for the novice in this field. Thus, while the
use of trough liquids may lead to easier sectioning, will prevent the 'rolling' TABLE 4.2
of sections, and can produce pleasing morphological details, the current Typical applications of ultra thin cryosections
concensus - coupled with improvements in cryoultramicrotomy apparatus Tissue/ Celf Technique
Application
and techniques - tends to suggest that they should only be used where their Reference
Various Protein A-gold
presence is unavoidable. Antigen localisation
labelling Roth (1982)

4.3.3a Contrast enhancement Bean cotyledons Immunoferritin


P haseolin localisation
labelling Baumgartner et al.
Contrast enhancement (staining) is used for morphological, but not microa· ( 1981)
Human lymphoid
nalytical, studies and is usually only applied to 'wet' sections. Either positi1e Double immuno-
A ntigen localisation
hssue
o r negative staining methods can be employed. T he freshly cut sections enzymatic label- Mason and Woolston
ling (1982)
sho uld not be allowed to dry out. F loating grids o n a solution of 0.2-4.0"
sodium silicotungstate or 1-2% neutralised potassium phosphotungstate al frog skin
X-ray analysis
Cellular electrolyte
pH 7.2 for about 10- 30 sat 35- 40°C (308- 313 K) provides good negati\C Rick et al. (1979)
concentration
staining. The precise time depends on the nature of the specimen and musr
Rabbu cardiac X-ray analysis
be established by experimentation. Positive staining can be effected using llllUcle Organelle elect rolyte
Hagler et al.
lead citrate (e.g. Reynolds 1963; 5.0 s- 1.0 min) or uranyl acetate ( 1.0-5 O concent ration
( 1983)
min) in the conventional manner (Lewis and Knight 1977) but for much Ulalase crystals
Low temperature
StructuraLeffects
shorter times than for embedded material as the stains have more direct a~ TEM Chang et al.
of freezing
cess to the structures with which they react. Tsuji (1974) has outlined (I 983)
lllous
method for simultaneous section collection and staining using a droplet A utoradiography
Localisation of diffusible
and X-ray analysis Baker and Appleton
silicotungstate. After staining, the grids should be rinsed briefly by floau substances
(1976)
on clean water and then carefully d ried on a piece of filter paper. Sta1m M.nmahan gut
A utoradiography
using osmium tetroxide vapour has also been suggested and may be prefer ..-iium Localisation of diffusible
Johnson and Bronk
substan9es
able if redistribution of intracellular components is to be minimi (1979)
although it does not work if sections are coated with a developed nuc Immunolabelling
Antigen localisation
of sections Beesley and Campbell
emulsion . ( 1984)
236 Robardi and Sleytr Low temperature methods in biological electron microscopy
Freeze-sectioning
237
Cryosectioning has produced some excellent images of chemically
untreated cells (e.g. Frederik et al. l982b, 1984; Chang et al. 1983; McDow. Appleton, T.C. ( 1974), A cryostat approach to ultra-thin f .
copy; a morphological a nd X-ray analytical st d J M' rozen sections for electron micros-
all et al. 1983) but it is in the general area of cytochemistry and microanaJy. App Ie ton, TC ( 19 u y, . icroscopy JOO 49
. . 78), The contribution of cryo-ul trami l , . . .
sis that the technique has a particularly significant contribution to make. in: E lectron probe microana lysis in bio lo D cro omy to X-ray analysis in biology,
don), p. 148. gy, .A. Erasmus, ed. (Chapman a nd Hall, Lon-
The application of cytochemical methods following frozen thin sectioning
falls outside the scope of the present book. The reader is referred to one Baker. J.R.J. and T.C. A ppleton (1976) A 1 h · r
• ec rnque •Or electron m ic .
(and X-ray microanalysis) of diff ' bl b . roscope a uto rad1ography
or more of the increasing number of papers reporting the use of cytochemi- J. Microscopy 108, 307.
us1 e su stances usmg free d · d f
ze- n e rcsh frozen sections,
cal (including immunocytochemical) applications following cryoultramicro- Barnard, T. ( 1982), Thin frozen dried cryoseclions and b.10 1 .. 1 X .
tomy (e.g. in Bullock and Petrusz 1982, 1983; Polak and Vamdell 1984; Ta- roscopy 126, 31 7. ogica -ray microanalysis. J. Mic-
ble 4.2). Frozen thin sections are also ideal starting points for the use of Baumgartner, B., K .T. Tokuyasu and M J Ch .
. · · nspee 1s ( 198 1) Jmmunoc t h · .. 1 .
autoradiograpby to localise radiolabelled compounds within cells (Williams lion of reserve protein in the endoplasm.. . • . Y oc em1ca locahsa-
. tc re11cu 1um of developing b " (Pl
cotyledons, Planta 150, 419. ean U/.ffo1us vulgaris)
1977; Table 4.2).
Beesley, J.E. and D.A. Campbell (1984) Th f I ·
However many frozen thin sections are used for determination of struc- · · ' e use 0 u trathm cryosections ~ I 1· · -
mOuenza virus antigens in infected vero cell 1 . H. .· · or oca 1satwn of
ture, it has to be admitted that their real usefulness resides in the analytical cu tu res. 1stochem1stry 80 497
Bernhard. W. ( 1965), Ultramicrotomie a bassc lem perature, Ann . Biol. 4, ;, .
information that they are potentially able to provide. It is therefore necess- Bernhard. W. ( 1971). Improved techniques for the . . .
Cell Biol. 49. 73 1. preparation of ultrathm frozen sections, J .
ary to adopt rigorous criteria to ensure that any elemental or soluble com-
pound redistribution is detected so that false interpretations will be avoided Bernhard, W. a nd E.H. Leduc (1967), U ltralhin frozen sections I
preservation, J. Cell Biol. 34 757. · ·Methods a nd ultrastructural
It is useful to prepare 'standards', such as a gelatin solution containing dis- Bernhard.W.and A.Viron(l971) Impo dt h .
. · r ve ec rnques fo r the prepa . f f .
solved salts (Chandler 1977), which can be frozen and sectioned using the sections. J. Cell Biol. 5 I, 732. ra •on o u1trathm frozen
same procedures as used for the biological tissue so that ion concentrations B1ddlccombc. W.H., D .M. Jenkinson SA M w ·ii ·
• · · c 1 iam s WA p N . h I
can be determined (see also Roomans and Seveus 1977; Somlyo and Silcox D.W. Dempster ( 1982) Preparatio n o f c . : ' · · · IC o son, 1-1 .Y. Elder and
' ryosect1ons with a modified Sorvall MT2B l
rotome and cryoattachment. J. Microscopy 126 u tramic-
1979; Zierold 1982a,b, 1984; Gupta and Hall 1984). Appleton (1974, 1978) 63
Boll. H.U .. J. Eschwey, G. Reb and R. Taugner, ( 1 ~75 A . .
has shown how it is possible to prepare frozen salt solutions so that any pos· ullracryotomy, Expericntia 31, 618. ), . sectJon stretchmg appara tus for
sibility of elemental diffusion is avoided. Bullock. G.R. and P. Petrusz (eds) (1 982) T- h . . .
demic Press. London). . , cc n1ques m immunocytochemistry. Vol. l (Aca-
811llock. G.R. and P. Petrusz (eds) ( 1983) T h . . .
demic Press London). . ' cc rnques m immunocytochemislry. Vol. 2 (Ae<t-
References
( handler, J.A. ( 1977), X-ray microana lysis in the electron m' .
in electron microscopy Vol
5 AM GI . icroscope, 111: Practical methods
Appleton, T .C. ( 1969), The possibiljties of locating soluble labelled compounds by cleclro ( hang J-J AW MD , . , . . , a uert, ed. (Elsev1cr/ North-l-lolland, Amsterdam)
.' " . c owall, J . Lepault, R . Freem· CA W I . . .
microscope autoradiography, in: Autoradiography of diffusible substances, L.J. Rolh • Freezing, sectioning and o bservation artefact an, . . a te1 and J. Du bochet (1983),
W.E. Stumpf, eds. (Academic Press, New York), p. 301. roscopy. J. Microscopy 132, 109. s of frozen hydrated sections for electron mic-
Appleton, T.C. (1972a), " Dry" frozen sections of unfixed and unembedded biological male lin,lenscn A K (1967) A .
' · · • simple way to cut frozen th" · · . .
for X-ray microanalysis of naturally occurring diffusible electrolytes: the cryostat appr iempcratures Anat. Rec. 157, 227. Ill sections of tissue at hqu1d nitrogen
J. Microscopie 13, 144. n~1ensen, A.K. (1969), A way to pre are froz . . . .
Appleto n, T.C. ( 1972b), " Dry" ultra-thin frozen sections for electron microscopy and X· microscopy, in: Autoradiography of d~f .bl .c~ thm sect mn~ of fresh tissue for electron
microanalysis: the cryostat approach, Micron 3, 101. 1-\cademic Press, New York), p. 349. I us1 c su stances, L.J. Roth a nd W.E. Stumpf, eds.
Appleto n, T.C. (1972c), Transmission microscopy a nd analysis of frozen sections, Proc: n11ensen, A.K. (1 971), Frozen thin sections o ff h .
Eur. Reg. Conf. Electron M icroscopy, Manchester, p . 56. tenp11on of pancreas and liver J C II B. l res tissue for electron microscopy with de-
Appleton, T.C. (1973), Cryoultramicrotom y, possible applications in cytochemistry. in lhopf FL • · e 10. 51, 772.
, .. . , G. Lechner, K . Neumann a nd H . Sitte (19 . .
tron microscopy and cytochemistry, E. Wisse, W.Th. Daems, I. Molenaar and P "an Ol9L dicn. J. Microscopic 13, 152. ?2), The cryoultram1crotomc Rei-
eds. (North-Holland, Amsterdam), p. 229. e. A.. R. Rick K. Gehring and K Th
. urau (1978), Preparation of freeze-dried cryosectio ns
Freeze-sectioning 239
' ~w temperature methods in biological electron microscopy
238 Robards and SIey tr '-"'
Karp. R .D. , J.C. Silcox and A .V. Somlyo (1982), Cryoultramicrotomy: evidence against melt-
.t t" e X ray microanalysis of electrolytes in biological soft tissues. Ptleugers Arch.
for quant1 a 1v - ing and t he use of a low temperature cement for specimen orientation, J. Microscopy 126,
Eur. J . Physiol. 373. 85. C It ·cro tom y· st udy of ice crystals and frcez- 157.
J d AW McDowall ( 1984), ryou ram1 .
Lewis. P .R. a nd D.P. Knight (1977), Staining methods for sectioned material, in: P ractical
D ubochet , . an . . Re . Conf. Electron Microscopy, Budapest 2, 1407. .
mg damage. Proc . 8th Eur. g . , d J L a ult ( 1982). Cryoprotectio n on o rganic methods in electron microscopy, Vol. 5, A.M. G lauert, ed. (Elsevier/Nor th-Holland Biome-
h J Aw McD owall. R. Freeman an · ep dical Press, Amsterdam).
Duboc et, .. , . . C n •r Electron Microscopy. Hamburg/, 19.
specimens, Proc. 10th lnt. 0 g · . F F . ks and M.H . Asquith (1977), Polymeric \.tason, D.Y . and R .E. Woolston (1982), Double immunoenzymatic labelling, in: Techniques
. p H le B Skaer, B.O .C . Gardmer. . ran . ' I f' in imm u nocytochemistry, Vol. I, G .R. Bullock and P. Petrusz, eds. (Academic Press I nc.,
Echlm, .. · · . f b" 1 . al ultrastructure. II. Phys1olog1ca c 1ects, J.
cryoprotectants in the preservation o 10 og1c London), p. 135.
M icroscopy I 10, 239. ) L t . perat ure X-ray microanalysis of the differ- McDowall, A.W ., J.-J. Chang, R . Freeman, J. Lepault, C.A. Walter and J. Dubochet (1983),
82
Echlin. P .. C.E. Lai and T. L. Hayes09 f L. ow e:or L J Microscopy 126, 285. Elect ron microscopy of frozen hydra ted sections of vitreous ice and vitrifi ed biological sam-
t" s ue in root tips o emna 1111 ., .
entiating vascu lar is . . f It . th. n freezing sectioning technique lo the ples, 1. M icroscopy 131, l.
Fernandez- Moran, 1-l. (1952), Apphca11on o u ra i h F ··k 4 47 1 Polak , J .M. a nd J. M . Varndell (eds.) (1984), lmmunolabelling fo r electron microscopy (Else-
'ti the electron microscope, Arc . isl ' . . . .
st udy of cell st r uctures w1 1. (198 1) Jee crystal damage in frozen t hin sections - free11ng vier. Amsterdam).
Frederik, P.M. and W.M. Busmg ' 'licholson, P.W. (1978), A device fo r s tatic elimination in ul tramicrotomy, Stain technol. 53,
· . t' J M icroscopy 121 , 19 1.
effects and their restora 1_on , · . ( a) Concerning the nature of the cryosec- 237.
Frederik , P .M., W.M. Busmg a nd A. Persson 1982 , Reid. N. (1974), Ultramicrotomy, in : Practical methods in electron microscopy, Vol. 3, A.M .
167
tioning process, J. Micro~copy ~ H· (l 982b) Frozen. hydrated and drying thin
125
• Glauert, ed. (North-Holland Publishing Co., Amsterdam).
Frederik, P.M .. W.M. Busmg and W .H . . ax . Reynolds, E.S. (1963), The use of lead citrate at high pH as a n electron-opaque stain in elect ron
. b ved in STE M J. M icroscopy 126, R PI.
cryosections o ser . · I-I ( 1984) Observations on frozen hydrated and microscopy, J. Cell Biol. 17, 208.
.k p M WM Busmg and W.H .A. ax ' 2 l Rick. R., A . Dorge, K. Gehring, R . Bauer and K. Thurau ( 1979), Quantitative determination
Freden . · ·· · · R C f Electron M icroscopy. Budapest . 411
. · t" s Proc 8th Eur. eg. o n · .
drymg thm cryosec 10n · · . d b dd"ing o f biological samples. m: PraCll- of cellular electrolyte concentrations in thin freeze-dried cryoscctions using energy-dispersive
4) F t' dehydration an em e
Glauert. A.M. (197 • ixa lOn, AM Glauert ed. (N orth-Holland. Amster- X-ray microan alysis, in: M icrobeam analysis in bio logy, C.P. Lechenc a nd R .R. Warner,
cal methods in electron m icroscopy, Yo1. -3 . . . ' eds. (Academ ic P ress l nc. , London). p. 517.
dam). . b X . . alysis of frozen hydrated sections 111th Rick. R., A. Dorge and K. Thurau (1982). Quantitative ana lysis of electrolytes in frozen dried
1979) The electron m1cropro e - ray an . Cp
G upta, B .L · ( · . . h r ·n· M icrobeam analysis in biology, · ..cclions, J Microscopy 125. 239.
new information o n fl uid transportmg e_pn e ia, I . 375 Roland. J.-C (1 978). G eneral preparation and staining of thin sections, in: Electron micros-
W· ds (Academic Press, London), P· · .
Lcchcne and R.R. arner, e · . . .. o f frozen-hydrated cryosecuons: prot>- copy and cytochem istry of plant cell s, J.L. Hall, ed . (Elsevier/North-Holland, Am sterd am),
BL d T A H a ll ( 1984), X-ray m1croana 1ys1 5 6 S
Gupta, . . an . . C f Electron Microscopy. Budapest 3. 1 6 . p I.
!ems a nd results, Proc. 8th. Eur. Reg. o n · . d L M BuJ·a (1980). Technique' f(\1 Poomans. G.M. a nd L.A. Sevi:us ( 1977), Preparation of t hin cryosectioned sta ndards for
B
H agler, H .K., K.P. urton, ·. ·
c A Q reico.. .L.E. Lopez ,111 · •
. h t dy of norma l a nd inj ured myocardium
.
quantitative m icroprobe analysis, J. Submicrosc. Cytol. 9, 3 1.
.
cryosecuomng an. d X -ray microanalysis 111 l e s u
Roth. J. (1982), T he protein A-gold (pAg) technique - a qualitative and quantitative approach
Scanning Electron Microscopy 1980, 2, 493 . . . d L M BuJ·a ( 1983) Standards [,11 for antigen loca lisation on thin sections, in: Techniques in immunocytochemistry, Vol. 1,
L JS Flores R.J . Lundsw1ck <111 • • ' • •
Hagler, H .K., L.E. opez, ·. . . . . 1 . f biological cryosections: vahdauon anJ GR. Bullock and P . Petrusz, eds. (Academic Press Inc., London), p. 107.
quantita tive energy dispersive X-ray m1croa_na ys1s o 131 221 ubennann. A.J . (1979), General considerations of X-ray microanalysis of frozen hydrated
. d' f cardium J Microscopy · · .
application to stu ies o myo .. · . r h , X ay m icroanalysis of cryoscctwn• u,;ue sections, Scanning Electron Microscopy 1979, 2, 607.
Hall, T. A. and B.L. Gupta ( 1982), Quant1ficat1on ior t c -r ubennann. A.J ( 1980), Application of cryosectioning to X-ray microanalysis of biological
J. Microscopy 126, 333.. . . (Academic Press. New York). ll•1uc. Scanning Electron Microscopy 1980, 2, 42 1.
Hayat M A. ( 198 1), Fixation tor elect ron m icroscopy n . of ..d ry.. ultra-thin Ir< ubennann, A.J. and P. Echlin ( 1975), The p reparation, examination and analysis of frozen
. . M S" " t .. ( 1974) An anti-roll plate for attenmg
Hellstrom. S. and · JOS rom · h)drated tissue sections by scanning transmission electron microscopy and X -ray microana-
zen sections, J . Microscopy 102, \ 97. . h . f r cutting ultrathin secll<' )Ms. J. Microscopy 105, 155.
. h· 11 (1970) Ultram1crotomy - a tee mque o
H odson, S. and 1· M a rs a · · J Microscopy 91. 105. bennann. A.J. W.D. Riley and R. Beeuwkcs (1977), Cutting work in thick section cryoul-
of unfixed frozen biological tissues for electron_m1crohscop~, ~ction sawing whilst cu111ng tram1crotomy, J. M icroscopy 111, 39.
arshall (1972) Evidence a gamst t roug -s
Hodson, S. and J · M · · llabennann, A.J., P. Echlin, P.O. Peters and R. Beeuwkes ( 1981 a), Application of scanning
the ult ramicrotome, J. Microscop y 95, 459. . . d"ography of a d1ffu11 dectron microscopy to X-ray a nalysis of frozen hyd rated sections. I. Specimen handling
Johnso n, LT. and J.R. Bronk (1979), Electron m1croscopc a~t::mtmalian intestinal cp1t~ la:hmques, J. Cell Biol. 88, 257.
intracellular constituent. using freeze-dned frozen sect10ns o
lium. J. Microscopy 115, 187.
240 Robards and Sleytr Low temperature methods in hiological electron microscopy Freeze-sectioning
241

Saubennann, A.J., R . Becuwkes and P.D. Peters (1981b), Applicatio n of scanning electron Williams, M.A. (1977) Autoradiograph . d ·
' Y an 1mmunocytochernistry · . p ·
electron microscopy Vol. 6 AM GI . • m. rac!Ica 1 methods in
microscopy to X-ray a nalysis o f frozen hydrated sections. II. Analysis of standard solutioni
Willison, J.H.M. and ~.J. Ro~e CI9.80 a~~rt;i~:- (E.lscv1e_r/ North-Holland, Amsterdam).
and a rtificial electrolyte gradients, J. Cell Biol. 88, 268.
Schiller, A. and R. T augner (1979), C ryo-ultramicrotomy after coating the cut surface of the
Practical methods in electron micrl:co p V~ls~adowmg and freeze-etchmg. techniques, in:
land, Am sterdam). py, · ' A.M. Glauert, ed. (Elsev1cr/North-Hol-
tissue block with carbon and metals before sectioning, H istochemic 62, 153.
Seveus, L. ( 1979), The subcellular distribution of electrolytes. A methodological study using Zierold, K. (1982a), Preparation o f biological er osc f
Ullramicroscopy 10, 45. y c ions for analytical electron microscopy,
cryoultram icrotomy and X -ray microanalysis. Thesis, University of Stockholm. Sweden.
Scveus, L. ( 1980), Cryoultramicrotomy as a preparation method for X-ray microanalysi~ Zierold, K. (J982b), Cryopreparation ofmammar1311 t' ~ .
J. Microscopy 125, 149. issues or X-ray microanalysis in STEM,
Scanning Electron Microscopy 1980, 4. 161.
Seveus, L. a nd L. Kindel (1974), Dry cryo-scctioning of human and a nimal tissue at a \CT) Ziernld, K. ( 1984), X-ray microanalysis of freeze-dried er . .. . . .
1a11ons, Proc. 8th Eur. Reg. Con f Elect M' yosec!Ions. advantages and IImi-
low temperature {l 33 K). 8th Int. Congr. Electron Microscopy, Canberra 2. 52. . · · ron icroscopy, Budapest 3 1695
Sitte, H. (l 982), Instrumen tation for cryosectioning, Proc. 10th Int. Congr. Electron Microi- Zmgshe1m, H.P. ( l 984), Su blimation rates o f · · . · ' ·
307. ice ma cryo-ultramicrotome, J. Microscopy 133,
copy, Hamburg / , 9.
Sitte, H. (1984), Equipment for cryofixa lion, cryoultramicrotomy a nd cryosubstitution in bio
medical TEM-routinc, Zeiss M agazine for Electron Microscopists 3. 25.
Sitte, H., K . Neumann, H . Hiissig, H . Kleber and G. Kapp! ( 1979), FC4-cryochamber for th
Reichcrt-ultramicrotomc UmU4-Ultracut, Arzncim. Forsch. 29, 1809.
Sitte, H .. K . Neumann . H . Hiissig, II. Kleber and G. Kapp! (1980), FC4-cryochamber for Rei·

chert-Om U4-ultramicrotome U ltracut. Proc. 7th Eur. Reg. Conf. Electron Microscopy, Th,
Hague 1, 540.
Sjost rom, M . and L.-E. Thornell (l 975), Preparing sections of skeletal muscle for transm1ss1or
elect ron analytical m icroscopy (TEAM) of d iffusible elements, 1. Microscopy 103. IOI
Sleyt r, U.B. and A.W. Robards ( 1977), Plast ic deformat ion d uring freeze-cleavage: a rev1e1<
J. M icroscopy 110, I.
Somlyo, A.V. and J. Silcox (1979), Cryoultramicrotomy for cleclron probe analysis, in: M"'"
beam analysis in biology, C.P. Lechcne and R.R. Warner, eds. (Academic Press. Londtll,
p .535.
Som Iyo. A.P., H . Shuman and/\. V. Som Iyo ( 1982), X-ray mapping. electron energy los' anab
sis and q uantitative electron probe analysis in biology. P roc. 10th Int. Congr. Electron M,
roscopy, Hamburg/, 143.
Steinbrecht, R.A. and K . Zierold (I 982), Cryo-embedd ing of small frozen specimens for en
u ltramicrotomy, P roc. 10th Int. Congr . Electron Microscopy. Hamburg 3, 183.
Steinbrecht, R.A. and K. Z ierold (1984), A cryoembedding method for culling ultrathm en
sections from small frozen specimens, J. Microscopy 136. 69.
Thorn burg, W . and P.E. Mengers (1957), A n a nalysis of frozen section techniques. I. Sw
ing o f fresh -frozen tissue. J . Histochcm. Cytochem. 5, 47.
T o kuyasu, K.T. ( 1973), A technique fo r ultramicrotom y o f cell suspension and lissues.J (
Biol. 57, 551.
Tokuyasu. K .T. (1980). lmmunocytochemistry o f ultrathin frozen sections. Histochem .I
38 1.
T okuyasu, K .T. and S.J. Singer (1976), Improved procedures for immunoferritin la~chn
ultrathin frozen sections, J. Cell Biol. 71, 894.
T suji. S. ( 1984), C ryoultramicrotomy: simulta neou s transference, "ionic fixa tion"' and 111!¥
staining of fres h frozen sectio ns by means of a silico-tungstic acid droplet. Proc. 8th
Reg. Conf. Electron Microscopy, BudapestJ, 169 1.
Tvedt, K.E., G . K opstad and 0.A. H augen ( 1984), A section press and low elemental 'u
for enhanced preparation of freeze-dried cryosections. J. M icroscopy 133, 285.

· ,.
Chapter 5

Freeze-drying

5.1 Introduction

• A.s discussed in Chapter 2, water is an important structural component of


most biolog1cal systems. Thus, for the examination of non-frozen biological
aterial in an electron microscope this water must be removed. For exam-
lc. 1f a soft specimen (e.g. a piece of tissue) is transferred into the vacuum
fa SEM, the tissue water firstly boils. Subsequently extraction of the latent
heat of evaporation causes the remaining water to freeze. Surface tension
recs prod uced during evaporation, . as well as the ice crystals genera ted
nng freezi ng, cause severe distortions and displacements in most speci-
ns. Although environmental cells (Butler and Hale 1981), in combination
th high voltage electron microscopy, provide another possibilit y for the
cwnination of hyd rated unfrozen specimens they are expected to be of only
tcd use in biolo gy.
In electron microscopy th e following drying methods are used:

Evaporative-drying from distilled water or volatile buffers.


Evaporative-drying from (highly) volatile solvents which have re-
placed the specimen water in a previous preparation step.
Freeze-drying o f hydrated specimens (lyophilisation).
Freeze-drying of specimens substituted with non-aqueous vola tile sol-
vents.
Critical poin t-drying (after replacing the specimen water with a n or-
ganic solvent) from C02 or a suitable halocarbon (e.g. Freon 13).

reviews, monographs and comprehensive papers have been publish ed


Ydration proced ures for biological specimen s, including Goldblith et
7S), Nermut (1 977), Umrath ( 1983) and McKenzie ( 1965).

243
Freeze-drying 245
244 Robards and Sleytr Low temperature me/hods in biological elec/ron microscopy
These drying artefacts cannot be av "d d
is air, dry air, cool air hot . o1 e , whether the surrounding phase
As illustrated in the phase diagram for water (Fig. 5.1), there are three , air or any other g· .
(Boyde 1980). The shrinkage wh· h . as or combmation of gases
principal physical routes by which specimen water can be removed. Air-dry- . d ic is seen as a ma .
during ehydration with solv t . . croscop1c phenomenon
ing (Route 1) involves passing across the boundary between the liquid and . en s is nothmg com d .
force which is exerted at the . I . pare with the enormous
vapour phases. In Route 2, the freeze-drying procedure, the specimen water . . . air water mterface durin d .
s10n, which 1s 73 dynes cm- 1( . x _2N m - 1) . t iso rymg
g c ,. by surface ten-
is first converted into the solid state by freezing (crossing the boundary line 7 3 10
son
. (1951) f estimated that the surf . a
ace tension forces du · 1or
. water.
d . Ander-
between liquid and solid states) and is subsequently sublimed (traversing the
c1mens rom water are of the o d f rmg air- rymg of spe-
boundary between solid and vapour phases) by reducing the partial pressure . r er o one ton per sq . h
Some llllprovement is obtained b d . uare me (160 kg cm-2).
of water over the specimen below the saturation vapour pressure of the ice surface tension (e.g. diethyl eth y lry7mg from solvents which2 have a1 lower
in the specimen. As can be seen from the diagram, the critical point-drying er, dynes cm - 1(I 7 I
tone,24 dynescm- t(2.4xI0 -2 Nm-t) In .. · x 0.- Nm- );ace-
route for water (Route 3) is of no practical interest since the specimen would these organic solvents (see Tabl ). additl?n, the high volatility of
be pressure cooked (critical point of water: 374°C (647 K) and 2 18 atm . e 5· 1a and b and F 1g 5 14 ·
the 1tme taken for the air/liqui·d · t ,. · · m §5.4.1) reduces
(2.21 x 107 Pa)) under these conditions. Consequently, for critical point-dry- . . m er1ace to pass th h
resulting m less shrinkage · The s t ructural pre roug the specimen ' thus
ing procedures, water must first be replaced by a solvent (e.g. liquid co, ·
improved by chemical fixation b ,. servat10n of specimens can be
or a halocarbon) whose critical point lies within a Jess damaging range. - e1ore exchang· th
lower surface tension Howe ,. mg e water for a liquid of
Studies comparing air-drying with other methods for removing water . ver, ior most s .
forces at the air/liquid interface pec1mens the surface tension
have convincingly demonstrated that air-drying is of very limited value
(Boyde 1974a; Humphreys 1975; Nermut 1977; Boyde and Franc 1981).
~5.6) will still be unacceptable (Boyde 1980).
· . are so great that th I·
e resu tmg artefacts (see
With the possible exception of very resistant and rigid specimens (e.g. in-
organic mineralised tissue, mature wood, pollen grains, exoskeletons of in-
sects) most native (unfixed) specimens shrink severely during air-drying.
.. TABLE 5.la
Bo1lmg point and vapour pressure of liq uids used for air-drying
Pressure (Pax 1.33)
6 Boiling point• Vapour pressureb
10 2 10 Liquid Formula

-~.~.~.- - ---~3
:
650 oc K mm Hg(Torr) bar Pa
? ••• - - -
~
~
200 I~ ' Liquid
D"tilled water
th~l alcohol
1iloroform
H20
C2H 5 0H
C HCl 3
100
78 .4
61.4
373
35 1. 6
334.6
17.5
44.0
160.5
0.0233
0.0585
0.2133
2.33
5.87
2.14
x
x
x
1oi
1oi
l04
e~ I •
~
(CH3)iCHCH2CH1 27.9 30 !.0 432.4 0.5746 5.76 x 10•
I
250 ~ 446.9 0.5939 5.96 x 10•
eGI .... CHiCH20CH 2CH3 34.5 307.7
.... C C l3F
CC IFi -~~:~ 296 9
·
64
1.3 0.8523 8.55 x 10'
' 191.8 24624.0 32.7240 3.28 x 106
1
10 llaUons of boili ng pomts
· for som e liquids va r
Pressu re (bar)
~Ille significa nce in the present context a n:y from .source to so~rce. Small variations are
rounded to 0. 1oc Boili . . ' or this reason, bodmg point citations h ·
F ig. 5. I . The pressure-temperature phase d iagram for water. This shows the different po~" apou · ng pomts are given at 760
r pressures cited at 2ooc (
293
K). mm pressure ( 1.0 1 x I oi Pa).
ave

ways of moving from the liquid phase to the vapour phase. Route I is air-drying; Route:
11~ dunng air-d rymg
· improve
· . incr ·
lo halocarbon 13). Adapted and expan~a~~g vapour pressure of t he liquid (from chloro-
with
freeze-drying; and Route 3 is critical point-drying. The dashed line shows the equivalent P1
e rom K lem and Stockem (1976).
diagram for carbon d ioxide, including the position o f the critical point, superimposed on
for water. (See also F ig. 1.1 .)
I
246 Robards and Sleytr LolV temperature methods in biological electron m icroscopy
Freeze-drying
247
TABLES.lb
Boiling and melting points of some liquids
5.2 Basic principles offreeze-drying

Liquid Formula Boiling point Melting poin1 An enormous stimulus for technical developments and the analysis of para-
meters involved in the freeze-drying process has come from the pharmaceu-
oc K cc K
tical and food industries (for reviews see Goldblith et al. 1975; Mellor 1978).
Carbo n d ioxide C02 (Sublimes) -78.3 194.9 Boyde (I 978a) has defined freeze-drying for electron microscopy as a pro-
Distilled water H20 100 373.2 0 273 cedure for getting rid of water by subliming ice into water vapour, using
Ethyl alcohol C 2H 50H 78.4 351.6 -114.7 158.5 a fixed or unfixed freshly frozen specimen. With fixed specimens, water may
Chloroform CHC!i 6 1.4 334.6 - 6.5 266.7 have been replaced with another solvent which is itself frozen and sublimed.
Isopentane (CH 3) 2CHCH2CH, 27.9 301.0 - 159.9 j 13.3
As shown in Fig. 5.2, the term 'freeze-drying' includes a large number of
Methanol CH 30H 64.7 337.9 -97.9 175.J
DMSO CH 30SCH, 192 465.2 18.4 29 1.6
variations of the basic procedure (wh ich is indicated by the heavy lines). Af-
Acetone C 3H 6 0 56. 1 329.3 -96. 5 176.7 ter drying, the specimen may be examined using a variety of methods,
Diethyl ether CH 3CH 20CH2CH, 34.5 307.7 - 116.2 157.0 including direct viewing in an SEM, CTEM or STEM; thin sectioning after
Halocarbon 11 CC 13 F 23.8 296.9 - j j 1.0 162.2 embeddi ng; X-ray microanalysis; LAMMA (§5.5.2); autoradiography or
Halocarbon 13 CCl F, -8 1.4 191.8 -181.0 92.2 histochemistry.
Diethylamine C 4 H 11N 55.5 328.7 -48 225.2
Benzene C6H6 80.1 353.3 5.6 278.8
Dimcthylformamide C 3H10N 70.5 343.7 -61 212 2 Pretreatment :
G lycerol HOCH 2CH(OH)CH,OH 290 563.2 20 293.2 e.g . trlll'lmin g, surface - pretreatment Cwash•ng),
fixation, cryoprotectlo n (non-pol ar solvents) ,
Amylacetate CH 3 COi{CH2).CH, 149.3 422.5 -70.8 202.4 negaUve staining
Isoamylacetate CH,, COOCH 2CH2 CH(CH2h 140 4 13.2 -78.5 194.7
Halocarbo n 113 CC 12 F-CC l,F 48 321.2 -36.4 236.~
Camphene C10H10 160 433 .2 45.0 318.2
Tertiary butanol (CH,)3COH 82.2 355.4 11.0 284 2
X-ra y microanalysis
LAM MA
. . . uids vary from source to source. For this reason. citation•
Temperature c1tat1ons for some hq . . . 'ficance for their use in freeze-dr)m~ Autorad iog r aph y
. 1 0 0 r 0 1oc. This has 11tt1e s1gm 1 . H istochem istr y
are restncted
freeze-substitution and related processes. m mg porn ts and melting points are cited at 760 mm
here to · · B T .
( see also Ch. 4)
pressure (1.01 x !05 Pa).

Fig. 5.2. Flow diagram of freeze-drying procedures.


t ce tension the specimen cannot be d ried dir
To avoid the effects of sur a , f Fig 5 l freeze-drying o
ectly from the liquid phase. A~ can be seen rom al~er·n~tive procedur~
critical point-drying (from smta~e solv~nt:)d a~~ a series of very com inc
Boyde a nd his collaborators have emon~ ~a el , . t dried specimens sho 52 l Theory of drying by sublimation
h t f en dried or cnt1ca pom -
ing experiments, t a roz. - . . I (Bo de 1980; Boyde and Fra
much less shrinkage than a1r-dned matena y r freeze-drying both h) f 2 la Rate ofsublimation of ice
198 l). T his Chapter describes in deta1~ metho~s fo ous volatile
rated specimens and specimens substituted with non-aque
ublirnation and condensati on phenomena can be understood in terms of
utions.
lhc kinetic gas theory (see also §6.5. J). The absolute rate of sublimation of
~
TABLE 5.2
~
Freeze-drying limes a t d ifferent temperature
.,"'c "
Vapour pressure Etching rate Drying time Drying time per µm a.
Tempera ture
~
(sµm- 1) y M D h m 'i'2
(nm s- 1)
°C K Torr Pa ~
::;

I"""
0.68 c
1.08 x 100 1.48 x 103 6.76 x 10- 1 ~
-60 2 13 8.08 x 10-J

-70 203 1.94 x 10- J 2.59 x 10 - 1 3.64 x 102 2.75 x 100


3.0

13
l.,;:
5.36 x 10 - 2 7.70 x 101 l.30 x 101 ~
- 80 193 4.02 x 10- •
~
4.53 )( 10- 2 6.54 x 10 1 1. 53 x 10 1
15 s.
c
- 81 192 3.40 x 10 - 4 18
5.54 x I O' 1.8 1 x 10 1 ~
-82 19 I 2.87 x 10- • 3.83 x 10- 2 21 :;·
3. 23 x 10 - 2 4.68 x 10 1 2.1 4 x 10 1
["'"
-83 190 2.42 x 10- 4 25
2. 72 x I0 - 2 3.95 x IO' 2.53 x 10 1
-84 189 2.04 x 10 - · 30 ~
2.29 x 10- 2 3.33 x 10 1 3.00 x 10 1
- 85 188 1.72 x 10 - 4
101 3.57 x 10 1
38 [
187 1.44 x Io-• 1.92 x l 0- 2 2.80 x
,...
- 86
1.60 x 10 - 2 2.35 x 10 1 4.25 x 101 tl 43 "'..,
- 87 186 l.20 x l0- 4 51
185 1.0 1 x 10 - • 1.35 x 10 - 2 1.97 x 101 5.08 x 10 1 ~
- 88 I
-89 184 8.40 x 10-s l.1 2 x 10- 2 1.65 x IO' 6.07 x 10 1 :s"'
9.32 x 10- 1 1.37 x 10 1 7.82 x 101
13 q·
- 90 183 6.99 x 10-s
~c
27
- 91 182 5.81 x 10-s 7.75 x 10-) 1.14 x 10 1 8.74 x 101 ~
1.08 x 102 45
- 92 18 1 4 .8 1 x 10-• 6.4 1 x 10 - J 9.51 x 100
1110 ,.9!1 • 10 ' 5 3 1 )( 10 7.89 x 100 1.27 x 102 2 7
93 J
94 \19 '2lJ11elO 4 ,9 "' 10 ' I'. S.' · 1O" I 5J >< 10' 2 33

9 17 K ::? 7 1 • Ill ) 6 1 • 10 ' 5.39 )( I()'' U !5 >< 10- 3 5


96 177 2.2 3 x 10 • 2.97 x 10 J 4.45 x 10• 2.25 x 102 3 45
- 97 176 l.83 x I 0 - 5 2.44 x 10 - 3 3.66 x 100 2.74 x 102 4 34
- 98 175 1.49 x 10-s 1.99 x 10 -l 3.00 x 100 3.33 x 102 5 33
-99 174 I 22 x 10 ' 1.63 x 10 ) 2.45 x 10° 4.07 x 10 2
-100
6 47
173 9.93 x t0- 6 1.32 x 10-l 2.00 x 100 4.99 x 10' 8 19
- IOI 172 8.07 x 10-6 1.08 x 10- 3 1.63 x 100 6.1 2 x 102 10 12
- 102 171 5.54 x 10- 6 8.72 x 10- 4 1.33 x 100 7.53 x 10 2 12 33
-1 03 170 5.29 x 10-6 7.05 x 10- • 1.08 x 100 9. 29 x 10 2 15 29
- 104 169 4.26 x 10- 6 5.68 x 10 - • 8.70 x 10-1 1.1 5 x 103 19 20
- 105 168 3.43 x 10 -6 4. 57x 10 - • 7.02 x 10- 1 1.42 x !Ol 23 40
- 106 167 2.75 x 10-6 3. 67 x 10- • 5.65 x 10- 1 1.77 x 101 29 30
- 107 166 2.20 x 10-6 2.93 x 10- • 4. 53 x 10 - 1 2.21 x 103
-1 08 165 1.76 x t0 - 6
36 50 ~
2.35 x 10-• 3.63x 10 - 1
- 109 164 l .40 x I0 -6
2.76 x 101 46 0 ~
1.87 )( 10 - 4 2.89 x 10 - 1 3.45 x 103 57 '1'
-11 0 30 ~
163 1. I I x I 0-6 1.48 x 10- • 2.30 x 10- 1 4. 34 x JOl 12 20 ':::.
~
- 11 5 158 3.34 x 10 1 4.45 x 10 -s 7.04 x 10- 2 1.42 x 104 3 56 40
- 120 153 9.3 1 X \O-R l.24 x 10 - s 1.99 x 10-2 5.02 x 104 13 56 40
- 130 143 5.54 x Io - ~ 7.39 x 10- 1 1.22 x I 0 - 3 8. 17 x 10 5 9 10 56
- 140 133 2.16 x 10- 10 2.88 x t0- 8 4.95 x 10 - 3 2.02 x 101 7 23 19

- 150 123 5.01 x 10- 12 6.68 x 10- 10 1.1 9 x 10 - • 8.40 x 101 27 2 5


- 160 113 6.02 x 10 - 1• 8.03 x 10 - 12 1.49 x 10 - s 6.7 1 x 1010 2157 3 10

Data ta ken from U m rath ( 1983).


"'
""
"'
250 Robards and S leytr Low temperature me/hods in biological eleclron microscopy
Freeze-drying
251
ice (J,) can be calculated according to the following equation (Knudsen
Dryi ng timea _
equation):

J,=qP,(1~r) o.s O<q < l (5.1)


100

where 3. ftchlng rate (nm a· •)

~
- 150
=coefficient of evaporation
P, =saturation vapour pressure of ice
M =molecular weight of water vapour (18.016) ,! 150 "
~
rp =gas constant (=Boltzmann's constant x Loschmidt's number; note
• -100 e..
that, although the gas constant is often expressed as R, to ensure
•..
E ...iiiE
consistency throughout this book it is cited here as rp)
...• 2 . Subllmallon rate (g cm·•a-1) 200
..
I-
-etching rate (cm s-1)
T =absolute temperature -50

J, =absolute rate of sublimation of ice (in g cm - 2 s - 1). 4. Drying time p er mm (•)


250
In deriving this equation two assumptions have been made: (i) that there 0
10~-~3~-=-:,0~-~s~---:,~-~1~~~~~..L...~~..L...~_J
1 0~,~~~10~-~1 ~-;;
is a dynamic equilibrium between the surface of the ice and its vapour, and 0 10·9 10·11 10· 13 10-15

(ii) that the number of water molecules escaping depends on the temperature Vapour pressure
- Sublimation rate
of the ice, whereas the number impinging on the ice surface and not escaping {
Etching rate

depends on the pressure and temperature of the vapour. The actual sublima- Fig. 5.3. T he relationship between· (I) sat uraf
mation rate (J g cm -2 s - •) Th t h.. JOn vapour pressure (P., mbar); and , (2) subli-
tion rate cannot be calculated but must be measured under vacuum condi- " . e e c mgrate(J /p cm - 1) · b · · · ·
11on rate by the density of water/ice (p g _ 3)'· . ' . s •s.o tamed by d1v1dmg the sublima-
tions. Several attempts have been made to verify the equation for the abso- , cm • it •s approximately equal n · II
sublimation rate a nd is a lso shown (J) e d • umenea y, to the
xpresse as nm s- • Thed · d d · . .
lute rate of evaporation and to determine the exact value of the coefficient (4) and s pm -1 (S) are indicated with vertical bars to sh~w enve .rymg times ms mm _,
of evaporation (17), which ideally should have a value of unity (Strasser et (min), hours (h), days (D) weeks (W) h ( longer penods of time: minutes
. , , mont s M) and years (Y). (Redrawn from Umrath
al. 1966; Davy and Branton 1970; Mellor 1978). The theoretical maxim um 1983.)
for the sublimation rate and also the observed rates derived from micro-
(nm s- 1 - 10- 7 cm s- 1) d h ·
balance techniques (Davy 1971; Davy and Branton 1970) are shown in Fig th k - ' an t e time required for subliming a layer l mm
6.32 in §6.5. l. As can be seen, at temperatures above about - 85°C (188 K). 5.~~ (see Table 5.2). These results are summarised in Fig. 5.3 and Table
the observed evaporation rate begins to fall below the theoretical rate. Val
ues less than unity for 'I at higher working temperatures have been
explained by a rate-limiting step prior to desorption in the evaporation pro· 5.2. Jb Rate offreeze-drying of biological specimens
cess and a decreasing mean free path (less than the dimensions of the
vacuum space). Despite these difficulties, calculations of the maximum In p.ractice, drying times close to the values calculated for maximum subl.
ma11on rates of pu t 1-
theoretical sublimation rates using the Knudsen equation are of great prac- cal s ecim re w~ er can only be expected for certain kinds of biologi-
tical value. Based on 'I= l and using the analytical expression for the saiu- mosi biol e~s s~ch a~ dilute suspensions of macromolecules (see §5.5. I). For
ration vapour pressure of water given by Washburn (1924), Umrath (J983l sues) th og1ca specimens, such as isolated cells or bulk samples (e.g. tis-
' e actual rate of freeze-dry · h . .
calculated for the temperature range from - I to -160°C (272. 1- 113. I Kl !he the · mg can s ow dramatic deviations from
the saturation vapour pressure and the corresponding sublimation rare processo:~t:c~l~alc~lations for pure ice. In frozen biological specimens the
(g cm- 2 s- 1), the decrease of the ice layer in relation to time of sublimalion which sub u imlah.on .le~ds to an increasing surface layer of dry mat~rial
sequent Y mh1b1ts further free sublimation.
Freeze-drying 253
hodi in biological electron micro.icopy
Robards and Sleytr Low temperature me/
252
vacuum conditions; and the presence, geometry and orientation of condens-
d th t for the water being sublimed at the ice sur- er(s) in the system. With the small specimens used for electron microscopy,
It must be remembere : ~cimen surface), an amount of energy quanti- the net rate of heat flux(es) generally requires that the specimen is cooled
face (and transported ~o the p h t f evaporation will be required. Thus,
. ·valent to its latent ea o during the freeze-drying process to avoid a (critical) temperature increase
tat1ve1Y eqm . . . b th mass transfer (of H 20 m o1ecu- resulting from radiant heat. With large specimens, or specimens which are
. . operation mvo1vmg o
freeze-drymg ts an f d . depends on the magnitude of resis-
t t nsfer The rate o rymg . . . protected by cold shrouds (where the radiant heat load is very small), the
Jes) and h ea ra · . . of these processes mcludmg res1s-
~ A schematic view ' heat transfer itself is the sublimation rate-controlling step when it takes
tance to these trans ers. . . . p· 5 4 There are two main
d heat transfer ts given m ig. . . place th rough a layer of dry material external to the frozen zone.
tance to mass an ' .. r duri'ng freeze-drying. Heat transfer The large differences in size, shape and structure of biological specimens
. . f s and heat trans1e
combmat1ons o mas h h the same path (dry layer), but in oppo- do not allow the provision of general rules for the selection of freeze-drying
and mass transfer may pass; roug t ke place through the frozen layer and
. d. · or heat trans1er may a . . procedures. It has been shown, however, that the drying time for a given
site uections, d 1 practice all possible combmat1ons specimen under optimal vacuum conditions is temperature-dependent (see
1
mass transfer through the ry ayer. nnd the a~tual transfers of heat and
h t t ansfer can occur, a Table 5 2), and so the shortest drying time is achieved if the specimen is kept
of mass and ea r . . t ucture and shape· the nature of the iust below the temperature at which disturbing recrystallisation of ice may
d the specimen size, s r '
mass depen on . .h h ort (temperature-controlled plate); the
contact of the specimen wit t e supp occur.
As discussed in Chapter 2, for most biological specimens with an average
water content, recrystallisation phenomena can be expected at temperatures
above - 80°C (193 K), but since recrystallisation is also a time-dependent
phenomenon, no detectable damage may occur during short drying periods
at higher temperatures. Frequently, only the outer few tens of micrometres
of the well-frozen surface layer of a bulk specimen will be examined (see
§S.5.2a). Co nsequently, it saves time to start by drying the surface layer at
a temperature which gives no risk of recrystallisation (e.g. -90°C; 183 K)
and then to fi nish by drying the inadequately frozen central part of the spe-
omen at a much higher temperature (e.g. -60°C; 213 K). As can be seen
from Table 5.2, specimens frozen as a thin layer can be dried within a few
minutes (e.g Table 5.2 shows that, at -95°C, IO µm of pure ice will sublime
t bl• (temperature-controlled)~~ within approximately 30 min) at a 'recrystallisation-safe' temperature (e.g.
~~~~""~"~"~~"~"~""""""""~ o Dry part ol epecimen
95°C; 178 K ). (For a discussion of the theoretical advantages of freeze-
drying excl usively at low temperature, see also MacKenzie (1965). It is con-
~Direction of man <H20> transfer
Sldered important that some bound water remaining in the tissue has a sta-
~ Direction o f heat tran•f•r §I Froxen put ol apecimen
hsing effect on protein structure. Furthermore, since gases are dissolved
__ _ __ _ - Drying bound UY d naer cellular water, some freeze-drying artefacts may be generated by tempera-
b ndary and th• con •
..it-- Reaietanca ol apace between drying ou ure- and pressure-dependent expansions of these gases.) Umrath (1983)
tor ma•• traneter .
c• and th• drying cs a survey of published results for a great variety of specimens, including
. . . . . . - Raaiatanc• of apace between th• heat eour
boundary tor heat transfer data on drying temperatures, specimen dimensions and vacuum conditions.
. rt s of combi nation of mass and Many of these results are difficult to compare since in most experiments no
Fig. 5.4. Schematic representation of the two maJO c ype . ta ces are shown for (leftl a
. Th heat a nd mass trans1er res1s n rat ICCurate temperature measurement of the drying specimen was made. Thus,
transfer during freeze- d rymg. e . · od contact with the tempe
d ( · ht) a specimen m go possibility of considerable differences between the temperatures of the
cimen with poor thermal contact, an ng . t f om Goldblith et al. 1975.)
controlled support. (Redrawn m par r
254 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-drying
255
specim en and of the specimen stage, and/or specimen support h as to be kept
in mind.
Knowing the maximum sublimation rates for pure water, it becomes
obvious that a I to 2 mm diameter tissue b lock cannot be frozen-dried +2o·c ~
within realistic times at a temperature below - 100°C (173 K). Nevertheless o·c z:!
such conditions have been described in the literature. This indicates that
-ao ·c e

Q.

there was either a poor contact between the specimen stage and the speci- I
I ....•
I
men or that the actual drying process took place during the, often slow , I
:t-93K -1ao·c
warming up period at the end of the freeze-drying schedule.
From this discussion we can conclude that both the recrystallisatio n tem- Fig. 5.5. Sublimation rate of freeze-dry· (b lk .
. mg u specimens) as a ~ · ·
perature for the specimen ice (see Chapter 2) and the maximum sublimation hne shows sublimation rate and the d 0 tt d . . unctwn of time. The solid
. . . e 1me shows specimen tc . .
rate for pure water (at the specimen temperature) are the basis for p lanning zone within which the specimen ini·i·· 11 mperature. A indicates the
ia Y warms to the fr d ·
freeze-drying experiments. The ratio between the time required for the com- drying occurs by the sublimation of ice· while i eeze- _rymg temperature; in zone B
,. n zone C d rying is by desorption. (After Gold-
p lete freeze-drying o f a defined specimen layer and the sublimation of an bhth et al. 1975.)
identical layer of ice can best be characterised by the so called 'prolongation
factor' (f), where:
frozen specimen (wh ich has been stor d 1
liquid nitrogen) is warmed up to the ti e adt a_ o w temperature; e.g. under
Drying time for x µm specimen layer reeze- rymg tempe t h .
tion ra te rises rapidly to a . ra ure, t e subhrna-
.f Sublimation time for x µm ice layer ' f ~1 . maximum value. In circumst·
zen specunen is surrounded by a layer of ic th .
h
ances w ere the fro-
maxim um theoretical sublimation rat . e, . is value may be close to the
By comparing different results from the literature, one encounters enor- front reced es from the surface to t h ~ giv~n m Table 5.2. As the drying
e mterror of the s · ( .
mous differences in recommended schedules for optimum drying (Umrath the drying rate generally decreases rapidly This . d pecm~en penod B),
1983). The course of freeze-drying under high vacuum conditi ons can be fol- water vapo ur and heat flow . th . : is ue to resistance to both
m e mcreasmg area b t h .
lowed if the water is exchanged against deuterium oxide (D20) by mass mation and the free specimen surface . Finall . e we_en t e site of subli-
spectroscopic analysis (Wildhaber et al. 1982). From accurate measure- 1ce crystals duri ng the third period (C) th y, with th_e disappea rance of the
ments on model systems (Nietsch and Jochem 1973) and tissue samples more rapidly and finally ap h , e evaporat10n rate decreases still
proac es zero Durin th· ,
(Stumpf and Roth 1967) it appears that the sublimation rate decreases expo- drying phase the · . · g is so-ca1led seconda ry'
. ' specimen is gradually warmed u .
nentially with the thickness of the dry layer on the specimen su rface. Thus ambient tem pera ture and th .d . . P to approximately
1972). 1
e res1 ua mo isture is desorbed (MacKenzie
fo r practical reasons, bulk specimens must be trimmed as small as possible
to keep the resi stances to heat and mass transfers low. R ealistic prolonga· At this point it is necess d. . .
tion factors for the drying of I mm tissue cubes can be expected to be in which fo rms on que h. arhy to 1st1~gmsb between the freezeable water
nc mg t e water m the · . '
the range 20-100 (Umrath 1983). Since heat and mass transfer resistances called (chemically or physica l! ) 'b d' specimen matnx, and the so-
are dependent on the specimen structure, it is understandable that, for d1f made between these t r Y oun water. No sharp distinction can be
wo iorms a nd the val · r
ferent but equ ally sized tissue samples, different drying times are required considerably depend· h ues given ior bound water vary
(Stephenson 1960; Pearse 1964; Stumpf and Roth 1967; Schunzel 1968). A1 Podolsky a nd K o tmg_on t e method of determination (Meryman 1966-
ns antmov 1980) Jn . . '
an example, Stumpf and Roth (1967) found the relationships of drying tunes toater comprises a bout 5- 10% of h. an average b10log1cal tissue, bound
ofliver:kidney:brain, to be 1:1.7:3.3. ID terms of g wa ter pe 100 o t_ e total wat~r con tent and is characterised
The sublimation rates during freeze-drying of bulk specimens generally lhat the amount f r g spe~1r:ien _dry weight. Muller ( 1957) has shown
o water remammg m th ·
follow the c urves shown in Fig. 5.5. During the first period (A) when die rtcze-drying depends h . e specimen after completion of
on t e freeze-drying temperature (F ig. 5.6) and the
Low temperature methods in biological electron microscopy Freeze-drying 257
256 Robards and Sleytr

of 100 (and preferably 1000) lower than the saturation vapour pressure for
ice at the specimen temperature (see Table 5.2 and F ig. 5.3). With a higher
water vapour pressure in the specimen area, the actual sublimation rate (J.)
is determined by the equation:

(5.2)

where
50 60 70 80 J. =actual sublimation rate
-2o·c Drying time (h) r; =coefficient of evaporation
. · - (o; d y weight) after freeze-drying at different lemper- P, =saturation vapo ur pressure of ice
F 5 6 Residual water in plant tissue 1 0 r ..
ig. - · . _ th h" her the residual water. (Redrawn from Muller Pc =partial pressure of water in the specimen area
atures. The lower the drying temperature, e ig
1957.) M = molecular weight of the water vapour
cp = gas constant (see Eq. 5. I in §5.2. la)
, t h "ch the secondary drying process is completed. To avoid T =absolute temperature.
temperature a w i . · · b
. . d h. ve maximum drymg, the specimen is est
contammation an to ac ie
c: d above ambient temperature before the vacuum This equation shows that the maximum sublimation rate considered earlier
warmed up to a iew egrees . . r
. b . d . g dry inert gas (e.g. Ni). Alternatively, a vanety o (§5.2. I a) resembles the special case where P 0 is much smaller than P,. The
is broken Y mtro ucm . · · ·h ·h
additional preparation steps, such as infiltrat10n wit~ resins (wit ~r w.1t - relationship between Pc and P, and the consequences for the freeze-drying
out revious fixation) , sputter-coatin-g, and shadowing and/or. rep_ 1cat1on. procedure are summarized in Fig. 5.7 (Umrath 1983).
Pb f rmed immediately after drying while the specimen is stlll ~nder Depending on the specimen temperature, other condensable gases in the
can e per o · d d ymg 1s
With some specimens exhaustive secon ary r system (e.g. C0 2, hydrocarbons) could theoretically mask the specimen sur-
vacuum (see .§5 .5·2) ·
accompanied by collapse phenomena (see §5.6). face and reduce the sublimation rate of water. However, with normal pump-
ing systems, such as oil diffusion pumps or turbo-molecular pumps in com-
bination with rotary pumps (see §6.3.2), even without cold traps, water
5 .3 Practical aspects offreeze-drying \apour is the main constituent in the residual gas (Table 5.3). Today most
~ommercial freeze-drying equipment includes cold baffles or shrouds which
can be cooled either by liquid nitrogen or by refrigerator cryopumps. Such
5.3.1 Basic principles
cold traps are kept at considerably lower temperatures than the drying spe-
· during the freeze-drying pro- amen (usually at - 196°C; 77 K) and, consequently, they dramatically re-
The water vapour released from t h e specimen . be effected
cess must be continuously removed from the syste~. This chan . al or me duce the partial pressure of condensable gases (see also §6.5.2). In such sys-
6 3 2) by adsorption to c em1c tems, the required factor of 102 to 103 difference (see Fig. 5.7) between the
by vacuum pumping systems (see § · · ' t consider
chanical desiccants, by condensation on a surface ~~ a ~~mp:~aal~:~ese meth· values for the water saturation vapour pressure at the specimen temperature
ably lower than that of the specimen, or by a com ma wn and the partial pressure of water vapour (and other gases) in the specimen
area is easily maintained. This is illustrated by reference to Table 5.2. For
ods. . . t for the specim
In a freeze-drying unit the maximum sublimation ra ell the conciensable Uample, the figures show that the water saturation vapour pressure of ice
water can only be obtained when the partial pressures of a t I ast a fact -80' C (193 K ) (which is a practical drying temperature for a specimen)
gases (especially water vapour) in the specimen area are kept a e 4.02x JQ- 4 Torr (5.36 x 10- 2 Pa). To provide the required factor of
258 Robards and Sleytr Low 1emperature methods in biological electron microscopy
Freeze-drying
259
3
Condensation -• ---+----..- Sublimation J0 difference the condenser must have a temperature of about - l l 5°C
(158 K), which
7
corresponds to a saturation vapour pressure for ice of
3.34 x 10- Torr (4.45 x 10- s Pa). This clearly shows that cold traps do not
necessarily have to be maintained at a much lower temperature than the
specimen to ensure an efficient freeze-drying process. For practical reasons,
however, cold traps are usually cooled down to 77 K with liquid nitrogen.
To allow the maximum access of vapour molecules to the cold trap (or
other p umping system), the pressure in the freeze-drying apparatus should
Non
inhibite d Inhibited Inhibited
be such that the mean free path of the residual gases is greater than the dis-
c ondensation condensation sublimation tance from the specimen to the cold trap (see Fig. 5.8). In practice, however,
the actual freeze-drying rate is more likely to be determined by the diffusion
resistance for water molecules within the increasing dry layer of the speci-
men than by the rate of water vapour transport from the specimen surface
_J----~---~,---1~0~0,------::1~0~-1,--~10-2
10
3
10
2
10
to the cold trap or vacuum pump (Stephenson 1960). Nevertheless, special

Pressure (Pax 1.33)


4 2
10 10 10° 10- 2 10-•
F . bet the parlla
. I pressure of water in the specimen area (P,) and
ig 5 7 Relationship
· ·· ween r:
(P) ences ''r r the freeze-drying p rocedure
the saturation vapour pressure o ice • f theUmrath
and consequ1983 ) 0
(Redrawn rom ·

TABLE 5.3
Composition of residual gases at I x 10- sTorr (1.3 x 10- ' Pa) MFP at 1 Torr
N2 = 0.065 mm
Gas Proportion Partial Components ( Molecular weights ) H20=0.034 mm
oftolal pressure 10- 5 '----1-~-'---...Jc..._~-1-~--L.J
2
(%) (Torr) 10 10-2 10-• 10- 6
Pr••sure (Torr)

H 20 70 7.0 x I0- 6 H,O + (18),0H • (l 7),0 2 ( 16),H 2(2) Fig 5.8. Mean free path (MFP) versus vapour pressure for nitrogen and water vapour (a t
CO, Ni 12 1.2 x J0- 6 CO+(28),N 2+ (28) C. 273 K). The mean free pa th (A.) is calculated using the eq uation:
C0 2 12 1.2 x 10- 6 ·co2 + (44)
H 2 +others 6 0.2 x 10- 6 H 2+ (2), C H-groups, noble gases

100 x I0- 5
Total IO
. . cuum unit at a total pressure of I _x
The partial pressures of the res1d~al gases m a va ump in combination with an oil d11Tus:i : mean free path (the average distance that a molecule will travel before striking another
molecule
Torr (1.33 x J0- 3Pa) after pumping with a rotary p) f room atmosphere (ions meusu
1 1 p (without a cold trap rom d : diameter of molecule
p ump or turbomo ecu a r pum t ) (From U mralh 1977.)
by mass speclrome ry · • : number of molecules per cm 3•

(Redrawn from Meryman 1966b.)


260 Robards and Sleylr Low 1emperature me/hods in biological elec/ron microscopy

Freeze-drying
attention must be paid to the function and design of cold (condenser) sur- 261
faces in freeze-drying un its. It is best to place the condenser system as close
as possible to the specimen, or even to have the specimen surrounded by
an 'optically tight' cold shroud, as provided in some freeze-fracture replica-
tion units (see §6.3.3b). As discussed in more detail in §6.3.3b, this type of
design has the distinct advantage that, in addition to their function as a
cryopump, the cold surfaces act as a very efficient anticontaminator and as
a total shield for the specimen against radiation heating from warmer sur-
faces in the system.
Instead of (or in addition to) cryopumps, chemical or mechanical desic- Dried
- - --
cants can be used to remove the water vapour released by the specimen.
Specimen
With chemical desiccants the water molecules undergo a chemical reaction.
The most common ones are phosphorus pentoxide, calcium chloride, cop-
per sulphate and sulphuric acid. They have the great advantage that within
the gas in contact with them , even at room temperature, the partial pressure Support

of water vapour generally approaches very low values (e.g. P 20 5, 1x 10-s


Torr; 1.33 x J0 - 3 Pa). For this reason, they may be used for long secondary
Increasing temperature
drying procedures to remove the last traces of water from the specimens.
Fig. 5.9. Heat input by conductio f .
Mechanical desiccants such as zeolite or the synthetic zeolites known as r.1 h n rom specimen s T
•I e I al can be expected across the d . . . upport. his shows lhe tempera tu
'molecular sieves' can be regenerated by heating. They are freq uently used · rying spcc1m h re pro-
'" the close vicinity of the specimen surface A h e~ w en Ihere is a cold condensing surface
where a higher residual water vapour pressure is tolerable (e.g. drying of men support. the surface will be at the I . s eat input occurs by conduction from the spe .
owest temperature (Red r c1-
an air flow carrying water vapour with it). At very low temperatures (e.g . rawn irom Meryman I 966b.)
96°C; 77 K), molecular sieves act both as a water. vapour trap and a vacuum
pump and can produce total pressures as low as 5 x 10- 11 Torr (6.67 x 10 • are those of the drying boundary (which det .
and of the still frozen part of th . ermmes the sublimation rate)
Pa) (Read 1963). To reach this value, the volume of the molecular sieve and
sure that the temperature is m~ _s~e~1men (which makes it possible to en-
the volume of the vacuum systeni must be in the correct proportions
eutectic temperatures). In pract · amb ained below the recrystallisation and
required by the specification. Freeze-drying without vacuum pumps at all ice, ecause of the 11 .
on IYthe tempera ture of th f . sma size of the specimen
has been performed using desicc_1µ1ts by Stumpf and Roth (1964, 1965). . e rozen and subse I ,
specimen can be measured. ' quent y of the dry, part of the
Irrespective of the freeze-drying system used, it is important to have the
possibility of accurate temperature control of the specimen stage and, ult1- ~3.2 F
mately, of the specimen itself. As mentioned before, poor thermal contact reeze-drying apparatus designs
between the specimen stage, the specimen support and the specimen itscl
Historically th
may lead to a considerable temperature gradient. Thus, it is strongly recom ' e vacuum for freeze d .
then by mechanical rotary pu - rydmg was first generated by water jets
mended that actual temperatures are checked within a given system to en pump . h mps an subsequent! b ,
sure that the specimen temperature is the same as the stage temperature or s wit ei ther oil diffusion o t b y y combining rotary
\'apou r ur o-molecular pu T
at least, th at there is a constant and predictable relationship between the . r more efficiently firstly ch . Id . mps. o remove water
two.
idc sir 1 '
. rca ge ' molecular sieves) and th I
enuca es1ccants (e h
.g. P osphorus pentox-
As can be seen from Fig. 5.9, a temperature profile can be expected acr1* tems (e.g. liquid nitrogen-cooled tr en» ater, low temperature condenser
the drying specimen. It is obvious that the temperatures of greatest intercsl : to use both cold traps and mol apt w~re used . Nowadays it is com-
1mum water vapour trapping ;~u ~ sieves (at low temperature) for
. e Jrst stage-cooling methods used
262 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-drying
263
refrigerator systems or thermoelectric devices (Bendet and Rizk 1974),
The most controlled freeze-d .
whereas current systems often use liquid nitrogen or cold nitrogen gas that rymg procedures c· b b
modern f reeze-etching units (see §6.3 3b) . . an e est performed in
can be further utilised in feeding the cold traps in the apparatus. In future, tely contamination-free envirom t ·. . ~1thm such apparatus, an absolu-
refrigerator cryopumps will undoubtedly replace liquid nitrogen-cooled sys- en 1s attamed by · , .
cold shrouds around the speci
· ·
w· h usmg optically opaque'
men. it such shr d · · ·
tems. With three-stage refrigerators, temperatures down to 163 K (- l 10°C) contammatrng molecules to ou s, It is impossible for
can be achieved. move over long d · t . .
wards the specimen without b . . is ances m a straight line to-
A schematic representation of the two main types of freeze-drying system eing mtercepted by a cold "
sur1ace (see, for
is provided in Fig. 5. 10.
a b
5.3.2a Simple vacuum sys1emsfor freeze -drying

The simplest and generally older freeze-drying systems (Fig. 5. 1Oa) use only
a rotary pump for evacuating the chamber during the drying process. The
specimens are placed onto a pre-cooled, temperature-controlled specimen
stage in the drying chamber. After evacuation, the chamber is sealed off and
the subliming water vapour is either trapped onto a condenser (not illus-
trated), which is kept at a temperature considerably lower than that of the
specimen, or by using a chemical desiccant (e.g. P20 5) positioned between
the specimen and the vacuum pump. Using such a system, the subliming
water vapour is often not very efficiently trapped since both the condenser
and the chemical desiccant tend to be positioned at distances from the speci-
men much greater than the mean free path of the water molecules a ttainable
under tb6 particular conditions of vacuum. A further disadvantage of such
early designs is the possibility of specimen contamination from back·
streaming vacuum pump oils (if no adsorption trap is present).

5.3.2b Modern vacuum systems for freeze-drying B Ad1orp ti on trap (molecular sieve }
St: Specimen stage
· ·;:;..:_-· - -:... D: De siccant (e. g. p )

Most modern desi-gns of freeze-drying unit have more efficient vacuum


-0 Ga outlet filter
C: Condenser
205

pumping systems, usually incorporating, in addition to a rotary pump, an


oil diffusion pump as shown in Fig. 5.1 Ob or a turbo-molecular pump a
~ Rotary pump 12 •l!'ge )

illustrated for the freeze-fracture replication unit in Fig. 6.10 (§6.3.2) Type Val ve: (VC) venting valve for c hamber

JI. Again, as with freeze-fracture replication units (see Fig. 6.10, Type Im • Ji' Ho t ClthOd• i I .
Oil tSiffu11ion pump Cfor oil vapour)

the technology of refrigerated cryopumps can be expected to replace tht


other systems because (working as 'closed systems') they provide the uln-
r11Jf r hermal
on aat1on 9•uge for chamber

conductivity gauge for chamber =t$ - Wate r cooled battle


5 JO. Schem· (
mate convenience. In the most recent designs of freeze-dryer, the condenser ~I ly o a ic representation of the two ma in t s of .
is placed as close as possible to the specimen, thus ensuring optimal trapping IUbJ i:ie f freeze-drying system. The Specime h ypebc . freeze-drying apparatus. (a) The
1ming wat · n c am r 1s pu d
efficiency for water vapour together with additional protection of the spea- dl"ic1ent pum . er is gene~ally trapped using a desiccant. (b) M~pe only by a rotary pump
men against condensation contamination. ng surface~~g) syst_em, including both rota ry pump and oil d"ff em freeze-drying units use
positioned closely above the t I us1on pump, and a cold con-
emperature con tr0 lied .
specunen stage (St).
Freeze-drying 265
thods in biological electron microscopy
264 Robards and Sleytr Low temperature me
Fig. 5.1 1 illustrates the simple module developed by Roberts and Duncan
h effective specimen transfer and air (1981 ) fo r freeze-drying specimens mounted on electron microscope grids.
example, Figs. 6.15-6.17). Furt ermdorehi,. h guarantee contamination-free
h been develope w c It consists of a circular base, which is divided into two parts so that its mass
lock systems ave d . 'd itrogen onto the pre-cooled spe- can be halved if necessary, and a cylindrical top. The top is recessed a few
. ns from un er 1iqm n . .
transport o f specime of condensation contammation is millimetres into the base, to which it can be locked. It can be filJed with lead
33 b) (The phenomenon
cime~ stag~ (§6 · ·. . · [n addition, special specimen supports and shot and has several holes which alJow the free passage of gaseous or liquid
explamed m detail in §6. 8 ·3 .) . . f ze-etching units have been de-
d t for freeze-drymg m ree . nitrogen, and there is a lifting mechanism within the evaporation unit. Heat
transfer ga ge s . . S .th 1980). Since freeze-etchmg proce- transfer between the top and bottom rings is reduced by an insulating ring.
976
. d (e.g. I w ata and Aita 1 bl' ' mi.; n of specimen .ice, the d es1gns
·
scnbe of In use, the cleaned top and bottom parts are separately cooled in liquid ni-
. 1 th ontrolled su imauo
dures mvo ve e c good thermal contact and enable
. orts and stages ensure trogen. The cooled grids with the mounted specimens are then placed in the
most specimen supp d . n temperatures. Radiation heating wells of the base unit before the cold, top section is attached by means of
't ·ng of stage an spec1me
accurate mont on . ·n models where cold shrouds sur·
eglected part1cu1ar1y I . the bayonet fitting. The loaded and assembled freeze-drying module is sub-
can genera11Y b e n ' . . d the frozen-dried specimens can bt sequently transferred from under liquid nitrogen to the vacuum coating unit
· Finally 1f require , .
round the specimen. .' . f tching units without breakmg tht
shadowed and/or coated within reeze-e and installed at a set distance from the evaporation source. The bayonet
clips are disengaged by turning the top section, and the lifting mechanism
vacuum.
1s engaged with the Jug on the lid. During pumping, the unit is allowed to
Simple freeze-drying devices for use in vacuum systems warm up slowly. The insulation ring between the base and top section caus-
5.3.2c
es the lower part, containing the specimens, to warm up more quickly than
. . devices are very simple indeed and can he 1he top. This maintains the required temperature gradient between the
Some versions of freeze-drying . 't used for electron microscop1ca
acuum coating um freeze-drying specimens and the top plate, which acts as a very efficient con-
fitted to almost any v Sk'll. . Morioka et al. 1980; Osatak
· (e.g. Heckly and 1 mg 1979 ' · an denser. From thermocouple measurements, the authors found freeze-drying
preparation ) S ch devices are inexpensive
80 R b ts and Duncan 1981 . u he to be finished when the bottom part is completely defrosted. This process
et al. 19 ; o. er . orksho facilities. Furthermore, they can normally takes about two hours but the time can be considerably varied by
easily made usmg routine w p l'ng systems because they have pr
. d d tly of permanent coo I . h changing the amount of lead shot put into the top and/or by using either
used m epen en . k the specimen cold dunng t e en
cooled metal blocks as heat sinks to eep adouble or single base plate. Finally, the upper section is lifted away using
aSimple crane mechanism to allow shadowing of the dried specimens. The
cal period of drying.
assembly is then left under vacuum overnight while it warms up to room
lanpcrature so that contamination of the specimens by condensation of
water vapour is avoided when the vacuum is broken. To ensure that the
correct physical conditions are fulfilled, is seems advisable to use thermo-
couples to check the temperature of the upper and lower halves during
fleez.c-drying.
In view of the success of the above approach, it is to be expected that
Bayonet clips light modifications to the standard Bullivant type II freeze-fracture replica-
;4;;;::=:;;::;=;::;;:;;~;--1nsulating ring
dcvice (§6.3.3c) - for example, by placing an insulating ring between
ease base plate and the tunnel cylinder and by closing the tunnel holes - could
Yprovide an effective freeze-drying module similar to that suggested by
4 cm loliens and Duncan (1981). In fact, it has already been shown by Severs
. . his module is con11nerc1ally a
collaborators that, even without such modifications, excellent deep
Fig. 5. 1I. Diagram of a simple freeze-drying dev1cbe. ~ d Duncan 1981.)
d . 2) (Redrawn from Ro erts a n
(Appell IX ·
266 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-drying
267
etching (i.e. surface freeze-drying) can be produced using the Bullivant de-
vice (Severs and Hicks 1977; Severs and Warren 1978). Nevertheless, all
freeze-drying gadgets working without continuous refrigeration only give a Venting valve

satisfactory results when the specimen is thin enough to dry before the base
plate reaches the temperature at which recrystallisation phenomena in the
specimen can be expected, or when only the surface layer of a bulk specimen
is to be frozen-dried below the recrystallisation temperature. The latter cri-
terion is adequate when specimen surface structures are to be examined in Frozen embedding
the SEM. Finally, these simple freeze-drying modules can be successfully medium
(monomer drop)
used to dry specimens in which water has been replaced by an organic sol-
Object plate
vent prior to freezing (see §5.4; Osatake et al. 1980).
Heating cartridge l'H~~tff';;:J•>f.i.!4,~;:;j.,LLFrozen cryotlxed object
Molecular sieve Temperature sensor
5.3.2d Simple freeze-drying device for immersion in liquid nitrogen fl':"-:+:-:~~- Immersion vessel
LN 2
An elegant but simple technique for freeze-drying specimens exclusively by
cryosorption pumping has been described by Edelmann (1978b, 1979). As
shown in Fig. 5. l 2a and b, the drying process takes place within a vacuum-
and cryogen-tight sealed container which is immersed in liquid nitrogen
T he actual freeze-drying apparatus (Fig. 5.12b) consists of a metal vessel
b
half filled with molecular sieve (in this instance, Zeolite 13X from Leybold- 1111\'l\\'l:~fr--Screw cover
Heraeus, Cologne, but o ther sources would undoubtedly also be suitable. ~e;.-B•ll b••ring
see Appendix 2). The vessel is sealed by a metal screw cover to which a plug
with five pins is fixed. The socket fo r this plug is fixed to a metal specimen
support with small wells for the frozen specimens. The cover is connected
to a metal tube, approximately 600 mm long, which houses the cables for
Ven el Support for specimens
the electrical supply to the heating element (power transistor) and the tern·
~~m~tij'--Heetlng element
perature monitoring of the specimen support. The upper end of the tube 1s
closed by a ventilation valve (Fig. 5. l2a).
The method used by Edelmann ( I 978b, 1979) for freeze-drying and
embedding chemically unfixed biological specimens is as follows:

(I) The specimen support (pre-cooled in liquid nitrogen) is loaded with


frozen specimens and small drops of evacuated Spurr epoxy resin
(Spurr 1968).
I- i
(2) The specimen support is picked up with the plug fixed to its cover. 30mm

(3) The vessel is immedia tely sealed and transferred into a 30 I container
of liquid nitrogen. ;.12. (a) General view of a simple freeze-d r in . . . .
a1 freeze-drymg takes place exclusive! b c Y g dev.1ce for immersion Jn liquid nitrogen
(4) The temperature of the specimen support - which rises during seali
I apparatus shown in (a) (El t . I y Y ryosorptlon. (b) Detailed view of lhe freeze-
and cooling of the vessel to about 133 K ( - 140°C) - is raised to the a di . . ec nca connections not sh ) (F'
agram kindly provided by H.Sitte. Fi own. ig. 5. I 2a red rawn from
g. 5. l 2b redrawn from Edelma nn I 978b.)
268 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-dry ing
269
chosen freeze-drying temperature. For drying 0.1 mm 3 pieces of mus-
cle tissue, the support was maintained for three days at 193 K (-80°C), Specimen

followed by six days at 213 K (-60°C). The final pressure


obtained with the cooled molecular sieve, together with the cold walls
of the container, has been estimated to be about 130 mPa. Under such
conditions, the actual partial pressure of water and other gases that
Water replaced by (volatile)
are condensable at the specimen temperature must be extremely low. organic solvent
(5) At the end of the freeze-drying process, the vessel is vented with dry
nitrogen by opening the valve at the upper end of the tube and the
support is warmed at a rate of about 10 K h- 1 to 258 K ( - l 5°C), at
which temperature it is maintained overnight At this temperature, the
Spurr resin liquefies and infiltrates the dried tissue.
(6) The support is then warmed further at IO K b- 1 to room temperature Air-drying
Critical Point Freeze-drying
drying Freeze-aubatitutlon
and the freeze-drying apparatus is removed from the liquid nitrogen (organic solvent +fixative)
''
container. I

® I

©
(7) After opening the assembly, the tissue samples are t ransferred into
'
(+Embtddlng) ' aee Ch. 7
fresh embedding medium prior to polymerisation.

This technique, which can be used with or without the embedd ing proce-
dure, has the great advantages of being very simple and of making it possi-
Examina tion by: SEM

'
SEM
TEM

Fig. 5.13. Schematic diagram o utlining di'ffie e t


TEM

.
SEM

r n specimen preparatio
.
I
TEM

h ·
replacemen t of water with an o rganic solve t t . n routes t at m volve the
ble to freeze-dry specimens with low cryogen consumption and at very 1011 . n a some stage d urmg proc .. · R
Air-drying from a n organic solvent· Rout css1ng. o ute I shows
, e 2 represents the nonnal ·t' 1 · .
reproducible temperatures. •ay: Route 3 uses freeze-d rying (possibly after f f . en ica pomt-drymg pa th-
reeze- ractunng) from ·
er lhan water; and Route 4 entails use of th f . . an organic solvent rath-
e reeze-subst.Jtutmn process (Cha pter 7).

5 .4 Freeze-drying from non-aqueous solvents


Critical point-drying (F' 5 13 .
f . . ig. . , route 2) p rovides a much better me
There are three methods of drying electron microscopical specimens which ~ga71d:~g ~he damaging effe_cts of surface tension forces than does air-d:~~
involve a step in which the constituent water is replaced by an organic sol aga.ms~ ane o~:~n~~e~~l:=~~r m a fixed o r unfixed specimen is exchanged
vent with the aim of avoiding or reducing the damaging effects of surfact DMP)) Subseq ti h (e.g'. etha?ol, acetone, 2,2-dimethoxypropane
tension forces (Fig. 5.13, routes 1, 2 and 3). As an additional benefit. !Ct · uen y, t e specimen 1s tra fl d ·
cssel wh · · . ns erre mto a closed pressure
crystal damage during freezing can be greatly reduced if not avoided ere it is first infiltrated and surrounded by liquid CO (Ande
1966) or Freon (C h 2 rson
1
route 3. The freeze-substitution procedure shown as a fourth route in th the vessel are then r~i:e~ etba . 1968). T he temperature and pressure within
diagram facilitates the combination ofrapid freezing of specimens with re5lll point) the I' .d a ove the critical values at which point (the critical
embedding. This is accomplished by using organic solvents to remove cruical poi~;u~ and gas phases have the same density. After passing the
from the specimen at temperatures below that at which specimen ice rec CIF (F (or C02, 72 atm (7.29 x J06 Pa) and 31.Joc (304 3 K)· fi
tallises. This procedure is described in detail in Chapter 7. As with air-drJ- I i reon 13), 38 atm (3.85 x 106Pa) and 28.9oC (301 9 K)) th. , or
sowly red uced so that d . · , e pressure
ing from aqueous systems, all solvents possess a finite surface tension whidl g is avoided. Critic I r~con e.n sa~1on of the fluid through adiabatic cool-
may cause deformation of the sample during fluid loss (Revel and Wolkcl fo r processin a . pomt-dry1~g is one of the most commonly used rout-
1973; Albrecht et al. 1976). it h g specimens (pa rticularly tisss ues) for the SEM Neverth
as artefact pro bl f. · e-
ems o rts own (Boyde and Franc 1981). A detailed
270 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-drying
27!
description of critical point-drying is outside the scope of this book: the a specimen infiltrated with ethanol th .h .. . .
reader is referred to Hayat and Zirkin (1973), Willison and Rowe (1980) and freeze-drying may be used (Fig 5 !3)'.f ~n eit .er cr'.ttcal pomt-drying or
Boyde (1980) fo r further information. in an SEM (Boyde 1980· Hu . h. I t e specimen is then to be observed
, mp reys et al 1974) A d. .
Freeze-drying from volatile organic solvents (Fig. 5.13, route 3) provides §6.3, specimens frozen in the hyd t d · · s iscussed m detail in
the possibility of avoiding both the damaging effects caused by ice crystal ra e state pred · I f
mem bra nes. This cannot necessa .1 b ommat Y racture along
growth during freezing and the surface tension forces during air-dryi ng. . n Y e expected after fi , f · 0
infiltra t10n wi th organic solvents Th f· . . ixa 10n m s04 and
. . · us, re at1vely little t 1 · .
1s found m such preparations becau th f . cy op asm1c detail
5.4.1 Procedure for freeze-drying se e ractunng occurs predominantly
Temperature (K)
73 173
Specimens must be fixed before replacing water with non-physiological or- 3
273 373
10
ganic solvents . Usually the dynamic processes of the living specimen are 10 5
hal ted by glu taraldehyde fixation. To keep structural changes to a mini-
• Melting point
mum, optimal conditions (temperature, pH, osmolarity, ionic composition)
of fixation must be chosen (Glauert 1974; Boyde 1978a). To reduce the solu- .. 101 ;:)
..,
bility of lipids in the substitution fl uids, post-fixation with Os04 , and possi- c
0 10 3
-.
>(

bly also uranyl acetate, is necessary. Generally, fixation schedules similar ~ ~


::J
.,"
...,:;•
10- 1
to those used prior to thin sectioning (Glauert 1974; Hayat 1981) or critical IP
a I 10 1
point-drying (Boyde 1980) are used (see also §2.3.3). I
:; I
1 Glycerine

Subsequently the water of the chemically fixed specimen is replaced with
..
0
Q.

10 · 3
I
I 2 DMSO
3 Water ..a
::J
a solvent such as ethanol, iso-amyl acetate or chloroform (Boyde and Wood > t I

1969); camphene (Watters and Buck 1971); or tertiary butyl alcohol I


I
I

I
I
I 4 Ethanol
5 Benzene
1 0- 1 ."'
0

>
6 Acetone
(Wheeler et al. 1975). These substitution liquids at 100% concentration soli- I I
I I I 7 Carbon dioxide
1 0-5 I
dify on quenching (for techniques see §2.4) to form a glass, or crystals of -200
8 Ammonia
-100 0 10-3
such small size that they can be neglected. Consequently, structural detail 100
Temperature (°CJ
within bulk specimens can be preserved which would otherwise be destroyed
if the specimen were to be frozen from the hydrated state.
Due to the high vapour pressure of liquids used just below their melting
points (see F ig. 5. 14 and Table 5. la in §5.1), freeze-drying can be extreme!)
rapid. As an example, freeze-dryi ng of a bulk specimen from amyl acetate
(melting point, - 7l "C; 202 K) at a temperature of - 75°C (198 K) and
pressu re of 5 x I0 - 3 Torr (6.65 x 10- 1 Pa) may take only about half an hour
(Boyde and Wood 1969). However, to have a safe margin to prevent evapo-
ration of liquid rat her than sublima tion of solid, Boyde (l 978a) recom
mends that specimens infiltrated with organic solvents should be frozen
dried at temperatures at least JO K below their freezing points.
Since no ice crystal damage occurs on freezing, internal structures of bulk
specimens can be exposed by fracturing before drying. This is best done
under liquid nitrogen with samples infiltrated with ethanol or hafocarboll
13 (Boyde and Maconnachie 1979b). If freeze-fracturing is carried out Oii
Freeze-drying 273
Low temperature methods in biological electron microscopy
272 Robards and Sleytr
5.5.la Specimen mounting and freezing
.t.on tissues become very brittle after
. d d tly of membranes. I n addI I ,
zn epen en . l t d may easily fracture during subsequent
·th organic so ven s an . for most mounting procedures it is advantageous to make the surface of
treatmen t w1 . . . the SEM Despite such d1fficul-
. t during examinat10n in · the support hydrophilic. For grids with carbon-coated support film s this is
processing s eps o~ th advantages of the method (no ice crystal
ties, for many routine purposes e . best done by glow discharge in air at low pressure or by radiation with ultra-
artefacts, rapid preparation) offset the disadvantages. violet light. Specimen adsorption may also be improved by the addition of
wetting agents (e.g. bacitracin) to the suspension (Gregory and Pirie 1972).
No 'wetting' problems should occur if specimens are applied to the surface
Freeze-drying procedures for different types of specimen of a freshly cleaved piece of mica or when the specimens are sprayed onto
5.5
a pre-cooled support. Supports can be positively charged with poly-L-lysine
5.5.l Freeze-drying of small biological specimens (see §6.2.2d) (Fischer 1975; Nermut and Williams 1977) or 1% alcian blue
8GX (Sommer 1977; Nermut a nd Williams 1977, 1980) to achieve a better
. b used in a variety of ways for structural
The.f reeze-drying technique can e . adsorption of specimens with a net negative charge. Selective adsorption
. ens (e subcellular fractions, macromolecules, v1rus-
of particles (e.g. viruses) may be obtained by coating the support with speci-
stud1es of small specun . .gl. _and eucaryotic cells and monolayer cut-
es membrane structures, smg e pro . -d . d . h- fic antisera before use. This is simply done by floating carbon-coated grids
, ll b. l ·cal specimens can be frozen ne wit on a dilute serum for several minutes (Derrick 1973). Except for cells (e.g.
F 5 15) Most sma 10 og1
tures; ig. . . h hemical fixation or cryoprotection, provided
monolayer cultures) directly grown on the specimen support, the majority
out any pretreatment, s~c t::cenough and the cryoftxation technique is ap-
of small specimens are mounted by one of two methods: spray-freezing
that t~e specimen laye,r I~ §~ 2 2 ) Once the specimen has been liberated
(Williams 1952, 1953); or adsorption (Nanninga 1968; Nermut et al. 1972;
propnate (see §2.4. l an d '. .t . aai) ice it can be shadowed, coated or repli-
Nerrnut 1977, 198la).
from the surroundmg (an m .ern ) when suspended in a negative-staining
cated. Some specimens (e.g. viruses , t . the TEM even Jn the spraying technique, the specimen is directly sprayed with an atomiz-
. ·d no ugh contras m
solution before freczmg, may pro~1. e e er (e.g. Effa Spray Mounter, Ernest Fullam; Appendix 2) onto grids or piec-
without any additional metal depos1t1on. es of mica firmly attached to a cold block. This is very conveniently achieved
by mounting the specimen support directly onto the temperature-controlled
specimen table ofa freeze-etching unit (Steere 1973; Nermut 1977) or freeze-
drying unit. For spraying, particles are best suspended in distilled water or
volatile buffer (e.g. 0. 1-0.3 M ammonium acetate, ammonium carbonate or
ammonium succinate) (Backus and Williams 1950; Williams 1953). Since
lhe melting point of the eutectic concentration of ammonium acetate buffer
1s below - 70°C (203 K), freeze-drying temperatures must be ::;:;-80°C
(193 K) (see Williams 1953). The schedule for the spray-freezing, freeze-
drying procedure is then as follows:
unidirectiona1
or ro ta r y
(I) Mount specimen support on the specimen table.
2) Evacuate chamber to high vacuum and cool specimen support to
-----· fNe-g-;ti;ely:st;1;ed1 about 123 K.
: Pteudo 1__ _ 1 preparations 1
'r•pUca \ L- - -- - - -- --- 3)
L- -- -- Vent chamber with dry nitrogen and maintain cold surfaces in
vacuum chamber as free as possible from hoar frost by continuing to
TEM, STEM
· ens by rrccze.Jf'
purge with dry gas.
D ifferent possible stages in the preparation or small spec1m
Fig. 5.1 5.
274 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-drying
275
(4) Spray specimen onto cooled support. requires more concentrated suspensions b t
(5) Stop dry nitrogen flow and evacuate chamber. . h b w· . . u mono1ayers can be compressed
wit a ar. ith virus particles spreading of 20 I (10-20 .
(6) Warm specimen table to selected freeze-drying temperature ( - 85°C; · d 'hi ' µ mgprotemmJ - 1)
m1xe wit 0 µI of 0.1 % cytochrome c (" M .
188 K) while protecting the specimen from contamination by using
0
m 0· 1 ammomum acet t )
vides a good yield (Nermut 1973a). a e pro-
a cooled condenser (either a shroud (see §6.3.3b) or a cooled micro-
tome arm (see §6.3.3a)). Irrespective o~whi~h mounting procedure is used before freezin the best
results are obtamed 1f the liquid surround· th . . g,
. ( . mg e specimen ts free of non-
voIat11e non-subhmable) constituents S .
If a freeze-fracture replication unit is used which provides for contamina- · . · pecimen supports loaded b
spreadmg or adsorption techniques must be transferred through I y
tion-free transfer of specimens from liquid nitrogen into the vacuum unit Ieas t th ree) was hes m
· · · severa (at
d1st1lled water or volatile buffer (F. 5 16)
(see §6.3.3c), spray-freezing can be performed on a separate cold stage. 1g. . to ensure
There a re problems when pathogenic organisms have to be examined using Filmed grid
the spray-freezing method and there is also the disadvantage that considera- / Monolayer

tr""'",.,""Lj
bly more material is required than for the adsorption technique. Furthermore,
special attention has to be paid to the spraying technique itself: the droplets a
freezing on the support should be as fine as possible and deposited as a Spe cimen spread
o n surface

0 0 0
monolayer to prevent extensive overlapping of particles after the ice has 0 0 0

been sublimed away (see §2.4. lc). As a n alternative to spraying small speci- Filmed grid~ ~
mens onto cooled supports, they may be sprayed onto supports at room Washing (repeated)
temperature. After repeated washing with distilled water, as with the b 0 °o ~ o i on distilled water,
volatile buffer
adsorption technique, and rapid freezing, specimens are transferred to the (or negative-
staining solution)
cold stage of the freeze-drying unit (Walzthony et al. 1983). Specimen In
suspension
T he adsorption technique is a relatively more convenient method for
applying small specimens to supports. In comparison to the spraying techni-
que, much less material is required and monolayers of particles are more
easily obtainable. As illustrated in Fig. 5.16, adsorption of the specimen to
the support can be achieved in two ways: by touching the support to the
surface of a suitable liquid (distilled water or buffer) on which a monolayer c
Shoreline of
of the specimen has previously been spread (Fig. 5. l6a), or by floating the freeze-fractured
support on a droplet of the particle suspension (Fig. 5.16b). With the float· molecules

ing procedure the required adsorption time depends on the concentration


of the particles in the suspension. Thus, even low concentrations can give
a sufficient yield after prolonged contact (e.g. 15 min). Fowler and Aeb1
(1983) obtained excellent results when a diluted solution of protein mol-
ecules (typically 10-30 nM) or supramolecular structures (typically 0.5
mg mJ - 1) was allowed to adsorb to a glow-discharged, carbon-coated cop-
per grid or freshly cleaved mica for 1 min. When coated grids are used as
supports, negative-staining techniques can be used to check the success or
this preparation step.
The spreading technique (Kleinschmidt 1968; Willison a nd Rowe 1980
276 Robards and Sleytr Low temperature m ethods in biological electron microscopy
Freeze-drying
277
that both non-adsorbed particles and masking material are fully removed. The
. major advantage of this proced ure 1s
. that .
Before freezing, the excess liquid is drained off with filter paper (Fig. 5.16). particles, such as viruses can b . macromo1ecules a nd small
. • e viewed unfixed 10
· d" .
To avoid complete drying of the thin film, this is best done just above the of their natural ionic environ men.t con rt10ns close to those
surface of liquid nitrogen. With hydrophil ic supports, freezing by immer-
sion (§2.4. I a) must be done within two to three seconds to avoid air-drying. 5.5.lb Freeze-drying and subsequent handling
An even faster transfer into the cooling bath is required if more hydropho-
bic supports are used. In order to produce a thin film of water prior to freez- Frozen specimens mounted by th .
e spreadmg or abs t'
ing, grids can be placed for a few seconds with the specimen side down on only be embedded in ice a few t f . orp ion technique will
a small stack of wet filter paper in a Petri dish. With hydrophilic supports, from Table 5.2 in §5 2 la und ens _o micrometres thick. As can be seen
· · ' er optimal condif 50
the filter paper only needs to be kept slightly wet, whereas for hydrophobic ice sublimes at -85°C (188 K) "th" . JOns a µm thick layer of
specimen carriers the paper must be soaked (Wiirtz et al. 1976; Kistler ei son, at - 9ooc (183 K) the same~1 m approximately 25 min. For compari-
. ayer would take an h .
a l. 1977). calculat1ons show that for thi t . our to sublime. These
. ' s ype of spec1m l ·
With some specimens (e.g. suspensions of cultured cells), the fo rmation limes are sufficient. Nevertheless 't ·11 , en, re a t1vely short drying
. , 1 w1 take longer t
of usable monolayers by adsorption may be inhibited by surface projec- from the specimen matrix (see §5 3) w· o remove all the ice
11 . . Jth some specim (
tions. If the pr~servation of the gross morphology of the cells is less impor- ce s, monolayers of eucaryotic ens e.g. procaryotic
11
tant, the area of adhesion can be increased by centrifuging the cells onto shown that prolonged freeze-dry~e s, membrane structures), it has been
mg can lead to cert · hr"
the support (e.g. mica). Iapse phenomena (Nei I 962 1973 . N am s mkage or col-
To examine the inner surfaces of membranes, tissue culture cells, firmly ro check whether better stru~tur I ' ermut_ 19~ 1a). Consequently it is best
. a preservation 1s achi d ·f b
attached to a support, are lysed by contact with distilled water. After the preparation step (e.g. replication) . ffi . eve I t e subsequent
h b . IS e ected immediate! ft
unattached membranes and cell contents have been removed with a stream as een hberated from the surro d" . · Ya er the specimen
of buffer (e.g. hypotonic phosphate buffer, pH 7.4, applied from a syringe As summarised in Fig 5 15 thunfi mg ice (~euser 1983).
· · , e reeze-drymg proced fi
fitted with a hypodermic needle), the inner surface of the plasma membrane mens can be combined with the foll . ure or small speci-
can be examined (Mazia et al. 1975; Nermut 198J a, 198lb; see also §6.9. lcJ Sliado1ving (unidirectional or taowmg preparation steps:
ro ry· see §6 4 I) fi T
As an alternative to the common adsorption technique, Heuser (1983) avoid secondary structural ch· , . . or E M. To reduce or
suggests mixing the molecules with an aqueo us suspension of tiny flakes of moisture, the shadowed spec· anges caused by the uptake of atmospheric
I... b·1· rmens mounted on supp t ti!
mica and then freezing and fracturing this suspension as with freeze-etching ut' sla I ised by evaporating th. ( or ' ms on grids can
If a m < 05nm)Ja f
techniques (§6.3.3a). During the subsequent freeze-drying process, which re· shadowing or replication of th fi . . yer o carbon from above.
. e rozen-dned spec· ·
sembles deep etching (§6.5), areas of unfractured mica and thus intact mac am bieat temperature the th . imen ts performed below
. n e specimen should b
romolecules are exposed (Fig. 5. l 6c). When specimens are frozen on cold ng l he coating unit to avoid c d . e warmed up before vent-
R /" on ensat10n of atmo h . .
metal surfaces (§2.4.2) the adsorbed macromolecules are not obscured b) ep 1cation (shadowing C sp enc moisture.
salt deposits even if they are s~s-pended in hypertonic solutions. An add \\'nh sma ll specimens (e g, c?a mg (§6.4.I) and cleaning (§6.6)) for TEM
tional advantage is the fact that the exposed macromolecules are only mrru- cas can easily be obtain~d. ~ruses,dsu~cellular fractions) satisfactory repli~
~I ) . reeze- rymg of larger .
mally frozen-dried and consequently retain their three-dimensional topo- s generates more pronounced t . specimens (e.g. isolated
graphy very well. Using freeze-etching apparatus for this procedure, freeze. during cleaning (see §6 6) f: olpograph1cal profiles. Specimen swelling
. requent y causes frag .
drying is best done at -102°C (171 K) for 3 min. After this superficial dry- s problem can be reduced if the cell mentat10n of the replica.
ing, the specimen is replicated by rotary shadowing with Pt/C and ..:en example, bacteria are best mo:~~=:r~zen as a _densely ~acked layer.
(§6.4.1). After thawing, the replica is cleaned by treatment with full stre pellet over a piece f . y smeanng a thm layer of a
chrome-sulphuric acid and full strength hydrofluoric acid (Caution: hydlO- llounted in this way cann~t ;;1ca c;; other suitable support. Since cells
fluoric acid is extremely hazardous).
6), lhe pellet must b was:ct by floating the support (see F ig
e suspended either in distilled water or in volatil~
278 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-drying
279
buffer to avoid the masking of structural details by the deposition of non-
volatile components (culture medium, salts, etc.).
Pseudoreplication (shadowing and coating) for TEM. Small structures
1- - - - - - - - -
(e.g. subcellular fractions) mounted on mica can be examined using the 1De ep..e tching I
1A r eplic ation :
pseudoreplica technique . With this type of preparation the specimen is not : se e se ction 1 - Water replaced by
1_ - - ~-: - - - : ------~~ organic solvent '
______\.
removed but remains attached to, or enclosed within, the deposited heavy ISee section 5:-4-:
metal and carbon layers. Pseudoreplicas reveal the same structural details 1& Fig. 5.13
---------- J
1

as shadowed specimens mounted on coated grids. Fixation with vapour


e.g. Os04 , aldehydes
Freeze-drying and negative staining for TEM. A procedure combining
negative staining with freeze-drying was developed by Nermut and Frank Vacu.um embedding
in resins
(1971) for structural studies on viruses. Negative staining is carried out after
Repli ca c le aning &
the last washing step and immediately before freezing (Fig. 5.1 6a and b). exami natio n of replic Thin sectioning
in SEM
Best results have been obtained using 2% phosphotungstic acid (PTA) and ____ rA-.,lo°"r;d-;;g-;a-p;;y- - 1
1 Cytochemistry 1
3% ammonium molybdate but 1% uranyl acetate may also be used (Nermut : LAMMA :

1973b). Following freeze-drying, the specimen is warmed up to 30°C (303 ~X_:'!~ microanalysis 1

K). Immediately after breaking the vacuum, the excess dried negative stain Fig. 5. 17. Possible steps in the preparation of bulk specimens by-f~~::::y~:g.
which remains as a powder on top of the specimen has to be removed using
in, with or without previous che . 1 fi .
a gentle flow of dry nitrogen gas or with a fine brush. Since the dried nega- . mica ixat10n unde
quent thm sections are suitabJ 10 ~ ' r vacuum. The subse-
tive stain is very hygroscopic, the specimens should be examined as soon e r morpho1ogica1 t d ·
procedures, a utoradiography (e g f, h . s u 1es, cytochemical
as possible after preparation and/or stored under vacuum (Nermut 1977). or LAMMA (laser microprobe ~ . ]or .t)e det~ct1~n of diffusible substances)
Sputter-coating for SEM. Examination of frozen-dried single cells and na ysis applications.
monolayer cultures by SEM is described in §5.5.2a, together with the prepa- 5.5.2a Examination oif bulk . .
ration and examination procedures for bulk specimens. speczmens m the SEM
Embedding. Frozen-dried cells may be vacuum-embedded for thin sec-
The effects of specimen pretreatment and .
tioning with or without previous exposure to fixatives in the vapour pha degree of similarity between th . freezing. For most specimens, the
(see §5.5.2c). . h e native structure and th
mt e SEM depends strongly th e structure observed
process - in particular the oh? e treatments preceding the freeze-drying
5.5.2 Freeze-drying of bulk specimens was mg and fixatio t
ent pretreatments as we]] as of different n s a~es. The effects of differ-
freeze-drying) on specime t dehydrat10n processes (inc1uding
Freeze-drying can be used to prepare bulk specimens for the study of both d h' n s ructure have been d. d.
an is colleagues (Boyde 1976, l978a . iscusse m detail by Boyde
surface and internal structures (Fig. 5.17). In most unfixed bulk specimen .An optima] fixation procedure Jdl~78b, Boyde and Franc 1981).
only a very superficial layer can be well frozen (see §2.3.1 ). To avoid ice cry m1c processes within thos wfou mstantaneous1y stabilise the dyna-
tal damage at a deeper level, the water must be replaced by an organic ol . e areas o the specime th t
tu died in the SEM. Usua11 . . n a are subsequently to
vent before freezing (see Fig. 5.13 in §5.4) unless high pressure freezing t particular, the plasma m b y, this is the surface of the specimen and in
niques are available (§2.4. ld). The surface structure of the frozen-dried em rane G1utara1dehyd fi d '
ua 11 Yshow less ice crystal dam~ . e- ixe specimens (§2.3.3)
cimen can be studied either in the SEM or in the TEM using the high re ol m~rane than do those frozen fr~e m froze~-dried images of the plasma
tion heavy metal/carbon replication technique. This technique resem ation may also e h h . m the native state (Boyde l 972 l 980)
that for deep etching specimens during conventional freeze-etching m~th d n ance t e res1stanc t d. · ' ·
n ucti~ity of a specimen. The latter ~ o ra 1at1~n damage and e1ectrica1
(see §6.5). Finally, the frozen-dried bulk specimens can be embedded 10 zen-dned specimens are to be studied.Is of some importance if uncoated
280 Robards and Sleytr Low temperature m ethods in biological electron microscopy
Freeze-drying
281
. . lo ical systems it is important to wash off,
For many SEM studies on b10 gh ·men surface This can be done
k. aterial' from t e speci · Irrespective of any pretreatments, washing in distilled water or in a vola-
or digest, 'mas . mg m .d
fixation to avo1 s a I I t bTsation of material that coatsh.the tile buffer before freezing is an essential step if 'clean' specimen surfaces are
before the pnmary I
f . t st Special attention o . t this step is necessary. The was Ing to be exposed by the freeze-drying process.
. h
structure
. o m ere ·
Id b . otonic buffere an p d d hysiologically appropnate. tofi t e As discussed in §2.4.1 , ultra-rapid quenching allows the superficial layer
liquid. shouSh e~is ces generated , b YsmaII J·ets of water from a syringe Itted of a bulk specimen to be frozen without any ice crystal damage that will
specimen. deer ·or edle or a senes . o f m1"Id centrifugations (e.g. for proto- be visible at the resolution of most SEMs. Nevertheless, special care must
with a hypo h IerIDic ne
t emove the mas k'mg material · Since such treatments may be taken that the film of overlying liquid is thin enough to allow high cool-
zoa) may e p o r h Id T possible be performed after a pre. ing rates to be obtained within the superficial layer while, at the same time,
cause structural changes they s ou , I et ~pithelial surfaces can be
F r example mucus on w . air-drying of the specimen prior to freezing is avoided . If samples are too
fixation stage. ? ;ic a ents after primary fixation with Os04, large, cracking artefacts may be produced by thermal stresses during freez-
removed by specific mucoly I' k? lutaraldehyde (Boyde 1980). Never- ing.
omitting the use of the cross- m_ mg ~II be performed with glutaraldehyde,
·mens fixat10n w1 FinaIJy, fixed or unfixed fully hydrated or substituted (see §5.4) frozen
theless,
· for most ~spec1 d by one or more other fixatives known from TEM specimens may be fractured under liquid nitrogen (see §6.3.3c) before freeze-
11
where necessary o owe 974) t render the specimen more able to drying. If the substitution was with ethanol or halocarbon (e.g. Freon 13),
processing schedules (Glauert l A, o example little loss of structural
h d h dration process. s an , d critical point-drying may be considered as an alternative dehydration proce-
withstand
. t e e Y d
h n glutaraldehy e- ftixed tissue samples are washe hm dure for freeze-fractured specimens (Figs. 5.13 and 5.17) (Boyde 1974a;
matenal occurs w e . uffer system) before freezing. On the ot er Boyde and Maconnachie 1979b).
water to remove the vehicle ~b . d to render lipids insoluble in organ- Freeze-drying and examination. Frozen specimens are stored under liquid
fi t. 'th OsO is reqmre d
hand, post ixa wn w1 4 I . th specimen water when freeze- ry- nitrogen until required for use. Transfer to the temperature-controlled stage
hi h used for rep acmg e
ic solvents
. ~
w cd fromare non-aqueous syst ems (see §5.4) (Boyde 1980). of a freeze-drying unit (§5.3) is best carried out under the protection of
ing is per orme . the first and only fixative used for man} hquid nitrogen to avoid the r isk of ice recrystallisation due to a critical tem-
Although glutaraldehyde will be d . 0 I M sodium cacodylate buffer.
perature rise in the specimen (Smith I 980; Heuser 1983; see also §6.3.3c).
~.g.
. . ( 2- 3% glutaraldehy em · h
applications . o ould be aware that it can generate c anges The individual parameters which determine the freeze-drying procedure
pH 7.2, for arumal tissues), we sh h b n studied in detail by freeze·
t Such artefacts ave ee d. for bulk specimens have been discussed in detail in §5.3. Since unmounted
in membrane struc ure. ( §6 8 3) but most of the changes (e.g. is- specimens (mos t bulk specimens) have only a poor mechanical fit and ther-
fracture replication methods se~ d. rticles) occur at a morphological
mal contact with the flat, temperature-controlled stage within the freeze-
placement of membrane-interca_ate -!~of the SEM (§5.6). Special atlen·
drying unit, special care is required to ensure that the chosen temperature
level beyond (below) the resolut10·~· h f the fixative vehicle. As with lixa maintai ned during the drying process. Boyde and Echlin (1973) and
tion has to be paid to the compost ion o hanges in the surface topograph
d for the TEM severe c and Boyde (1975) suggest that the specimen should be enclosed during drying
tion procedures use . . , I' d at a physiological temperature. Within a gadget such as that illustrated in Fig. 5.18a. This consists of a cylin-
can occur if the fixative i.s not app ie fl .d . the tissue or cell environ-
drical metal dish with a lid with a diameter larger than that of the dish. The
pH and is not isotonic with the ambient ~1 ':ecipitates on the specif'llCI
lipper part of the sides of the dish has several 'cut-outs' to facilitate the
ment. Furthermore, buffer systems pro~ucm(g p phosphate buffers) should
fleape of subliming water vapour. The dish is deep enough to carry over
surface during fi1xat10n
· and/or dehydrat10n
. . e.g. d d that whenever feat-
d 1980) Generally it is recommen e ' d 11il IDme liquid nitrogen from the Dewar flask in which it is loaded with the
be avoided (Boy e .. d s) in lutaraldehyde is allowe . lpccimen. This design has the advantages that it protects the specimen from
ble, a long fixation period (sev_eral a:rder agnd mechanically more stable
tllnning above the desired freeze-drying temperature during transfer and
has been shown to make the tissue h B d 1972· Humphreys 19
lu. so long as good thermal contact of the dish with the temperature-con-
~
· t
which is of particular impor ance for SEM ( oy e se a 'degradation in
·
. . parently do not cau platform of the freeze-drying unit is maintained, the specimen will
These long fixat10n times ap
appearance of the tissue fine structure (Hump rey
h s 1975).
~II protected from radiation heating. In other words, specimens will
s in biological eleclron microscopy Freeze-drying 283
Robards and Sleytr Low temperat ure method
282
2 µrn rnin- 1 (see Table 5.2 in §5.2.la). Thus the thin layer of ice covering
the specimen surface will have sublimed away after 5-10 min. Considering
a much slower drying process (lower sublima tion rate) below the cell bound-
ary (cell membrane, specimen surface) freeze-drying for 1-2 h at a 'recrystal-
lisation-safe' temperature is sufficient for good structural preservation of
the surface topography.
Since most frozen-dried specimens are very fragile, freeze-drying should
ideally be performed with the specimen already mounted on a stub of the
type used in the SEM or on a support that can be a ttached to an SEM stub
without risk of specimen damage. The latter approach is possible with cul-
tured cells tha t have been grown, or mounted by adsorption, on coverslips
b
a or other appropriate substrates (see §5.5. la). Supports carrying the frozen-
. d . b lk specimens. (a) A gadget described b)
dried specimens are best attached to stubs using electrically conducting
. S imen containers for freeze- rymg u . . . . . h· glues, such as silver or graphite paste. For direct mounting of small speci-
Fig. 5.18. pee . . . d . 1metal d ish with a hd. The castle-like s ape
1
Boyde and Echlin (1973) consisting_ of_a cy m nca The relatively geometry of the lid/base mens (e.g. bacteria or tissue culture cells) Boyde (1975) recommends the use
. . h . f subhmmg water vapour.
of the dish allows l e escape 0 d' t heating during the drying process. (b) Plan of hemispherical specimen supports, such as round-headed aluminium riv-
· · rotected from ra 1an
ensures that the specimen is P .h . proved pathway for the escape of water
. · · 1 t 0 (a) but wit an 1m ets, which have been micro-roughened by treatment with an abrasive unit
of a similar design pnncip e . ed so as to allow easy escape of water vapour
th dish and hd a re arrang . (see Appendix 2) to render them clean and wettable.
vapour. T he bames on e . h tin by radiation from external sources. (F1~
while totally protecting the specimen from ea g d 1975) Most frozen-dried specimens are very hygroscopic. Thus, if the vacuum
5. 18a redrawn after Boy e ·
must be broken for mounting, special care is required that a ll manipulations
. . d at the desired temperature during the whole take place in an atmosphere of low relative humidity for the shortest possi-
only 'see' surfaces mamtame . roach based on the same princ1· ble time. Isolated specimens (e.g. organisms, tissue samples, larger cells) can
· dure An alternative app , .
freeze-drymg pr~ce . h ~ the escape of water vapour, is illus- be mounted in the frozen-dried state by carefully attaching them to double-
ple but with an improved pat . way or 1ded sticky tape which has previously been firml y attached to the SEM
trated in Fig. 5.18b. . depend on the type of spec1- tub. With large specimens the area of contact can be increased by mounting
Freeze-drying schedules for bulk spec1m~ns . d . §5 2 lb the them on silver or graphite paste. Special care must be taken to prevent the
. . · ired As d1scusse m · · ·
men and on the ~cie~t1fic mforrnatl~~:~:ine~ the upper temperature hm1t solvent of such pastes from being soaked (by capillary actio n) into the speci-
critical recrystalhsatl~n temperatur dried For SEM purposes, samples arc men area of interest. The evaporation of the solvents from adhesive pastes
at which the ~rea of mterest can b~ ( - 100 to - 500C) but even higher tern can make the specimen damp and may even cause secondary shrinkage arte-
usually dried m the range 173- 223 l (B d 1980· Boyde and Franc acts (see §5.6). To avoid these potential hazards, Boyde (1980) suggests the
. e satisfactory resu ts oy e ,
peratures may g1':' . . .d rule for freeze-drying procedures. use of rapid-setting epoxies as adhesives. The viscosity of these resins can
198 1). It is also difficult to give a ng1 . h t be observed' be adjusted by chosing the time at which the material is used and, as a
. . ~ bulk specimens t at are o
particularly the drymg time ?r . u rficial layer in most hyd unher advantage, the vapour pressure of the polymerised resin is so low
SEM As pointed out earher, smce only a s pe t dry r.hc
the . 't akes little sense to try o t ii can be neglected.
rated specimens can be well froze.n, 1 m ere the ice crystals will, in an Frozen-dried biologica l specimens are primarily composed of consti-
more central regions of bulk spec1~en:n <7:
avoid recrystallisation (~ydc bients with low atomic number which behave as electrical insulators. To
case be large) at a temperature c os . l la er of water is prc- cnt charging under the electron bean in the SEM and to enhance
1980) provided that only a thin (5- 10
. .
µm?
superfic1a y
At specimen temperature o
f - 85 C (I llcondary electron emission, the following measures can be considered:
sent during spectmen free~mg. ,a d . this layer of pure water
K), for example, the 'drymg front procee s m
284 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-drying
285
(i) coating the dried specimen with a layer of electrically conducting Finally, particularly for the stud f" .
Yo irozen-dned s · .
material; gular topography, even replicat1·0 t h . pec1mens with less irre-
n ec rnques can b ·
(ii) depositing meta l within the specimen; thick evaporated layer is deposi't d ( e considered. Firstly a
. e e.g. of carbon d Id) '
(iii) using a very low beam accelerating voltage and/or current; and the specimen (see §6.6) it is poss.bl b . . an go ·After digesting
. ' i e to o tam h1 h I .
(iv) replicating the specimen surface and studying the replica. of the replica side originally in con . g reso utwn SEM images
1978). tact with the specimen surface (Revel
Coating the specimen with a thin layer of an electrically conducting mate-
rial is still the most commonly used procedure. For high resolution work 5.5.2b Specimen replication/or the TEM
this layer should be as thin as possible, have a very small crystallite size and
give a sufficient yield of secondary and/or backscattered electrons. As with Frozen-dried bulk specimens can b . .
. . . e exam.med m the TEM . .
replica techniques used for the TEM (see §6.4.1 a), the actual resolution limit lut1on rephcat1on techniques (Fig. 5 17) T . usmg high reso-
is primarily determined by the grain size of the evaporated layer (Peters )tructural information that can b b. . . his reveals much of the overall
. e o tamed by SEM (§ 5 5 2 .
1977, 1979; Blaschke and Boyde 1978; Jakopic et al. 1978). For morpholog1. lion, allows the resolution of stru t 1 d . · · · a) but, m addi-
. c ura eta1ls dow t h 1. . .
cal studies the most common coating materials are gold, alloys of gold and by the high resolution shadowing t h . ° n t e 1m1t imposed
palladium, platinum and silver (Echlin et al. 1982; Kemmenoe and Bullock l~sis of stereoscopic images m kec .nique ~§ 6 .4. I b). In particular, the ana-
. . a es it possible to 0 bt ·
1983). For X-ray microanalysis, specimens are coated with carbon, alum1. d1mens1onal information on both th am excellent three-
. e gross morphol d
nium or chromium (Marshall 1980). The technology for evaporating coat· 1ails of the frozen-dried specime (S ogy an structural de-
. n teere 1973· H 1980
ing materials by resistance heating and electron-beam evaporation is dis· High resolution replication of ti d . ' euser , 1981).
. rozen- ned spec· ·
cussed in §6.4.1, and that for sputter-coating in §3.2.3, §3.3.6 and §3.4.6. cly earned o ut in freeze-fracture 1· . . imens is almost exc!usi-
rep 1cat1on uruts h · h h
As an alternative to specimen coating, procedures have been developed cont.rolled stages (§6.3.3). The h d w ic ave temperature
sc e u1e for free d ·
for depositing osmium within the specimen before dehydration (OTO and specimens largely resembles the f; ti ze- rymg and replicating
reeze- racture re r · .
OTOTO methods - see Kelley et al. 1973; Munger 1977; Sweeny and Sha deep etched specimens (§6 5) w1'th th . . p icatwn techruque for
· e maJor d1ffere th f;
piro 1977; Murakami 1978; Murphy 1980). These techniques involve th r non-fractured specimens are - . nee at reeze-fractured
mamtamed at slight! h. h
risk that the benefit of improving specimen conductivity is accompanied b\ cg. -90 to - 95°C· 183 t 0 178 K) . Y ig er temperatures
. . . ' to achieve comp) t bl. .
a limitation in resolution arising from metal precipitation on the specimen specimen ice within a practical . d f . e e su imatwn of the
peno o time (Tabl 5 2 ·
surface which may thus produce results that are not essentially differen mens that are not fractured unde e . m §5.2. ! a). Speci-
r vacuum befo b · .
from surface-coated specimens. protected from hoar frost d .. . re emg dried need to be
. epos1tion durmg tr ti
Examining the specimen at very low accelerating voltage (e.g. 1-2 kV) lpec1men stage. If no prote t. ans er onto the pre-cooled
c ive transfer or · I k .
the SEM may be a further alternative to the deposition of an electrical 3Jb and c), a simple metal lid b air oc system is available
conducting material, although high resolution images are then difficull IS lifted out of the liqw·d .t can ~ placed over the specimen before
n..__ Ill rogen contame (H .
obtain. Reducing the beam voltage limits the penetration of the electroas VllQ: a suitable vacuum (e g 10-1 r euser and Kirschner 1980)
) has been reached th . . to lQ - 4 T . I 3 .
but provides the possibility of working in the range where the secondalJ . . orr, . 3 x 10- 1 to 1.33 x 10- 2
Piia . , e protective lid can be re d .
electron yield is close to unity and, as a consequence, no charging occun led via a vacuum lead th h)
- roug and the
move (usmg a lever mani-
·
(Reimer 1978). With frozen-dried specimens, charging is also not a problcll ng temperature. If highl 1 specimen wanned up to the
llruc:t y scu ptured surfaces d .
when a beam voltage of 3-5 kV is used at TV scanning speed. This IS ures, are exposed by th d . . , an particularly fibrous
-.... h e rymg process 1t ca b d
particular interest if specimens are to be micro-dissected in the SEM ---1 s adowing and re 1· t' , n ea vantageous to use
P 1ca ion proced (§6
which case it would not be sensible to deposit a surface conducting la structure. With some s . ures .4.1 f) to reveal the speci-
8-crand Kirschner . 1980) .
pec1mens (e g fro
. .
d .
zen- ned cytoskeletons. see
(Boyde 1974b). (See further discussion of charging problems related to . ' it may even be p "bl . ,
temperature observations in the SEM in §3.2.1.) specimen as a pseudoreplica ( . h ~ss1 _e to examme the frozen-
w1t out d1gestmg the specimen away).
ds in biological electron microscopy 287
Robards and Sleytr Low temperature me /ho
286
study it has been shown to be essential to use ultra-rapid freezing techniques
. . d frozen-dried specimen must be warmed up to
In this case, the rephcate ' k. the vacuum to avoid water vapour (e.g. freezing on cold metal surfaces; §2.4.2) to ensure that the ice crystal
. ture before brea mg size in the area of interest is at an absolute minimum. The potency of this
ambient tempera . h d t" n Replicas of rough and hetero-
. d ec1men re y ra io . technique is demonstrated in Fig. 5.19, which shows coated pits and cyto-
condensation an sp . h tendency to disintegrate during the
d · d specimens ave a skeletal elements. Although the study of these replicas does not necessarily
geneous frozen- ne H r and Salpeter (1979) suggest that
. d I · g procedure. euse require stereoscopic techniques, they are nevertheless highly recommended
thawmg_ an c eam_n hould be transferred, while still cold, to a small
the rephcated specim~ns ~ . I) ff zen methanol which is then allowed as they provide a better appreciation of the three-dimensional structures
scintillat1on via o ro .
receptacIe (e.g. a b . t ) the material at the same time. After which often closely resemble the images obtained (certainly at considerably
h and to infiltrate (su st1tu e . . lower resolution) in the SEM (Heuser 1981).
to t aw b. t t mperature they are then rmsed in
. . h · ens to am 1en e ' .
bnngmg t e spec1m h al cleaning solutions (e.g. sodmm hypo-
water and transferred to t e usu 5 5.2c Freeze-drying and embedding for thin sectioning
chlorite; §6.6). . of frozen-dried tissue culture cells
. · licas part1cu1ar1Y . Methods which combine freeze-drying with embedding have been developed
High reso 1ut10n rep . ' . h ve revealed remarkable new mforma-
·ri d toplasm1c protems, a as alternatives to the standard chemical fixation and organic solvent dehyd-
and pun 1e cy . t" f the cytoskeleton and membrane-
t and orgamsa ion o ration techniques preparatory to ultramicrotomy. The aims are:
tion on the struc ure . H 198 l 1983). For this type of
associated structures (for reviews, see euser '
(I} to achieve a more ' natural' preservation of the native specimen struc-
ture by avoiding or reducing artefacts caused by chemical fixation and
solvent dehydration; and
( 11) to avoid or reduce the extraction or displacement of labile or diffu-
sible substances (e.g. during specimen fixation or chemical dehyd-
ration).
It was soon realised that this alternative procedure involves its own set
of artefacts (for reviews of early developments see Sjostrand 1967; Rebhun
1972). However, freeze-drying in combination with plastic embedding offers
unique possibilities for special electron microscopical applications such as
auloradiography of water-soluble substances (Stirling and Kinter 1967; Wil-
ms 1977; Pfaller 1979; Edelmann 1980b); cytochemical analysis (Gersh
973); ion localisation by X-ray microanalysis (Chandler 1973, 1977;
ngram et al. 1974; Edelmann J 978a, l 980a, 1981 ; Dudek et al. 1981 ); laser
llleroprobe mass analysis (Edelmann I 980a, 1981 ); and investigations of the
lkllccular structure of cell compartments (Sjostrand and Kretzer 1975). As
With most other low temperature techniques, the success of these prepara-
methods depends primarily on optimal cryofixation of the specimen
lld many past failures can be attributed to this problem. Again this will
more of a limitation for high resolution studies, where the aim is optimal
flelcrvation of ultrastructure, than fo r cytochemical or microanalytical in-
. 'di frozen deep-elched. rollfJ
. 5 19 Micrograph of a high resolution replica of rap1 y ed by the method of ga1ions. Finally, infiltration of the frozen-dried specimen with the
F ig. · · · d keletal elements prepar ~
<lowed cells showing coated pits an cytos H d reproduced with his penn• -.bedding medium can cause considerable structural damage, some of
. 1 ·d d by J euscr an
(198 1). (Micrograph kind y provt e .
288 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-drying
289
which may be attributable to surface tension forces created by the advanc-
ing plastic (Hanzon and Hermodsson 1960). For these reasons, in many stu-
dies the dried specimens are fixed before embedding. This can be done by
exposing the specimen to Os04 or aldehyde (e.g. glutaraldehyde or formal-
dehyde) vapour. It is obvious that vapour fixation is contrary to the original
concept of using cryofixation as an alternative to chemical fixation but it
may be tolerable for some experimental approaches (e.g. localisation of dif-
fusible substances).
A reliable procedure involving a simple apparatus for freeze-drying and
embedding small pieces of unfixed biological tissue was described in §5.3.2d
Using pieces ( <0. l mm 3) of frog sartorius muscle or human blood platelets
as model systems, it was convincingly demonstrated that, with optimised
freeze-drying and embedding methods, the quality of structural preserva-
tion is directly related to the effectiveness of the freezing stage (Edelmann
1978a,b, 1979). As summarised by the author, the following steps seem to
be particularly important in ensuring good structural preservation of chemi-
cally unfixed biological specimens:

(i) 'Pseudovitrification' of the specimen area of interest.


(ii) Freeze-drying below the recrystallisation temperature.
(iii) The embedding procedure should be performed at low temperature
(see also §7.5). Using Spurr's resin (Spurr 1968) as an embedding
medium, a temperature rise from 258 to 293 K leads to a much poorer
preservation of muscle tissue structure. Similarly, Edelmann has 1
shown that better structural preservation of the unfixed specimen was The technique of di~ectly infiltrating the frozen-dri . . .
obtained when the polymerisation was carried out at 40°C (313 K) in has the advanatage that one th f . ed specimen with resm
been put into the dryi)lg,ch eb e rozen specimen and frozen resin have
comparison to 70°C (343 K). The recently introduced resins, specifi- am er, very few subseq t . .
rtquired. To avoid bubbling of th . . uen mampuJat10ns are
cally designed for low temperature infiltration and polymerisation. e resm on warmmg ( · 11 . .
\acuurn infiltra tion), it must b d ~p especia y if usmg
should prove ideal for such applications (§7.5).
directly in the freeze-dr i~ uenitegassed befo~e freezmg. This can be done
ntrod ucing the specime y g before cooling the specimen stage and
Although it might be expected that embedding at reduced pressure would
ctdure has the advantag: ~~~~ :hs:~;rate vacu~m apparatus. The latter pro-
lead to better infiltration of the frozen-dried specimen, Edelmann (1979
be put into the freeze-dr . h ozen. specimen and the resin beads can
obtained improved structural preservation when infiltrating at atmosphe
lpecimen is to be vapour ~::d cth:~~~~ simultaneously. lf the frozen-dried
pressure. The schedule for embedding pieces of frozen-dried tissue
low that of the meltin . ts ~ust be done at a temperature be-
Spurr's resin is given in §5.5.2c and for the low temperature resins in §7
lledium can fi rst be d g po~nt of the resm. Alternatively, the embedding
At present this technique represents the simplest approach for achievnta
lien specimen from a s~e~:sr~ u;der va~uum and then poured over the fro-
optimal structural preservation of unfixed, frozen-dried specimens. M
Ian. the dried specime b o the mam vacuum chamber. Using this sys-
example of the results is provided in Fig. 5.20. n can e vapour fixed at ·
!JOStrand 1967· S ' .. t d any appropnate temperature
For . . ' JOS ran and Kretzer 1975· Pfaller 1979)
111 situ vap fi . ' ·
our ixa tton, the frozen-dried specimen has to be sealed
~---------------------
290 Robards and Sleytr Low temperature me(hods in biological electron microscopy Freeze-drying
291

. h ber Different techniques have been used


from the vacuum of the mamhc da~ ·(P"aller 9) which is commercially fixation times would be required (e.g. 24 h; Rebhun 1972). Postfixation can
. 5 21 hows one sue es1gn ,, 197 be performed (e.g. for one hour) with formaldehyde vapour, which is gener-
and Fig. . s d h. h may also be suitable for adaptation
available (see Appendix ) ~n w .~c It consists of receptacles for both the
2 ated by heating paraformaldehyde crystals to 80°C (353 K) (Bo ndareff
to other types of ~reeze-dr~n~ ;m
.d a flexible (vacuum lead-through) rod
1967).

resin and the fixative ~e.g. s 4 ;~which can be connected to the specimen For freeze-drying and embedding, special specimen tables have been
with a glass flange (at its lo':er e\ the specimen can either be exposed to designed (Hanzon and Hermodsson 1960; Sjostrand and Kretzer 1975). I t
table. Through the con~efictlmg tud e~t'h the resin which has been previously is particularly convenient for the embedding procedure if the frozen speci-
ti · our or m 1 trate w1 mens can be placed (under liquid nitrogen) directly in gelatin capsules (e.g.
the 1xat1ve .vap
h t . .
pump. F1xat1on w1·th OsO4 at reduced pressure is 5.8 mm in diameter) which fit into indentations in the holder (see Fig. 5.22).
degassed wit a ro ary hours At atmospheric pressure, longer
usually carried out for one to severa1 . To avoid a critical temperature rise during the transfer of the table on to the
cold stage, specimens are best covered with liquid nitrogen. After the speci-
mens have been infiltrated with the resin, they are removed from the freeze-
Pump drying unit for the final polymerisation step. Embedding of frozen-dried
Resin
:::@ - specimens can a lso be carried out outside the drying chamber but, especially
11·1th the more hygroscopic unfixed specimens, care will be required to avoid
rehydration.
For the analysis of membrane structures, attempts have been made to
.ombine freeze-drying with embedding at low temperatures (Sjostrand a nd
oso4 (or aldehyde) Kretzer 1975). Using this procedure, the material is first stabilised by means
f brief cross-linking with glutaraldehyde and is subsequently infiltrated
11.ith 30% ethylene glycol to prevent ice crystallisation during freezing. After
rceze-drying at -80°C (193 K), specimens are infiltrated with pre-eva-
cuated hydroxypropyl methacrylate at -30°C (243 K). Polymerisation by
means of ultraviolet radiation is performed at -27°C (246 K). It remains
vacuum
unit o be seen whether the polar and non-polar methacrylate-based lo w temper-

a b ::·-:=:~r-~~- Gelatin capsule +


frozen specimen
r: vacuum lead-through rod

fl: glass flange d-through ro d)


c : condenser (LN:;r-cooled on vacuum 1ea

st: specimen stage· b ddinJ


T see A endix 2) fo r vacuum em e
Fig. 5.21. Apparatus (Balzers BUO 20 004- , . . p~the components during freczc-d
vapour fixation of frozen dried specimens .. (a) Pos1t10n od . filtration with resin . (D1agralll 5
. t the fixative vapour an m 1 2i A typical specimen hold er for freeze-drying and vacuum embedding. T he ho lder
(b) Position dunng exposure o d d by courtesy of Balzers AG.)
drawn and repro uce can conveniently be made from aluminium .
292 Robard~ and Sleytr Low temperature methods in biological electron microscopy
-------------------·
Freeze-drying
ature resins developed by Carlemalm and collaborators (§7.5) will provide 293
1977; Barnard 1982) auto d.
similar or improved membrane images. • ra 10grapby (F d .
and Appleton 1976) and io . re enk and Klepper J976· Bak
Sections of frozen-dried , unfixed embedded specimens can be stained by n microscop y (Stika et al. 1980). , er
floating on uranyl acetate (e.g. 5% uranyl acetate, 5 min) and lead citrate
(Reynolds 1963) (see F ig. 5.20). For microanalytical studies, it is importan t
to ensure that the benefit of trapping diffusible specimen components by 5.6 !reeze-drying artefacts and
mterpretation problems ofspecimen
freeze-drying is not lost during the infiltration process (Ingram et al. 1974)
Jn this respect, structural preservation m ay even have to be sacrificed in
As already indicated in the . .
favour of reduced translocation of diffusible specimen constituents (e.g. . . previous sections s e .
ferences in the1r 'sensitivity' t th . . . , P cimens can show great d"f- 1
electrolytes). This is certainly a field in which only systematic work on o e md1v1dual t f
dures. The most common reasons fi s eps o freeze-drying proce-
defined model systems will help to resolve the numerous questions. or artefact formation are:

5.5.2d Freeze-drying of cryosections 5.6.l Fixation artefacts

Freeze-drying of cryosections (§4.3.2) may be performed either at atmos- Specimen fixa tion may be .
req uned as a prer .
pheric pressure or under vacuum (Christensen 1971; Appleton 1974; Bar- of bulk specimens. In particular gl t I muna ry step for freeze-drying
· , u ara dehyde fi t" .
nard and Seveus 1978). In either procedure, the dry sections (cut without specimens more resistant to d . ixa IOn is used to m k
. . amage by ice er t I d . a e
a trough liquid) are first picked up on coated grids. Drying at atmospheric shrinkage followmg ice sublimat· d . ys a s urmg freezing and to
h !On urmg dry · (§ 5
presure is most conveniently done directly in the cold nitrogen atmosphere c anges p roduced by fixation a re r t d . mg .5.2). The structural
of these artefacts w ill not be reveale~ ef m fTa ble 2. 1_2 in §2.3.2, but several
of the cryochamber with the specimen grid attached to a temperature-con- a ter reeze-drymg.
trolled cold block, or within a holder containing molecular sieve. Drying
under vacuum involves a contamination-free transfer of the cryosection {62 Adrorption artefacts
from the cryomicrotome to the pre-cooled stage of the freeze-drying unit
(Spriggs and Wynne-Evans 1976; Geymayer et al. 1977). During this Strong electrostatic interactions b t
branc vesicles) and the support ( e ween_ very flexible specimens (e.g. mem-
transfer, special care is required to maintain the specimen below the critical ca· §5 5 I ) e.g. positively cha d
recrystallisation temperature. After freeze-drying (for drying times see Tablt .... a may induce collapse d. . rge , coated grids or mi-
and Williams 1977; Nermut l98la~r istort1o n during adsorption (Nermut
5.2 in §5.2. la), the sections can be carbon-coated within the drying unit t
In contrast colla se . .
reduce rehydration from atmospheric moisture during subsequent transfer bse , ~ may aid better structural . .
into the electron microscope (Zingsheim 1984). Zingsheim (1984) also found quent preparation steps (T .. k . preserva tio n during the
nmc and Elliott 1982) .
that the sublimation rates of ice from cryosections in a cryoultram icrot~
6 J Washmg artefacts
were about four to five orders of magnitude slower that they would be under
vacuum. This means that there is less worry that such cryosections will inad-
Since (pseudo) t ·
vertently freeze-dry. lim . eu ect1c layers generated around .
F rozen-dried cryosections of most unfixed biological specimens are ~ e ~t a lower rate than the ice the r _1c~ crystals during freezing
suitable for high resolution morphological studies, but they are of uniqll r With the non-volatile d b . ' emammg eutectic network to
Iller et al. 1983) T e ns, can mask the frozen -dried s .' -
value for depicting the distribution of diffusible substances (such as elccUO" · o reveal 'non con . pec1men
ng. specimens must be rinsed . d-. . tammated' surfaces during freeze
lytes or specimen components which are lost during chemical fixation
ng. Especially if unfixed s m_ istiiled water or a vola tile buffer befor~
dehydration with organic solvents as used for conventional thin secti artefact (l · pecimens are used th . ·
procedures) by X-ray microanalysis (Baker and Appleton 1976; C s ys1s, structural alt , . .' is technrque can lead
h chan · erations, extraction f ·
ge m pH, temperature differe o_ constituents, etc.)
nces, osmotic shock and differ-
294 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-drying
295
ences in ion distribution. Therefore, washing should be for as short a time
5.6.5 Freezing and recrystallisation artefacts
as possible. If specimens were originally suspended in solutions of high salt
or sugar content (e.g. from a sucrose gradient), repeated washing in (double) As discussed in detail in Chapter 2 th '" .
distilled water will be required to ensure the complete removal of non-vola- · , e iormatwn of ic I .
freezing may cause severe morphological chan . . e crysta s durmg
tile components. On the other hand, suspended specimens mounted by cur if initially well-frozen specim ges. Similar artefacts can oc-
adsorption (§5.5. la) may become detached during washing with disti lled ens exceed the recryst II" .
during the transfer to the cold sta e of h ~ isat1on temperature
water due to changes of ionic strength. This can be particularly important period of drying (e.g. due to radi·agt. ht e :reeze-drymg unit or during the
if statistical data are to be accumulated. A parallel preparation using nega- ract betwen the specimen and th
wn
Id
eatmg and/o h
r a poor t ermal con-
tive-staining methods after each preparation (washing) step will detect this specimens will only be disturbinge.fco stalge). lee crystal formation in bulk
' desorption' artefact. . l crysta s grow thro h th .
face (ice crystal puncture artef:act ) .f f ug e specimen sur-
s or 1 racture f; .
freeze-drymg techniques. aces are examined by
5.6.4 Air-drying artefacts Dependmg on the specimen texture sha d .
mg may occur during freezing Th. 't '" pe an_ size, thermal stress crack-
With most specimens, the surface tension forces created at the vapour/liquid . · is ar e1act, which c b .
lar.ly with large specimens is caused b th d. _an e expected part1cu-
(water) interface during air-drying cause severe structural changes (e.g. col- . ' Y e 1fferent1a l d 1· ·
in ice and ice-encapsulated water t . mens1onal changes
lapse, disruption, aggregation, dimensional changes). Specimens dried b1 a various tempe , t h
ness of the specimen (Boyde 1978 . ) Th ra ures t rough the thick-
solvent evaporation generally appear smaller than the equivalent frozcn- . a · ermal stress er· k ·
happen wi th hydrated specimens b t I . ac mg may not only
dried specimens (Fig. 5.23). Air-drying can occur in the interval between · u a so on freezmg ·
had their water replaced with 0 . specimens that have
rganic so1vents Th . .
specimen mounting and freezing or during transfer of the frozen specimen duce the specimen size. · e o n 1Y soluti on 1s to re-
to the cold stage of the freeze-drying unit, where drying is preceded by speC1 At too high freeze-drying temperatures m It'
men melting. Special care is required with specimens that are frozen as g the ice crystals may occur F , _e m~ of the eutectic surround-
t hin layer (§5.5. la). With bulk specimens (§5.5.2a) it must be ensured that . or most brologrc I ·
lemperature for eutectic melti"n . l a specimens the critical
a thin film of water (or other volatile solvent) covers the specimen surfa g is c ose to - 35°c (2 38 K
hove the critical recrystallisati·o t ) and thus far
at the moment of freezing. n emperature Eut t' .
oughout the frozen portion of th . . ec ic meltrng occurs
liquid. This means that evapor· t" e spefc1men and liberates large volumes
. a ion o water oc ·
sublimation of ice In comm .b . curs simultaneously with
· on wit a1T-dry" f
a b c d e surface tension forces created t th . , mg o non-frozen specimens
100% ..., h a e mtertace b t h '
80% _. I e molten eutectic can cau e wen t e vapour phase
2· 1975a; Boyde 1978b). se severe collaps h
35% e P enomena (MacKenzie

- r4?1 r4?1~ .6 Dehydralion artefacts


~~~
I
,,,,.,,,.,,,,.-
,,.
Specimen Freeze-dried Critical Air-drie d Air-drl•• I kinds of artefact can be a ttributed t
at start from water point- fro m F 11 3 from pse phenomena (frequently . o the freeze-dehydration stage
wtttr associated with th h · ·
s ) r.epresen t the ma1· t
dried
e s nnkage of bulk spe-
n s ructural change th
Fig. 5.23. Diagram, drawn to scale, showing cubes of tissue (a) frozen-dried from WI
atton abo ut them h . . at can be expected. Valuable
ec amsms mvolved h b b .
critical point-dried (c), ai r-dried from a volatile non-polar solvent (trichlorotriOu~ s and, particularly by f. II . as een o tamed from model
Freon 113) (d), and from water (e). Data were obtained from the results of Bo)dc ~ith specially desi~ned l~gh~wi~g the freeze-drying process by obser-
(1981) and from Boyde a nd Maconnachie (1979a,b). (Redrawn from Boyde and Fr.lllC 1975a, 1976· Nei and F .. microscopes or in the SEM (MacKenzie
' UJ1kawa 1977a,b; Boyde 1978b; Boyde and
~---~-----------------
d~ in biological electron microscopy
Low temperature metho ·
1 Freeze-drying 297

296 Robards an d SIeytr very much when frozen-dried at -90°C but became distorted if warmed
h d' t vacuum sublimation of ice and much above - 80°C (Heuser and Kirschner 1980). Furthermore, shadowing
F ranc 1981). It was shown that t eh tTech condensed amorphous barriers, and replication of frozen-dried specimens is best done with the specimen still
f water vapour t roug JI
the disappearance o . 11 intact Cracking, co apse and at this, or even lower, temperature. Less rigid structures, such as bacterio-
· structure essentia Y ·
leaves the specimen d minantly after the passage of the phages (polyheads) or cylindrical protein self-assembly structures (Kistler
b hown to occur pre 0
shrinkage have een s . f water (e.g. during the release and Kellenberger 1977; Kistler et al. 1977; Sleytr 1978; Kellengberger and
during the desorption o
freeze-drying f ron t . h' 'nt it should be remembered that Kistler 1979) may collapse during adsorption, during freeze-drying, or dur-
f om protem) At t 1s poi d
of unfrozen wat er r · ecimen temperature an water ing shadowing and replication. A pseudoreplica of a frozen-dried self-as-
· is dependent on sp .
desorption beh av1our . ( lative humidity). Thus, shnnkage sembly cylinder composed of bacterial cell wall surface proteins (Sleytr
. the specimen area re f d.
vapour pressure m . d b 'dable if the partly rozen- ned 1978; Sleytr et al. 1982) partially collapsed onto the support is illustrated
t10n shoul e avot
artefacts due to desorp t which freeze-drying is conducted in Fig. 5.24. The non-attached part of the cylinder is extensively shrunken
specimen can be held at the temper~thure a ect to that temperature (Boyde and shows less structural detail than the part that has collapsed onto the
· humidity wit resp
and at 100% re1attve . k wrinkling folding and collapse) support film. Such an elevated, dehydrated, protein framework can be
1978a). Artefact problems (sh~m age, d ther s~all biological structures expected to have poor thermal conductivity and frequently it may collapse
f
related to the reeze- r
d ying of vtTuses an °
1977 1981 a). In practice, these small spec1- or even disintegrate from the beat load imposed during shadowing and rep-
have been reviewed by Nermut ( ' d . d at _90°c (183 K) than at ltcation. Similar effects are known from freeze-fracture replication techni-
· k when frozen- ne
mens exhibit Jess shnn age tures do not appear to collap~c ques and should be considered in this context (§6.8.3). Even wi thout this
- 7ooc (203 K). Similarly, cytoskeleton struc
energy input, extended, protruding areas of frozen-dried specimens (e.g.
bacterial flagella) are likely to vibrate under vacuum due to thermally

''
''
''
....
/~,
''

5.25. Schematic diagram showing the changes in linear d imensions of specimens during
'-it-drying. Al the beginning, at the left, ice is outside the s pecimen. The specimen reach es
ongmal true specimen dimensions at t he second vertical arrow from t he left. T hese dimen-
are reduced by 6% linear when all the ice has sublimed o ut of the specimen and by 7%
tllm the remaining fi rmly bound water hl\s left the specimen . The diagram indicates that the
of shrinkage may be related to the po re volume within the tissue created by the ice crystals
lflidi are larger at deeper levels within the specimen. However, this different rate of shrinkage
also be related to the change in the mass of the ice within the s pecimen. A smaller mass
~would cause less self-cooling by s ublimation . (Redrawn from Boyde and Franc 198 1.)
298 Robards and Sleytr Low temperature me(hods in biological electron m icroscopy
Freeze-dry ing
299

induced movement (Anderson 1954). This may either lead to an irreversible


collapse onto the support or may cause the formation of blurred 'shadows'
(Kellenberger and Kistler 1979). However, collapse (flattening of structures)
may be advantageous if regular arrays of the macromolecules which make
up hollow structures are to be evaluated by image ana lysis methods (Kistler
et al. 1977). Chemical fixa tion (e.g. with glutaraldehyde) can render flexible
structures more resistant to collapse (Nermut l98la).
Especially with bulk specimens, an overall dimensional change can be
expected during freeze-drying. Volume changes of biological samples during
different p reparation procedures for the SEM, including air-drying from
water, air-drying from organic solvents, critical poin t-drying and freeze-dry-
ing have been carefully analysed by Boyde and his colleagues using d irect,
quantitative on-line measu rements (Boyde l978b, 1980; Boyde and Macon-
nachie 1979a,b, 1980; Boyde and Franc 1981). A variety of biological sam-
ples (mostly glutaraldehyde-fixed) have been shown to sh rink linearly by
about 7% (20% by volume) during freeze-drying (Campbell and Roach
1983). As can be seen from Fig. 5.23, this is still considerably less than can
be expected from the alternative dehydration procedures. Using l .O mm
L·ubes of glu ta raldehyde-fixed mouse li ver as a model system, it was shown
1hat most of this shrinkage ( ::= 6% li near) is associated with 'secondary' d ry-
ng when the firmly bound water, which can be up to about I 0% of the spe-
cimen dry weight, is removed (Fig. 5.25).
Freeze-dryi ng artefacts may a lso arise from the pressure- and tempera-
urc-dependen t expansion of gases dissolved in the specimen water prior to
rcezing. Finally, at too high freeze-drying tempera tu res, macromolecular
particles may be lost from the specimen surface or fro m the withd rawing
let fro nt along with the dynamic stream of subliming water molecules (water
pour wind) (Anderson 1954, 1956) and isolated particles that have been
nm berated from the ice may d iffuse la terally on the frozen surface and re-
arrange into aggregates (Lepault and D u bochet 1980).
. ~ d with mouse leukemia · virus as visualised
Fig. 5.26. Structural features of. cells m ecte a Cells freeze-d ried for SEM are rUSOlll" 6.7 Rehydration artefacts
freeze-drying (a, d) a nd critical pomt-:rym~b, ~~~u~ (arrows) and t he protrusions an~
bly well preserved. Some fract ures an crl~c n~ I point-dried for SEM show well-p
seen in (b) are largely destroyed. (b) Ce s en 1ca I 'st.ble (c) After c ritical-point drytlll rozen-dried specimens must be warmed up to ambien t temperature before
p rotrusions a nd fibres. Aga1~, t. e v1r
· h · uses are c ear Y vi · dd g
) are still faintly visible but early bu an king the vacuum of the freeze-drying unit so that water condensa ti on
T
EM the viral surface proJect10ns (knobs I . . hampered by the dehydrauon
( • · ble Reso utton is f voided. Most frozen-dried specimens are very hygroscopic. A potential
kno bs assem bly) are rarely recogmsa . . .th critical point-d rying (c). recz
involving organic . soIven ts. (d) In comparison
. r TEMWlwork . K nobs, even .m ear Iy buddanl e of specimen rehydra tion is the relatively hig h water vapour press ure
clearly gives better structural resolut10n fo I I at higher magnification in the low;M is inevitably near to the surface of any man ipula tor (forceps o r other
( arrows) a re easily observable (shown more c eardy . ·is the method of choice for T
ling devices). Specimens are therefore either transferred immedia tely
graphs). 'In conclusion, for t h is
· specim reservation. (Micrographs ki ndly
· en , freeze- II rymg
. . · b tter for SEM a nd overa P . .
critical pomt-drymg is e d 'd with his penmss1on.)
by H . Fran k and repro ucc
300 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-drying
301
into the microscope or are stored in a good desiccator until use. Rehyd. microscopy: a morphological and X .
' -ray analytical st d J M.
ration can be slowed down by coating the specimen with a thin layer of car- Backus R. C. and R.C. Williams ( 1950) Th
. . . ,
f u!• · 1croscopy 100, 49.
e use o spraymg mcth 0 d d f .
bon. mg med ia m the preparation of specim r
1
.
ens 1or e ectron m1crosco
s an o volatile suspend-
J
Baker, J.R.J. and T.C. Appleton ( 1976) A t h . · PY, · app1. Phys. 21, 11.
. • ec niquc for electro · , .
(and X-ray microanalysis) of di ffu si.bl b n microscope .iutoradwgraphy
5.6.8 Shadowing and/or replication artefacts · e su stances using f , d ·
J. Microscopy 108, 307. recze- ned fresh frozen sections.
Barnard, T. (1982), Thin frozen -dried er osecti . . .
Structural alterations caused by thermal load during shadowing and replica- copy 126, 317. y ons and biological X-ray analysis, J. Micros-
tion are discussed in §6.4 and §6.8.3. Bamard, T. and L. Seveus (1978), Preparation of b. . .
1
of diffusible elements, J. Microscopy I 12, 281. io og1cal ma ten al for X-ray microanalysis
Bendel, U and N . J. Rizk (1974) T hermo •1 t . fi .
5.6.9 Freeze-drying: structural comparisons with other dehydration • e ec nc reeze-drymg J M '
Blaschke, R. and A. Boyde ( l9l 8 ) P a . . . • · 1croscopy /OJ , 311.
procedures , r1IC1e size m conduct' .
SEM , Scanning J, 64 Ive coatmg and resolution in the
Bondareff, W. (1967) Demonstration of ·
' a n mtercellular sub t ·
Comparative studies on bulk specimens have shown that specimen shrink- z. Zellforsch. 81, 366. · s ance in mouse cerebral cortex,
age during critical point-drying is always greater than that fro m freeze-dry· Boyde, A. (1 972), Biological specimen preparation for the . .
overview, Scanning Electron Microscopy , _ scanning electron microscope: an
ing (Boyde 1978a,b, 1980; Boyde and Maconnachie 1979a) (see Fig. 5.23) 1972 257
Boyde, A. (1974a), Freezing, freeze-fracturing and f d . . . .
Nevertheless, for many specimens it appears useful to compa re the struc ration for the SEM Scan ning Electron M. rceze- rymg In b1olog1cal specimen prepa-
tural preservation achieved by both techniques. Fig. 5.26 shows an example ,. ' .. 1croscopy 1974, 1063.
Bo1dc, A. (1 974b), Volumenanderung biologischer Gewebe b . ..
of structural differences in leukaemic cells after freeze-drying or cnhcal Be1tr. elektronenmikr. Direktabb. Oberfl. ?, 1.
23
et Entwasserung und Trocknung,
point-drying. Whereas protrusions and fibres from the virus-producing cell Btl)de, A. (1 975), F reeze-drying in the preparat' f . .
appear to be better preserved after critical point-drying, the different bud
SEM, in: Publication no P l (Edward H'gh v' ion
o. d A . s I
° anunal tissue a nd cell specimens for
acuum Ltd Crawley)
"''Y e. · (1976), Do's and don ' t's in b. 1 . 1 . " ·
ding stages are more clearly depicted in replicas of frozen-dried cells (Fran~ · 10 og1ca specimen pre . t' ~
Electron Microscopy 1976, / , 683 . para ion or the SEM, Scanning
personal communication). ll<'}de, A. (I 978a), Pros and cons of critica l point d . d
Flectron Microscopy 1978, 2. 303. ry1ng an freeze-drying for SEM, Scanning
Boi de, A. (1 978b), Volumcnandesungen biologischer Garebe . .
nung. Be1tr. elcktronenmikr. Direktabb Obe fl be1 E utwasserung und T rock-
References Bt11de A (1980) R . . · r · ll,23 1.
. . . ' ev1ew of basic preparation techni . .
ro•copy, Proc. 7th. Eur Reg Conf El t M' ques for b10log1cal scanning electron mic-
Albrecht, R.M., D .H . Rasmussen , C.S. Keller and R.D. Hindsi!I (1976). Preparauon ofc Bo1dc. A. and P. Echlin ( ;973). F _-. ec ron tcroscopy, The Hague 2, 768.
tured cells fo r electron microscopy: air drying from organic solvents, J. Microscopy /OJI . • reezmg and freeze-dry' . .
xa.nnmg Electron Microscopy 1973, 759. . mg. a preparative technique for SEM,
Anderson, T.F. ( 1951), Techniques for the preservation of three-dimensional structure in
Ide, A. and F Franc ( 1981) F . d . .
paring specimens for the electron microscope, Trans. N.Y. Acad. Sci. 13, 130. o-.:opy 122, 75 ' reeze- rymg shn nkage of glutaraldehyde fixed liver. J . Mic-
Anderson, T .F. (1954), Some fundamental limitations to the preparation of three-d1mcn"
de. A. and E. Maconnachic (l979a) Volume .
specimens for the electron microscope, Trans. N.Y. Acad. Sci. 16, 242. cm~ryonic tissue for scanning elect '. changes dunng preparation of mouse
Anderson, T .F. (1956), Electron microscopy of micro-organisms, in: Physical technique de. A. and E M . ron microscopy, Scanning 2, 149.
biological research, Vol. 3, G. Oster and A.W. Polister, eds. (Academic Press, Ne" Y . . aconnach1e (1979b), Freon 113 free - . .
P). Scannmg 2, 164. ze drymg for scanning electron micros-
p. 177.
e. A and E. Maconnachie ( l 980) Treat . . .
A nderson, T .F. (1966), Electron microscopy of micro-organisms, in : Physical techniqur •on shrinkage of gluta raldehyde Ii , d . ment _wit h lithium salts reduced ethanol dchyd-
biological research, Vol. 3, A .W. Polister, ed. (Academic Press, New York), p. 319 A. and C W xe tissue, H1stochem1stry 66 18 l
· ood ( 1969) p · • ·
Appleton, T.C. (1973), Cryomicrotomy; possible applications in cytochemistry. in E ~~I • reparauon of animal tissues for surface SEM J M'
• . 1croscopy
microscopy and cytochemistry, E. Wisse, W.Th. Deams, l. Molenaar and P. van DutJIL
EP. andK .F Hale ( l98J) D . . .
(North-Holland Publishing Company, Amsterdam), p. 299. methods in electron micro~co~~a~~lex:e~";,'.;nts in the elect ron microscope, in: Practi-
Appleton, T.C. (1974), A cryostat approach to ultratbin "dry" frozen sections for tcrdam). ' · • · · G lauert, ed. (Elsevier/ North-Holland,
302 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-drying
303
Cam pbell, G.J. and M .R. Roach (1983), D imensional changes associated with free ze-drying
Geymayer, W., F. Grascnick and Y. Hodl ( 1977) S. bT ·· .
of the internal elastic lamina from cerebral arteries, Scanning 5, 137. drying, J. Microscopy fJ 2, 39. ' ta 1 ismg ultrat hrn cryosections by freeze-
Chandler, J.A. ( 1973), The use of wavelength dispersive X-ray microanalysis in cytochemistry.
Glauert, A.M. (1974), F ixation dehydratio d b . .
in: Elect ron microscopy and cytochemistry, E. Wisse, W.Th. Deams, I . Molinaar and P. van ' n a n em edd1ng of biolo · l ·
tical method s in electron microscopy V
13
A gica specimens, in: Prac-
D uijn, eels. (North-Holland, Amsterdam), p. 203. dam). , o. ' .M. G la uert, eel. (North-Holland, Amster-
Cha nd ler, J .A. (1977), X- ray microanalysis in the electron microscope, in: Practical methods Goldblith, S.A., L. Rey and W W Roth ( d
in elect ron microscop y, Vol. 5, A.M. Glauert, ed. (Elsevier/ North-Holland, Amsterdam). · · mayr e s.) (1975) F d ·
technology (Academ ic P ress, London). ' reeze- rymg and advanced food
C handler, J.A. and S. Battersby (1976), X-ray microana lysis o f ultrathin frozen and freeze.
Gregory, D.W. a nd B.J.S. Pirie (1972) Wetti· r
dried sections of human sperm cells, J. Microscopy 107, 55. . ' ng agen ts •Or electron ·
specimens, Proc. 5th E ur. Reg Conf Elect M. microscopy of biological
C hristensen, A.K. (197 1), Frozen t hin sections of fresh tissue fo r elect ron microscopy, Wllh · · ron 1croscopy Manch 1
Hanzon, V. and L.H. Hennodssen (1960) F d . ., es er, p. 234.
a description of pancreas and liver, J. Cell Biol. 51, 772. • reeze- rymg of tissue for r ht d I
copy, J U ltrastruct. Res. fl, 332. ig a n e ectron micros-
Coh en, A.L., D .P. Marlow and G.E. Garner ( 1968), A rap id critical poi1 .. method using nuor- Hayat, M.A. (1981), Fixation fo r electron mic
ocarbons (Freons) as in termediate a nd transitional llu id s, J. M icroscopie 7, 33!. Haya!, MA. an d B.R. Zirk in ( 1973) Critical pro.sctodpy (Academic Press, New York).
Davy, J.G. ( 1971), Vacuum m icrobalancc apparatus for rapid determination of low-tempera- • om rymg meth od in· p · · l
es of electron microscopy Vol. 3 MA H . • · nnc1p cs and techniqu-
t u re vapourizatio n rates, Vac. M icrobalance Tech. 8, 155. p. 297. , , . . ayat, ed . (Van Nostrand Reinhold, New York).
Davy, J .G. and D. Branton (1970), Subliming ice surfaces: freeze-etch electron microscop1 Heckly, R.J a nd D. Skilling (1979) Device to r . . Ii
Science 168, 1216. . ' 1ac111tatc rccze-drying · r
electron microscopy, Cryobiology 16, 196 . specimens 1or scanning
Derrick, K.S. (1973), Q uantita tive assay for p lant viruses using serologically specific electron
Heuser, J.E. ( 1980), 3-D visualiza tion of coat d . C
1
m icroscopy, Virolo gy 56, 652. Heuser, J.E. (1 981), Preparing biolooical s· e I ve~1c e orm, J. Cell Oiol. 84, 560.
Dudek, R.W., A.F. Boyne and N . F reinkel ( 198 1), Quick-freeze fi xation and freeze-dryrn~, .,. amp es 1or stereom1croscopy b · k f
e1ch, rotary replication technique in· Meth d . . Y q u1c - reezc. deep-
isolated rat pancreatic islets, J . H istochem. Cytochem. 29, 32!. . ' · o s m cell biology Vol 3 JN T
dem1c Press, New York), p . 97. ' · ' · · umer, ed. (Aca-
Echlin, P., B. Chapman, L. Stoler, W. Gee and A. Burgess ( 1982), Low voltage sputter coaling Heuser, J.E. (1983). P rocedure for freeze-d .
Scann ing E lectron M icroscop y 1982, /, 29. Biol. 169, l 55. ryrng molecules adsorbed to mica Oa kes, J. m ol.
Edelmann , L. (1978a), Visualization and X-ray microanalysis of p otassium tracers in frm Heuser. J.E. and M.W. K irschner 0 980) F'I . .
dried a nd plastic embedded fro g muscle, Microscop ica Acta Suppl. 2, l 66. • 1 ament o rgan1sat1on revea led · I ·
offreeze-dried cytoskeletons, J. Cell Biol. , _ m P atmum replicas
86 2 12
Edelm an n, L. ( 1978b), A simple freeze-drying technique for preparin g biological tissue \\1lh Heuser, J.E. and S.R. Salpetcr ( 1979) 0 . .
chemical fixation for electron microscopy, J. Microscop y J12, 243. • rgarnzat1on of acetylchor 1· · k
etched. and rotary replicated To. d . me n qwc -frozen eleep-
Ed elman n, L. ( 1979), Freeze-drying of chemically unfixed biological material for electron m 1pe o post-synaptic membra J C II B' ,
Humphreys, W.J. (1975), D rying soft biol . ·I . ne, . e wl. 82, 150.
roscopy, Mikroskopie 35, 3 1. og1ca tissue for scanning cl 1 . ·
nmg Elcctron Microscopy 1975, 707 _ ec ron m1croscopy, Scan.
Edelmann, L. ( 1980a), Preferential localized uptake of K .,. a nd Cs + over Na + in the A-ba
Humphreys, W.J., B.O. Spurlock and J.S. Johnson l . . .
of freeze-dried embedded muscle section : detection by X-ray microanalysis and laser m1.: mliltratecl cryofractured biolog· I , . ( 974), Cn tical pom t drying of ethanol
probe mass ana lysis, Ph ysic!. Chem. P hys. 12, 509. El . ica specimens for scanning el ·t .
ec1ron Microscopy 1974, I , 275. ec ron microscopy, Scanning
Edelmann, L. (l 980b), Potassium binding sites in m uscle: electron microscopic vi ~uah1at
gram. F. D.. M.J Jngram and C.A.M. Ho ben 1974 .
of K , Rb and Cs in freeze-d ried preparations a nd a utoradiography a t liquid nitrogen t t'lllbcdded electron probe sp . g . ( ), A n analysis of the freeze-dried plastic
ec1men preparat10n · . M' . '
peratu re using 86 R b a nd 134Cs, Histochemistry 6 7, 233. T Hall, P. Echlin a nd R Kauf d , m. icroprobc analysis as applied to cells
Edelmann, L. (198 1), Selective accumulation of Li ~, Na+, K +, R b • and Cs-+ at prolem · mann, e s. (Academic p L d ,
•ta. II. and S. Aita (1976) F . . ress, on on , New York), p. 11 9.
of freeze-d ried and em bedded muscle detected by LAMMA, Fresenius z. Anal. Chem od fi ' reezc-dryrng technique for sm· II b" I .
Ill I ied freeze-etching device J E l t M. a 10 og1cal materials using a
218. opic E A B ' . ec ron icroscopy 25 305
. .. , . runegger, R. Essl and G. Windisch (1978) ' . .
F ischer, K.A . (1 975), " H air' memb rane en richmen t: verification by electron micro<.cop1 l'OScop1c preparation, Proc. 9th Tnt. Con r E . ' A sputter source fo r electron mic-
ence 190, 983. kcncnbcrgcr, E. and J. K istler ( 1979) T h gh l~ctron Microscopy, Toron to 1, 150.
Fowler, W.E. and U . Aebi ( 1983), Prep aration of single molecules and supramolecular Ill electron microscopy for mo le , I· e p ys1cs of spec1_mcn preparation, in: U nconventio-
plexes for high resolution shadowing, J. Ultrastruct. Res. 83, 319. 1'~ !Friedr. Vieweg a nd Sohn Wcubardstru)cture determmation, W. H oppe and R. Mason
Frederik, P.M. and D . K lepper (1976), T he possibility of electron microscop ic auto -,. R 0 , 1es a en , p. 49. ,
· ·· R.A. F. Dekker and J G Bl · k .
phy of steroids after freeze-dryin g of unfixed testes, J. Microscopy 106, 209. tpphcation in coating biological~ . . uem1; (1973), Ligand -media ted osmium binding: its
Gersh , I. (ed.) ( 1973), S ubmicrosco pic cytochemistry. I. P roteins and nucleic acids. (A la 45, 254. pec1mens o r scanrnng electron microscopy, J. U ltrastruct.
P ress, London ), p. I. '-menoe BH
' · · and G. R . Bullock ( 1983) St
Iler-coated lilms a co . . . ' ruct ure analysis of sputter-coated and ion-beam
mpara t1ve study, J. M icroscopy 132, 153.
~--------------llllllm ........
304 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-drying
305

Kistler. J. and E. Kellenberger (1977), Co11 apse p h en omena in freeze-drying, J. U ltrastruct Nei, T. and S. F ujikawa {I 977a), Observation of the freeze-drying process of biological materi-
als with a scanning electron microscope, Dev. Biol. Stand. 36, 243.
Res. 59,
Kistler 70. (
J., Y. Aeb and E. Kellenberger l 977 ,
) Freeze-drying and shadowing a two-dimensio.
Nei, T. and S. Fujikawa (1 977b), Freeze-drying process of biological specimens observed with
' . . · n J Ult rastruct. Res. 59, 76. a scanning electron microscope, J. Microscopy 111, 137.
nal periodic spec1me , · .. . b. logischer Objekte fiir das Rasterelektro-
Kl . Hp and W. Stockem ( 1976). Praparauon io Nerrnut, M.V. {1 973a), Freeze-drying for electron microscopy, Balzcrs High Vacuum Report,
em, . · · · ica Acta 78, 388. . . VB2.
nenm1kroskop, Microscop . f tu . g and etching artefacts m particulate sus.
Lepault J . and J. Duboehet (1980), F reezing, rac rm Nermut, M.V. (1973b), Freeze-drying and freeze -etching of viruses, in : Freeze-etching, techni-
72 223
pensions, J. Ultrastruct. Res. • · . . observations during freezing, Biodyna- ques and applications, E.L. Benedetti and P. Favard, eds. (Soc. Fran~aisc de Microscopic
MacKenzic, A.P. (1964), Apparatus for m 1croscop1c Electronique, Paris), p. 135.
Nermut, M .V. (1977), Freeze-drying for electron microscopy, in : Principles and techniques o f
K 9, .2 13.
mica . t h e mechanism
A p (1965) Factors affecting . of transformation of ice into water electron microscopy, Vol. 7, M.A. Hayat, ed . (Van Nostrand Reinhold Company, New
Mac enzie, . . , . rocess Ann. N.Y. Acad. Sci. 125, 522. York), p. 79.
vapour m the freeze-drymg P · ' h . n the freeze-drying process, Cryobiolog)
MacKenzie, A.P. ( 1967), The collapse p enomena I Nermut, M .V. (I 98 la), Advanced methods in electron microscopy of viruses, in : Newer techni-
ques in practical virology, C . Howard, ed. (A. Liss, New York).
3, 387.
MacKenzie, .
A.P. ( 1972), Freezing, freeze-d rymg
. and freeze-substitution , Scanning Electron
Nermul, M.V. (l98 lb), Visualization of the " membrane skeleton" in human erythrocytes by
freeze-etching, Eur. J. Cell Biol. 25, 265.
Microscopy 1972, 2, 273. . f d ·ng _ qualita tive and quantitative aspect'
. A p ( 1975) Collapse durmg rccze- ry1 \lermut, M.V. and H. Frank (197 1), Fine structure of influenza A2 (Singapore) as revealed
MacKenzie, · · • · h 1 SA Goldblith, L. Rey and W.W. Roth·
in: Freeze-drying and advanced food tee no ogy, . . by negative staining, freeze-drying and freeze-etching, J. gen. Viral. IO, 37.
d (A demic Press New York), P· 277.
." ... les of freeze-dry ing, Transplantation Proc. 8, 1 ~ 1
'\ermut, M.V. a nd L.D. Williams ( 1977), Freeze-fracturing of monolayers (capillary layers)
mayr, e. s. ca
MacKenzie, A.P. (1976), Prmcip . . . t" n technique for biological X-ray nm of cells, membranes and viruses: some technical considerations, J . Microscopy 110, 121.
Marshall, A.T. ( 1980). Frceze-subst1tul!on as a prepara io '\ermut, M V. and L.D. Williams, (1980), Freeze-fracture autoradiography of the red blood
. S . El ·tron Microscopy 1980, 2, 395. ceU plasma membrane, J. Microscopy 118, 453.
roanalys1s, canning ec . ) Adhesion of cells to surfaces coa ted with pol) II
M -. D G Schatten and W. Sale (1975, '\ermut, M.V., H. Frank and W. Schafer (1972), Properties of mouse leukemia viruses. JJJ.
azia, ·• · . C II Biol 66. 198.
sine. Applications to electron microscopy. J. _e . demic Press London), p. 3. Electron microscopic appearance a s revealed after conventional preparation techniques as
M II r J D . ( 1978) Fundamentals of freezc-drymg, (Aca 's . 130 6?8 11ell as freeze-drying and freeze-etching, Virology 49, 345.
e o ' . , . i n freeze-drying without vacuum, c1encc , - .
Merymann, H.T. ( 1959), Subhmat o. . - b. I HT Meryman , ed. (Academic Pr~' '\1etsch, W. and E. Jochem ( 1973), Gcrust- und Grcnzf1iichenwidcrstand bei der Gefriertrock-
Mcrymann , H .T · ( 1966) • Prccze-drymg,
· m: Cryo io ogy, . - nung dunner Schichten, Vakuum-Techn ik 22, 73.

New York), p. 609. NL L· . a l (! 9 83), Artefacts associated '"'h O~takc, H., K. A toh, A. M itsush ima, K. Tanaka and H. Takahashi ( 1980), A simple method
Miller K.R. , C.S. Prescott, T. L. Jacobs a nd · · assign of freeze-d rying for scanning electron microscopy, J. Electron Microscopy 29, 72.
' . d f _ d · g J U ltrastruct. Res. 82, 12 3 ·
quick-freezing an recze- rym ' . . (1 980) A simple freeze-drying mcth1• Pearse, A.G.E. {1 964), Ein thermoelektrisches Geriit zum raschen G efriertrocknen biologischer
. H Ozasa T. K ishida and A. Suganuma , . Gewebe, Die Kalte 17, 75.
Moro1ka, H., · ' . J Electron Microscopy 29, 64.
for electron microscopy of bactcnophagcs, .· . thode an Pflanzenzellen, J Uhr Pttcrs, K.R. (1 977), Stereo surface replicas of cultured cells for high resolution electron micros-
Miiller, H.R. ( 1957). G cfriertrocknung als F1X1erungsme copy, J. Ultrastruct. Res. 61, 1 15.

truct. Res. I , 109. f - nductivity in electron microscopy 5,,1nni Pttcrs. K.R. (1 979), Scanning electro n microscopy at macromolecular resolution in low energy
Munger, B.L (1 977), The problem o specimen co mode on biological specimens coated with ultrathin metal fi lms, Scanning Electron Micros-
48 copy 1979, 2, 113.
Electron Microscopy 1977 · I , 1. . . · of biological specimens for 0
. T ( 1978) Tannin-osmmm conductive stammg •lier, W. ( 1979). Freeze-drying and vacuum embedding of spray fozen unicellular organisms,
M urakami, · · I 127
M1kroskopie 35, 37.
coated scanning electron microscopy, Scanning , d ~iological specimens conduct11~ I
h J A (1980) Non-coating techniques to ren er Podolsky, M.V and J.A. Konstantinov ( 1980). A study of the final period of freeze-drying
Murp y, · . ' · 1980 I ?09
update, Scanning Electron Microscopy • ,- . . l subunit of Baci/111.1 whtt/11 p llld determination of residua l moisture of dry biological materials, Cryobiology 17, 585.
. . N · (1968) • The conformation of the SOS nbosoma
Nanmnga, d. P.L. (1963), Sorption pumping at high and ultra-high vacuum, Vacuum 13, 271.
natl. Acad. Sci. U.S.A. 61, 61 4. . of microorganisms, J. Electr1•n Rcbhun, L I. (1 972), Freeze-substitution and freeze-drying, in: Principles and techniq ues of
Nei, T. ( 1962), Freeze-drying in the electron microscopy electron microscopy, Vol. 2, M. A. Haya!, ed. (Van Nostrand Reinhold Company, New
York), p. 3_
roscopy 11, 185. . . of microorganisms on the basis nf •
Nei, T. (1973), Some aspects of freezing and d rying er, L. (1978). Scanning electron microscopy: present state and trends, Scanning /, 3.
water. Cryobiology 10, 403.
Freeze-dryinK 307
Low temperature methods in biological electron microscopy
306 Robards and Sleytr
Strasser, J., R . Heiss and P. Gorling ( 1966), The water vapour a nd heat trans port during
· f r physicists and engineers, Scan- vacuum freeze-drying with reference to optimal process conditions, Kaltetechnik 18, 286.
1978) Biological scanning electron microscopy o Stumpf, W. E. and L.J. Ro th ( 1967), F reeze-drying of small tissue samp les and thin frozen sec-
Reve I, J .P . ( ,
78 1 829
ning Electron Microscopy 19 • • · . t. of the underside of cells in culture, Exp. tions below -60°C, J. Histochem. Cytochem. 15, 243.
Revel, J.P. and K. Wolken ( 1973), EM investlga ions Sweeney, L.R. and B.L. Shapiro (1977), Rapid preparation of uncoated biolo gica l specimens
Cell Res. 78 , I. . . , hi h Has an electron-opaque stain in electron for scanning electron microscopy, Stain Technol. 52, 221.
!963) The use of lead citrate at g P
Reyno Ids, ES. ·( • Trinick. J . and A. Elliott (J 982), Effect of substrate on freeze d ried a nd shado wed protein struc-
microscopy, J. Cell Biol. 17, 208. 82) Quantitative analysis of electrolytes in frozen dried ture, J. Microscopy I 26, 151.
Rick, R ., A. Dorge and K. Thurau (19 ' Umrath , W. ( 1977), Prinzipien der Kryopriiparationsmethoden: G efriertrocknung und
239
sections, J. Microscopy 125 • · . d vice for freeze-drying electron microscope Gefrieriitzung, Mikroskopie 33, 11 .
Roberts, LM . and G.M. Duncan (1981), A s1mp1e e Umrath, W. (1983), Berechnung von Gefriertrocknungszeiten fiir ele ktronenmikroskopische
24 295 Praparation, M ikroskopie 40, 9.
specimens, J . Microscopy 1 • · d p It' -Kaskade fur die biologische Gefriertrock-
Schunzel, G. (1968), Uber den Einsatz er e ier Walzthony, D., M. Bahler, Th. Wallimann, H.M . Eppenbcrger and H. Moor (1983), Visualiza-
n ung, Acta Histochemica 30, 20l. f· eplicas of rat bladder lumina l membrane. tion of freeze-dried and shadowed myosin molecules immobilized on electron microscopic
Hicks (1977) F rozen-sur ace r
Severs, N ..J an d R ·M · ' films, Eur. J. Cell Biol. 30, 177.
J . Microscopy Ill, 125 · . f mbranestructurcin the t ransitionalepith- Washburn. E.W . (1924), The vapo ur pressure of ice and o f water below the freezing point,
C W en (1 978) Ana1ys1s o me
Seve.rs, NJ. and R. . arr l Th~ luminal membrane, J . Ultrastruct. Res. 64, 124 .. Monthly Weather Rev. 52, 488.
eh um of rat urinary bladder. · . r , d tissues Vol. 1, InstrumentatJOn and Watters, W B. and R.C. Buck (1971 ), An impro ved simple method of specimen preparati on
s·ostrand, F.S. ( 1967), Electron microscopy o ce11s an , for replicas or SEM, J. Microscopy 94, 185.
J d · p New York London), P· 188.
techniques, (Aca em1c ress, ' f dry·111g technique applied to the anal};" Wheeler, E E., J.B. Gavin and R.N. Seelye ( 1975), Freeze-drying from tertiary b utanol in prep-
F K ( 1975) A new reeze-
Sjostrand, F.S. and · rctzer . ' d . d hloroplast membranes, J. U ltrastruct. Re' aration of endoca rdium for SEM, Stain Technol. 50 , 33 I.
of the molecular structure of m1tochon na an c Wildhaber, 1., H. Gross and H. M oor ( 1982), The contro l o f freeze-drying with deuterium ox-
53, I. r ' lecules on bacterial cell walls: structure ide (020 ), J. U ltrastruct. Res. 80, 367.
Sleytr, U.B. (1 978), Regular arrays o macromo \\11lemer, H (1975), Low temperature freeze-drying of aqueous and non-aqueous solu tions
. d r fon Int Rev Cytol. 53, I.
chemistry an unc 1 • · . · p ( l 9 ) Periodic surface structures on proca1 using liquid nitrogen , in: Freeze-drying and advanced food techno logy, S.A . G oldblith , L.
B p M ssner p Sch1ske and D . um · 82 ' Rey and W. Rothmayr, eds. (Academic Press, London), p. 461.
Sleytr, U . .. . . e , . Electron Microscopy, Hamburg3, 1.
yo tic cells, Proc. I 0th. lnt. Con gr. . ~ I t on microscopy' J. U ltrastruct. Rei - ' \\1lhams, R.C. (1 952), Electron microscopy of sodium deoxyribonucleate by use of a new
Smith, P. R. (1980), Freeze-drying specimens ore cc r freeze-drying method, Biochim. Biophys. Acta 9, 237.
\\11l1ams. R.C. (1 953), A method of freeze-drying for electron microscop y, Exp. Cell Res. 4.
380. . . l ss J Cell Biol. 75' 245a.
Sommer, J.R .. ( 1977), T o cat1orose g a, '1976) Observations o n the production of frozen-dncd 188.
Spriggs, T.L.B. and D . Wynne-Evans ( in , unfixed fresh liver, fast frozen without cr)OP" \\olhams, M.A. (1 977), Autoradiography and immunocy$((chemistry. in: Pract ical m ethods in
thin sections for electron microscopy us g electron microscopy, Vol. 6, A . M. Glauert, ed. (Elsevier/North-H olland, Amsterdam).
tectants, J. Microscopy 107, 35. . b dd. g medium for electron m1cro"l'r \\ 1!111on, J.H .M. a nd A.J. R owe (1980), Replica, shadowing a nd freeze-etch ing techniq ues, in:
1968) A low viscosity epoxy resin em e in
S purr, AR. ·( , Praciical method s in electron microscopy, Vol. 8, A.M . G lauert, ed. (Elsevier/ North-Hol-
J. Ultrastruct. Res . 26, 31. . . n freeze-etch freeze-fracture, frozen ,urr... land, Amsterdam).
Steere, R.L. (1973), Preparation of h igh resoluhtlo d le and t~e use o f stereo electron m1cr V.unz. M., J. Kistler and T. Hohn (1976), Surface struct ure of in viuo assem bled bacteriophage
. - · s'ngle free ze-etc mo u • d
and freeze-dned rep11cas in a 1 .. F etchi'ng techniques an ·1P lambda polyheads, J. mol. Biol. IOI , 39.
. . f . · from them in . reeze- •
copy to obtain maiomum m ormat1on . , F . . de Microscopic Elcctrun•~ ng,heim. H.P. (1984), Sublimation rates of ice in a cryo-ultramicrotome. .J. Microscopy 133,
. EL Benedetti and P. Favard, eds. (Soc1ete ranya1se 107.
cations, . ·
Paris), p. 224. . · h freezing and drying of b11•1
Stephenson, J.L. (1960), Fundamental physical ~r~~msn1gn ~eS. Parkes and A.U . Smi1h
. I . . Recent research in freezing an ry1 ' .
cal ma ten a s, .m. . . . l 22.
(Blackwell Scientific Pubhcat1ons, Ox~~rd),(~980) Diffusible ion localization by ton 111"
Stika, K .M., L.K. Bielat and G.H. Momson . , f f eze-d ricd, unlixed "'er
. of chemically prepared and fast- rozen, re
copy: a comparison
tions, J . MicroscopBy JKJ8, :09(.1967) High resolution radio a utography of galacto>c-'H
Stirling, C.E. and W. . in e1 ' .
mutation in the hamster intestine, J. Cell Biol. 35. 585.
Chapter 6

Freeze-fracture replication

6.1 Principles offreeze-fracture replication andfreeze-etching


techniques

Freeze-fracture replication and freeze-etching techniques both involve the


prod uction of a vacuum-deposited replica from a fracture plane in a frozen
,pecimen. Generally, the cleaned replica is examined in a conventional
transmission electron microscope, although methods are also available for
1udying the topography of the coated fracture face in a scanning electron
microscope (§3. 3) if a resolution of less than about IO nm is sufficient. The
more common tenn for 'freeze-fracture repl ication' is still 'freeze-etching'
but because not all methods developed involve, or allow, sublimation of ice
etching) from the fractu re face before replication, o ne should distinguish
between 'freeze-fracture replication' o r 'freeze-fracturing' and ' freeze-frac-
unng and etching' or 'freeze-etching' methods. In this Chapter we will use
'tt:e-fracture replication as a term which covers both methods and freeze-
hing fo r methods where controlled freeze-dryi ng (etching) of the fracture
fact is an essential part of the schedule.
Nowadays, freeze-fracture replication is an established method in elec-
n microscopy, especially in cell biology and membrane biology but also
some areas of materials science, such as colloid research. T he main rea-
for the remarkable expansions and success of this preparation techni-
since the early 1960s are that:

In freeze-fracture replication, chemical fixation is replaced by freeze-


lixation, thus the molecular architecture of hydrated and in some in-
stances even potentially living, cells can be examined.

309
Freeze-ji·acture replication 311
310 Robards and Sleytr Low temperature 1nethods in biological electron microscopy

oved to be superior to Steere' s f reeze- f racture pr Th


(ii) Although the precise fracture path within a cell or tissue is generally · iihle!haler equipment the mod . . ocess. e Moor and
unpredictable, it preferentially follows membranes or other structur- ' ern version of which ·th .
me, is still commercially suppl· d b B , wi a simpler micro-
ally weak planes, snch as phase bonndaries. In consequence, the three," . ie Y alzers AG ( A ·
ed rehable, routine preparatio t b . see ppendix 2), al-
dimensional information derived from freeze-fracture replicas has led . ns o e made in many I b .
t time and contributed signifi ti h . a oratones for the
to an nnderstanding of the molecular architecture of cells, and parti~· can Y to t e widesp d
of freeze-fracture techniques. rea use and applica-
cularly of membranes, which is not obtainable with any other meth ··
he complexity and price of the M oor d ev1ce . subsequ tl ·
of specimen preparation. ber of electron microscopists t d 1 . . en Y sllmulated a
o eve op simpler in t
to fulfil. essentially similar fu nc t.tons. T he most .s rumen ts and
. appar-
.
Nevertheless, as with all other methods in electron microscopy, there sand stimuli for further develop t important simphfica-
very few instances where the morphological details in a freeze-fracture men s came from:
1ica represent the actual 'true-to-life' structure. Specific artefacts can 0
applying simpler fracturing methods t . I .
at each preparation step and great efforts are still being made to learn tome device, and consequent} ma, _no Invo_ ving a complex micro-
about their generation. However, since it is only a relatively beam-s faces of the fracture, i.e. the co~ le;mg possible the study of both
replica, containing information about the topography of the fracture Moseley 1969), and p entary fracture faces (Steere and
that is examined in a transmission electron microscope, there are few
lems arising from possible specimen beam damage. · d that it is possible to look at f rac t ure 'aces
showing c that h b
0 btame by freeze-fracturing und r 'd . ave een
The present-day freeze-fracture replication techniques are based ¢ ,ration unit (Bullivant and Ames ~r9~~m mtrogen outside an evapo-
pioneering work of Hall (1950), Meryman (1950) and Meryman and .al. 1968). ' Geymayer 1966; Bullivant et
(1955), who were concerned with the production ofvacuum-deposite
cas from the surfaces of frozen specimens. Subsequently, Steere
•recent technical developments wh· h h .
modified and improved these methods. His approach was to freez . nt, have been stimulated by,atte~ t atve agam led to more complex
of tissue in a cold chamber on a brass block. In the isolated cha flexibility, and reproducibility as ~;II o work with maximum au to-
specimen was cleaved, at atmospheric pressure, with a cold scalp n of contaminants on s . as by efforts to reduce con-
a dissecting microscope and subsequently transferred to a high va ly at very low temper~~~~7enC fratctu~e f~ces before replication,
equipped with a cold stage and cold trap. After superficially fre . · on am1nation can b d d
1nated, by using either hi hl . . e _re uce , or
the fracture face, a replica was produced by evaporating chromiu · gy (Gross et al 1978a) o bg y. sophisllcated ultra-high vacuum
bon onto it. Although, compared with present-day techuolo around ti . . r y usmg very efficient cold traps (cold
•· ie specimen area (Stee 1969 S
method has its limitations, he was the first to apply the freeze-fr 978). (Cold h d _re ; leytr and Umrath 1976·
s rou s are sometimes termed ' . . ,
cation and etching method successfully to biological specimens t there is no direct line ofs· ht th optically llght'. This
strating the structure of virus crystals within frozen cells. direct access by a molecule ig rough them and, therefore there
The actual breakthrough in the application of freeze-fracture> th .d . or parl!cle movmg ma straightr )
e w1 e vanety of methods and e . rne.
techniques, both for biological and non-biological material, ca re replication techniques no 1{mpment developed so far, all
development of the rather sophisticated 'freeze-etching' appar steps: rma y mvolve the following major
by Moor et al. (1961). The basic design consisted of an e
incorporating a liquid nitrogen-cooled microtome for speci
and a specimen cold stage which could be maintained at a pr.
ture. The microtome, which was originally designed to prod g;
tions, failed in this purpose (Miihlethaler 1973) but, since·
) sublimation of·ice f ram the fracture face (etching);
trolled fracturing of a frozen specimen under vacuum for: ,
Freeze-fracture replication 313
t ethods in biological electron microscopy
312 Robards and Sleytr Low tempera ure ni

the final two-dimensional image of the final replica, it is essential to under-


(v) shadowing and replication; and . stand fully the principles involved in the various preparation steps (see re-
(vi) cleaning the replica free of specimen residues. . views by Benedetti and Favard 1973; Sleytr and Robards 1977b; Rash and
udson 1979; Willison and Rowe 1980). An annual survey of the freeze-
. . h . th individual steps of freeze-fracture replica(<
A schematic diagram s owing e . . . h ---"/_, racture replication literature was published for some years by Balzers AG,
. .d d. p·g The heavy ]mes md1cate t e sequence:
tion procedures is provi e in 1 . 6 . 1. . . . . h f . ,"-__;:
f . steps As can b e seen, the greatest variation 1s in t e ractur1ng·-"'
ttal . .>:
or~~::: itself. Fr~eze-fracturing can either be performed under_ va~uu~- i~;
P
the evaporation . .
unit or out s1.d e at atmospheric pressure under hqmdf
mtr :-,
Specimen preparation
. · gas to protect the fracture faces rom cont
( h hum) or dry mtrogen '"'
ge.n ot.r e Both' fracturing techniques allow the use of either simple fract •.
mina 1on. bl" Sampling and pretreatment
in methods or complex cryoultramicrotome assem ies. . , .
. chapter dea 1s par t.icu 1a rly with the practical aspects of the. md1v1d
gThis
en using freeze-fracture replication techniques, the main purpose of all
. f f -fracture replication. Nevertheless 1t must
reparative steps o reeze . . .d d' piing, pretreatment and cryofixation methods is preservation of the I
pernp h asise
. d th a t , for the correct interpretation of information prov1 e '.
"led molecular organisation of the specimen, as close as possible to that II
vo. These requirements are common to all cryotechniques and have
dy been discussed in §2.3.

Specimen supports and specimen mounting

are many different designs of holder for mounting specimens for


I=.;;:.:~---..J-------1 atm.
V Fracturing at
pressure fracturing. The best design can be judged from the specimen cooling
Liquid or dry gas (N2, hat can be obtained and also from the ability to facilitate the specific
!!i
nical operations that are essential during processing of freeze-fracture
ens. As shown in a detailed study by Costello and Corless (1978), spe-
:::-_Cooling rates can be increased by decreasing the specimen mass, dec-
<:-the sample holder mass and increasing the surface area of the speci-
contact with metals of high heat conductivity. A further improve-
cooling rates of specimens frozen in liquid coolants at high immer-
·ocities can be expected by designing the specimen and holder shapes
'ze heat transfer (§2.2.3). Consequently, specimen holders are made
s possible and are constructed from metals with high thermal con·
, such as gold, silver, copper and brass. As discussed in §2.2.6, heat
Replica cleaning & mounting
:be extracted from the specimen surface during cooling and, due
· thermal conductivity of water, adequately high freezing rates can
_-hieved in a very thin superficial layer in non-cryoprotected speci-
Examination
1). Therefore, the specimen holder of choice depends not only
V 71 Under e of specimen (e.g. bulk specimen or suspension) but also on
tLL'
fryoprotectant is used. A selection of commonly used specimen
Fig. 6.1. Freeze-fracture flow diagram.
314 Robards and Sleytr . nie thod~. in biological electron microscopy
Low tempera/we Freeze~fracture replication
315

. 1 . view'
nd side elevation . together with the way thatd they
. are .,· how combinations of supports can be used for special methods such as
holders, m P an a . . S f these designs can be use dlfectly ·.'
. ·11 t led m Fig 6.2. ome o . bl N
ndwich-freezing, or complementary replica techniques, which allow the
loaded, is i us ra · . dapted to fit the specimen ta es of;
. hile others are eas11Y a _-. udy of both fracture faces obtained from a single fracture (see §6.3.3a).
for fractunng, w lication apparatus. Fig. 6.2 also shows,
different types of freeze-fracture rep r bulk specimens, hat-shaped holders are used (Fig. 6.2; types C, F and
This fomi prevents the tissue from being lost or pulled out during the
A
cturing process. Sand-polishing, or scratching the surface of the specimen
Ider with coarse emery paper, reduces the likelihood that the fracture
e will follow the metal surface of the holder.

2a Single cells and cell components

ensions of single cells or cell components are best applied to the holder
a modified Pasteur pipette. The original fine tip of a pipette is pulled
by hand, over a small burner, so that the diameter is reduced to about
.lts original size. With this reduced bore, small volumes (about 0.3-0. 7
can be accurately placed as a hemisphere or less on type A-D speci-
olders. To be pipettable, the sample must not be too viscous. If denser
s are used, they are best applied from the top of a needle which is
by touching the surface of the pellet. As with all specimens, mount-
i1 1

d subsequent freezing must take place as quickly as possible to avoid


,

1 used specimen . h o lders·, in plan and side elevation:.:,_: 'ng of the sample. Compared with this drop-freezing method, much ;.11
Fig. 6.2. A selection of common Y S ts for freezing, and microtome
A· B Ba C and Ca) uppor -, · .terial is necessary using the sandwich method (Fig. 6.2; holders B-E, i'i,1
loading methods. (A, a, ' ' ,·al products of Balzers AG; Appen :1:11'
1 1961 and commerc1 ) and this can be a great advantage if valuable and/or scarce materi-
of suspensions (Moor et a . . 1( coppe>) foils. I-folders B and Care ma ;11
h d t of thin meta e.g. · bl f er study. In addition, even if it is not necessary to obt:iin comple- :11
discs in A are pone e ou . dwich-freezing of suspensions su1ta e or 'I
from gold. (Bb to Bh) Assemblies fohr ~an (Mu''hlcthalcr et al. 1970, 1973; Nermut replicas, the amount of replica obtained from a small drop is dou- 111
lica tee n1ques
ture and complementary rep .
EM gnds or a nng ca
. . n be used as spacers between
( g this technique. A small amount is placed between the two holders I'
cl et al. 1978). One or two , . . . fsuspensionssuitableforsinglerepu 11'
bl f , ndwich-freez1ng o • ) ( xcess is drained off with a filter paper immediately before freezing. I
(Dato Dd) Assem y or sa d 1 f thin metal foil (e.g. copper
es The discs (Da an
d Db) are punche ou 0
bl f 0 r 'unidirectional' propaneJe - r
· t ti sandvvich methods and mounting procedures for preparing com- 11' '

· (Dd) s dwich assem Y Th 11'


and Costello 1978). an . tely µm thick) and a flat e .. fracture faces (§6.3.3a) a 'spacer' in the form of an electron mic-
. · g (approx1ma 20 ( I
ing of a metal dtsc, a spacer nn . . 1 1981) Espevik and Elgsaete~ :- id or small metal ring is inserted (Fig. 6.2; methods B, D, E and I
sheet (Pscheid and p 1attner
1980· Pschetd et a·
' . f
· d ·h
·ng (Ea to Ee) San W<C
I '

b. d' ctional propane Jet- reezi . Opane jet-freezing (§2.4.1 b), a gold 400 mesh grid, 2.3 mm in dia- 11 '
I,
the same assembly for t- ire . d Eb holders manufactured by Baiz
12 µmin thickness, can be used (Miiller et al. 1980b) or, alterna- I 1'
bi-directional prop~.ne jet-freez1n;O~~a (~~ Hin~ed complementary replica ho! . I ,,:,
ers described by Muller et al.
Steere and Moseley (1970). (Ga a.n
1: d. Gb) Assembly of thin mica or metal fo1.
h . (Nermut and Williams 1971:
pproximately 20 µm thick spacer-ring (Pscheid et al. 1981).
method of mounting spherical cells so that selective fracture
11'
11'
11'

,,i
1

t replica tee niques . d I


freezing and complemen ary . 1 for freezing suspensions an-<: btaincd has been described by Peng and Jaffe (1976) (Fig. 6.3). 11 I'
bl of hollow nve s b) A
1980). (Ha and Hb) A ssem y dU th 1974a b). (Ia and I I 11il

1970· · Sleytr an mra ' (Ch is suitable for cells that are larger than 70 µm. Cells are loaded '
ceplica techniques (Sleytr
.
a,
pies for comp e1n
1 entary replica techniques.
of a thin nickel screen (Perforated Products; see Appendix 2).
I 1111

llil
rivets for freezing tissue sam h 1974 ) (Cross-hatched or stippled areas
vant 1970; Sleytr and Umrat . ba . holders indicate spacers.) have many closely spaced, round, funnel-shaped holes into I /11

. dashed hncs etwecn


s fit To prevent the eggs from being lost during subsequent I ' lljl
,'ii

I 1':1

1·,1.
I
I
ii I

I
': '
Robards and Sleytr Low temperature me thods in biological electron microscopy Freeze-fracture replication
316
317

b c
a transfer should be as fast as possible and it is important that the specimen

~~~~~;~~·~ ~ ~
- _'c\\\\\\ --- -+-~
~0---
.. .,.,
· not unduly squeezed or squashed. The trimmed tissue can be mounted
irectly, using the adherent liquid to filJ any remaining gaps between it and
e holder. When necessary, gum arabic or a suspension of yeast can be used

~0 ,--,,
H

a glue to stick the .tissue to the holder. The specimen is usually gently
' ... _... .. ated with the suspending medium or placed directly into a droplet of this
Nickel screen p-face E-face .
eon a specimen mount. To avoid osmotic and other effects, such addi-
nal media must be adjusted to the pH and osmolarity of the buffer and/or
f . se1ect1ve
oducing . fracture planes through
. .
sphencal
h . ce'.
(
. 6 .3. Mounting method or fractured
pr protectant in which the tissue has previously been immersed.
Fig. d ass cells (c) can be produced using this tee mque,. :
P-faces (a), E-faces (b) an er dr~wn after Peng and Jaffe 1976.) or some fragile specimens, which cannot be trimmed to a suitable size
out excessive damage before freezing, special mounting techniques have
.n developed (Johnson 1968; Johnson et al. 1976; Willison and Brown
. positively charged by coating
manipulation, the surface of the slcreeln l~ne coating procedures see §6.2. ). Such specimens are first frozen by pouring a liquid cryogen over
a layer of poly-L-1ysme.· (For. po y-L-
. . ysd against either the fill'
1 mg s1'de or by plunging them into cold liquid propane or Freon (§2.4.Ja). Sub-
. 'hat' is pos1t10ne d h ntly, the frozen samples are quickly transferred to liquid nitrogen
A specimen mounting . . . 6 3b) of the screen an t e ass
h os1te side (Fig. . , . f . pre-cooled tweezers (Fig. 6.4). Pieces that are smalJ enough for mount-
6 .3a and c) ' or t e opp
. . a liquid . cryogen or by propane iet- reezmg b ....·.
is frozen by immersion m . , F. 6 3c can be placed etwee. conventional specimen holders (Fig. 6.2; types C, F or I) are then
.
tably sized rmg-spacer, as shown
. in ig.fractured
· ' ·
by pullmg t he h a.t from the rapidly frozen specimen (Fig. 6.4a). Meanwhile, specimen
'h t' The specimens are d'f' t ts carrying a droplet of cold, but stilJ just liquid, butyl benzene (Willi-
screen and the a · atus By using two 1 1eren: 'I::
. freeze fracture appar . p d Brown 1977) or trichlorethylene, kept at a temperature of 188 or
from the screen ma : thods of assembly, E-faces, -
h 'th different me . d ( -85 or -80°C), respectively, or a low temperature 'glue' (Table 4. I
of screen, toget er w1 §6 8 2) can be obtame u
cross- fract ured cells (for nomenclature see ..
technique. b Specimen

,,,, '""'"""
6.2.2b Bulk specimens ------- Freezing in LN 2

. l'
. animal or plant tissues, the chem1ca Fracturin) with blade
With bulk matenal, such as ll ried out on pieces that 188 K
. t ps are usua y car . .

~
Spo~o;mo.0~ ~~
and/or cryoprotect10n s e . d freezing. Final trimmmg 1
. d f r mountmg an d
than will be reqmre o f dental wax in a few rops Bb
bl d on a sheet o ·
with a sharp razor a e t d Whenever poss!0
. h f e was last treae · - - - Freezing
ution with which t e issu d' ens1'ons of less than 1 mrn in LN 2
· h uld have 1m ·
from bulk specimens s o a cylinder. A stereo mi
h t as a cube or as 188 K Complelentary
bly much less, w en cu . h t' ue Particular care has to replica
~ t immmg t e iss . . . .
be a special he1p or r t evaporate during tnm techniques

. h plate does no . f nting technique for fragile specimens involving embedding the frozen specimen
the liquid on t e wax . . the concentration o c
d · ble increase 1n utyl benzene (Bb) prior to fracturing. Fragments are removed from a rapidly
would lead to an un .es1ra f the trimmed tissue to the sup ,:n while immersed in liquid nitrogen (a) and are pushed into viscous butyl ben-
or other solutes. Transfer o e eo le are sufficiently skd.1 either to a disc (b) or to a rivet-shaped specimen support (c) maintained at 188
done with fine tweezers but som p pf ck for this delicate bly is then transferred back to liquid nitrogen before continuing with the
0

. t ed wooden apphcator s 1
die or a porn .ring process. (Redrawn and modified after Willison and Brown 1977.)
318 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-fracture replication
319

in §4.2. l; §4.3. lc; §4.3.2c) are prepared (Fig. 6.4b). The supports are main, · .0
;:,; -~
tained at this temperature with simple cold blocks as used in the spray-freeze• > ~
ing method (§2.4.Jc; Fig. 2.54). The frozen specimen pieces (S) are quickly,·; " 0. "d
"0
transferred, with cooled tweezers, from liquid nitrogen to the droplet al'• .s"" "" e 0 .0""
0
·a :~ 2 0
5
butyl benzene (Bb) or trichlorethylene, which acts as a glue, and the who!~·
" c c ·a
0
"" E -~""
assembly is frozen in liquid nitrogen. This method allows the specimen ·
0

""'t E
0
u
0
u -50" """
.a ~ ""
-
~
"u a0 2
be fractured with either a cryomicrotome assembly (Fig. 6.4b) or a comp· 0
;i
"@-
"
u uu -5"
" s -~
0 "d
c"
E
"d
"
u
0
'51
"'
;;
'O'
mentary replica device (Fig. 6.4c). For complementary fracturing, the g · "" E 0 -"
u ""
c::"
0 u
droplet containing the frozen sample is frozen as a sandwich between 'ii
"d
"
.a ·= "d"0
"d u
s £l
;;; "'
Eu" .s"
suitable specimen holders.
"
.0
c" 0 -~ "' -""
"" E
" .0-"
;;;'"
u 0
0 0
'ii"
"d
"
:c u"" ·s
0 0 0
...i ;-;: ;-;: u
'ii 0
u
6.2.2c Mono/ayer cultures "'
,,
A variety of special mounting and pretreatment procedures has been '

oped for cell and tissue cultures, which allow the manipulation of thei
in their in vivo positions while still attached to the support on whic
were grown (Table 6.1 ). "d
1'1'
I
Collins et al. (1975) describe a simple technique for the in situ freez :;;" ' 11
2
~
0
ture of cultures grown on 3 mm diameter gold carriers of the type B
~
i'1'
~
u "
0
in Fig. 6.2, coated with a thin layer of vacuum-deposited silicon mO' ·
A silicon layer is a suitable substrate for the growth of many differ
"
0
'ii "d
0
"d
0
""
"
""
-"
0
u
>,
""
0
''1' ,1

After fixation (the authors used a two-step procedure involving par "'""
0
"" B " -~ "du -~ ;£
"O

-" 0y" "'E TI 0 TI e


"d J;-
0
"'
.:;;""
1'
,1
'1

dehyde and glutaraldehyde), incubate the cells in cold 30% glycerol, "' u"0 u0 u">0 u0"> 0"0 "> =0 0
"
>,
"00 "
0
0
I
u 0 u
the carriers from the dish and carefully drain-off the excess liq :·
spray the carrier (Fig. 6.5) with a 1% suspension of 9.7 µm polystyt'
"0
~ ~
Oi u
"" -~ -~ .2 "u ;;:"
" 0 ~ 0
u '51 ~
E
0

£
"
.n
s §
s f':"
u
0

beads (see Appendix 2). Place a small copper hat on top of thd '" '"
freeze the whole assembly. The polystyrene latex spheres functiorr I I
between the cell monolayer and the copper hat and compens 1
i ,' 1
I
urreven surface of the cell culture. To increase the probability Ii I

'' 1'i
1.'111'
GO ,11''
00
~OPP°'
~
~
Latex /··1··· ·.\ hat °' °' 00
~
s
00 s ,: 1,\
spheres---==-.:.:.\:. I.I
" °' °'
00 I'
/J \ MonolayeA
E
"
0

§"
°' 0
-" I !,I l il

Monolay., 'q-n
1:. :·\-: ·
· .\ \ I cells+latex
op he,.,
u
u
0 GO
"O

·=
~
~

-a
,.:. .: ::g )5!':' "
0"
0
u 'I ·I
1'1 llij

.,"
u ~
°' ·c
5 0
-
'c
o a
'-::i-,J
rye;'6i:l:'.'i
'--i-
4 ,

-
u
""-u °'
::: "'
"O
°'
::: oi
f:'
~ ~

"'
:;;
~~
E
00 " '

"
0
"
-"
>,
1
I 11'.
1I~
' "
0 0
0.
"' ""
0
"O "O
Freezing "" oi
"" E "
0
" "" " -5""d
0 0 "d
0 -3' 1'1
ill'
·8 ]" "
3"
1§" -~ 16
Fig. 6.5. Mounting method for in. situ freeze-fracture of cell monolaye· """ ·c"ft ~ ~- /,11

i
0 "O 1
0
1
0 1i
latex spheres as spacers. (Redrawn after Collins et al. _19:7. u -" s ]
0.
"" " ~ "
"' CJ ;:i 0. 1't1,
11
'i
,,
'
320 Robards and Sleytr Low temperature nu,th ods in biological electron microscopy Freeze-fracture replication
321

h the monolayer o f ce11 s, beads with a smaller . dia- Joe.; Appendix 2). The following procedure is recommended for preparing
Planes passing throug h face of the specimen earner and
b d and both t e sur . ·1· he mounting medium:
meter (4.7 µm) can e use . . h s the advantage that st JCot(
r
of the hat can be po ts
bed Th•s techmque a
· •
h <•
. h t inducing any structural c ang,, 20 ml of IOOf;: glycerol is added to 80 ml ofO.I M phosphate buffer,
T ed at 132"C
monoxide can be ster11z .
wit ou ethod of Pfenn1nger
. h th
. an d Rin'\.:·
. ··'
pH 7.3, containing 0.14 M NaCl (phosphate buffered saline; PBS).
. .· t possible wit em . h
es somethmg that ts no tt ched to a gold gnd as t es The mixture is heated to 368 K (95°C) and stirred.
, d l\agen layer a a
derer (1975) who use a co h d ntage that the collagen layer• 20 g of Vino! 205 S powder is added slowly, with constant stirring,
h. method hast ea va . . .
strate. (However, t ts n be examined m the hght micr until the powder is completely dissolved in the hot PBS/glycerol mix-
transparent and consequently the ce11 s ca . ture.
scope). . for in situ freeze-fracturing of rnonolayer The residual white foam on the Vino! 205 S mixture is removed from
Another versatile method d . ltures grown on standard pl the surface with a pipette attached to a water-jet vacuum pump.
b d for cell an tissue cu .
tures, which can e use .b db Pauli et al. (l 977) (Fig. 6.6). This m The mixture is cooled to room temperature while gently agitating with
tic coverslips, has been descn e y lture can be used for both light a magnetic stirrer to prevent 'surface skin' formation.
that the same cu .·
od has the advantages . . ssible to preselect areas tq:· A crystal of thymol is added to prevent mould growth and the mixture
. nd also that it ts po . ·•· is stored at room temperature.
electron microscopy a ' ' . , erslips are fixed in s1tu (the a
d C Us grown on p1ast1c cov d 4"
freeze-fracture · e d ( H 7 3) 2;;: glutaraldehy e at ,
M , dylate-buffere P · 0 . • escott and Brightman (1978) developed a method for controlled frac-
ors used 0.1 caco d .th 25 01 glycerol by 1mmers10
b ntly infiltrate WI lo . h .g of mono1ayer cultures in which the cells are stained with ruthenium
30 min) and su seque . l 5 m squares (selected by hg t
5 hat 4oc. For freeze-fractunng, l. m 1 solution is drained off with ultures are grown ·on glass coverslips which have first been coated
d the excess g ycero ,c Collodion and then with collagen. Fixation is in glutaraldehyde to
scopy) are cut out an . f 1· uid over the culture. The sq
. ly a thm layer o iq , 1 · ruthenium red has been added. After fixation the composite three Iay-
paper, so leavmg on . d on a drop of polyvmy
d . th the cells facmg own, Af glycerinated, cut into pieces less than 2 mm in diameter, mounted
are then place , WJ d. the specimen holder.
. . f g me 1um on " ecimen holder and frozen. During the fracturing process, the red-
(PVA)-contammg moun m a ti on of fractures throug
ing this PV A layer encourages t~e pro~at~e mounting medium or cells can be distinguished from the underlying Col\odion and overly-
ze~ cell monolayer rather than tl rodug Elvanol 51-05 (DuPont de
. I the ongma stu y, . . ·
t using a light microscope (magnification about 200-300 x ). The :1
'
l'I

specimen carrier. n f g medium but this ts rt nee of pink flakes on the edge of the microtome blade serves as an 'I (I

d . 2) s used in the moun in ' f n that the fracture plane has passed through the monolayer culture. ;i_',1
see Appen ix wa . l (1979) suggest the use o 1·,·.1
·1 hie Bendtcht et a . dC ethod of mounting (Fig. 6.7) described by Griepp et al. (1978) (see
commercially avai a_ . Vino! 205 S (Air Products an ..
cally identical polyvinyl alcohol, p and Bernfield 1978) was developed for producing large areas
led, frozen-fractured, embryonic heart cell aggregates and layers 1',l
1.'.l
e potential for successful application to a variety of specimens. I
Coverslip layers are first fixed in situ with glutaraldehyde and are subse-
~Cells pregnated with 25f;: glycerol, both chemicals being made up in
~v ,,
4··::
Mounting '
sic salt solution (EBSS). An area of approximately l mm' is 1,
around each aggregate and each aggregate layer is then removed 1, I

Freezing bstraturn by aspiration and placed in a dish containing a warm


Specimen holder
!i% gelatin in the glycerolfEBSS mixture (Fig. 6. 7a). After solidi-
· ·t freeze-fracture o f ce11 monolayers using_
h" gelatin is fixed at 4"C in a solution of 2.5f;: glutaraldehyde in
Fig. 6.6. Technique for the m st u f ·t res through the ce!l layer. T t.S
in the mounting medium to encouragfe rac_; actured. (Redrawn after Paull S. After l-5 h, small cubes of the hardened gelatin, containing
prese\ection of the area to be reeze r layers, are cut out (Fig. 6. 7b and c) and mounted in the correct
322 Robards and Sleytr Low temperature methods in hiological electron microscopy
Freeze-fracture replication
323
a
Culture

·~ a·1"h
Foll membrane

Cells
Foil membrane
c
0
Glue

,,
'I '
,, '

Fig. 6.7. Method for the orientated mounting of cell aggregates and layers. Pieces: df
aggregates or monolayers that have been fixed and cryoprotected in situ are embedded'[
tin (a). After fixation and solidification of the gelatin by fixation with glutaraldehyde,
(b) or blocks (c) of the hardened gelatin containing the aggregate layers are cut out. The
can be mounted for single or co1nplementary fracture (f, g) or for freeze-fracturing using' .8, Mounting method for cells grown on a~ .
rotome (d, e). (Redrawn and modified from Griepp et al. 1978.) to selected areas of the foil membran (b . oil 'mcn1brane (a). Specitnen supports are
tthe specimen/foil membrane/,upp et ) us1ng super-glue' (cyanoacrylate). After cut
l d" or assembly ( ) , · -
u mg co1nplen1entary replica tech . ) c , Vanous sandwich mounting meth-
niques can be used (d) ·
orientation, in relation to the predicted fracture planes, on specimen plunging (e). (Redrawn from U h pnor to rapid freezing by
mrat and Isenberg 1980.)
(Fig. 6.7d and!). The samples are frozen by immersion in liqnid F
propane. For fracturing, both single (Fig. 6.7d and e) and comp!
(Fig. 6.7g) fracture methods are possible. en used with equal success· Hystoacr l
endix 2)· and CA-JS (D · Y blue (Braun-Dexon GmbH .
The method ofUmrath and Isenberg (1980) for the observation , ' e1o GmbH . s . .,
tse after a few seconds and t l ., . ee Appendix 2). Both types
fie areas of monolayercultures is illustrated in Fig. 6.8. Cells are o erate hrgh c I'
he specimen holder wt'th th t h oo mg rates (Heinzmann
Petriperm (registeredctrade mark of W.C. Heraeus GmbH.; see ·.. ' ea tac edfo'l b
underneath is dissected out . I mem rane and the adher-
2) tissue culture dishes (Fig. 6.8a). The bottom of each dish co ' us mg small sci (F ·
are removed. Subsequent] ssors rg. 6.8c) and sur-
25 µm thick fluoTethylene-propylene foil membrane, which is pe rs facing each other . y, t~o h~Jders are mounted, with the
-gases but not to liquids. Before use, both sides of this membr ,maspecralpat fD .
.her suitable tool for sandw' h . r o orceps (Frgs. 6.8e and
hydrophilic by a 12 s treatment in a glow discharge, at 10 Pa PJ: re -mountmg Aft th
brougbt into contact , th e assem b ly rs. . er e opposing layers
fr ozen bY tmmersion
.
650 V, 4 A m- 2 at a distance of 14 mm from the electrodes. Tm .4.la). in the
then sterilised by ultra-violet or electron irradiation in the n
l rn~tbods of mounting can be u . .
Monolayers are either chemically fixed and glycerinated, or frq
atrve to the assernbl f sed, as illustrated m Fig. 6.8d.
pretreatment. For specimen mounting, hat-shaped specimen ··· . Y 0 two monolayer ·
en agamst a blank hold h. h s, a smg1e monolayer
Fig. 6.2; types D, E or H) are roughened and carefnlly degrea er w 1c can b .
ethod of Collins et al (1975) e covered wtth latex beads
re medium may be re~laced ~ o~ othe.' spacers, and glycerol or
tion in ethanol or acetone. With greatest care, the specime_'.. '
.I alcohol (Pauli et al. 1977) ·~ th.er. connectron liquids', such
glued to selected areas of the foil membrane (Fig. 6.8b). Two
' l variations in the fracture p1ane
----------------~~
Freeze-fracture replication
Robards and Sleytr Low temperature me th 0 ds· in biological electron microscopy 325
324

Specimen mounting for complen1entary replica methods

arious designs of specimen holder can be used to produce complementary


acture faces as illustrated in Fig. 6.2. To ensure the proper alignment of
vet-type supports (Fig. 6.2; types H and I) during specimen mounting and
eezing, special forceps (Fig. 6.9) can be used (Sleytr and Umrath 1974a).
r some complementary replica methods, cell monolayers or membrane
ctions are sandwiched between two thin sheets of metal (M iiller et al.
Oa) (Fig. 6.2; E), mica (Nennut and Williams 1977), or a combination
mica, metal plates and glass (Nermut and Williams 1977; Fisher 1978)
·g. 6.2; G). Each of these techniques has the advantage that the specimen
be rapidly frozen and that only very small amounts of material are
ed (1-10 µ1). Furthermore, if one side of the sandwich is coated with
lycationic material, it is often possible to achieve uniform fracturing of
olayers.
·sitively charged, smooth surfaces of glass, freshly cleaved mica, or car-
d · d to ensure correCt alignment of paired rivet-typeh sr. :Coated metal sheets are produced either:
Fl·g 6.9. Special forceps esigne · . (R dca"'n from Sleytr and Umrat -
·
holders .
during specimen · and freezing.
mounting e "

by a glow discharge in amylamine vapour (Dubochet et al. 1971 );


. void excessive and/or uncontrolled squ by coating with poly-L-lysine (Mazia et al. 1974; Fisher 1975, 1978);
are desirable. Fmally, to a db t n the specimen holders. r
can be place e wee . 1·
the cells, a spacer . l l developed for uni-direct10na .• ycoating with Alcian blue (Sommer 1977; Nermut 1982a).
. cedure part1cu ar Y . db
A mouutmg pro 11 ltures has been descnbe y
freezing (§2.4.lb) ofmonolayherdce tcul (1981) Cells are grown . ·ne- or Alcian blue-coated supports have been used for a variety of
1980) d Psc et e a· · .
and Plattner ( an . 2 mm thick (Lux Scientific; see · ·ncluding controlled fracturing of erythrocytes and bacterial purple
thin Thermanox plastic sheets, . 0 h d out from the plastic )les (Nermut and Williams 1977; Fisher 1978; Fisher et al. 1978;
2). Small discs with cell layers are punc_ een holder (Fig. 6.2; typ '1982a)
h d low mass spec1m .
mounted on a U-s ape .' mm gold sheet. Me t and WiUiams (1977) and Nermut (1982a) suggested the following
0 l m thick copper or . 0 3
made from . m It ative to Thermanox._ To a_' for coating glass coverslips and either freshly cleaved or carbon-
. 2) , be used as an a ern
Appen d tx can d . h d during the cutting or p a with poly-L-lysine. Pieces of freshly cleaved mica or glass (e.g.
. damaged or etac e .
layers becommg tely shaped plastics ape shown in Ob in Fig. 6.2) are stored in chromosulphuric acid
b rown on accura
cedures, cells can a1so e g d r ring is placed betw~ '11Uric acid in saturated potassium dichromate, diluted 1:2 before
f being squeeze ' a space 62
tectthe ce 11 s rom hect The sandwich (Fig. . ; tl thoroughly with glass-distilled water and dried by touching one
men holder and the Thermanox s th . etal side only (§2.4.1 b) •.; dter paper. A drop ofpoly-L-lysine (2.5 mg ml- 1; Sigma Chemi-
. ·d propane onto e m ,
by squirting I1qm . . . bl ot only for the examt!J. ' MW about 4000; see Appendix 2) is placed on the mica. After
t ge that 1t ts smta e n
que has the a d van a . s ofmacromo1e:SP: ss poly-L-lysine is washed off with a stream of glass-distilled
grown as monolayers bnt also for suspens10Kn noll et al. (198J. .l_
..
oated mica is either used immediately or is allowed to dry and
11 ·n suspens10n.
lar fractions or who 1e c~ s ~ . t of specimens withq~'.t 'bin 20 min. To avoid adsorption to glass surfaces, the poly-L-
show excellent cryofixat1on tn a var1e y -Y(
t1 must be stored in plastic bottles and, -to ensure uniform coat-
pretreatment.
326 Robards and Sleytr Low temperature meth ads 1·n biological electron microscopy Freeze:fi'acrure replication
327

l d l·n the process m U st be scrupulously clean. d f t In practice, many different combinations of sheet-like material can be
ing, all too s use ·n alternative proce ure or coa"'.
Fisher (l 982c) recommends the fo 11 ow1 g used for sandwich mounting. Nermut and Williams (1977) and Robards et
. covers l·ps
ing 1 with poly-L-lysme. 'al. (1980) have developed special gadgets for cleaving apart 10 x 15 mm or
x 10 mm sized sheets of mica which are frozen together in a cross-wise
. in orcelain staining racks, in hot t angernent (Fig. 6.2; Ga). An alternative technique has been developed by
(I) Wash coversltps, preferably p t Sonicate for 60 s (e.g. with.
water plus a few drops of detergen . 'sher (1975) where the specimen material is sandwiched between disc-
kH 2 80 W bath-type sonicator). aped glass and copper sheets.
i • • in tap water for 30 s. Specin1en mounting can be done in different ways for preparation of com-
(2) Wash covershps with runn g.
(3) r .n chrom1c-su1p h u ric acid (4.0 kg concentr. ementary fracture faces (Fig. 6.2). As an aid to fracturing suspensions in
Immerse covers tps I (F" her Scientific Co.; see Appen
H,S04 + 25 ml Chromerge is right plane, one or two gold finder grids (grids with marks so that indivi-
and clean at 60-70oC for .1-2.h. . g ion-exchanged distilled 1squares can be recognised and recovered; see Appendix 2 for suppliers)
(4) Rinse the rack of covershps m runmn be placed between two adjacent specimen holders (Miihlethaler et al.
. . lass distilled water. 3; Nermut 1973). If one grid is sandwiched (Fig. 6.2; Bg), more even
and then rmse mg - k .th lean forceps and quickly
(5) f from raC Wl C ture planes may be obtained, with the additional advantage that the rep-
Remove covers 1ps . The cleaned coverslips s .
a stream Of compressed nitrogen gas. . from one fracture face frequently remains attached to the grid through-
be used within 30 min. . ( ol -r-lysine hydrobromide, MW •the cleaning procedure. If two finder grids are frozen in register (Fig.
(6 ) Apply 5.0 mM poly-L-lysme PC y) MW 1500-8000 (Miles . .·. h), replicas of both fracture faces may be easier to locate. Indexed (two
11 (Sigma Chem o. or .d with their meshes aligned; see Appendix 2) gold grids are best attached
4000, Type · h drophobic pipette to one s1 e
. ) A pendix 2) from a Y · ) G h other using a thin layer of Vaseline which also increases the likeli-
tones see P . 1 for l l x 22 mm covershps . yr,
clean, dry covershp (25 µ f 30 s at 20-22oc so t that fracturing will occur between the grids. Since handling of the two
. in
covers1Ip . the horizontaldplane or : ust be done under a dissecting microscope, this method is frequently
coverslip is uniformly coate d. l lysine with glass-distille... ,ered to be too laborious. As another aid to identifying complemen-
unboun po y-L- eas, morphologically distinct, easily identifiable, marker cells (yeast
(7) Wash away excess, r .th a burst of dry nitrogen
for 30 s and dry the covers rp w1 < throcytes or pollen grains) can be mixed with the suspension under
fore freezing (Sleytr and Umrath 1974a).
(8)
air gun. r p should be nsed within 30.
The poly-L-lysine-treated covers I 'al mounting methods are used for spray-freezing (Bachmann and
.c.Fumian 1973; §2.4. lc), or for fast-freezing on cold surfaces (Heuser
. . . 1Y charged surface by,:
lied to the pos111ve
The specimen IS best app . h 10 µl) over the surface .6; §2.4.2). A mixture of spray- or jet-frozen droplets with butyl ben-
. ens1on (less t an be transferred and mounted on almost any type of specimen holder
a drop of specimen susp d 5 ·n depending on the con
, _. . seconds an mi · with a cold table. The droplets can be enclosed in holders suitable
it for between a ,ew . . d by washing in e1
. E ater1al is remove ngle-fracture methods, such as complementary replica techniques,
the suspenswn. xcess m d r e (Fisher 1975; Nermut a
of water or in phosphate-buffere sa m or cryomicrotomy. Dempsey and Bullivant (1976a) used speci-
ted on thin Teflon supports for freeze-substitution and on more
1977). . ts positively charged using Ale'
T ke specimen suppor d. 2) semblies for freeze-fracturing (Dempsey and Bullivant 1976b).
o ma . 74240, Serva; see Appen ix ,
Alcian blue 8GX p.a., CJ. mM solution for 5-10 sand,.
brought into contact with a 5.0 f m a pipette) for appro,
. ( as a fast stream ro th
with water e.g. f 1 dry nitrogen gas, e
s. After drying in a stream o ~ ean, .th a net negative G \hods have been described in detail in §2.4. The storage methods
. 11 Y a d8 orb specimens WI
will preferentla §2.S arc perfectly suitable for specimens to be used for freeze-
1977).
328 Rohard~ and Sleytr Lov..' temperature methods in biological electron microscopy Freeze~fracture replication
329
fracture replication. No changes in structure have been noted during storage Fracturing
of frozen specimens in liquid nitrogen. Frozen samples are generaJly stored--:;-
in 25 litre containers. A special storage container for freeze-etch specimenS(-~<, General considerations
which reduces the loss of liquid nitrogen and allows the storage of over 3o(J"
different specimens in a 30 litre container, is recommended by Krah ef'it~- he fracturing processparts the 1rozen
c .
specimen al th 1
(1973) (see Fig. 2.65 in §2.5). ··· rs least resistance to the appl' d c . ong e P ane which of-
1e 1orces Jn practic f .
n is done either by tensile stres h' .h e, racturmg of the speci-
6.2.4 Transfer offrozen specimens . s, w rc produces . I I
with a cold scalpel or microto k .0 . a smg e c eavage plane
me m,e which d t h ,
During the various transfer stages, the specimens must not warm up a s conchoidal flakes in front of th k ·c , ue o s ear forces, pro-
. e nr.e edge Both · I ,','
the recrystallisation temperature which, for most biological specime turmg processes can be accom 1. h d . · smg e and multiple "'
. P is e either under
around 180-190 K (§2.l). Consequently, all transfer devices (tweezers, rc pressure (see Fig. 6.1 in §6 I) F f vacuum or at atmos-
. . · · or reeze-fracture 1· ·
tainers, etc.) must be carefully pre-cooled before use. The various fr · ng frnctunng methods have been de I rep icatron, the fo!-
tail. ,ve oped and are descnbed in §6.3.3
fracturing methods involve different transfer procedures but, as a ge ·
rule, the speci1nens should only be exposed to conditions where hoar
will be formed for as short a time as possible. If specimens are not fi reeze-fracturing under vacuum:
frozen, they are first transferred from their large, long-term, storag~­ Methods for single fractures usi d .
tainer to a small (0.5-1.0 litre) container which should have comparf complementary, fracture faces. ng ev1ces for producing single, or
to keep (at least 3) specimens separately. Depending on the par 'Freeze-fracturing with a scalpel 0 r a m1crotome
.
assembly.
freeze-fracture replication method, specimens are then either tra
directly with a pair of cooled tweezers from this interim containef::-"' eze-fracturing under improved vacuum conditio ·
technology or cold shrouds. ns usmg ultra-high
cold stage of the freeze-fracturing apparatus, or they are first
under 1iquid nitrogen on a special gadget, such as a cornplementar ze-fracturing outside the vacuum unit u n d er a l'1qu1d
.
e_

gas·
device, and subsequently transferred to a cold stage. The loading or .

reeze-fracturing under liquid nitro en . . .


of fracturing devices under liquid nitrogen is best done in small -s microtome assemblies. g us1ng s1ngle fracture methods
(expanded polystyrene) containers which have excellent thermal'
eze-fracturing under liquid helium.
properties and can be handled with less care than glass Dewar fla'
specimens are frozen in cryogens with a melting point higher than. e·.fracturing under a dr "t
y n1 rogen atmosphere using a cryoultramic-
Freon 12 or 22; see also §2.2.3a, Table 2.6), remnants of the liq1f
around the specimen when it is plunged into liquid nitrogen "fi
This can create difficulties if specimen handling devices are 1 cterise the fracture planes obtained from .
liquid nitrogen. As an alternative, loading can take place und e, it is useful to disting . h b the different techniques
·- . u1s etween gross d fi
or 22) or specimens can be frozen in cryogens with a lower r is defined by the t f an me topography.
. ex ent o coarse undulat" ( I
(e.g. propane or melting nitrogen). Before specimens are tran ean level of the fract wns > .0 µm) in rela-
the liquid nitrogen storage container to a cold stage, it is a g on the replica with d .ure p 1_ane, whereas the latter defines struc-
1mens1ons much less than 1.0 µm.
to 'lubricate' the stage with liquid Freon (using a small br
hoar frost and to ensure good thermal contact). This also im topography of the fracture plane
sitory thermal contact of the specimen with the stage duritt
mounting procedure. On the other hand there is the di y low temperatures ( < 200 K) .
impurities in the Freon may cause contamination of both fracturing with a . , , simple cleavage induced by ten-
mrcrotome kmfe both lead to the production
tern and the fractured specimen (§6.5.2).
330 Rohards and Sleytr Low temperature method~ in biological electron microscopy Freeze-fracture replication
331

of identical fine topography in the fracture plane. In heterogeneous sped< his. has the advantage that larger areas of the su f
mens, the fracture plane is determined by the distribution of chemical bonds tnnal angle to the shadowing so . . r ace can be placed at an
licas. urce, so 1eadmg to the production of good
and predominantly proceeds along intermolecular paths requiring the least.'
energy to break these bonds. By lowering the microtome bl a d e b etween cons t"
. en chips can be fractured aw' C ecu ive strokes, small spe-
The evaluation of replicas of complementary fracture faces has b ay. onsequently th f 1
shown to be particularly useful when analysing the fine topography of not neces~arily represent th d ' e Ina fracture plane
·. . e pro uct of the last t b
fracture plane. As will be discussed in detail in §6.8.3, the fine topogra ·. chips from a deeper level th th cu , ecause conchoi-
an e plane of th ·
of freeze-fracture replicas of biological specimens indicates that potenti en two consecutive cuts may h b c microtome advance
ent of the blade. Furtherm ave een .produced by the preceding 11,'
'brittle' and 'ductile' regions alternate along the cleavage plane. Ex . ore, most replicas of ·
ments on model systems have shown that many polymers, biological as with a microto1ne show areas "th k . specnnens smoothed-
. WI n1fe marks or 'fl t ,
as synthetic, deform plastically during freeze-fracturing at temperatu by localised friction between th e k n11e ·c and the s ,· a areas gener-
of localised surface melti·n d . pecimen surface. These
low as 4 Kand that considerable energy must be dissipated during the g are ev01d of · f .
turing process. Heat dissipation during local deformation process kable that such knife marks c· b. any m onnat10n. It is
. 1
an a so e observed h f
cause a rearrangement of molecules in adjoining regions which may,_. done vvith a microtome unde r .d . w en reeze-fractur-
. . r 1q m mtrogen (§6 3 3 ') Th 1
brittle fracture characteristics. Alterations to the fracturing behaviour· by fnct10nal heating can be th' · · · c· e ayers des-
-. very in and subs ·
reveals structures that sho . equent etching some-
particularly the deformability of biological structures, observed afte w no c1ear sign of h t d
ent pretreatments, such as fixation or infiltration with cryoprotect fragments produced during th f . ea amage. In addition,
e ractur1ng proc , f
errna1 contact with the m· t k . ess are requently in
be interpreted as the result of such interface (surface) phenom icro ome n1fe unl f
changes in intermolecular bonds (for reviews see Sleytr and Robard der the protection of a liquid as (e . . es~ reeze-fracturing is
erated by the trimming proce:s withg;~1qmd mtrogen). Frequently,
1982). fractured areas of the s . f e microtome fall into deeply
In biological specimens, the plane of weakness is often predeter ' pec1men racture pla h
the disposition of membranes, which have been shown generally to by the final strokes of the knili (St h r new ere they are not
, these small chips and f e ae em and Bertaud 1971). Con-
along a central hydrophobic region (Branton 1966). Other struct ragments attach to th ·
may have an influence on the path of the fracture plane include li -sent a potential source of subl" . e microtome blade
1m1ng water for d ·
such as plant spherosomes, which frequently show a concentri en surface (Staehelin a d B con ensation onto
.. n ertaud 1971· D l .
hony et al 1982) I . ' un op and Robards
men! of bilayers. In addition, regular aggregates of macromolec · · n contrast to mic t f .
ture inethods as needed c h ro o~e ractur1ng methods,
proteins (e.g. protein crystals, microtubules and flagella) can act ' ior t e productio f
oduce no or very little . . n o complementary
tial cleavage sites. In procaryotic cells the fracture path is fre ' , contam1nat1on of th f
lected, not only by the cytoplasmic membrane, but also by othe . face n1elting or sublimation of. e racture face from
matic examinations ha b ice f~om fractured fragments.
as the outer membrane of the Gram-negative cell envelope or
_de edge or any otl1 ~e een carried out but it is evident that
endspores. Similarly, the multi-layered cell walls of some eucar ' er microtome k ·fi ·
.the hard frozen specimen (Parish 1~~ e, is very rapidly damaged
cleaved by the fracturing processes.
nsions of cell fragm t . I 3). With simple specimens
. . en s, 1so ated cells -0 h . '
implication for the to o r omogeneous tissues,
6.3. l h Gross topography of the fracture plane
::h:~~r:c~:~osr,altikee microtome
!1a~a~~isas~i:~~;:
1

As opposed to the fine topography of the fracture plane, dif bl


c~:~::~:, fiah~;
described by R b d sem Y can be useful such
two razor blaode~r osn:~:l. (19.70) and Robards and Parish
topography can frequently be observed between microtome
tured surfaces. In general, at low temperatures, the cha
s), In addition it .h b r prehmmary cutting and the other
. a less highly sculptured fracture plane are better when a Ill , as een shown that different knife angle
332 Robard~ and S!eytr Low temperature meth ods I·n hiological electron microscopy Freeze-fracture replication
333
. . ,, fl nee on the final gross topography of the·:
settings can have a s1gn1fic~nt in ue th ugh surface of a fracture through. a single fracture method. The possibilities for stndying small samples of ho-
F moothmg down e ro ..· ·mogeneous material or monolayers of cells, for obtaining complementary
fracture plane. or s . h commended that a few fine cuts...T'
s tissue it has een re -, Jracture planes and for reducing the hazards of contamination by condensa-
a very heterogeneou ' h ld he made and that the knife an
. 1 J00-500 nm steps s ou · 'on of water, all favour single fracture methods. However, for fracturing
in approximate Y . __ o (Parish 1973; Robards 1978h
b d d to approximate1Y 45 55 ,, _xtremely heterogeneous tissues, a microtome can be advantageous
should ere uce
r .nforcement proce d ures (§ 6 .4 .2) ' particularly the sll: hough, even here, replica reinforcement methods can help to overcome
If no rep 1ca rei h 1978 ) are available, the excessively ro
technique' (Robards and Umrat 't, can make it difficult or impos , problems associated with the unduly rough surfaces created by a single
f fa heterogeneous issue cturc method. Fracturing with a microto1ne is also required if replicas are
fracture ace o th. down the fracture face by th
. . t licas After smoo ing . --.: be prepared from the limited, well-frozen zone of specimens that have
to obtam mtac rep .· , f , entation of the replica dunng th
oval of small conchmdal chips, ragm bl 'en frozen on the surface of a cold metal block (§2.4.2; Heuser 1981),
m . (§ 6 6) ill generally be less of a pro em.
ing and cleamng · w. _ , th gross topography of the fra
. l' ttl or no difference m e
. 1e f racture methods or microtomes~
Vacuum units forfi-eeze-fi-acture replication
There ts 1 ,e d by either s1mp
plane, as expose . f cells or cell fragments are exa
tissues suspensions o ' . spective of which freeze-fracture method is used, deposition of the rep-
ho1nogeneous ' d ous complementary rep 11ca t
, 1 f ture methods an ana1og _onto the specimen fracture face must be done in a vacuum coating unit.
For smg e. rac , with flat ends can be used (see Fig. 6.2; types,
ques, specimen holders 1 thin layer of a suspension or ,degree of complexity and performance of different commercially avail-
E and H). With these holdebrs, on y tahe holders and consequently a units vary considerably and examples of different vacuum syste1ns used
· d'hdetween ' reeze-fracture replication are illustrated in Fig. 6.10. The schematic
layer culture is sa_n w1c e tial hazards from microtome fractu
fracture plane, without the poten __:_J: '-ngs show the essential parts of the most common systems and also the
obtained. nal parts required if specimens or evaporators are to be interchanged
an air-lock system (the area in Fig, 6. 10 indicated by a dashed line).
6.3.1 c Conditions for freeze,fracturing st usual vacuum units consist of a bell-jar, or a similar working
r, pumped by a two-stage system comprising an oil diffusion pump
. -fractured under vacuum using singl~_; , by a rotary pump (Fig, 6. 1O; system I), The operation of such a sys-
Most specimens are freeze d to 77 K. The first .-
. t cs at temperatures own -> been described in detail by Willison and Rowe (1980), As an alter-
methods or micro om B t (l 973) who tried to fret<Z,
l' tion was by ran on a diffusion pump, some manufacturers offer systems incorporating
from this genera isa . t 16 K Freeze-fr
d urn with a m1crotornc a . olecular pumps (Fig, 6.1 O; system I!). These have the advantage
specimens un er vacu d 10 K has become less of a prob\e:
· temperatures aroun . . can be operated more easily than diffusion pumps which must al-
specimen · _.
fa new generation o f'fr eeze-fracture rephcat1on u
d k below the critical backing pressure to avoid the possibility of gas
intro uctwn o . , d he Balzers UHV apparatus
et aL 1980; Niedermeyer 1982af, ant t, g temperature to have breaking through the oil jet and stopping the pump action. A
d' 2) Tl e lowest rac urm ith low partial pressure of hydrocarbons is achieved more readily
see Appen ix '. 'fli uid helium (Sleytr 1974; Sleytr and U
so far has been that o q d complementary fr 'molecular pumps than with diffusion pumps which also require
using a fracturing technique to pro uce ximately 180 K oled with liquid nitrogen to condense organic vapours. It must,
(§6 3 3c). Within the range tested, from appro h (§6.8 e remen1bered that turbo-molecular pumps have to be vented
1. h, . b hown that in contrast to the/me topograp y , t with dry nitrogen gas) to about I kPa (7.5 Torr) when the system
t as een s _ ' . h of fracture planes ;
cant differences m the gross topograp . y f t 'ng is done "·.·. so as to avoid back-streaming of oil from the rotary pump.
1 d hen freeze- rac un designs of freeze-fracture replication unit use vacuum systems
cal specimens arc revea e : th under vacuu1n or liquid
or lower temperatures, or~ e er b freeze-fractured sll:ti§__,, g refrigerator cryopumps (these are cryopumps that are cooled
In summary, most spccunens can e _- <- refrigerator system: see Fig, 6. 10; system III), adsorption and
334 Robards and Sleytr Low tempera t ure ,,,'et!-ods
' in biofoo-ical
6
electron n7icroscopy
Freeze-fracture replication
335
Additional parts for
Parts for @ @@ in different vacuum systems
·sputter-ion pumps (Fig. 6.10; system IV), or combinations of these.
I
air l~c~ ~~-:_m_
-, [s~~~~ Type System
' ' c
Besides the inherent limitations of some vacuum systems (e.g. oil back-
I
I
I
. trean1ing), frequent causes of contamination are degassing gaskets, grease,
Oil diffusion
I
I '" I pump and
cooled baffle
f- ii and finger prints (hence use gloves!) in the equipment, which can create
TM
I™ poor vacuum over long pumping periods.
I
I
Different values are given by manufacturers for the final pressure achiev-
'I Turbo-molecular Ie with their systems. The cited 'working pressure' frequently varies
I II pump f- 4
I
I tween l x J0-- and 1 x J0- 7 Pa (7.5 x I0- 7 and 7.5 x 10- 10 Torr) but, as
I
Cryo-pump 'scussed below and in §5.2.1, this value can be misleading, since it repre-
with air lock:
m v 4 not required= f- ts the total pressure of all residual gases in the system. The pressure which
closed pipes
, f interest for freeze fracturing is the partial pressure of the gas compo-
Adsorption-pump
ts which condense at temperatures down to the fracturing temperature
(LN 2 cooled)
& Ion-getter pump
fhe specimen and thus can lead to condensation contamination. There
]][
with air lock: also considerable variations in the degree of automation and in the
v 4 not required=
closed pipe ber of fail-safe devices with different pumping systems. Complete auto-
'on of a vacuum system is generally expensive but gives the great advan-
of fast and correct sequential operations. So far the highest degree of I
ation has been achieved by the Leybold-Heraeus Bioetch 2005 freeze- I 1

Oil diffusion pump (3 stage) Adsorption pump (LN<["cooled) g unit (see Appendix 2) where the automatic vacuum system is inter-
with a fully automatic freeze-fracture replication process (§6.3.3b). I'
Turbo-molecular pump (wllh valve for nnling
ii al 3 tand•till to avoid oil back-atreammg Cooled balll"
:new freeze-fracture units have at least a fully automated pumping sys- I'
from rotary pump)

Rotary pump (2-•lage) Adao•plion trap for oil vapour f.• most frequent cause of poor results with freeze-fracture replication
contamination of specimen fracture faces with, particularly, water
condensing from either the vacuum system or parts covered with
Cryo-pump
st. Thus, excellent vacuum conditions, with a minimum of condens-
in the neighbourhood of the specimen during exposure of the
Ion-getter pump
lanes for etching and replica deposition, are essential. There are
. diagram
. . in g the basic principles unitse-u,s;
of. vacuum d/or rent approaches to obtaining an excellent vacuum environment
F. 6 IO Schematic illustrat . .
ig. . . . . m onents that are reqmred if specimens an , 'nimurn of condensable gases in the specimen area. The simplest
fracture replication. The co p , h I "d within the dashed hn
. 1 k t m are s own enc os.., iquid nitrogen-cooled traps which act as very efficient cryopumps
interchanged via an atr- oc sys e . . . three-stage oil diffusio
ing system compns1ng a ydrocarbons and water vapour. Cold shroud (and specimen stage)
commonest, two-stage, pun~p T II a turbo-molecular pump may be
by a rotary pump; alternattvely ( ype ) . t ms frequently use either-r_._--·, res as low as about 10 K have become easily obtainable by using
a diffusion pump. Ultra-high vacuum pump1n~ syspcumps (Type IV). V1 to V , 'gerator cryopump systems which do not require a continuous
f
P umps (Type III) or ad sorp ion and sputter-ion
k hamber· y, high-vacu quid helium. The other method of improving the vacuum condi-
· · 1 k system to wor c ' ~' ··'·.
valves: Vi, valve connecting au- oc V ! for pre-evacuating the work c·'··.
for evacuation of air-lock system; 4, va ve( k hamber)· VA• valve f6t.
s complicated ultra-high vacuum (UHV) technology. For UHV
. h h. h- cuum pumps or war c
for pre-evacuating t e ig va
' .
IM . h t athode ionisation gauge
.. ,ure replication, not only must the vacuum around the specimen
Vc, valve for venting the wor~h~=~;~~duc~~~tyov:cuum gauges. roved, but the whole system must be evacuated to the same level.
ry technology requires a system in which every part can be
336 Rohards and Sle,vtr Low temperature method~ in hiofogicaf electron microscopy Freeze:fracture replication
337

baked to between 200 and 400'~C (473 and 673 K) to desorb molecules ·· primarily designed so that replication take I· . .
·ng and, therefore a · s Pace immediately after fractur-
trapped on metal surfaces and within metals. In such systems the total pres--, , ' specimen temperature control i h. .
sure can be reduced to approximately 1.0 x J0- 8 Pa (7.5 x 10- 11 Torr) or ther crude or non-existent Due t th . .
·
. . or etc ing is usually
o elf s1mphc1ty th · ·
even lower. Elaborate air-lock and specimen transfer devices are necessat~i evices are easy to b UJ·1d an d can be fitted . t ·' · e maionty
. of such
1
to maintain such a vacuum (Gross 1977; Gross et al. l978a; Niederme~OJ;·'. daptcd for use in freeze-etching un1·t h' h nlo ex1s:1ng coating units or
e assembly. s w 1c a ready me orpora t e a micro-
·
1982b) and the whole unit must be baked out again after it has been ven:t~tl{
(e.g. to reload the evaporation sources if no special air-lock system for ,, • A simple single fracture method 1nvo · 1ves f reez1ng
. the · b
evaporators is available). In such a syste1n, the main sources of gas are o specime.n holders of the type u d c - . specimen etween
se JOr producing comp! t
faces bearing hoar frost (particularly on the specimens, their supports e faces (Fig. 6.2· types B-I) Th f . . emen ary frac-
' · e rozen sandwich 1 'th
the transfer devices) and the water vapour released from warming ice fj y onto a pre-cooled cold stage in . s e1 e.r inounted dir-
I
t (Nickel et al 1978) . fi a coating or freeze-fracture replication
ments during fracturing (Walzthi:iny et al. 1982). The hoar frost ca . . ' or is rrst fixed to a transport late d . . I
removed by sublimation al temperatures from about 173 K ( - l00°C ogen and subsequently transferred to th . p un er hqmd
the released water vapour will adsorb to metal surfaces in the unit and- ,O; Sleytr et al. 1972) Afte th k. e stage (Steere and Moseley
, · r e wor tng vacuum has b b ·
'-sp.ecimcn has reached. the correct temperature, fracturi~~nt~k~;1n1:~ea~d
I
seq uently lead to a poorer vacuum unless it is trapped by a colds
(Escaig et al. 1980). ln practice, it has been shown that good shielding mg off the free specimen support from the fixed one . p . y 1:1
an optically tight cold shroud, that is kept at a temperature at least ethod, a replica can only be produced f f (Fig. 6.11). Usmg
f . rom one o the cornple1nentar i"I
below that of the specimen, is the most efficient protection of the f re aces smce the other is sacrificed with the b ro k en-o ff specimen . sup-y 1:1
surface from contamination (Steere 1973, 1982; Sleytr and Messner
Steere et al. 1980; Sleytr et al. 1981; Niedermeyer 1982a). Mounting & freezing Fracturing Shadowing &
Apart from the vacuum conditions, there are other important c ' with lever replication l,[I

the selection of freeze-fracture replication methods and apparatus (


cimen temperature control, evaporation methods) as discussed belo

6.3 .3 Instrun1enls JOr.freeze-jfacturing


Cold stage
There are several ways of fracturing frozen biological specimen$-'~
6.1 in §6.1 ). Freeze-fracturing can be carried out either under vac)l
atmospheric pressure, and the fracture process itself is done eit
ing tensile stress or with a cooled knife. This diversity of methods
in the many different types of commercially available apparatli
shop prototypes' developed in the last fifteen years. It is not po
cuss all variations in this Chapter but the major procedures will
in detail.

'\vo basic methods for producing sin le (non- .


6.3.3a Freeze-fracturing under vacuum Specimens frozen b. .g complementary) fracture surfaces for
. · c1ween two nvet-shaped h Jd .
Imcn support from th fi d . o ers, arc fractured by breaking
c 1xe one Alternat J (h) ·
(i) Single fracture methods frozen in pre-scored ca ill· . . ive y spcc1n1cns (particularly sus-
.-Pt/C indicates the d. pt. ary tubes which are subsequently broken apart The
These are the simplest freeze-fracture replication te1oh11ique! ncc ion of plat" ; . b . ·
labelled C h , . . . inum car on shadowing while the arrow
the pioneer work was carried out using them. Single fract s ows ca1bon deposition onto the fracture face.
338 Robards and Sleytr Low tetnperature methods in biological electron 111icroscopy Freeze-fracture replication
339

port Replication takes place either immediately after fracturing or after , The lower block contains a heater t o raise . the spe ·
etching. Such methods, using tensile stress to separate the two specimen-,::t etching. The Bullivant and Ames t II . I czmen temperature for
. . ype smg e fracture de · (§6 3 3 ) .
holders, are suitable for both cell suspensions and bulk tissues. 1nal1y designed for fracturing und 1. .d . vice · . c , or1-
er tqu1 nitrogen d.fi d
A single fracture method for fracturing frozen suspensions or cells sand>_::;~f !ear and Kreutziger (1967) to JI f .
a ow ractur1ng of th
' was mo I ie by Mc-
·
wiched between two pieces of aluminium foil has been described by Winkel.·· cuum by means of a spring-loaded u er b . e specimen un~er
vice Using the spring release th ppbl lock which houses a cuttmg
mann and Meyer (1968) and Winkelmann and Wammetsberger (1968). T ' e upper ock rotat f
frozen sandwich is clamped in a two-part brass block that has been p en and the tunnels are brought z'nt , . . es, ractures the speci-
o a 1ignment with th ·
cooled in liquid nitrogen and is then transferred to the evaporation u --. ·. r shadowing and replication. c evaporation source
When the fracturing (or etching) temperature is reached the specimen An apparatus for fracturing either bulk specimens
. or fro·· e , ·
fractured by tearing away tbe upper foil. der vacuum, developed by Stolinski (1975) b z n suspens10ns
In other single fracture methods one stroke of a blade or a wedge mec · ting-units. Fracturing is done us· I , ~an e fitted to conventional
' mg a c eavmg me ·h · (F"
nism is used to fracture the specimen under vacuum, as in the sliding form of a double-sided converging wedg h. h c amsm zg. 6.13) in
the vacuum chamber Th 1 . • e w ic is operated from out-
block freeze-fracture device originally described by Bullivant and A·. . e ocat10n of the fracture plane can be roughly
(1966). Based on this principle, Akahori et al. (1972) and Akahori and
hiura (1972) developed a simple device which provides good anticontan>i'
tion shielding and can easily be constructed in a laboratory workshop;
shown in Fig. 6.12, the device consists of two matching wedge-like ------
-
''i
blocks. The specimen support holder is inserted into the lower block. 3 2

upper block, which contains the knife for fracturing, is held in positid --- --- '
' I

'
,'1
I< I 11
a stop. When the stop is removed, the upper block slides down the in 1'1
plane, fractures the specimen and the two shadowing tunnels coni
,,1: I

alignment over the fracture face so that heavy metal and carbon (i,1
'· !

ration can take place. As in the Bullivant and Ames type II device ( ,ri I._!

the narrow tunnels of the upper cold block provide excellent pr


against contamination of the fractured speciinen surface by cond

, '-'-"'"'- St age

'Device for fracturing bulk s ecime , . .


-as the t\vo cleaving w d p ns or cell suspens1ons under vacuum. The frac-
e ges converge leaving mo 't 0 f h 1.
a-) Top view of the cleave Th . ' ~ s t e racture surface
r. e c1rc1ed numbers sho th l .
tion to the moving clea Th . . w e re at1ve positions of the
ver. e specimen is loaded · t h
Opening in the cleaver (posi( I) 0 . rn o t c specimen holder
ron · nee under vacuum th I ·
of the arrow so that ( .. , e c eaver ts n1ovcd .'I
· pos1tton 2) the two wedges · d f 'I
Specimen · Shadowing and ca b . . in ucc racturing as shown
. r on c1epos1t1on take place "th th. . .
Ver is moved out f th . . WJ c specimen 1n position
Fig. 6.12. Sliding cold block freeze-fracture device providing good anticoh_ 0 e way (pos1t1on 4) so th t th f
be .re111oved (R d · a c ractured spccin1en can
ing. (Redrawn from Akahori et al. 1972 and Akahori and Nishiut · e rawn from Stolinski 1975.)
340 Robards and Sleytr Low ten1perature methods in biological electron microscopy Freeze~fracture replication
341

predetermined by pre-aligning a chosen region of tissue prior to freezing in principle, the problem of aligning the f rac
. t ure t:aces wa 1
ine wh en methods were perfect d f' f . " s on y properly over-
the specimen holder. This method has the advantage that the cleaving wedg- . e or ractunng u d 1· · .
elmm (§6.3.3c). n er 1qmd rntrogen or
es only come into contact with the outer edges of the specimen, so leaving
most of the surface untouched (Fig. 6.l3b). The cleaver, which is main-. The second and most comm on met h od for p d .
acture faces under vacuum inv 1 h ro uc1ng complementary
tained at a lower temperature than the specimen, also functions as an anti,..< o ves t e use of hi dd .
oseley (1969) were the first to d . ngc ev1ces. Steere and
contamination shield for the freshly fractured specimens during etching, , 1
eve op a hmged h 0 Id .
ntary fracture faces (Fig 6 14.) Th . er to yield eomple-
The temperature of the specimen stage can be controlled by balancing coolt;• . · a · err compleme t' f
eere and Moseley 1970· Steer 197 ,). . nary racture device
ing using evaporating liquid nitrogen with heating using a small bl . ' e -' rs now available f D
pen d.ix .2). Subsequently ' oth er d es1gns
. ,
of com ] rom enton
. (see
heater. A modified form of this freeze-fracture stage has been incorporat'
e descnbed by Miihlethalcr et al (197 p ementary replica device
b) and by Sleytr and Umrath Cl974~; ;~; ~e Balzers apparatus (Fig.
into a Balzers BAF 400D freeze-etching apparatus (Stolinski et al. 1984).
Single fracture methods for use under vacuum have also been adapted f,
ersaJ electron microscopical . oth the Leybold-Heraeus
use in more elaborate apparatus for freeze-fracture replication. In . prcparat10nplant EPA 100 .
Steere-type freeze-etch module (Steere 1969, 1973), the specimen is fro
5 automatic freeze-fracture repl.iea t.10n umt. (Fi 6 14) and the Bioetch
plementary replica device (D g. · c · Both the Steere
onto a metal support, mounted on a temperature-controlled cold stage empsey et al. 1974) and the Sleytr and
fractured under vacuum with a cold steel blade, mounted on a
(§6.3.3b(i)) c

(ii) Devices to produce complementary.fracture/aces


The develop1nent of freeze-fracture devices which allow the retenti
both faces produced by a single fracturing process was stimulated by ·.
lems concerning the interpretation of membrane fracture planes, and
ture and mechanism of artefact formation (see reviews by Sle
Robards l977a, 1982). Within one year, four separate groups indepe
developed methods for obtaining complementary replicas (Steere an
ley 1969, 1970; Wehrli et al. 1970; Chalcroft and Bullivant 1970:
l970a), applying two basic principles.
Using one method, the so-called 'broken-capillary' or 'snapped;t
d
vice, the specimen is frozen in a small tunnel-topped metal capil(
1:--,
which is scored to predetermine the point of fracture. The tube i
the cold stage of a freeze-etching unit and fractured, after obtain·
[BBf
-.-.J ~
vacuum, with the arm of a microtome. The broken-off part fa
down, into a cage alongside the lower part, and both fracture fac
~~g~
cated simultaneously (Wehrli et al. 1970). A similar method, ada
sues, was later devised by Tonosaki and Yamamoto (1974). Al
quality replicas of con1plementary fracture faces were obtaine
methods, they had the disadvantages that the specimen hol
;---------L_ _ _ ___l_
variety of devices for the production o
--
---
.._,.

(1969); (b) MUhJethalcr ct al. (1970)· ~ con1ple~nentary fracture faces. (a) Steere
too complicated, that the broken-off part had only poor th
·,-(e} Nc1mut and Williams (1977)· (D(c) Sleytr and Umrath (1974a); (d) Milller
with the cold stage and that both fracture faces were not ace
1976). The modes of operation of~h -~ob~rds et al. (1980); (g) Mcnold ct al.
with respect to one another during shadowing. Using ese ev1ces arc described in the text.
342 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-fracture replication
343

Umrath device (Slcytr 1975) ean be adapted for use with the Balzers freeze- aterial (metal, glass) aligned in a cross-wise
. fashwn. (F'
rozen sandwich is mounted on th ig. 6 ·2; type G). The
etching unit. e pre-cooled device d f .
The complementary replica device for the Denton modular freeze-frac,, . ace under vacuum by re1easin th h an racturzng takes
. . g e catc A mod'fi 1 t'
ture replication unit (Fig. 6.14a) is loaded under liquid Freon (or liquid nF. Williams device, which allow f . · tea 10n of the Nermut
. s ractur1ng of two sandwich ,
trogen) which is cooled with liquid nitrogen to a temperature close to · ' n d es1gned by Robards et al.
0980 ) (F' . es at once, has
6 14
etch 2005 unit. With this mod'fi . tg. · 0 for the Leybold-Heracus
melting point (Steere 1973). Up to six specimens, frozen in the hinged ho 1 1Cat1on the fra 't · d . .
ers illustrated in Fig. 6.2 (type F), can be mounted in the fracturing devi h specimens under liquid nit ' c uring ev1ce rs loaded
rogen and released . t ·
Fracturing takes place under vacuum according to the Steere-type free uum by an electromagnet Th. d . au omat1cally under
· is ev1ce can th ·1
dard evaporation units. us casr Y be adapted for
fracture replication routine (§6.3.3b).
For the hinge mechanism developed by M uhlethaler and collabora enold and collaborators (1 972 1976) d . ,
evice' suitable for preparing ' escribed a drop-freeze-fractur-
(1970), the specimens are mounted and frozen between a pair of speci comp1e1nentary fra t f
supports of type B, C or E (Fig. 6.2). The device shown in Fig. 6.14b lets of suspensions The bas· . . cure aces from small
6. l 4g. A drop (0 5_ j 0 m . Ide. prmctples of the method are shown in
sents a slight modification of the original design which is comme · · min 1ameter) of suspe · · f
available from Balzers (see Appendix 2). The mechanism is loaded, n.. ds of two wires which are att h d ' · nsion is rozen between
ac c to a hi d
(see §2.4.lf). Under vacuum th d . ngc copper block mecha-
liquid nitrogen, with three frozen specimen 'sandwiches' and subseq . e rop is fractured b 11'
transferred to the pre-cooled cold stage (see Fig. 6.15 below) of the B part. To avoid heat transfe t th d , Y pu mg the two
. r o e rop from the
unit. The device is opened under vacuum by pulling the upper pa during freezing wires wr.th 1 th warm parts of the
' ow ermal c d · ·
using the microtome arm or lever. A hinged device similar to the M A.fter the drops have been ra 'di f on uct1VIty (tungsten) are
.. . P1 Y rozen the copper bl k
aler design, but especially adapted for type E holders (Fig. 6.2) w · .'·.n. ltquid nitrogen. The devi·ce, wh.IC h IS. preferabl
' oc s are cooled
suitable for sandwich- or propane jet-freezing of suspensions, has be flange, is then transferred t 0 . y mounted on a sup-
. any type of coatm · Th l'I
and described by Muller et al. (1980b) (Fig. 6.14d) (see also Miil\/ 1ncri::-:ase of the specimen d . . g unit. e slow tem-
_red with a thennocouple. udnfng pu1~p1~g down the vacuum unit
1984). an racturrng rs p f. d
The Sleytr and Umrath (1974a) hinged complementary fractu peraturc has been reached (M Id er orme when a suit-
eno et al. 1976).
designed for the Leybold freeze-fracture replication units con·
hinged plate with an adjustable cover (Fig. 6.14c). The cover ,-;eezeji-·acturing ·with a microtome
spring-clip holders to ensure good thermal contact between the _ successful and still most co1nmonl d .
cturing \Vas developed b M y use ' microtome assembly for
and the upper specimen holder after fracturing and, therefore, eq , , Y ooretal (1961) · d · ..
gmally described by Port d Bl. usmg a es1gn simtlar
of complementary fracture faces. The hinged plate is held to , er an um (1953) Th .
the Balzers type BA 360 M fr . : e microtome is
clamps, against the force of a spring and can be released either
s BAF 300 and BAF 301 Theeze~etchmg umt, or one of its suc-
or with an electromagnet incorporated in the universal specirrte>,,
which is used as a knife . Fi e mrc~otome arm and the attached
Leybold-Heraeus Bioetch 2005 freeze-etching unit (§6.3.3b), Si bout 77 K (-l oC) C ( . g. 6 .lSa), can be cooled with liquid
closed rivet-type specimen holders (Fig. 6.2; types H and l) can 96
done using e1.the ; utthmg (1.e. fracturing) of the frozen spe-
device is suitable for the production of complementary fract - r a mec an1ca1 th
ecimen is transferred from Ii . ~r erma1 advance mecha-
either tissues or suspensions. The hinge mechanism is loade ,
"Cooled (approximately 123 ~m~mtr~gen (storage Dewar ves-
nitrogen and transferred to the cold stage. ates both liqu'd . '. 150 C) specimen stage. The
Nermut and Williams (1977) described a hinged, springol . r nitrogen cooling a .
rntained at te s we 11 as a heating element
freeze-fracturing cell monolayers (Fig. 6.!4e). The gadget mperatures from 77 K ( 19 "
an accuracy of + K D . - 6 CJ to above 293
the Balzers cold table but can easily be adapted for ot 0 01
tured in one run "th. . _epend1ng on the number of sam-
units. The specimen is frozen between two pieces of ' ere are different designs of specimen table
344 Rohards and Sleytr Low temperature met/rods in biological elec/ron micrmcop_i Freeze-fracture replication 345

F ig. 6. l 5. Balzers freeze-etching unit. (a) Microtome in models BAI- JOO anJ JO
model BAF 400; (c) microtome in model BA F 400; (d) evaporators and anll•'OD
shielding in model BAF 400. (Photographs reproduced by courtrsy or Bali
rf ods in biological electron micro.scopy Free=e-fraclure replictttion 347
346 Robards and Sleytr Low temperarure me ' .

. th sample (Nickel et al. 1978). T he most although not completely closed, liquid nitrogen-cooled anticontamination
and different ways of mounting d ~ the Balzers unit are types A-E (Fig. shield over the specimen a rea. The unit can be equipped with a complemen-
· en supp orts use m .
common specim t hoar frost from the specimen stage tary replica device, a rotary shadowing specimen table (§6.4. l f) or a liquid
ood thermal contac ,
6.2). To ens ure g d . J' 'd Freon 22(or 12) before thespeci- nitrogen-cooled microtome which, compared with the previous model (Fig.
is washed off with a b rush soakefi m iqudi (BA 360 M) cutting is manually 6. J Sa), is considerably smaller and consequently more economical in liq uid
. ~ d to it In the 1rst mo e1 .
men is trans erre · (BAF 300 and BAF 301) cutting can be nitrogen cons umption. Furthermore, it does not have a thermal advance
t ·n the later modes 1 .
contro 11ed b u 1 t lled Experience wi th many speci- rnechamsm. As with the earlier models (BAF 300 and BA 301 ), the micro-
f lly or manua11y con ro .
both automa ica h · . 1 dvance alone is usually adequate. tome can be controlled either automatically or manually. To avoid hoar
that the mec amca a
mens h asshown Id proceed as follows. Cool the specimen frost fo rmation on the liquid nitrogen-cooled cold shrouds and specimen
. l freeze fracture run wou ~
A typica - K ( - l SOoC). After venting the umt (pre1erabh stage, the frozen specimens, attached to the specimen table, can be trans-
123 ferred from liquid nitrogen to the specimen stage against a counter-flow of
stage under vacuum to h f specimen quickly to the specimen
. ) transfer t e rozen
with dry nitrogen • d h vacuate the chamber agai n. Dunng dry nitrogen gas thr<;>Ugh a simple door or more complex air-lock system.
· led forceps an t en e
stage using coo . d down to liquid nitrogen temperature Thus, when equipped with electron guns, several preparations can be made
. the microtome is coo 1e .
evacuation h f turing temperature. Fractunng, etch· without warming up t he cold parts for defrosting, as was necessary with the
· 1·s warmed to t e rac
and the specimen d t a working pressure of about 2 x 10 earlier models. The pumping system is automatically controlled and can be
ing an d rep ica
r ' tion are then . one. . a t equired the specimen temperature equipped with an oil diffusion pump, a turbo-molecular pump or a refriger-
. 1 5 10- 6 Torr). If etchmg is no r , . 10 tor cryopump system (see Fig. 6. 10 in §6.3.2). The specimen can be
Pa(· x . . K(intherangeofl53 tol63K,- 1- to
173
is usually mamtamed ~elo.w t d onto the surface tmmedia tcly alter observed through a window in the work chamber and an opening in the cold
oC) d the replica is evapora e shroud, using a binocular microscope attached to the front door (Fig. 6. I 5a
- 110 an . . uired (§6.5) fracturing takes place at a spec11ncn
the final cut. If etching is req f h fi al cut place the microtome arm nd b). Due to the large antico ntamination shields over the specimen area,
173 K A ter t e m •
temperature of a b out . . to act as a cold trap to redu the new Balzers freeze-fracture replication unit gives excellent results even
'b\e above the specimen
as rapidly as poss1 f h. remove the knife and evaporn hen freeze-fracturing is done close to 77 K.
. I t . nation A ter etc mg,
any poss1b e con am1 . b b eking layer (§6 4). Agam. pla A basically similar microtome design to the early Balzers model (BA
. t ·al and the car on a ~
the sh~dowmg ma en . efore admitting air to reduce hoar fro~t J60M) was used in the NGN freeze-etching unit developed by Preston and
the knife above the specimen b . . oves the chances of the rcpl lamett (Barnett 1969; Robards and Cooper 1971). The essential difference
mation on the replica. This precaution imhpr vent the bell-jar and wa that the blade holder of the microtome arm is moved directly across the
h · g fluid Afte r eac run ,
floating on the t awm . . . d specimen table). The sin 1men and not in a circular path as in the Balzers unit. In addition, spe-
led parts (microtome an d
up a nd defrost a II coo h . tome knife is replace a kmfe systems for this unit have been devised by R oba rds et al. ( 1970)
h icb is used as t e micro ,
edged razor bl a d e, w f lly degreased before use h use two razor blades on the microtome a rm for rough and final cut-
. and must be care u
each fracturing process 'd ontamination of the fra and for varying the knife settings. Other instruments wh ich resemble
chloroform) to avo1 c
washing in acetone or The Balzers apparat Balzers unit were developed by Koehler (1966, 1972b) and Elgsaeter
. , . hydrocarbon vapours.
surface with o utgassm g . d electron guns. Ocpost ) There is now a variety of commercially available apparatus (see
· · aporators an
available with both resistive ev d . a qua rtz cry.stal d1x 2) incorporati ng less complex knife mechanisms for controlled
rates during evaporation can
· be ch ecke us mg
. ' ts there is the option o
r hi pie specimen fracturing.
(§6 4 l e) On later Balzers freeze-etching um for the ,acuUJI
eithe~ a~ oil diffusion pump or a turbo-molecular pump Free:e-fracturing under impro ved vacuum conditions
· · unil
tern. freeze-fracture repltcat1on
T he latest version of the Balzers &'. • t es as the earher sation of material onto the fracture face of the freshly cleaved spcci-
. JI the same iea ur
BA 400, incorporates essentla .Y . . 6 l Sb c and d) ha an the period between fracturing and replication is one of the main
(BAF 300 and BAF 301). This urut (Fig. . •
Freeze-fracture replication 349
. biological electron microscopy
Low temperature met I10ds m
348 Robards and Sleytr

The Steere-Denton freeze-fracture replication unit with liquid nitrogen-cooled dou-


"1rnud . (a) View of the module consisting of a metal chamber (ch) with a glass cover-plate
lhc fracturin g assembly sits within o pposi ng hollow tubes (t), serving as cold shrouds,
have apertures (a) through which the evaporation sources (e) can deposit the shadowed
(b) Double shroud with complem entary replica cap (c) in place in outer shroud (os)
rlmcnt. (c) Double shro ud with complcmcnlary replica cap in inner shroud (is) comparl-
hl) Diagram of specimen tube and specimen stage, with specimen holder clamped in
nJ ~urrounded by inner chamber of shroud. Left 10 right: units in position for fractur-
l(d-down pos1t1on after fracturing (e.g. during etching); aligned for shadowing and car-
bon coating (Reproduced , with permission, from Steere and Moseley 1970.)

s of artefacts (§6.8 .3) since fine structual details can become masked
anced by specific decoration. Therefore, it is important that the
is as good as possible. Two different approaches have been made
ning an excellent vacuum environment, with a minimum of condens-
aases in the specimen area during exposure of the fracture plane. One
350 Robards and Sleytr Low temperature me thods in biological electron microscopy
Freeze~fracture replication
351

. f ompletely enc1osmg . the ·speciinen before fracturing


method consists o c inwards until the specimen is encJosed in the inner chamber of the shrouds.
. . . (or hehum- coo 1e c old trap or cold shroud (Steere.
. ) , d
within a hqmd mtrogen- . d r l 982a) The other method -Specimens are fractured by rotation of the shroud and the specimen tube
1973· Sleytr and Umrath 1976; Nie ermeye 1 . 'n opposite directions, which lifts the bar attached to the specimen cup and
'
involves more comp1ex ultra-high vacuum techno ogy. urns the upper part of the hinged specimen holder through 180" (Fig. 6. l 4a
nd 6. l 6d). If etching (§6.5) is required, the specimen stage is warmed to
(i) Cold shroud devices . . . 'th a liquid nitrogen-cooled bout 173 K while the fracture faces remain enclosed by the shroud. For
f 1 ephcatJon unit w1 adowing (§6.4), the specimens are recooled to 123 K ( - l 50' C) or colder
The first freeze- rac ure r . , developed by Steere (1969) 0

, d th specimen area was d the specimen tube and shroud are rotated to the 45" shadowing posi-
ble shroud aroun e h t (cor detailed description see
1 (1970) T e sys em '' . n. The specimen temperature is then raised to 163 K ( - 110°C) for evapo-
Steere and Mose ey . V I corporation (see Appendix 2)..
. I f Denton acuum n . _,- ·on of a carbon backing film. Finally, the vacuum system is brought to
1983) is availab e rom f plication module consists
. . 6 16 the freeze- rac1ure re
illustrated m Fig. · ' ( ) and two opposing hollow t. ospheric pressure with dry nitrogen gas and the comple1nentary replica
) ·th , glass cover p 1ate P . ice is removed from the vacuum chamber.
chamber (ch wi a 'double chamber shroud' consisting of an i .
(t) which can be rotated. A d th moveable hollow tubes. Theo cold trap-type freeze-fracture replication unit with a high degree of au-
h ber 1s attache to e . ation was developed by Sleytr and Umrath (1976) and Umrath (1978)
and an outer c am large opening, . throu gh which the specimen 1s mou .·
h manufactured by Leybold-Heraeus as the Bioetch 2005. The apparatus
chamber (os) as one . t b and two small apertures (a)
. t· ge on the inner u e, , 6.17a) consists ofa cylindrical vacuum chamber equipped with quick-
on the specimen s a . d the cold shroud are cooled_
. B h th specimen stage an . . ection flanges for the specimen cup transport device and the evapora-
evaporation. ot e th outside and after introduc1n
. . h gh tubes from e ' and a liquid nitrogen tank, which acts as a vacuum baffle. Two concen-
liqmd mtrogen t rou .h t to each other so that the
b rotated wit respec
specimen, they can e b cold surfaces during fracturing a~d e iquid nitrogen-cooled cylinders act as cold shields for the specimens;
men is completely enclosed y 1 osed to warm surfaces dunng these cylinders is fixed to the nitrogen tank, the other to the moveable
Thus th~ fracture planes ar~ on ~t~x~o}d surfaces is only provided en stage. A pneumatically driven bellows-sealed cylindrical air-lock I
lion. This complete protection w . used (see Fig. 6.14a). Int moved over the outer cold shield to avoid ice condensation on the I
I ,
f t e specimen cup 1s during venting of the air-lock for specimen and evaporator
complementary rac ur
r , f unit rac ure caces
. f t
i;
can be also obtained
ge. In the schematic drawing (Fig. 6. l 7a) the air-lock is shown cut
freeze-fracture rep ica ton . , seal el blade mounted at the end o
ing' the specimen surface with a ~ n cannot be surrounded b h the central line to demonstrate the opened and closed positions.
pull rotary probe, although the spec1me perature of the specimen stage and of the specimen cup (Fig. 6. l 7b)
hie shroud while this takes place. r r run using the complem ,:·accurately controlled using two independent platinum resistance
For a typical freeze-fracture rep ica ton h'nged holders (Fig. 6. eters. The specimen stage and the inner cold shield attached to it
. h · ens are frozen 1n 1 . rotated to predetermined positions during the preparation cycle
lica techmque, t e spec1m t d under liquid Freon 221
· are then moun e
The frozen spec1me~s ~- 14a and transferred with special fo 'lilotor. For routine preparation, specimens are frozen in specin1en
men cup illustrated m Fig. 6. Id The unit is pumped do of type H or I (Fig. 6.2) and mounted under liquid nitrogen on
3 K· l 50oC) co stage. lementary fracture device (Fig. 6.14c) in the universal specimen
pre-cooled (12 ' - r uid nitrogen temperature. A!1
cold shrouds are cooled tol iq10-1 Torr) is obtained, the n in Fig. 6. l 7b and c. Before use, the unit is evacuated and the
(] 3 10-3 Pa < x .
vacuum . x l~OoC) and held at this temperature Jin, Ogen tank is filled. This automatically cools the two cold shrouds
warmed to 173 K ( - c has d1'sappeared from th.e cimen stage. The specimen cup, with the specimens, is now trans-
. d d ing the trans.er . . liquid nitrogen to the preparation plant using a bellows-sealed
which has forme ur l d down to liqmd nttro
. Wh the shrouds have coo e vice (Fig. 6. 17a). When a vacuum of 10.6 Pa (8 x 10- 2 Torr) is
specimens. en _, Pa ( x _6 Torr) or better
2 10
ture and a vacuum of 2.6 ~ (- l SOoC) or colder and the sh
10 the air-lock, the air-lock is opened automatically and the speci-
specimen is cooled to 93 be inserted through the aligned openings of the outer and inner
352 Robards and Sleytr Low temperature nwt Iro ds m
. hiolor;ical electron microscopy
Freeze-fracture replication
353

cold shields into the socket in the pre-cooled specimen stage. The transfer
a lco"•" '"'"" •o• ;device is then locked in the upper position shown in Fig. 6. l 7a. Completion
I !I "mol• """""' fthe preparation run is now fully automatic. The individual operating par-
eters, such as specimen temperature and vacuum status during the frac-
ring process, time of etching and evaporation conditions, can all be preset
an electronic process control unit.

Filling pipes
The first automatic action in the cycle is that the specimen stage, together
h the inner cold shield, is rotated in relation to the outer cold shield so

Main air-lock.
valve drive
--;-J
Motor for spe:eiii('
stage and inn&r
t the specimen is in a completely enclosed position. The two liquid nitro-
, -cooled shields now act as an opaque screen against any contamination
condensable gases. This is the position in which fracturing and etching
cold shield -
Outer (stationary) cold shield performed after the preset criteria for temperature and pressure have
Inner (rotatable) cold shield
achieved. Only immediately prior to evaporation is the specimen stage
Evaporator
(Pt-C,C) ed so that the cold shroud is opened again. To reduce any sudden out-
"ng during evaporation, both resistive sources (or an alternative electron
To the high
vacuum pump evaporation system) are degassed by automatic pre-heating at the be-
'ng of the cycle while the specimen is still enclosed within the cold
c ds. Finally, after the specimen has been rotated back to its original
b
Double replication device, on, the specimen cup can be removed via the air-lock using the transfer
spring loaded,
magnetic release
Specimen
support Specimen tray ead of a double replica device (see Fig. 6.14), other gadgets can be
~-+-7"'1 Magnetic armature
ted in the universal specimen cup, such as a device for contamination-
Magnetic coil 'ansport of specimens from liquid nitrogen into the vacuum chamber
Pt-resistance Coaxial connector J 7d). This is particularly useful when fracturing or handling speci-
thermometer
der liquid gases (§6.3.3c) such as is called for when using freeze-
Specimen cup
n combination with cytochemical techniques (§6.9). When equipped
.oil diffusion pump, the attainable pressure of the cold-trapped
.is in the range of 2-3 x I0- 6 Pa (1.5-2.25 x 10- 8 Torr) as measured
e cold shields. Within the liquid nitrogen-cooled shields the partial
Of water is considerably better and should reach the order of
d
----------------------------
_Leybold-Heraeus Bioetch 2005 freeze-fracture replication unit with sophisticated
:---~tem and various devices for the study of specimen fracture faces obtained under
'd nitrogen or liquid helium. (a) Cross-section of the apparatus. (b) Detailed view
cup with complementary replica device. (c) Complementary replica device after
Device for contamination-free transport of specimens freeze-fractured under
or liquid helium into the vacuun1 cha1nbcr. The anticontan1ination hood is

LN -fracture (He-fracture) trans fer device


~fractured specimens under liquid nitrogen. (Redrawn fron1 Slcytr and Umrath
2 1976andS!eytreta1. 1981.)
354 Rohards and Sleytr Low temperature metI10d~,.m biological electron microscopy
Freeze~/Tacture replication
355
11 4
•l.2 x 10- Pa (9 x J0-i Torr). Thanks to the air-lock system, the cold
.;·Shrouds remain continuously under high vacuum and thus maintain their
··umping capacity during specimen and evaporator exchanges.
A helium-cooled cold shroud device suitable for installing in a standard
alzers 301 vacuum uhit for freeze-fracture replication at about 10 K has
n developed by Niedermeyer (l 982a). A schematic drawing and illustra-
n of the cooling device are shown in Fig. 6.18. Cooling of the whole de-
e is achieved using a continuous flow cryostat in which the coolant is
_awn, by a vacuum pump, from a liquid helium container through the ;·1
ling coils. The required temperature is adjusted by changing the helium
-rate using a valve situated between the cooling coils and the pump. As ii
8 -_-y
trated in Fig. 6.18a, the vacuum-isolated helium transport pipe (2) is ('

ected to the cooling device which is on top of the vacuum chain ber. (1

pipe inside the vacuum chamber (6) leads directly to the inner shroud !;I

A---i---• here the lowest temperature is achieved (approximately 6.0 K). Subse- .?1
I
tly, the cryostat (11), with the specimen stage (10) and specimen holder ,,;
' I
's cooled. After cooling the cryostat (ll), the helium cools the outer :ii

;'1
d (7) which reaches a temperature of approximately 20 K. Finally, the
passes through a cryopump (4) to improve the vacuum in the
ber before it leaves the apparatus (I).
a-----:,1 I I
allow rotary shadowing (§6.4.11), the specimen stage can be rotated
::090 ---
.-J.<-1.~./• ~/./~
7
A an external motor via a rotary lead-through (14). The connection to
: '==-1 ' 1 / 8 ostat is achieved with a strong permanent magnet (12). To allow the
-1'--ltt+--9 hoar-frost free transfer of the specimen holder through a swing door,
le air-lock system has been developed which can easily be connected
.chamber with a flange (Niedermeyer 1982b). At the start of the cool-
ess, the vacuum is in the range of 13 x l0- 6 Pa (9.7 x I0- 8 Torr).
the cooling process, this improves to 3 x J0- 6 Pa (2.2 x J0- 8 Torr).
13 possible to measure the actual vacuum within the shroud but the

Liquid helium-cooled cold shroud (developed by Niedermeyer 1982a) for the Balz-
um unit_ This attachment allows freeze-fracture replication at about 10 K. (a) Sche-
g of the system. I, lfelium outlet; 2, helium inlet; 3, vacuum seal from the nonnal
atus; 4, cryoputnp; 5, tubing around cryopump; 6, helium pipe; 7, outer shroud;
ud; 9. specimen holder; 10, spccirnen table; 11, cryostat; 12, permanent magnet;
-Pillar for the shroud; 14, rotary lead-through; 15, base of the vacuu1n chamber.
~:~,B',helium pipes leading from A to A' and B to B' (dashed lines, not drawn to
:otograph of the device showing the position of the swing-door (SD) for specimen
e position of the quartz crystal. (Illustration kindly provided by W. Niedermeyer
and reproduced with his permission.)
356 Robard~ and Sleytr Low temperature methods in hiological electron microscopy
FreezejYacture replication
. 357
author estimates a vacuum of about 1.3 x 1o- 8 Pa (I x I 0- Torr). Analysis
10
approximately 13.3 Pa(! x 10-1 T .
of the residual gases outside the cold shrouds, before and during replication b f orr) is obtain d· b k
car. ons rom the pump (6) t th e ' ac -streaming of h d
by mass spectrometry revealed that the contamination-free vacuum condi: sieve trap (8). The specimen ~tag: (c~a)mbehr
(hl) is prevented by a mo!e~u~~;
tions were equivalent to those of ultra-high vacuum systems, even down to ' w Jc can be c oo Ied down to 80 K
a specimen temperature of 12 K. The effectiveness of the helium-cooled ..
cryoshield and the cryopump in the standard high vacuum apparatus wa; Position of cylind
freeze-dryin er ( 12) during
9 or etching
confirmed by analysis of the quality of yeast plasmalemma fracture fa ·
as test specimens.

(ii) Ultra-high vacuum technology b


Relatively complex ultra-high vacuum (UHV) technology provides an
native means of obtaining an excellent vacuum around the specimen. U a
pumping units are designed to pump out all gases in the system and not 0
those which represent the potential hazard of leading to condensation c.
tamination when the specimen is sufficiently cold. Due to its complex;:
UHV technology has rarely been used for freeze-fracture replica'
although some relatively early work was done with such systems (Kreu
1970). Koehler (1972a) described the first attempt to use a UHV freeze
ture replication unit. The apparatus incorporated a titanium getter
and an orb-ion pump. The roughing system consisted of a Gast Cl!rb
turbine pump and two cryosorption pumps filled with Linde molecula
material. Fracturing was done with a simple freeze-fracture block de
single fracturing under vacuum. Since the block was loaded with sp
under liquid nitrogen external to the apparatus, the problem of frost
tion during transfer into the unit does not seem to have been ov d

~ e~··
this experimental system. /'··"
UHV freeze-fracture units with air-lock systems have been dev~r
Escaig and Nicolas (1976), Gross et al. (1978a) and Balzers U
apparatus from the French group (Escaig and Nicolas 1976), ill
Fig. 6.19, is manufactured by Reichert-Jung (see Appendix 2). Th,, Rotary shadowing
device
~ ..
Complementary .
replica device
0 evice for specimen
system comprises a rotary vane pump (6), a liquid nitrogen-cool 1
rotary shadowin°; fracturing with
pump (7) and an ion-getter pump (4). The manual and electric · Ultra-high razor blade
< . vacuum unit for freeze-f .
stainless steel bellows and Vi ton gaskets and the rotary feed-th( g and Nicolas 1976) (a) Th . . " racturc replication at temperat d
work chamber are also bellows-sealed. The system can be bak ption pump (7 an~ . e unit incorporates a rotary pum u_res_ own to 12
d via . ) an ion getter pump (4) S . p (6), a liquid nitrogen-
lures up to 250"C. An ultimate pressure of 2.67 x 10-- 6 Pa an an-lock system (9) F I . pecimcns attached to the h Id (I
e position f . or exp anation of oth o er 0)
is reached in 4 h without baking and a pressure of 4 x !Oc, eeze d . o the cooled cylinder (I2) in relatio t er numbers see text. (b) Detail
Torr) in 10 h with a baking cycle of 4 h. For the primary· e~c~ ~~;ng. (c-e) Several types of specin1en hol~ o t11c sp~cimen stage (I I) during
chamber is pumped by the rotary pump (6) and the liquid n:if cturing ~s~nco~plementary replica device for ro~~:yrc ~vad1lab_Jc, including a rotary
g a razor blade (e) · (Red rawn from · . s a . owing (d) ' and ad evice .
sorption pump (7). The rotary pump is used alone until courtesy ofR . 'h ong1na1 illustrations provided b
ew ert-Jung.) y
358 Robards and Sleytr Low temperature methods in biological electron n1icroscopy Freeze:frarture replica/ion
359

with cold nitrogen gas or to 12 K with liquid helium, can be tilted, using using five pumps . . working Jn . series.
, . . a rotary vane u . .
an external control, to any shadowing angle. The stage is partially encased pumps, a hqmd nitrogen-cooled baffle 1 . p mp, two ml diffusion
in a cylinder (12) which can be cooled with liquid nitrogen to trap water _sublnnation pump A Vito k . p us Meissner trap and a titanium
. n gas et is only u d f: h .
vapour liberated from the specimen if etching is done. A modification of amber, all other flanges b . 1 . se or t e mam flange of the
. emg sea ed with gold · .
the cooling system for the shroud, in which the flexible nitrogen-supply nng, shadowing and coating f . wrre. The unit a11ows frac-
o specimens at 77 K ·
pipes are replaced by flexible copper braid, has been suggested by Allain an~;:; 33 >< 10-1 Pa (! x 10-' Tor) S . . ma vacuum of about
. r . pec1mens are froze .
Ardonceau (1983). This has the effect of increasing reliability while retairr:;:; ,bes I mm m diameter (see Fig. 6. 11 b in " n m scored capillary
6
ing the high quality of replicas. The electron guns for heavy metal (14) a ... trogen (Fig. 6.20) in a spe · . § .3. 3a) and loaded under liquid
. cnnen cartndge which . h
carbon (15) evaporation are water-cooled. It is possible to adjust an exter ulator. Together with the t "d . IS t en screwed to a ma-
car n ge a cyh d .h
control so that the electrodes of the electron guns are moved up to 10 manipulator which allow t r, n er wit a lid is attached to
s rans1er of the C' t ·d .
This means that several evaporations can be made (depending on the th' er the protection of liquid .t . ar ri ge into the air-lock
m rogen The an I k 1.t If .
ness of the evaporated layers) without the need to vent the work cha chamber system. In the fi t h b. - oc se consists of a dou-
. rs c am er the 1· "d · .
to change the evaporants or adjust the guns. A quartz crystal oscillator removing the lid from the 1· d ' 1qu1 nitrogen is poured out
cy m er and the , t ·d ·
the second chamber Whe car ri ge IS then transferred
allows deposition rates to be monitored (§6.4. le), while a quadrupole · n a vacuum of ·
spectrometer (19) provides information about both total and partial r) has been achieved a gat , approximately 4 Pa (3 x 10-2
' e connecting the · h
ber is opened and the cartr'd .
1 ge is mam c amber to the UHV
sures in the system. . screwed to th .
The sample holder, having been loaded with specimens under liquid . The specimen temperature is th . e pre-coo1ed spec1men
, . en raised to - IOO'C (173 K
gen, is introduced into the air lock (9) which can subsequently be p n oi both Freon (if it was used,1or freez1ng) . and · , · .) for subli-
down to 1.3 x 10- 1 Pa (1x10- 3 Torr) within 2 min. During this t' cooled back down to -196oC 77 K ice_contammants, and
temperature of the specimen holder rises by about 20 K. When th rutrogen-coo!ed lever which b k( ). Fracturmg is done with a
. rea s the capilla · d
ens Jn the region of the . . ries an the contained

al
(13) is opened to allow the specimen to be inserted into the cham score-mark (as md1cated in Fig. 6.1 lb). The
pressure within the work chamber only increases to about 1.3 x
(1 x J0- 6 Torr). The pressure falls to 4 x 10- 6 Pa (3 x J0- 8 Torr) as,_ Screwdriver
6
the valve (13) closes and reaches approximately 1 x I0- Pa (7 x LO Ii
Specimen capillary
j·1

@!~):··
after about 5 min. This good vacuum, together with the very limif ermined / / Manipulator flange

sure increase during evaporation, means that fracturing can take pf re zone--0 ·· ··:,:·:·: ."·.:· .::: Manipulator shaft 11
Q ~ ;'.: 0 :·.; - - .·.·. . ··· Specimen cartridge Ii
out any problems of contamination down to a temperature of 12 Specimen ;.::.. , :;:·· LN ;.:. ·.. : .:·: Cyhnder
cartridge °' •• 2 .. · · ··
----- lid
et al. 1980).
A variety of specimen holders has been developed for the Esc ·
tus. In addition to the standard hinged complementary rep·
revolving specimen supports for rotary shadowing (Fig. 6.19c), c
tary replica devices (Fig. 6. l 9d) and devices for fracturing with a ..
(Fig. 6.19e) are available. During freeze-drying or freeze-etchi
the cylinder (12) is cooled with liquid nitrogen and maintaine
Inner gate CartridQe screw
tion shown in Fig. 6. 19b to trap on the internal walls the su of airlock
vapour molecules, which would otherwise destroy the go
'Specimen mounting and two-chamber air-lock s . .
vacuum. vacuum unit (from Gross 1977 d ysten1 for inserting specimens into
1 · an Gross et al 1978 ) s ·
The UHV equipment designed by Gross (1977) and Gross, ary tu_ bes (see Fig. 6.11 bin §6
d I . 33
··aw
,) h' · a · pec1mens are frozen
rcharcloadcd · t I ·
is based on a converted Balzers BA 350 U unit. The system er igrnd nitrogen. (Redrawn afte Jn o tie specimen cartridge
r Gross 1977 and Gross et al. 1978a.)
360 Robards and Sleytr Lo;v temperature methods in biological elertron microscopy Freeze:fracture replication
36!
broken-off parts are collected in a liquid nitrogen-cooled baffle, thus pre- ,into the ·high-vacuum lock with ou t f rost contam1natio
. .
A
venting contaminative outgassing. Heavy metal and carbon are evaporated increase 1n specimen ternperatu d . n. n unacceptable
. re unng transfer th h h .
with outgassed electron guns. With the exception of a very short ( < 1.0 s) lock is prevented by cooling the cold t bl . . roug t e high-vacuum
a e with hquid "t ·
pressure increase during fracturing, the whole process, including evapor"' vacuum lock. The temperature of this tabl . m rogen m the high-
ation, is performed in a vacuum of approximately 2.6 x 1o- 7 Pa (1.95 x lW~ d 123 K ( + 20 and -1 SO"C) bl' e can be adjusted to between 293
' ena mg freeze-dry· t b
Torr). The deposition rate is about 0.2 nm s- 1during platinum/carbon sha~, od vacuum conditions The v 1 b
· ave etween the h" h
Ing o e done under I'; I

do wing (§6.4.1 ). HV vacuum chamber is ope t d . ig -vacuum lock and the


.
vice. ra e automatically b Y th e specimen
. · .
insertion
Balzers Union has produced a commercially available UHV freeze-frifei
ture replication system (Type BAF SOOK). The automatic vacuum unit c
prises a refrigerator cryopump system (see also Fig. 6. I 0 (type Ill) in §6.. Freeze-fracturing outside the coating unit under a liquid gas
which produces an ultimate vacuum of 5 x 10-s Pa (3. 75 x 10- 10 Torr);
gether with a catalyst trap to remove oil vapour and an integrated clO Freeze-fracturing under liquid nitrogen
. .for the evaluation of fractur e t:aces 0 b ta1ned
circuit hot water system for heating the chamber to 80°C. A gate val · thods . under r 'd .
positioned between the cryopump and the UHV vacuum chamber an e ongmally developed independent! b . iqm mtrogen
pass pumping is possible using a turbo-molecular pump. An auto' mayer (1966). y y Bullivant and Ames (1966) and
pumping system control monitors the different vacuum systems and he routine preparation procedure usin th .
but the individual elements can also be manually controlled. The spe trated in Fig. 6.21 Th g. e Bullivant-Ames device is
. . e apparatus consists of th r .
table is refrigerator-cooled and, in combination with a heating elemen each of 70 mm diameter (Fi . 6 21 ree. cy mdncal brass
peraturcs between 8 and 293 K (-265 and + 20"C) can be achieve wer tier (1) attached to a sm!ll .ol a)~:efore fractunng the specimen,
an accuracy of ~ 0.2 K. A tilting device on the cooled UHV specim e two other parts (2 3) f th p YP pylene (or metal) container (4) ;':
. . ' 0 e assembly are all · d
allows for continuous adjustment of the evaporation angle betwee in liquid nitrogen in an expand d I immerse and cooled
90°. Complementary replica devices (Muller et al. 1984) and specirrf fracture planes the frozen e_ po ystyrene container (5). To obtain
' specnnen with its hold .
ports for TEM grids (e.g. for freeze-drying) can also be used. The tr! se plate and is fractured . er is mounted onLo
mechanism on the fracturing device is cooled to ~53 K (-220°C 6.2lb). For complementarusmgla scalpel or single-edged razor blade
·. Y rep 1cas the spe · · .
first cooling stage of the refrigerator cold table. Three autom llow specimen holders (Fig. 6.2· t ~ . c1men is mounted within
cooled electron-beam evaporators allow up to eight freeze-fract fashion The f . '. YP I) positwned together in an end-
. rozen Specimen IS fractured d I° . .
tion runs per UHV cycle. Since the evaporation is in an upward the holders apart, and the holders un er _1qmd mtrogen by
not only platinum/carbon but also tantalum/tungsten, silver, gold 2!c) in the base plate S . are then pos1t10ned side-by-side
. pec1men holders of t e G (F
substances can easily be evaporated (§6.4. I). A manually controlf y two spring clips attached t 0 th I YP ig. 6.2) can be
; · e ower block and f ·
matically operated shutter is cooled to ~ 53 K ( -220"C) and itrogen is performed b II. ractur1ng under
the quartz crystal film-thickness monitor. The high-vacuum I.Cf g, the middle cylinde;:i~h i:e t~; ou~f~r ~heet (Brown 1981 ). After
men transfer can be pumped with the turbo-molecular pump w; to position on a pin in th I : m) shadowmg tunnels is
•· h f e ower cylmder and fi 11 h
vacuum of ~3 x 10- 5 Pa (2.25 x l0- 7 Torr). If the pressure iJ1 unctions as a protectio 1·d . ' ina y, t e upper
n 1 , is placed on t f h .
< 5 x 10-s Pa (3.8 x 10- 7 Torr), then the pressure in the tJ The whole assembly tog th .h op o t e middle
' e er wit some of the r .d .
chamber rises from <I x 10- 7 Pa t_o ~ 5 x 10- 6 Pa (3.8 x lQc out of the expanded o1 s . 1qu1 nitrogen, is
specimen transfer but recovers within 5-10 min to the origi hit (Fig 6 2ld) wh _P. yfityrene contamer and transferred to a
. . ere it is Jtted t h b
high-vacuum Jock allows UHV cycles ofless than 15 min. Al nnels in proper alignment with on o t e as.e plate with the sha-
using a counter-flow of dry nitrogen gas allows the spec· ected with a m t· I . the evaporation sources. The top
ea Wiretoasmall I t. I
immersed in liquid nitrogen and then passed through a rot ator (Fig. 6.2le) th b 11 . . e ec nca ly operated crane in
' e e -Jar is placed in position and the unit
362 Robards and Sleytr Low temperature me lh od·~ in biological electron microscopy
Freeze-fracture replication
363

under these moderate vacuum conditions (Bullivant 1980b).


Specimen
Rather different from the Bullivant-Ames technique, which can be
adapted to almost any evaporation unit, is the Geymayer (1966) method,
willch was especially developed for the Leybold-Heraeus EPA 100 EM
reparation plant (Geymayer 1967). This is now manufactured by Paar
a .G. (Austria) as the EPA IOI (Appendix 2). The system uses the facilities
ovided by a revolving specimen stage and a cylindrical cold trap. Speci-
ns are mounted and frozen on specimen supports of type C, Hor I (Fig.
) in such a way that the material is slightly protruding (Fig. 6.2; Ca). Fro-
specimens are mounted on a carrier plate immersed in liquid nitrogen
fractured with a single-edged razor blade attached to a simple Reichert
omicrotome (Fig. 6.22a). After fracturing, the specimens are transferred,
g a special transport device (Fig. 6.22b), while protected by the liquid

d~
ogen trapped between the hood and the carrier plate, into the evapo-
e . n unit (Fig. 6.22c). This is often done through an air-lock system to
'd hoar frost formation on the cold shields. During pumping down, the
==:::__~---~
'men is kept very close to a cold trap. After a high vacuum has been
d ~ frccze-fractunng . un der liquid nitrogen using the. Bullivant-A b bl•.· ined, the specimen is either replicated immediately or, if etching is
Fig . 6 · 21. Proce ure or . d apparatus comprises a ase-
9661 (a) The three-here
vice (Bullivant and Ames I . 'ddl 'ecc with tunnels (2); and a lid (3). ired, firstly allowed to warm up. Platinum/carbon, or other heavy met-
11 ·ontainer (4)· a n11 e p1 d d.
which is fixed to a sma c ' d' (b c) Speciincns are loa e 1n n be evaporated at any angle between 0 and 90° and the carbon back-
. 1. ·d nitrogen before 1oa ing. ' .
assembly is cooled 111 iqui . b . a knife (b) or by tensile str
. . d th n fractured either Y using d yer is deposited while the specimen both swivels and rotates. The Gey-
under liquid nitrogen an e . . b -block (1). (d) The complete an
. t · l ylindncal hole 111 the ase fi technique was later adapted (Sleytr 1970a,b) to allow the production
are placed in the cen ra c . .. . fill d .th liquid nitrogen and trans e
. h , . n in position. is J e w1 . . plementary fracture faces.
apparatus, with t e spcctme . t (J) is attached to a lifting
. . ( ') The hd of the appara us
high-vacuum coating unit. c l . cl-through) and the work cham cial device (see Fig. 6. ! 7d in §6.3.3b) has also been developed (Sleytr
. , t or a rotary vacuum ea .
(e.g. a simple electric ino or . . ('t usually solidifies dunng evac )981) for the contamination-free transport of specimens freeze-frac-
· has subhmed away 1
cuated. (f) When the nitrogen I'd. rft 'd away and the frozen-fractll:'. under liquid nitrogen into the Leybold-Heraeus Bioetch 2005 freeze-
the chamber h~s atta1.ne (Pt/~) and then replicated with carbon (C.) (Re-_:a··
. d high vacuum the l IS l c
.e replication unit. The lid mechanism, mounted on the Bioetch
is shadowed with platinum/carbon . d Am 1966 )
Bullivant an cs · l specimen cup, is closed under liquid nitrogen and acts as a cold
for the specimens during transfer to the air-lock system. After the
, n cup has been mounted on the specimen stage, the concentric cold
d ressure the nitrogen first soH
is rough-pumped. Under reduce pf I 33 ~ 10-1 Pa (1 x lo-i T. are closed (§6.3.3b(i)). The lid is opened electromagnetically at a
then sublimes away. At a pressure o . d once a vacuum of int in the automatic preparation cycle of the Bioetch 2005.
. d to the vacuum an ,
fusion pump opene
-ip IS(1x10-sTorr ) 1s
. reac bed , the lid is lifted (Fig. 6...
.eeze-Jracturing under liquid helium
1.33 x 10 a. 1 . sh a <lowed and replicated.. M
f acture p ane 1s .
exposed specimen r . n lifting the ltd ts to develop methods which allow the evaluation of fracture faces
have shown that the specimen ~e1npe~atur~ o Re.suits obtaiti: by freeze-fracturing at temperatures close to absolute zero were
K (-140oC) and thus no etchmg ta es pace. ld- hro .
by the problem of the temperature-dependence of plastic defor-
t d that the narrow co s _,.
simple device have demo~stra e , . t ondensation contanl' uring the fracturing process (see §6.8.3 and Sleytr and Robards
provide excellent protection aga1ns c
<:heoretica11y, as has already been pointed out, freeze-fracturing
364 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-fracture replication
365

close to zero Kelvin can be accomplished in one of two different ways: frac-
Specimen_ turing can either be done under ultra-high vacuum or under the protection
of a liquid gas with a very low melting point. For fracturing at 4 K under
vacuum, highly complex UHV equipment with an air-lock system
(§6.3.3b(ii)) which prevents hoar frosting during specimen transfer, is
required. Since there are few such units commercially available, it has been
'mpler to develop methods for fracturing specimens under liquid helium
nd then to transport them to liquid nitrogen before loading into the
a acu um unit.
A simple method for fracturing specimens under liquid helium and trans-
rting them to a liquid nitrogen container while avoiding contamination
the fracture faces, has been developed by Sleytr and Umrath (1974b).
is system also allows the production of complementary fracture faces.
cirnens are frozen in type Hb or Ib specimen holders (Fig. 6.2), as used
the production of complementary fracture faces, and inserted under
uid nitrogen into a fracturing device (Fig. 6.23). This device basically con-
s of two stainless steel tubes which can be twisted and moved relative
b ,each other into certain defined positions (Sleytr 1974). In the position
wn in Fig. 6.23a, the specimen can be mounted onto the spring holder
e inner tube. The tubes are then twisted into a position where the speci-
is completely surrounded by cold surfaces and the whole device is
c rted i11to a container filled with liquid helium (Fig. 6.23b). Fracturing
Transport device
·· s place by pushing the inner rod downwards to knock off the upper spe- Ii
Air lock n holder. The fracturing device is then removed from the liquid helium 11

'ner, transferred back to a container ofliquid nitrogen (Fig. 6.23c) and 11

ecimens are collected for further processing by any of the procedures 11

ped for observing fracture faces obtained under liquid nitrogen


Carrier plate ~c(i)).

Freeze-fracturing outside the coating unit at atmospheric pressure

y and Bullivant (1976b) made controlled fractures at atmospheric


with a cooled glass knife, using the Cryokit attachment of the LKB
Evaporator
--- -----------------·-
(a) Gcymayer device for fracturing under liquid nitrogen. (b) Transport mechanism
ing the fractured specimens to (c), the Leybold-Hcraeus EM preparation plant
;now manufactured by Paar KG (Austria) in a slightly modified form as EPA IOI.
a-and b redrawn after Geymayer 1966; Fig. 6.22c redrawn from an illustration kindly
To the high
provided by A. Aldrian.)
vacuum pump
366 Robards and Sleytr Low temperature methods in biological electron niicroscopy
Freeze~fracture replication
367

turing was then performed at a specimen temperature of 93 K (-180°C)


and a kmfe temperature of 113 K ( - l 60"C) with a 1-3 µm advance per
stroke. Immediately after cutting, the specimen holder was removed from
the microtome arm and placed in a dish of liquid nitrogen. Dempsey and
Bullivant used the Bullivant and Ames block device to carry out shadowing
and carbon coating of the fracture faces, but any other method which allows
i:ontan1ination-free replication of specimens fractured under liquid nitrogen
§6.3.3c(i)) would be suitable.

Shadowing and replication

Evaporation and evaporators

Fig. 6.23. Device for fracturing specimens under liquid he~ium a~d subsequent conta~#' Formation ofthinfi/ms
tion-free transfer to liquid nitrogen. (a) Specimens are frozen ma pair of oppo:5ed hollow r
and are loaded, under liquid nitrogen, into the spring-holder attached t_o thc_1nner tube O
fracturing device. After twisting the tubes into a position where th~ spec1m.en '.s co~pletel replication process should copy the relief produced by specimen frac-
rounded by cooled surfaces, the device is inserted into the container of h~u1~ helium (b . g (§6.3.1), or by fracturing and etching (§6.5), as faithfully as possible.
pushing the inner rod downwards, the specimen is fractured and. the device is then re most usual method for replicating frozen specimen surfaces involves a
to liquid nitrogen so that the specimens can be released (c). (Specimens are then pro , -~step process. Firstly, an electron-opaque metal is evaporated onto the
any of the methods used for the preparation of specin1ens that have been fractured undet''
nitrogen.) (Redrawn and extended after Sleytr and Umrath l 974b,)
imen at an oblique angle (usually around 45°) to the 'average' orien-
of the fracture plane to provide contrast. The patches of shadowing
·al are subsequently bound together and reinforced by deposition of
ultramicrotome (see Appendix 2) and this procedure can obviously electron-scattering supporting layer of carbon or silicon monoxide
used with other cryoultramicrotomes. Specimens were mounted on is evaporated perpendicular to the fracture plane. Most limitations
men holder composed of an LKB silver pin holder inserted into a ial.resolution and in depicting the accurate surface geometry by repli-
brass plug which was suitable for fitting to the Cryokit attachment G ar1se from metal vapour-specimen surface interactions during sha-
Fig. 4.6). Freezing was done according to the method of Van H_arre g. The technical problems, as well as those of interpretation of high
Crowell (1964) by pushing the specimen, mounted on the spec1m<cn on shadowing of fracture faces, have been discussed in detail by
onto a liquid nitrogen-cooled copper block (§2.4.2). The frozen nn et al. (1972) and, with special emphasis on membrane structure
and holder were then inserted under liquid nitrogen into the LK heim and Plattner (1976) (see also Willison and Rowe 1980). '
specimen holder and transferred in a dish of liquid nitrogen to e~icaliy, if the randomly arriving atoms of heavy metal vapour
cryochamber on the ultramicrotome. The chamber _was kept. fil .· immobilised at the point of their first contact with the specimen,
nitrogen to avoid condensation of water on the specimen dur1~g resolution close to the dimensions of a few atomic layers of heavy
and fracturing processes. Before fitting the specimen to the m1ci; ould be possible. Such resolution has not yet been achieved since
the temperature of the arm was held at approximately 103 K ( "'C igrate laterally while dissipating their thermal kinetic energy until
the temperature of the knife was usually 113 K ( - 160"C). ieventually trapped, either by strong binding sites on the specimen
r by previously condensed atoms, crystallites, or clumps of the eva-
368 Robards and Sleytr Low tempera I ur e methods in biological electron microscopy
Freeze-fracture replication
369

. terial During. con1·inue d condensation of atoms, these .. first nu-


poral!on ma · , . t bl nuclei with no lateral mob1hty and High resolution shadowing and contamination problems
1 . tres grow so iorm1ng s a e
c cation cen .' . hb t form a continuous film. Since such
eventually merge w1:~i;;';or ~~:~ r~solution shadowing, the evaporation For high resolution shadowing in freeze-fracture replication, elements with
films wouldstbe too here individual metal aggregates are melting points over 2000 K are used. The most common are platinum and
be stopped at a stage w .. iridium, which are evaporated along with carbon and a tantalum-tungsten '11
,,
process
·11 ct• rmu ·shable Thus m . prac t.!Ce, the main problem of obta1n1ng fine.",, , I
alloy. A number of relevant physical properties of these elements are given
grain shadowing is to decrea~e t h e sur f ace mobility of condensing------,,
st1 is 1ngu1 · ' in Table 6.2. I

andA to inhi~it g_rowt~ of~~c~~~ a consequent increase in shadowing resohi'<


reduction in grain stz
Besides their ability to condense with minimum grain size, the elements
listed in Table 6.2 have the necessary advantage that they are, in pure form
lion can be obtained by: nr as alloys, inert to the chemicals commonly used for replica cleaning
(i) ' lowering the specimen
. temperalure (and thus reducing the lateral (§6.6). Simultaneous evaporation of platinum and carbon, introduced by
fusion of adsorbed atoms); · radley (1958, 1959), is still the commonest method for high resolution sha-
(ii) reducing the deposition rate; . owing. In the last twenty years, various technical modifications have been
(iii) using higher melting point sh a d owmg m-etals with stronger veloped (see Reimer 1967; Glitsch 1969; Abermann et al. 1972) with the
in objective of producing films with a consistent thickness and optimum
binding capacity; or . combination, of heavy metals with tinum/carbon (Pt/CJ composition. Under ideal conditions, the propor-
(iv) simultaneous evaporation, or
hon. n of carbon will be just sufficient to suppress the surface mobility of the
tinum atoms, so leading to electron scattering deposits with little crystal-
In addition to these nucleation parameters, local topological an~ phy
stru.cture. Layers with too much carbon show almost no granularity but
. erties of the specimen itself also have a strong in uen

~~::~~:~::~~he latera~=~~;~~cl~~~::~:n;~:;,::~:,s;,,:~~ss ::h;:r~ie


contrast and resolution. Pt/C can be evaporated either by resistance
ting (§6.4.lc) or by electron-beam evaporation (§6.4.ld). Good contrast
which 1s known as prefe metr of fine structural details low granularity is also observed when a platinum-iridium alloy is eva-
difficult to interpret accurately the gheo s·:ulate topological featu(' ted along with carbon (Moor 1973a) using the same methods as for Pt/
r . h andom nuclear growt can I aporation.
rep ica. SU~. r
. , .,-·
r However local differences in b1nd1ng
lead to m.1s1n~er~r~tag ~~~~rmation about structural or physicochem' sides Pt/C, one of the most suitable materials for high resolution, fine-
shadowing is tungsten-tantalum (Ta-W) alloy, although such refrac-
may .help . · m o tamm .
th pecimen surf ace w hi.ch would not be detecta
contlnu1ties on e s the mass distribution of condensing alloys must be evaporated using the more complex electron-beam heat-
metry alone were to govern lanes W ystems (§6.4.ld) or Penning sputtering (Penning discharge) techniques
Typical examples of this are crystal surfacels or clea;a!:t~l nuclei pie et al. 1978; Peters 1980; Wildhaber et al. 1984). Since the size of
h linear accumu at1ons o
to decoration P ~~omena, . .. hi ch are not detectaP ain in the shadowing layer decreases with increasing temperature of
sites of discontinml!es (crystal lattice steps) .wbl t distinguish bet g of the material, tungsten (Table 6.2) should give deposits .with the
. h d . It is almost tmposs1 e o
resolution s a owing. d . ·th fine struct grain. However, the use of pure tungsten as a shadowing material in
t. henomena when ea1ing wr
dowing and decora wn p . , 1 1972). In o fracture replication is limited due to its chemical instability if replicas
in replicas of biological specimens (Abermann. et a . reater b.
be cleaned in aqueous solutions. However, deposits of a Ta-W alloy
the smaller the structural feat~re .on the r~pltc~, th:;terial se~'·': ration occurs roughly in a 1:2 tungsten:tantalum ratio) have a fairly
t ·nty that the mass distnbut10n of shadowmg . th .··.
uncer a1 . e screen indicates · e emical stability and also have a small granularity (Abermann et al.
dimensional projection on the ~~cro:~op are of particular i
metry (topography). These cons1 era wns
arising, high melting point metals can be evaporated using elec-
rotary shadowing at small angles (§6.4.lf).
heating with good reproducibility and the deposited films show
370 Robards and Sleytr Low temperature me'h ods in biological electron microscopy
Freeze-fi·acture replication
371
' N M M 0 M

" ;::'.
0
::::
0
M

0
:':
0
::::
0 high density, fine grain size and no recrystallisation during exposure to the
electron beam in the microscope. Ta-W shadowing is superior to Pt/C if pre-
N s:; ~
cise surface geometry is to be recorded, but with Pt/C, periodicities or dis-
~ M

"'
M ci
"' "' continuities of the fracture face become visible through decoration effects
(Abermann et al. 1972). The decoration effects with Pt/C and Ta-W sha-
:: "<
~ E"
~
~
i! g .
0

°'
N
0
°'
°'
N
0
°'
~
N
M
M
0
~
0
M
00
M
dowing can easily be demonstrated by comparing the structural information
revealed from specimens onto which metal has been deposited at 90° (per-
" pendi.cular shadow) or ~45° respectively (Gross et al. 1982; Bachmann et
~
I !
I'
7 al. 1984).
s 'sx
I
i' ,'
.- x
N . In addition to complex specimen-metal vapour interactions, other factors
determine the degree of similarity between the geometry of the fracture
'.
I'

"'
M

1
00

"'
7 Jane and the final mass distribution of heavy metal on the replica. Vacuum
1"'

s s 's s
x
..
x
ci
x
~ M
M
x
nditions, particularly the partial pressure of condensab!e gases in the spe-
. en area during the period between fracturing and the end of replication,
e of particular importance (§6.5.2). Condensation contamination can
M
~

.. . ..
::i:
M
~
~
M
~
~
M
M
~
M
~

~
ask, enhance or specifically decorate structural details (Gross et al. 1978b;
alzthiiny et al. 1981, 1982). It is frequently overlooked that the evapora-
I ,f,

'
"

. ..
s themselves can cause a deterioration of the vacuum since, during eva-
M 0
~
0
~
N
M
0 N
~ °'
~
~ ation, not only heavy metal vapour but also condensable gases may be
~ N M M
N
II elled. To outgas the evaporators, it is best to pre-heat them under
um close to their working temperature and then to cool them down be-
§' 0
0 s 0
0
0
~
use. Generally, carbon evaporation should start as soon as the shadow-
~
~
M
'° '" ~
rocess is finished. Residual gases in the vacuum system can condense
·he heavy metal deposit and cause a separation of the two evaporated
during thawing and cleaning. This particular artefact is only observed
or vacuum systems and is easy to recognise since parts of the thin sha-

b '
x
E "" ~
N
ci
.
N
N
~

"';::; '°
M

::::
g layer will either be lost during replica cleaning or be seen curled up
carbon backing layer.
ctural alterations in the fracture face itself may arise from thermal
"'"' uring evaporation. As can be seen from Table 6.2, high temperatures
N

~
N vi
;o: "' °'
00
M Uired during vacuum evaporation of heavy metals. An energy flow
Pecimen is caused by:
"'
;":

·ssipation of heat to the specimen surface during condensation of the


pourised metal (heat of condensation);
rmal radiation from the evaporation source (heat of radiation);

u or electron bombardment when poorly designed electron-beam


porators are used.
372 Robards and Sleytr Low temperature method s in biological electron microscopy
Freeze:fracture replication
373
;;u
su The amount of energy flow to specimens under different evaporation condi-

b
~
'

::: ., .,
~ 0 M
N
0
2
0
M ::: tions has been calculated by Zingsheim et al. (l 970a, b) and relevant data
~ 5 are given in TabJe 6.3.
"''B C' The heat of condensation varies with the shadowing material and is about
"" ,' su
'
- "" ~

.,
~

;;; the same for deposition of equal thicknesses of tantalum-tungsten and Pt/C;
1l
~ 5 "' "
M ~
N
~

"' for all practical purposes this amounts to less than 5% of the total thermal
load (Bachmann et al. 1969; Zingsheim et al. 1970b; Abermann et al. 1972).

"" " ""'s


{I .2 u ~
0
0 c: ~ ~
0
he intensity of radiant heat for a given element is determined from the Ste-
n-Boltzmann law and depends on the temperature of the evaporation
u "
N
0

"5 <
urce and the ratio of the source radius to the source-specimen distance .
s the equation below shows, the radiant heat is proportional to the fourth
"
.2
§' ~ N
., 'ower of the source temperature.
"
'il
c ~
5
s M
N
~
.,
~
:!

"''
11

§ -
c
\iu
0
~
~
N
0
0
0
0
0
0
0
0
(6.1)
~
~ ~ 0 ~ 0 0
":;; ~
N N M
II
M M
II
M
II
"
-" ,::: =intensity of the radiation reaching the specimen
M "•
d
u
u
=source temperature
~
~ =source radius
"'.., 1'l ·"q~ ~ - ::; 0
~
~ 0 0 ~
0 M N N
=source-specimen distance
"';..-< s =total emissivity
._§"" "
.2 =Stefan-Boltzmann constant
~

~
\i0c
~
§- c
>
i ~
~ ~ ~ ~

n be seen from Fig. 6.24, the vapour pressure of the shadowing materi-
"-"" "'
"' reases much more steeply with temperature than the released energy.
,,, equently, the energy flow as radiant heat can be considerably reduced

"-
s 8
'E
§
"s"
"" "
u
d
"u
.-ng a smaller source or by increasing the specimen-source distance

~u .D"s -~~
d d
0
6.25). In other words, it is recommended to use the highest possible
~ 0~
~ §-
0
"
.D
tlu s "
-~ "" i:il
._§
bo ;3
0
-~·~ ·E "' · temperature, together with the smallest possible source diameter and
>
i:il 0
" ~ 0
" " -"""
h u ,D
"' -" ,D
"' "'-" a test possible source-specimen distance (Zingsheim et al. l 970b).
er. there are practical limits to the reduction of radiant heat, since
·on rates are affected to the same extent as the radiation intensity.
oo N adowing periods (5-10 s) are often necessary in order to guarantee
0
bilisation of the specimen surface. This is of particular importance
e faces are replicated at etching temperatures (§6.5.1 ), or if a freeze-
replication method is used in which it is difficult to maintain conta-
·free conditions (§6.5.2).
-~
' u ;:; _ent designs of electron gun are commercially available, but most
"" 0: if: •
;.. u u
a source diameter of about 2 mm and a source-sample distance
374 Robards and Sleytr Low temperature me th od~ in biological electron microscopy
Freeze-fracture replication
375
Temperature ("C)
1700 2200 2700 2470 K
20
Pt Pt

10 p

.~
15 I
;;; ~ 'E I
"! 0 E u I
I- u I
;i:
x 10'
~1 10' ;: I
e:,• , ! ! I
2560 K

,~
••

0.
..;;
~
~"
10
I
I
• c ;; \
••
2640 K
•c • \
0. 10 10
-1
/ 10
~
"' 5
\

,;• '' 2725 K


• ''
" 2470 K-1/D 2
Heat of condensation
-2
10
0 10 20 30 40
2000 2500 3000 Source-specimen distance D (cm)
Temperature (K)
The total heat flux at the specimen as a function of source-specimen distance (D).
. It o the fourth power of the source temperatur . ulated for a constant deposition of 0.1 nm s- 1 of platinum (solid line). The deviation from
d. · ·
roport1ona 2
Fig. 6.24. Heat ra 1at1on is P ( )d ds much more strongly on the tern /D dependence (dashed line), which would hold for constant tc1nperature. is very sma!L
fan~Boltzmann law). e vapou
Th r pressure P epen
df
.
t. um and for a source-to-spec1men
a"
s calculate or p 1a in . r instance, the source is placed at a distance of 30 cm instead of 20 cm, heat flux and
([) I
than the radiant energy · wa re and hence the evaporation r sition rate are approximately halved. The source temperature must be increased from 2640
of 10 cm. Between 2000 and 3000 K'.th~ vapourgpyr~::~ only by a factor of 5. (Redra 2725 K to restore the original deposition rate. This rise in temperature increases the heat
0 f 1000 whereas the radiation ener
by a factor ' Zingsheim et al. I970b.) the specimen again, but only by 10%. At large distances the total heat flux diininishes :11

ards the value for the heat of condensation. (Redrawn from Zingsheim et al. l 970b.)

(§ 6 4 Id). Although total elimination of radiant' Met hods o.f resistance heating
of about 14-15 cm . : . a ertures between the eva
is possible by introduc1~t rot~t~~;. Jidrian and Horn 1974), o.ll
source and the specimens obrn , d ced In contrast to the · ance heating can be used for both Pt/C shadowing and carbon repJica-
. t have een pro u ·
types of this appara us . d 1 t ns) created by electron • Willison and Rowe 1980). It involves passing a current between
tion, charged particles (tons and e .etch.ro a v~ry thin surface layer, ctrodes the resistance of which causes heating, so leading to evapo-
t. are adsorbe WI In .
ment evapora ion, t the fracture face. Dis~:,; of the desired material which may either be the electrode itself (e.g.
a much greater threat of heat damage o tor have been overcoUI\) ) or a substance attached to it (e.g. platinum). Common electrode
of early designs of electron-beam ev~p~r:n systems which tota, • ·es for Pt/C evaporation are illustrated in Fig. 6.26. In most freeze-
ducing effective charged-partlcle de ech1. t al 1970a}. As replication units the evaporators are positioned above, or at the
f h al load (Zmgs eim e .
this source o t erm .h h ell-designed electron-be .. el as, the specimen. This limits the choice of resistance evaporation
Zingsheim et al. (1970b), wit sue a wd . hi 5 10 s with onl.'f.:· . to types (a) and (b) (Fig. 6.26). In contrast, the very simple and
tor, Ta-W shadowmg . can b e performe
. . wit n heated
- Pt/Ce ·• resistance platinum/carbon evaporation method, developed by
heat load produced by a commercial res1stance-
(1969) (Fig. 6.26c), is only usable if the evaporators can be installed
ble 6.3).
376 Robards and Sleytr Low temperature me tiJods in hiological electron microscopy
Freeze-fracture replication
377
- , Electrodes are usually located at least 15 cm from
below the specimen stage. - t th' ·s the evap-oration procedure <level- former. With the type (a) (Fig, 6.26) carbon rod assembly, 5-7 cm of 0.1
. I An exception o IS I . .
mm diameter platinum wire gives a shadow of sufficient contrast. With the
the specimen Pane. ( 1976) (Fig 6 26b) which, in combmat10n
d b B k nd Robertson · · - type (b) (Fig. 6.26) assembly, 8 mm of 0. 127 mm diameter platinum wire
ope Y an a , t 1 1977 ) allows high resolut10n Pt/C
with a resistance monitor (Steere e a ··men d'1.stance of only 7 cm. Results is coiled around a cone-shaped (0.5 mm diameter tip) carbon rod, This rod
b d 't d at a source-spec1 is forced by spring tension against the opposite rod which has a flat end and
layers to e eposi e _ demonstrated that this short dis-
. d "th delicate specimens c1ear1y . >< an extended edge. Both rods in this system are made from 3.0 mm diameter
obtatne
tance hadw1no deleterious
. ef"'iect s o n the apparent resolution of specimen----/ raphite rods. To obtain good results with both type (a) and type (b) (Fig,
(.'

.26) assemblies, it is essentia1 to maintain firm contact, using spring pres- ,1,' I

topography (Steere
1 · evaporate d f rom o pposed spring-loaded carbon rods_,<_'
· 1982). :!_',I
,,
When p atinum ts . i s to accommodate the platinum wire have.-
re and correct alignment of the rods, during heating, Also, special empha-
different waysd of0 shapmg . must be given to the correct spring tension. Even so, with the possible
e tip may ep re d uce d to a cylinder (using a small lathe___._-•••
the bl
ception of the design suggested by Steere et aL (1977) (Fig, 6.26b), consi-
been suggeste
- 1 'sharpener') of a b out 1 mm diameter , while the other- is forme<J
- n
rable variation in the quality of shadowing is often observed. Part of this
or ,specrn
t , 1y, b o th t'1P s may be cone-shaped
cone Alternat1ve , with flat,
f L_ -- oblem can be overcome by pre-melting the platinum under vacuum. After
to a ds-eep t t --ps (Moor 1959) (F'1g. 6 ·26a) · An asymmetnc assembly o car_ c_._;_.--

ntirig the vacuum chamber, the carbon rods can then be realigned in such
mm rnme er t d , d f r automatic regulation of shadowing vi
bon rods (Fig. 6.26b), es1gne o db Steere et al (1977) (see also Ba way that the platinum, which has melted to a blob, points towards the
- ·t was suggeste Y · cimen.
a resistance rnoni or, _ d c r evaporation (obtainable_
1976) The p 1atmum use ''0 ;> he Glitsch-type evaporator (Fig, 6.26c), used in the Leybold-Heraeus
and Robertson : A dix 2) is usually in a tight coil w
pure wire from sdupbphern ~i~ege th:~~;e around a cylindrical, I mm dia ze-fracture replication unit, consists of a 3.0 mm diameter carbon cruci-
can eastly be ma e Ywin fixed between two spring-loaded tantalum cheeks. The crucible is loaded
4--6 mm of 0. 1 mm diameter platinum wire, coiled into a small ball,
can be used for at least three evaporations when reloaded for each run
-the same amount of platinum. This type of evaporator is unsurpassed
se of operation, reproducibility and fine grain of the deposited film.
ever, it has the limitation that it cannot be adapted for use with most
freeze-fracture replication units since it can only evaporate in an up-
direction as the melted platinum would flow out if the evaporator were
:mounted upside-down.
mmercial units are used, then evaporation is best done according to
nufacturer's specifications. However, a good starting point for choice
ditions is to use pointed carbon electrodes (LO mm diameter tips), a
of 55-60 A with a voltage of about 7.0 V, and a 5.0 s evaporation
b
" e can be selected either simply by turning the appropriate switch on
@ill(] or, on some machines, by using an automatic timer), For the type
, 6.26) assembly, a resistance monitor with sensors and automatic
lies for platinum/carbon resistance eva~ ut-off have been developed to control the thickness of the deposited
Fig 6.26. Common electrode assemb . f b n (C) rods to accommoda eere et al. 1977).
· f h· · g the ttps o car o
Two different methods o s ap1n 1 consisting of a carbon c_r
.
of platinum (Pt) wire. (c
) 'Gl"t
1
sc
h'-type evapora or , d-:'.
k · s The crucible is Joa e_ cond stage in the replication process -involves the formation of a
in -loaded tantalum (Ta) chee piece . . ··/:.;t us carbon backing layer on top of the heavy metal shadowing lay-
between two spr g of tightly coiled platinum wire. .-
n evaporation using resistance heating is generally Jess of a prob-
378 Robards and Sleytr Low temperature method~ in biological electron microscopy
Freeze-fracture replication
379
lem than is the shadowing procedure. The electrodes can be trimmed to the
improving the focusing of th I
same shape as for platinum evaporation but untrimmed thin carbon rods ··. e e ectrons This ·
w1th1n a smaller area of th t. f . Increases the temperature
with flat ends are also often used satisfactorily. As for platinum evapo- . e Ip o the anode I dd ..
matenal can be mounted 1 . n a Ihon, the evaporation
ration, good spring-controlled physical contact of the rods is essential for · on a ow thermal d · ·
a stainless steel. tube to . con uct1v1ty support such as
successful evaporation which is normally accomplished within 6-12 s using ' Increase the temper t . '
For Pt/C evaporation a sm II 1 . . a ure gradient (Tesche 1975)
60 A and 7.0 V. Both Pt/C and carbon evaporations are best done through a P atmum cyl d · fi ·
)ip of a carbon anode rod (F' 6 28 m er is llted into a hole at the
shields with apertures to reduce contamination of the work chamber. For ·.·the specimen the anode musitgfi. . t ba). When the guns are positioned above
checking the correct alignment of the apertures in relation to the specimen}_:-_<, • Irs e heated 1 th ·
6.28a) to pre-melt the platinum ( th . n e upnght position (Fig
a small current, just sufficient to produce light from the carbon tips, is:: o erw1se the pl f ·
The electron gun can usual! b fi . a mum would drop out).
passed through the electrodes so that the point of evaporation can be seen::;• ,. Y e ired several hm · ·
istently good quality and with Pt/C . es, g1vmg shadows of con-
Care should be taken during all alignment checks and evaporation proce~· :: . a ratio ofap ·
y microprobe analysis (Moor 1973 ) proximately 95:5, as shown
<lures that the bright source is only viewed through dark filters (e.g. neutr~f . . a . Although close . .
at th e gra1n size is about th . exam1nation reveals
density filters). · e same as with g d ·
is both the reproducibility and 1· b' . oo resistance evaporation
· re Ia ility that k I '
ion more attractive for Pt/C h d . ma e e ectron-beam evapo-
6.4.ld Methods of electron-beam evaporation de is prepared by winding a ts at lowmg .. For Ta-W evaporation, the
. , . an a um wire arou d th
YIman cal tungsten rod (Fig. 6.ZSb) On n e reduced tip of
Electron-beam bombardment heating can be used successfully for b lts and a Ta-W alloy is ' d .. electron bombardment, the tip
heavy metal shadowing and carbon replication. Since no materials are a lOrme This eva .
ral thnes. The tantalum t . poration source can be used
able from which crucibles can be formed to allow resistance heating of con ent of the depo .t d fi
oat the first evaporation d . si e 1lm, which is about
sten and tantalum to give adequate evaporation rates, the developm_e , ecreases w-ith sub
obvious changes in grai . . sequent evaporations with-
electron guns was particularly important in allowing improvements in_ a). n srze or chemical stability (Zingsheim et al.
<lowing methods. Electron-beam heating was used in electron micro
avoid condensation contamination on th f
for shadowing as long ago as the 1960s (Bachmann et al. 1960; West. e racture faces, electron guns
and Lorenz 1960; Bachmann 1962) but it was about ten years befo
were used successfully for high resolution shadowing of frozen sp
Deflection of charged particles with an electrostatic field was sho High
voltage
particularly important in avoiding serious heat damage to the fract
(Zingsheim et al. 1970a,b). The basic design of an electron-beam eva •

3-·
~
0
c
is shown in Fig. 6.27. It consists of a heated tungsten cathode fil
which emits electrons. The electrons are accelerated and focused to
"'
tip of a rod-shaped anode (2) which is made from the material to b~ 2
ated. The transferred energy causes the tip to melt and eventually -So----...j__j
.::
Low
voltage &
ate. A focusing aperture (3) prevents electrons from reaching the r--~"""'=----...__j Filament
4 heating
of the anode and its holder. The front aperture (4) has an opeu.i
is small enough to protect the specimen from radiation emitte
heated filament. Finally, ion-deflection plates (5) deflect charg~ 5
generated by electrons on their way to the anode. For conv:
plate is kept at ground potential and the other at the anode p
bhematic diagram of an l
possible to decrease the amount of energy needed for ev , e ec~ron-beam evaporator. (For details see te
after Z1ngsheim et al. 1970a.) xt.) (Redrawn
380 Robards and Sleytr Low temperature meth o d ~in
. b1·atogical electron microscopy
Freeze-fracture replication
381
a b c
Pt Ta by the amount of platinum wire wrapped around the carbon electrodes (Fig.
6.26a,b) or placed in a crucible (Fig. 6.26c). With the exception of Steere's
technique (Steere et al. 1977), platinum evaporation from pin-pointed car-
bon electrodes is less reproducible than from a crucible (Glitsch 1969). The
advantage of the Glitsch method is that the direction of evaporation is
etermined by the shape of the crucible. Again with the exception of the
ethod suggested by Steere et al. (1977), settings for resistance Pt/C evapo-
c w tion from a crucible lead to total platinum evaporation.
It is possible to deposit a film of the desired thickness by comparing the
ntrast in the final electron microscope image with the density of shadow-
F . 6 28. Anode assemblies. f th electron-beam evaporation of various metals. (a) material seen as metaJJic grey deposit on a piece of white paper or porce-
tg. · or e
(b) Tungsten(W)/tanta1um
(Ta) source. (c) Shaped carbon
num(Pt)/carbon(C) source. . t mperature gradient. ragment placed beside, or around, the specimen during evaporation.
improve e
er one or two initial trials, it is then quite easy to maintain reproducible
dow thicknesses in subsequent runs. A drop of high vacuum oil (e.g. dif-
·on pump oil) is placed on the surface of a white porcelain fragment. The
d der vacuum. D ur i· ng this cycle all parts of the as_s.
should be pre-heate un I t that reached during the actual red area appears to remain 'clean' (white) during the evaporation and
h t pera tu re c ose o
bly should reac a em h de electron guns can also be allows a comparative estimate of the thickness of the evaporated film
.h bon rod as t e ano ,
poration. Wit a car . 1 To improve the temperature g made (Reimer 1967). Although this only provides a rough estimation,
h bon backmg ayer. ·
to evaporate t e car ' k' a few millimetres below the ti · till the most frequently used method for monitoring resistive Pt/C sha-
d ·th a narrow nee . ·
ent, carbon. ro6 s28w1) However, ,or ·ng. Instead of comparing the 'greyness' of Pt/C deposits on an indica-
-"
re asons of economy, a comb1na
be used (Fig. . c . . d a resistance source for carbon aper or white enamel, the thickness of the deposits can also be judged
an electron gun for shadowing an '
e interference colours seen on gold-coated metal or glass sheets (Reiter
ration is often used. -" f ze fracture replication are c . A mid-purple colour is close to an ideal Pt/C evaporation. Carbon
ators ior ree - . ,
Electron-beam evapor . d Cressington (Appendix 2). g is generally more reproducible than platinum shadowing if the pre-
. bl f Balzers U mon an . h
cially availa e rom . that described by Zmgs ined voltage is applied for a given time (usually between about 8 and
. ll the same design as
guns have basica y dd.f to the freeze-fracture appat A carbon layer of IO nm thickness is just noticeable as a difference
(1970a) and Moor 0971). In a :a:nalso been fitted to many ot , sity between the oil-covered area and the rest of the white porcelain
which they were designed, they d . 1600--1900 V and 6 and a carbon deposit with sufficient stability for freeze-fracture rep-
. ll guns are fire usmg .
vacuum units. Genera y, . t·ng deposition rates. A s1 < pears light grey-brown. With gold-coated metal sheets or glass the
. . bl hadowmg or coa i . L
6-13 s to give smta des· d by Hagler et a 1. (1977) for use with tu_._._.e·• g interference colours relate to the thickness of the carbon layer
tron-beam gun was. e~1gne . Th·s evapor-ator is easily built in a_\
· 1961)•
freeze-fracture replicatton unit. I
C Pt/C evaporal!on .
sm· --
but only gives deposition rates adequate or 20±4.0nm
at only 1000 V. ,24±4.0 nm
35±3.Snm
6.4.le Monitoring film thickness .. 42±3.Snm

. 6 4 Jc) is used with standardiS#t carbon deposit for normal freeze-fracture replicas wiJI give a red
When resistance evaporat10n (§ . . f the shadowing layer\•.· o blue) interference colour. Due to the low electron-scattering
of voltage and current, the thickness o .
arbon, differences in the thickness of the carbon layer are much
382 Rohards and Sleytr Low temperature methods in biological electron microscopy
Freeze-:fracture replication
383
less obvious in the microscope than the heavy metal shadow. Self-shadowing arises when the sh d .
For accurate measurement of both film thickness and deposition rate of and eventually casts its own shad a owding material piles up on a structure
heavy metal shadowing and carbon coating, quartz crystal film-thickness . d . ow an consequent! th. .
ga1ne In resolution if fractu .c: Y no Ing IS generally
re 1aces are shado d t
monitors are used (see Appendix 2 for suppliers). Quartz crystal monitors than 45 (Zingsheim and Plattner 1976) s· we a ang1es much lower
0

measure tlie weight per unit surface area (mass thickness) of deposits and mens have great variations in sur.c: . Ince fracture faces of most speci-
are mainly used in combination with electron-beam evaporators. The . •ace geometry (to h )
shadowmg angle cannot be obt . d pograp Y , the optimum
· aine over the whole surface.
increasing thickness of film deposited onto the crystal changes its frequellti~.{
of oscillation which can be calibrated to give a direct indication of fiJntc: 4.Jf Rotary shadowing
thickness. If accurate film thicknesses are to be determined, it must bei
remembered that, with Pt/C evaporation, e. mixture of approximately is no~ possible to provide contrast in all dire . . . .
(w) carbon and 95% (w) platinum is evaporated simultaneously. With adow1ng. Furthermore the sh d ct1ons using unI-d1rectionaI
. ' a ow cast by 0 t
W, a mixture of about 20% (w) tungsten and 80% (w) tantalum evapor aller neighbouring structures T ne s ructure may obscure
from the anode tip (Fig. 6.28b in §6.4. ld) but, with each successive ev . . · o overcome these bl .
erent 1n uni-directional shadow· H . pro ems, which are
ratiOn, the proportion of tungsten increases. For accurate measuremen . mg, emmets (1949) · t d
ue of rotary shadowing whi h . . m ro uced the tech-
. c 1nvo1ves rotating th · .
film thickness, it is essential that the quartz crystal and the specime porat1on. Generally speci·m e spec1mens during
' ens are rotated at 60-150
both positioned in a precisely defined and similar alignment with res _nute 1,-vith the evaporation so revolutions per
•.· urce at an angle 0 f 6-45"
the evaporation source. If this is not done, incorrect mass distributio face. The method is most fre . to the specimen
quent1y used m I t .
be measured due to differences in the directional distribution of the onstrate isolated filamentous e ec ron microscopy to
vapour (Simpson 1979). It is best to evaporate using a shutter . . mo1ecu1es such as DNA (Kl . .
) but 1t 1s also applicable to th h d '. emschm1dt
.. e s a owmg off t , '
between the specimen and the electron-beam evaporator. This sh ; MargantJs et al 1977· H rac ure ,aces (Elgsaeter '
. ' euser and Sa!peter 1979 H
opened once the gun is providing a constant evaporation rate and i a,b; Hirokawa et al. 1982 . Heu d C ; euser 1980, 1981,
when the desired film thickness (=mass density) is reached. The ' ser an ooke 1983· E .k
1984). A special rotary repl1·cat· . ' , speV1 and Elg-
shutter has the additional advantage that the specimen is protect 10n umt •Or th B 1 f
atus has been described by M .. e a zers reeze-fracture
heat load and discharged gases during the heating-up phase of the : . argantJs et al (1977) d · i''
lly available (Appendix 2 ) d · an is now com-
. an a completely n d ·
tor. If a quartz crystal film-thickness monitor is not available, it is.s _Ith a rotating specimen stage has be . ew esign of freeze-etch ii:
11.',
ble to produce reasonably consistent film thicknesses with electr ·. _ reeze-etching units now h 1 en descnbed by Elgsaeter (1978).
. ave at east the p t f I
evaporators if constant evaporation times are used and if speci a stage for rotary shadowin ( §6 o en ia to be modified to ,,
given to positioning the anode with respect to the coils of the h. ·
g see .3.3.). ,,
g T4 polyheads, erythrocyte ghosts and ch
ment. In the absence of a film-thickness monitor, an indicator p · ysterns, it was shown by M .. loroplast membranes as
. argantJs et al (1977) th .
lain fragment or gold-coated metal (aluminium) foil is used to t improved by rotary shadowin W . at resolut10n per
thickness as described above. _ron-beam gun 2 5 nm . d' . g. hen Pt/C was evaporated with
, . per10 lCit1es were res 0 I d . b
For routine freeze-fracture replication, a heavy metal film -0, nd rotary shadowed sp . ve in oth uni-direc-
. ecJmens Neverthe1es d
is used. The optimal thickness of the carbon backing layer is str cal contrast and enhanced d . . s, ue to the radially
. ecoratlon pheno
dent on the type of specimen and the required stability during . ec1mens, the subunit structure of f 1 mena on rotary sha-
cleaning, but is generally in the range of 10-35 nm. Ho . :-:_ore clearly in some areas. par Ic es seemed to be demon-
necessary to note that, for highest resolution shadowing (Ta- pted out by Heinmets in 1949 h ·
thickness of0.7-1 nm will give sufficient contrast. An incre ds on the geometry of the b
t e optJmal angle for rotary shadow-
ness increases the contrast but, simultaneously, resolution · Ta-W as the shad . o ~ect. With ferntm as a model system
result of a decrease in relative contrast variations and self-sh 'wed that the appear::~:~t;t~ri;l, Neugebauer and Zingsheim
g o u ar molecules can vary dramati-
384 Rohards and Sleytr Low temperature meth ods in hiological electron 1nicroscopy
Freeze-fracture replication
385
call depending . on the sh·a d owing . angle and the thickness of. the metal d film.
y . p· 6 29 ferritin appears more and more hke a oughnut (6-12°) Pt/C rotary shadowing, seem to be mostly the result of specific deco-
As illustrated m ig. l . d ' reases and the metal deposit increases, while the ration of small protruding structures and self-shadowing phenomena.
as the shadowing ang e ec · h d'f'
b k nd granulanty · conco mi·tantly increases. By varying t e I 1erent Finally, in interpreting images of rotary shadowed specimens, it is necessary
ac grou parameters, the apparent diameter of ferritin can be to be aware that very few areas of freeze-fractured (and etched) specimens
shadowing 1 changed
h "'· .·'
re either flat or parallel to the top of the rotating specimen stage and also
from 12 nm to more t h an 2 6 nm. These model experiments . clear Y s how d t.,e. c' at, when more than one specimen is used, the specimens are not a11 rotated
.
Potential danger o m1sm e f .. t rpretation of results obtained
. by rotary
f s a. ow<.
·1'·>.
. . d R ) Such considerat10ns are o part1cu out their own cep.tral axis.
· g (see also W1lhson an owe 1980 · . f. at;.<
in
importance if . attempts are mad e to interpret the. subunit structure o tntr~'C;, In addition to its use in the study of fracture faces, rotary shadowing is
membraneous particles (Branton and Kirchanski 1977). . particularly suitable for demonstrating the fine structure of filamentous
In summary it . appears th a t ' u nder very well-defined expenmental . cytoskeletal) and other very sculptured, structures (e.g. coated vesicles)
'
f ns rotary shadowing . may pro vi de useful images of. fracture . . faces. but osed by deep etching (§6.5) (Heuser and Kirschner 1980; Heuser 1980,
io technique
' . 3a,b; Espevik and Elgsaeter 1984).
the cannot b e seen as a substitute for routine uni-directional
<lowing Rotary evaporation . wt'th Ta-Wat an angle of ,.,,45 0
. has shown. A method of 'double-axis rotary replication' has been described by Shi-
. .
th genuine structura e at s 1 d t ·1 of a fracture face are sometimes . depicted m._: la et al. (1984). In this method, as well as the specimen rotating, the sha-
e·1 b ys ymmetrical metal deposition. The images obtained by low a
eas1y Wing angle is changed continuously during replication. The authors con-
r that this leads to improved replicas from deep etched material.

Jg Bi-directional shadowing

50nm
n aJternative to either uni-directional or rotary shadowing, Willison
Moir (l 983) have suggested that bi-directional shadowing can be used
onstrate structural features of fracture faces. They showed the value
portrait shadow-casting using a 90° inter-shadowing rotation of the
n (Williams and Smith 1958) as well as the classical technique for
size analysis using a 180" inter-shadowing rotation (Williams and
ff 1944; Kahler and Lloyd 1950).
1,',

Replica strengthening methods i'·


45° 30° 20° 10° 1',
I·.
F' 6 29 Appearance of ferritin molecules after rotary sh a d owing under. di I, I,',
of heterogeneous tissues frequently disintegrate into small pieces
_1g. F. . .,. (30 µg ml-1) in distilled water was dried onto freshly cleaved mica
ttons. ern in . 1 · a vacuum un: hawing and cleaning (§6.6). The explanation for this is that there
. . nts of tantalum-tungsten at vanous ang es 10 Ji
with var~1ng amou ount of metal deposited (0.2 nm, 0.6 nm, L rential dimensional changes in the specimen and replica arising
(Zingshe1m et al. 1970a). The am . ·n the s ecimen table.re
!he plane of the support surface. Dunng rotary shadow1. g p , ated ort: Hing and shrinkage (Robards 1978a). More stable replicas are pro-
. . I·
240 revs m1n-1. A carbon ayer, app .
roximate!y 3 nm thick, was evapor
. "d (2 h) A molecule,
:~ th a thicker carbon backing layers but this limits the resolution of
. R licas were cleaned in 70% sulphunc ac1 . . .,,·, .a. Consequently, methods have been suggested which involve the
dow1ng layer. cp . h ft The series of molecules
laterally at 450 is shown for companson on t e 1e . k t x 36 000 n of an additional strengthening layer to provide support for the
M.
is also shown with reversed contrast. tcro grnphs were tad en da ith pemii
,
to Gaussian focus with a Philips EM 301 at so_ kV. (1:-epro uc)e , w
al/carbon replica during thawing and cleaning but which can be
bauer and Z1ngshe1m 1979.
way again afterwards.
386 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-fracture replication
387
6.4.2a Silver technique is vented with dry nitrogen gas and the fro . .
still frozen sample which is . . zen solut10n is lowered onto the
·. mamtamed at about 153 K (- o .
One of the simplest and most efficient means of strengthening the replica 6.30b). The chamber is then 1 . d . 120 C) (Fig.
. . c ose again and rou h d
is to use an additional layer of silver or gold (Robards and Umrath 1978) 1me while the temperature of tb t . . g -pumpe for a short
which is evaporated under high vacuum at the end of the normal freeze-frac. elting at about 213 K (- 60"C) \~eag~;ssraised to 218 K (-55"C). After
ture replication procedure (§6.4.1). Although this technique was develope~,}/ op of the specimen while the hi, } y tyrene solut10n spreads over the
. c oro,orm evaporat Af b
for the Leybold-Heraeus Bioetch 2005 unit (see §6.3.3b(i)) it is adaptable IQ'·· e specimen is removed from th es. ter a out I 5 min
e vacuum chamb d '
other freeze-fracture units which either provide enough space for the instai-" oled container in which is f . er an transferred to a
lation of a third evaporator which evaporates upwards, or possess an af hich the specimen was treatad r(ozen solut10n of the original liquid with
· e e.g. a glycerol soluti ) B I ·
lock system for exchanging evaporators or specimens. . '"··· ntamer under a lamp (Fig 6 30 ) . . . on · Y P acmg this
Silver is used as an evaporant because it is easy to evaporate, provici:··:o wly while, at the same ti~e .th c 'ilt is possible to allow the tissue to thaw
' e po ystyrene layer d ·
sufficient strength, is relatively inert to the commonest cleaning soluti . Stolinski et al. (198 3) sub nes on top of the rep-
sequent1y cleaned th .
and can easily be removed later from Pt/C replicas. The evaporation p einforcing film uppermost e specimen by floating
' on a sequence of cl · .
meters can be preset in the unit. For a range of test specimens (inclu 0% sodium hypochlorite 25% d. th ID e.amng so1utwns such as
plant tissue) prepared in the Leybold-Heraeus unit, 18 mg of silver eva also §6.6). After moun;ing o imFe y ormam1de and 50% nitric acid
on a ormvar-coated ·ct · h
ated from a molybdenum boat provides a supporting layer of sufficien ene layer uppermost the . gn wit the poly-
' rep 1ica and film a d . d
bility. After removing the silver-backed replica from the apparatus an& ersed in amyl acetate to dissol th . c . re ne and then gently
. . ve e rem,orcmg la Af d .
·solving the organic material, the silver backing-layer is removed using en is ready for examination. yer. ter rying, the
nide solution (§6.6(iii)) and the replica is tben mounted on a grid ce fragmentation of replicas often occurs dur' .
rbon evaporation has been t d b mg thawmg (§6.6), a two-
6.4.2b Other reinforcement (and replica-stabilising) methods plicas of delicate specimens 7~ t~ e useful for obtaining large in-
. er eavy metal shadowing, approxi-
Both to keep replicas floating during thawing and cleaning and to
a b
additional strength, Bullivant (1973) suggested that the replica side. c
cated specimens should be dipped in a 1% solution of Collodionf Wire loop
with handle
Frozen
nitrate) in amyl acetate. An alternative method is to apply the polystyrene
solution
solution to the replica using a loop. After cleaning and mounting
(§6.6 and §6.7), the Collodion backing-layer is gently dissolved b
-- ----,
Replica--------
successive loops of amyl acetate (Bordi 1979). The Collodion m Tiss~;~S~ecimen
~holder
the disadvantage that, to apply the strengthening layer, the spe( Cooled
metal
be removed from the vacuum and condensed water may theref~ .. block Cooled specimen
stage in free:oi:a-
firm attachment of the backing-layer to the replica. tracture replication
unit
Stolinski et al. (1983) found that polystyrene is superior to<,:,
a strengthening layer because is is considerably tougher whet\. , Melting glycerol solution

more resistant to cleaning solutions, such as acids, alkalis an4:: _Application of a strengthening layer of pol st rene .
mamide. A 2% (w/v) solution of polystyrene (e.g. from plastic;, en. (a) Freezing a thin layer f l y y to a rephca before digestion
Ithe frozen reinforcing soluti~ po ytstyhrene solution on a cooled metal block. (b)
in chloroform is used. A flat layer of this solution is picked UJ"lS v Af
.£'\.. ter wanning to 213 K
n on o t e factured and 1· t d .
rep 1ca e specimen main-
diameter copper wire loop and is frozen on a metal block coo~" - under vacuum to
tansferred to a melting glyc I I .
It h
me t e polystyrene solution the
ero so utron (c) a d th · C . '
nitrogen (Fig. 6.30a). When the freeze-fracture run is finishecl~ radiant heat. (Redraw ft S . ~ e rein arcing layer is dried by
n a er tol1nsk1 et al. 1983.)
388 Robards and Sleytr Low temperature meth ods in biological electron microscopy
Freeze:fracture replication
389
I 2001 f the IotaI carb on Ia Ye r is first evaporated. The specimen II is
mate y I10 ° ed up to 263 K ( - !OoC) or room temperature to · ow
a immediately ambient vacuum, then sublimation will take place. In contrast,
if the partial pressure of water vapour in the vacuum is higher than the satu-
then slow .Y warm ti
t 0 0 ccur and then, 1na11Y'
an additional carbon layer 1s eva-
freeze-drymg ·1d . f t !ability and to connect together ration vapour pressure of ice in the specimen, then condensation of water
b replica o grea er s .
porated to m up a . d ·ng drying (or thawing) (Gnepp , vapour from the vacuum onto the specimen surface will result. Consequent- ; ,j
1
cracks generated in the first rep ica ur1 ''-'->
ly, the two parameters of prime importance are specimen temperature,
al. 1978). . any o f these r ather more laborious replica stai.•. which determines the vapour pressure of the ice (water), and the partial
. f ith
Before expenmen ressure of water in the vacuum around the specimen. As can be seen from
. d 1·t h Id be remem b ered that shrinkage or swelling:..0.•.r..'
mg w
bilisat1on metho s, s ou . d/or cleaning can often be minim·,, ig. 6.31, as well as in Table 5.2 in §5.2.la, the saturation vapour pressure
. · d ring thawmg an . f ice varies markedly with temperature.
delicate specimens u . th r with the replica, directly mto a
by placing the frozen specimen'. toge e
For pure ice, the net rate of sublimation is a function of the ice tempera-
alkali or freeze-substitution flmd (see also §6.6).
e and the partial pressure of water above the ice. In the presence of its
n vapour at a pressure (P), vapour molecules will impinge on the ice sur-
e with a flux expressed by the equation;
6.5 Etching
l; = P(2nmrp T)- 05
(6.2)
r
The specimen surface ca: b: ated directly after fracturing. Often,
r:~ ;:rther enhanced by allowing a limit
ever, the surface to.po gr p y h t lied sublimation of volatile
h. This involves t e con ro
=impinging flux (dimensions: molecules per unit area per unit time)
gree of etch mg. . n surface un d er vacu um . This superficial freeze'
als from t e spec1me of the ice crystals exposed during fra =mass of vapour molecules (H,O = 18.016)
process lowers the surface h' en in the final image of the· =gas constant (Stefan-Boltzmann constant)
ti the effect of etc mg as se h =temperature (K)
Consequen y, . . . n of non-volatile components int,
is determined by the d1stnbul!o h oeutectic network produced, =water vapour pressure (Pa)
d ·ally the eutecl!c or YP . .. J'
men an , espec1 • . . . etching is a particular1Y~-~·
(§ 2 1 2) For b10log1cal specimens, .' . the water vapour pressure (P) in the system is equal to the saturation
freezing · · · d ctures such as membranes,,
process since additional faces an s1ru ' 1d --:,,>:_-'. r pressure (P,) for a given temperature, the total departing flux (J,)
faces and cytoplasmic structures, may be revea e .
equal to the impinging flux (.f;). In other words, the constantly eva-
g and recondensing water molecules wi11 maintain a state of equi1i-
6.5.1 Physical conditions for sublimation and condensation ...· lf the partial pressure of water vapour in the vacuum system is
, this state of equilibrium is disturbed and a net rate of sublimation
. . I conditions for the sublimation and co~(·
The basic phys1ca r d . th Chapter concerned wit~; nt to J,-.f;occurs (Southworth et al. 1975).
water have already been _out ine inrtic:lar aspects that must b_~:: modern freeze-fracture replication and etching units have liquid ni-
ing (§5.2). Here we c~n~i~::e~~~d~;ing of a superficial layer of'ai ooJed cold traps around the specimen to reduce the partial pressure
account when contro e d tion contamination of: vapour in the specimen area (§6.3.3b(i)). At 77 K, the saturation
cimen is required and when con ensa ressure of ice is of the order of 1.33 x 10-22 Pa (Ix J0-24 Torr).
ture faces must be avoided. . . that the physi .lecules leaving the specimen are readily trapped providing that the
The process of accurate etchmg reqmres II d (Du 's sufficiently good and the condensation (trapping) coefficient is
i
. 1 monitored and contro e --
the specimen are prec~e 977) If the saturation vapour pl."l'
Robards 1974; Umrat ' . f 1 ressure of watet
h 1983). At 1.33 x J0- 3 Pa (Ix 10-5 Torr), the mean free path of
· ules is greater than 600 cm and most subliming molecules will
the specimen is higher than the par rn p , peded to the trapping surface. Since the water flux from liquid
390 Robards and Sleytr Low temperature methodJ. m
. biological electron microscopy
Freeze-ji-acture replication
391

proportions when cryoprotective agents are present. Nevertheless, as can be


2
Solid Liquid seen from Fig. 6.32 and Table 5.2, under ideal conditions a layer approxi-
10 ,•
./· mately 100 nm thick will sublime from the surface of pure water (ice), main-
Triple point: •
p=4.579 Torr O Gas
tained at 173 K ( - lOOOC), in 1 min. This is a useful rate as it is controllable
--;:: 10°
0 I nd provides adequate· etching for most specimens. However, as can also
t 10-2 I e seen from Fig. 6.32 and Table 5.2, minor fluctuations in specimen tem-
,~
10
-2 •
: 10
-4 I erature cause considerable changes in the saturation vapour pressure of
ater and, therefore, in the etching rate. Even a change of as little as 1.0

10
-4
~
~ 10- 6 .i alters the rate by about 25%; at 183 K etching is excessively fast (approxi-
1
tely 800 nm min- ), while at 163 Kless than 15 nm sublimes within a
0
0. !
10
-· •
>
c
0
10
-6
I
i
·nute. It is thus obvious that, to obtain specified rates of etching, the speci-
n stage temperature must be controlled very accurately and calibrated
10 -· -; 10-10
, larly (Dunlop and Robards 1972). Depending on the actual amount of
ing, it is usual to distinguish between normal etching ( ~ I 00 nm) and
-10
10

-12
10
«I
"' 10

10
-12

-14
I

-150 -100 -so 0 +
50
etching (up to a few µm).
he effects of normal and deep etching on fractured specimens composed
1,','

th etchable (volatile) structures, such as ice crystals, in a non-etchable


Temperature <°CJ

Fig. 6.31. V . ( n of th~ saturation vapour pressure o f ice with temperature. Temperature (°C)
ana 10 _., fiom Umrath 1977.) -50 -10 -90 -100
10•

'
nitrogen-cooled cold,traps can be Practically neglected, as seen fJ . 'c ''
equation: e '
''
10'
'c
~ 10' e

~ E
c
J ""J,~P, (2 nm<p T) -0.5 ~ ,!!
"" ~
• 104 ~

o; o;
. the net water vapour flux (Jnet)is (;)" 104 c
c
(terms as defined m Eq. (6.2)), Th maximum rate of ice ev,
mined by the specimen temperatured. b ~avy (1969) is shown in "' "
!!
. ture as derive Y .
.. ~" .!
c
c 10'
in relation to tempera d . ati"on of the maximum .2 ~
. d. §52 etermm .·• •
As already discusse m . , "b d above can only be app ~ 10'
0
( h
etching) rate by t e m
ethod descn e
lure difference between. (.
. h ~• 0.

. large tempera b ,• w
>
w 10 2
ice and where there ts a Th theoretical values for su..
ding cold traps. e . · T bl
and the surroun d _ 1 oc) are given m a. 10 2
d ll3K(-60an 60 . • 220 200
between 213 an 11· . ns surface cooling, 180
h as gas co 1s10 , .-, Temperature (K)
A variety of factors, sue . I Jar properties assoc.i
.. d special mo ecu d
ness, impunl!es . an affect the sublimation rate (Davy an •
vapourisation of ice all . Mellor 1978) which may be reduc
Southworth et al. 1975,
392 Robards and Sleytr Low tempera I ure met hods in biological electron microscopy Freeze..jracture replication
393

Fracturing No etching Etching Deep etching (i) the general condition of the vacuum system;
(ii) the specimen temperature;
(iii) the position of the fracture plane in relation to outgassing warmer
a surfaces which act as water vapour sources;
the effectiveness of any cold traps which are kept at a lower tempera-
ture than the specimen itself; and
the nature of the fracturing method (simple fracturing or using a
knife).

b efficient cold traps are used in the vacuum system then, even in a relatively
or general vacuum, the quantity of condensable gases can be very low
unlop and Robards 1972; Robards 1974; Umrath, 1977). As shown by
ass spectrometer measurements, water vapour is the main source of speci-
FP

~
Fracture plane

Non-etchable material
n Deep-etched particle
(not completely replicatedf{:"
n contamination in a vacuum system pumped by routine methods (rotary
p and oil diffusion pump, or turbo-molecular pump) without cold i,
I

'1
s. Under ideal conditions, this water comes only from atmospheric
I
Deep depression
12] Etchable material (not completely replicat~
isture adsorped to surfaces during venting the unit. In practice, however, I

Heavy metal-carbon replica er sources of water need to be considered, such as:

F . 6 31 . . 1 and deep etching on fractured specimens ·comp


The effects of conven t 1ona
tg.etchablc
· -'· structures in a non-etchable matnx,· . ~n d ' (b) non-etchable structures man- degassing surfaces which have become covered with hoar frost during
(a)
exposure to moist air;
matnx.
leaks in the system; and
localised outbursts of small amounts of water vapour arising from
heat generation in water-containing systems (e.g. specimen fracturing
matrix (e.g. a pseudoeutectlc . networ k) ' a nd non-etchable
. . struct
. _, or cutting, hoar-frosted moveable parts of the apparatus, specimen
d d by an etchable matrix (e.g. water) are illustrated m Fig. 6. chips generated during fracturing and in poor thermal contact with
ro~h=n using deep etching, it is important to understand s, t~at cold surfaces).
ma become exposed which, for geometric reasons, cannot e
repYI.1ca t ed (F1·g · 6 ·33) · The incomplete replicas
. of such structures
D · d potential sources of specimen contamination are hydrocarbons Iiber-
lapse or get lost during the replica cleaning procedure. un~ om gaskets, finger prints (always use gloves) and back-streaming
it is also possible to 'lose' small liberated structures, before t er, ii diffusion pumps into the vacuum chamber (although this should
them to be replicated, through the effects of the water vapour W1 aproblem in a well-designed system using liquid nitrogen-cooled baf-
worth et al. 1975). ve the diffusion pump).
iscussed earlier (§6.3.3b(i)), liquid nitrogen- or helium-cooled
6.5.2 Etching methods and contamination phenomena provide ideal protection for a specimen during etching. Such cold
pt at a temperature below that of the specimen, efficiently prevent
T understand the phenomena of etch mg . and contaminatio
. lion of the fracture plane both from gases produced by desorp-
ta:t to note that the nature and amount of contamman\ane warmer surfaces and from gases in the system that condense at
. . o n the specimen during exposure of the fracture p
arr1v1ng .Ures lower than that of the specimen itself. Such cold shrouds,
394 Robards and Sfeytr Low temperature methods in biological electron microscopy
Freeze-fracture replication
395
which are only opened to the outer vacuum during replication, are, for
fracture plane. If the microtome assembl is .
example, installed in the Denton and Leybold freeze-fracture replication for contamination shielding th "t . . y used both for fractunng and
, en I is important t 1· h
units and, in a different form with a permanent opening to the evaporators, may be deeply fractured dur1·n . .. o rea tse t at the specimen
g an m1trnl passage of th k ·c .
in the Balzers unit (Figs. 6.15-6.17 in §6.3.3). With snch designs, the drastic lowing pass(es) may fail to ex ose . e m,e while the fol-
reduction in condensable gases in the specimen area has been demonstrated 6.34). Condensation contaminapt. nefw faces m these deep regions (Fig.
ion o exposed part 0 f th .
by long-term exposure experiments. No detectable contamination from con- g_reatly reduced by the liquid ni·t s e specimen is
. . rogen-cooled hood i t· ll d ·
densation onto exposed fracture faces was seen when specimens were kept mt illustrated in Fig. 6. l 5c and d i § ns a e m the Balzers
for up to 15 min at temperatures close to 100 K (Sleytr and Messner 197R; n 6.3.3a. Nevertheless, the possibility
Sleytr et al. 1981). In the Bullivant freeze-fracture device (§6.3.3c(i)), the spe-
cimens seem to be well protected by the cold block, although it is constant\)';
open to the outer vacuum through small evaporation tunnels. Since o~L~)
those gas molecules moving along a path parallel to the tunnel axis reach
the exposed fracture plane, the contamination rate is very low even wn.e-tf
the apparatus is in a relatively poor general vacuum. Attempts to pert\

(~~
controlled etching (§6.5.2) with the Bullivant type of apparatus have
generally unsuccessful since there is no simple way of maintaining a __
stant specimen temperature. Deep etching can be obtained with this ap
tus by using one of two methods. With the first, the specimen is frac
as normal under liquid nitrogen and its temperature is monitored wf ''
thermocouple positioned in the lower block. On average, it takes abo
min for the lower block and specimen to reach 173 K ( - I OO"C). The
necessary to give a good depth of etching must be determined by exper
(Bullivant 1973; Dempsey and Beever 1979). Alternatively, the speci!l{
be fractured with a spring-loaded blade when the correct etching te>
ture has been reached within the chamber. This gives more reprodu ;
suits because the start of etching is determined by the actual fract '
cess (Bullivaut 1980b).
In some freeze-fracture replication units, such as the Balzers a
(§6.3.3a(iii)), a cooled protection surface is provided by the micro!.
and, in the latest unit (§6.3.3a(iii)), by an additional cold shield. F
it is important to stop the blade and holder, immediately after t. Diagran1 illustrating the fracturing of f .
ur from the specimen chi . . a rozen spec11ncn and how sublimation of
in a position over the specimen so that the blade lies about 2.Q ps remaining on the edge of th . ..
e exposed surface The h' . e microtome knife n1ay con-
above the specimen with the edge of the knife about 4.0 mm f . ·
at1on of both deep and shall
· c 1pp1ng of the speci
f
h .
mens own in stages (a) and (b) leads
shown by Staehelin and Bertaud (1971), the knife should not b ow racturcs Durin th C 11 .
·.stages
( (c) and (d)) which d h . g e o owing pass of the micro-
' pro uccs t e fracture I (FP)
ward or backward during etching or else water subliming fr away while the deep , .
011 es ren1a1n unaffected At th
P ane
.
, the shallow fractures
.
chips which are in poor thermal contact with the knife edge r-from the chips the k ·~ · e same time the sublimation of
f h 011 111 e edge leads to the contam·
o t e advancing knife (stage ( )) S .~
c f h .
ina ion o t e specimen sur-
contaminate the fracture plane (Fig. 6.34). Also, a large ga c · u1i.aces of deep f t h
th e Iast pass of th · . rac ures t at are not cleaved
knife and the specimen decreases the effectiveness of protecti e lllJCrotorne knife before r r ' .
fd)). Similar events can lead to et h. . , ep 1cat1on can therefore be contami-
cules desorbing from warmer surrounding surfaces will be a:9 drawn after Staehel1· c indg81n supposedly unetched specimens.- (Re-
n an ertaud J971.)
Robards and Sleytr Low temperature meth ods m
. biological electron microscopy
396 Freeze~fracture replication
397
. f ater vapour genera1ed by frictional heating ..as the
of contaminat10n rom w . be considered (Walzthony et network that forms during freezing. To avoid this masking effect, attempts
knife edge moves over the specimen must s1111 have been made to use volatile cryoprotectants (Buchheim 1976, 1982;
al. 1982). . Schiller and Taugner 1980; Sawada and Yamada 1981). For example, infilt-
"th Id shroud devices, u Itra- high vacuum units allow con-
In common w1 co - " ·nee the amount of adsorbed ration of glutaraldehyde-fixed animal tissue with 30-40% ethanol provides
f t hi g of fracture iaces, st good cryoprotection and also has the advantage that, on etching at 173 K
lamination- ree e c n . tly reduced by baking at 500-700
. th system is grea
gases at all surfaces m e f imen transfer (to reduce the to 178 K ( - 100 to -95°C), in addition to the normal membrane faces (see
Kand by using an air-lock system or spec . . m (§6.3.3b(ii)). Fig. 6.38 in §6.8.2), extended areas of cell surface are revealed (Schiller and
!, t the specimen to a m1n1m u
ma tion of hoar ros on . r roduced under or exposed Taugner 1980). Frozen ethanol has a higher vapour pressure than ice (see
d that fracture ,aces p , '
It has been assume k d 'th a 10 nm layer of adsorbed mate_n•_. Table 5. la in §5.1 and Fig. 5.14 in §5.4.1) and this necessitates very accurate
(H N ) become mas e w1 d' > control of the specimen temperature to avoid excessive surface drying of the
liquid gases e, 2 b h. g under the best vacuum con le<
. t b removed even y e1c in . I ··'"<·· ,specimen. Using muscle as a model system, Sawada and Yamada (1981)
al which canno e . . differences in structura resQ:;::·
1973b) However m practice, n 0 . .. <:found that the concentration of the ethanol has a significant influence on
tions (Moor · '. h been observed when such spec1m(
lution or in the effects of etch1ngf avte d under improved vacuum co -he fracturing behaviour of the specimen. The fracture plane through tissue
d . h th0 se freeze- rac ure
are compare wit : a· Sleytr et al. 1981; Steere 19 filtrated with 100% ethanol (which obviously causes considerable lipid
1 1980
lions (Gross et al. l 97~a~l~~~~a:nd Wal;thony et al. (1981 ), the conde
1 traction) is flat, like a mirror. The lower the concentration of ethanol, the
As shown by Gross et a . . t .s usually determined by the ph eater the likelihood of the tissue splitting preferentially along membranes,
t t h main contam1nan ' 1 . h in specimens cryoprotected with glycerol. However, by lowering the
tion of wa er, e d f t re face Such condensation p
t . of the expose rac u . I 1 ncentration of ethanol, the chance of ice-crystal formation is increased.
chemical proper ies . d"f" t model systems (Deamer et a· .,
b observed m ' .eren d · -· wada and Yamada found that infiltration with 65% ethanol provided a
mena have een M 1971· Walzthonyet al. 198l)an
Staehelin and Bertaud _1971; oor 6 9 7J or to the enhancement of de -sonable compromise between these two factors.
lead to specific decorat10n effectsh (§ . . 'm1'c membrane-intercalat uchheim (l976, 1982) used pure p-dioxane or p-dioxane/water mixtures
hT . eas wh1c may mi ..·• e suspending medium to study the surface structure of casein micelles.
obviously hydrop i 1c, ar . ·1 1 ]though not yet demonstr
1 I 1978b) S1m1 ar y, a . ) mmon with the results from ethanol-treated specimens, changes in the
tieles (Gross e a. . . 'fi decoration (contaminat10n
might be expected that thereb1s spect1h1c vacuum Finally, warts or ring pattern associated with differences in dioxane concentration were 111,.,:
. b hydrocar ons in e . b . To sublime the dioxane, etching is done at 173 to 163 K (-100 to 1y,) i
rophob1c areas Y . d particles (§6.8.2), can e !
. 1 mernbrane-assoc1a1e 1.1.1 •,1
of a similar size ho d. . (Davy and Bran t on 1970) . These structure__•.-. °C). The specimens are then cooled back down to 143 K (-130°C) '

in replicas of etc e ice t but may reflect crystallogr lication. Despite the potential value of these techniques, the forma-
. d · nl!es of the wa er, f f specific artefacts (especially lipid-extraction artefacts) such as with
be ascnbe to impu b d less frequently in spray- r
c .
per1ectlons. Such warts are o serve ) latile cryoprotectants, must constantly be borne in mind.
mens (Zingsheim and Plattner 1976 .

6.5. 3 Etching of' non-aqueous systems Thawing and replica cleaning

. ens if they are suspended


It is also possible to etch spec1md_ ch as ethanol or dic>xan< e freeze-fracture replication process has been completed, the whole
. 1 ueous me ia su chamber, or the air-lock chamber, is vented and the specimen hold-
with smtab e non-aq , !"cable to ice. (Freez
lime,under similar conditions_to thosebapdp '§5 4) hole fracturing device (e.g. complementary replica device), are
lvents is descr1 e in ..
non-aqueous f rozen so f I cerol the most .. The subsequent thawing and replica-cleaning processes are among
ur pressure o g Y ' .·
Due to the 1ow vapo 1 th extended une elicate steps in the whole procedure. Heavy metal/carbon replicas
cryoprotectant, deep etching only revea s e ' uired thickness are extremely fragile and brittle and are strongly
398 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze~ffacture replication
399
attached to the surface that has been replicated. Consequently, dimensional a dish can be almost continuously altered b . .
changes in the specimen during thawing and cleaning can easily cause frag- neously, one for adding and anoth ' Y. usmg two pipettes simulta-
. er <Or removing liqmd Ifd'f'
mentation of the adhering replica. This is generally less of a problem with 1ng reagents are used th fi . . . s. I lerent clean-
' e lfSt so 1ut10n is diluted 11 th
suspensions than with tissues and the effects can be reduced by using one water before the other one is add d . . . a e way down to pure
e in an increasing co t · .
of the replica strengthening methods described in §6.4.2. It is best to thaw chambers to provide a continuou . ncen ration. Washing
. s concentration gradie t th h .
the specimen in a solution with the same composition as that in which the solutions have been designed a d n roug cleaning
n constructed by J
specimen was originally frozen, although, in some instances, it is advanta- by Hohenberg and Mannweiler (1980). ensen et a.1 (1973) and
geous to thaw the specimen in the first of the cleaning solutions. To inhibit
disintegration of replicas during thawing it is possible to transfer the still Cleaning schedule for cell }i-agments and soft biological ' .
frozen, replicated specimen to a plastic scintillation vial '"'"ntaining frozefr/. With most cell organelles and macr 1 1 . peczmens
methanol. The methanol is then allowed to thaw so that the specimen is> after 1-2 h treatment at room temp omto ecu .es, clean replicas are obtained
• era ure with 6% sod· h h .
simultaneously freeze-substituted. The specimens are then rinsed once -ih;':::, :a 30% aqueous solution of chromic acid. Red bl od ium ypoc Ior1te or
water and transferred through sodium hypochlorite and any other cleanihg>f readily dissolved in 6% sodium hypochlorite . o.o cells: for example, are
solutions until the replicas float off. · . 1975) To cl r withm 2 0 mm (Southworth et
. . ean rep icas of adhering viruses Nermut (197
Thawing, cleaning and final washing are best done in small Petri dish at the replica is floated on alternating solutions of 6% s d'3) re~ommends
or in depressions of porcelain 'spotting' plates. Spotting plates have the and 30% sulphuric acid. Using this roced . o o mm ypochlor-
vantage that the white background makes it easier to see replica fragme e usually clean after 2 h F fi dp . ure, replicas of unfixed viruses I•
. or ixe viruses whrch ma r . I I
Whereas replicas of suspensions often float when they are detached fr aning, combined treatment with 700/ s I h , . 'd y equ1re anger
/o u P uric ac1 or 40°/ ch · ·d ,,
'
the support by immersion at an angle of about 45° to the fluid surface, r 2 h) followed by hypochlorite (for 1-2 h) . /o rom1c ac1
is necessary.
es of tissue have a tendency to sink. Spray-frozen samples must fir iii
ii;
immersed in pure acetone to dissolve the butyl benzene (Bachmann;-"', '.) . Clf'aning schedule for microorganisms
Schmitt-Fumian 1973). The replicas, which float in the acetone, are hcas of most bacteria, fungi and algae can b I d "'
1!1

quently transferred, using a pipette, into a 10-20% acetone/water 0% I t' e c eane successfully with ii'
o sou ion of chromic acid within 1-3 da s , I-!'.·
where they spread out on the surface. The rest of the cleaning pro
for the replica can then follow any suitable schedule. Replicas which
ter~:~aning :ith chromic acid solutions is achi:Ve:~~~~:~:~~~r:;~~ 11' I
i-1 1 .1

er met ~ds involve treatment for several hours firstl with a 1 o


11'
.
ue to float during thawing are best transferred from thawing to um hypochlonte solution followed by 70% sul h . ' .· y O% ,,
I I,!

fluids using a 3.0-4.0 mm diameter platinum loop made of0.2 mm .'ce versa (Moor 1970· Lickfeld 1977) S o . p unc acid (Sleytr 1970b)
wire. If the replicas sink, they can be carefully transferred with a pi vein 70% H SO ' . pec1mens which are difficult to
. . o 2 'are often more readily dissolved in 40% h . , .
Unless completely water-soluble specimens are replicated, adheri ic ac1d treatment h th o c rom1c acid.

rial must be cleaned away from the replica before examination. Thi % sodium hypochlori~: is ~s!:~~r:~ta:::e~~:;~. that further cleaning
by oxidation and hydrolysis in acids and alkalis, which do not co
replica, or by enzymatic digestion. A variety of replica-cleanin _<Cleaning schedules for plant tissue
has been proposed and it is often necessary to employ an em:J?: .of the mature tissues f h. h
proach to determine the best cleaning agent and the correct lertg! red with most othe o . ig er plants are extremely heterogeneous
more firmly att· rhspdeczmens. Fracture planes of compact cell walls
for replica cleaning for a particular biological specimen. When· ac e to the replica than i
one cleaning reagent is used, care must be taken to minimise- vacuoles. This variaf . .. ' or examp1e, the large
lling or shr' k. f10n m compos1t10n, and the resulting differen-
mentation during transfer of the replica. Particularly if H,S04 · 1n ing, o ten causes fragmentation of the r 1· d .
transfer through a gently graded dilution series can be crucia and cle · Th ep 1ca urmg
plicas i:~~ng. .erefore, on~ .of the best methods of retaining large
floating and sunken replicas, the concentration of the cleanin:
provide an addlt10nal supporting layer (Robards and
l-1,

I
400 Robards and Sleytr Low te1nperature me tiio d~. in biological electron n1icroscopy
Freeze-fracture replication
401
b ved once the organic material has
Umrath 1978; §6.4.2) that can e rehmod to silver-backed replicas are best other plant tissue which yield comparatively large replicas have been de-
d , y Specnnens attac e f
been cleare awa · . . 0-40% h omic acid for 1-2 d'1ys. A !er scribed by Platt-Aloia and Thomson (1982), Dissolution of the tissue
d by immersion 1n 3 o c r . . d
cleane . away
d' '11 d water (three ch anges) , the strengthened replica 1s treate involves the sequential use of alcoholic KOH (12% KOH in 95% ethauol),
rinsing m. 1st1 e ,
. . 9% potassium , 'de and 10%o hydrogen perox-
cyan1 sulphuric acid (70%) and chromic acid (half-strength glassware cleaning sol-
with a mixture containing ~
. ution).
. f l-2 h to dissolve the silver.
t 0 ck solutwn) m water or
ide (3% aqueous s ried out at room t emperature · The solutions are then
All treatments are car r is transferred directly into clean (iv) Cleaning schedules for animal tissue
either diluted with water or the rep ica
Replicas of most soft and unfixed tissue can easily be cleaned with 30-40%
water. ·1· . ~ evapora t.mg a backi.ng layer are available, tb:e.-·... • -_:.· ~:.-;: chromic acid or 6% sodium hypochlorite in 5-24 h. As with other speci-
If no technical faci illes or . · by Fineran (1978) should mens, shorter cleaning times are required when a Cr0 solution is used at
d. fn 1 lant mater1a1 given - -, 3
typical schedules 60°C For tissues that have a tendency to shrink, it is recommended that
b t .forclean
igesreplicas
g. P wit . h.in a, few hours he suggests that, after. •. .·.·•.· .'· · ·
be used. the cleaning procedure is started in a mixture of 10% nitric acid and 10%
. Tooh am
t under the rep 1tea is t
. , first gradually replaced over half a.11.·.·c:
' J>',--
thawmg, t e wa er . . 'd d th left for 1 h. The concentrated acid sulphuric acid, or a 5-10% solution of caustic soda (Stolinski 1977),
t ated n1tr1c act an en . _._>-
hour by concen r , . . 'd h'ch is used for 1 h either at fOOQ'Ii epending on the type of tissue, this initial treatment lasts for a few hours
. 1 d by fuming n1tr1c ac1 , w 1 . -:<\
1s then rep ace . th acid is slowly diluted with water.•.c1;< overnight Subsequent cleaning steps involve strong acids (40% chromic
t 75"C On coo1mg, e d f <
temperature or id or 70% H 2 S04) or alkaline solutions at room temperature or 60°C The
I , 1· schedule (Fmeran 1978) , the replica is transferre
a . . . , , a.,.
In an a terna ive . "d for 4 h The acid rs timal time for the final cleaning step must be checked experimentally for
, '11 d t to 75% sulphunc ac1 ,
thawing in dist1 e wa er, . . o c: d into a mixture of 5% chr erent specimens. The more complex and prolonged cleaning procedures,
10% nd the rephca is trans.erre .
diluted to < o a . . ht As for the cleaning procedure n above, are mainly used for difficult unfixed tissues, such as skin or
and 5% nitric acid and left ovfierntl1g b~ placed in bleach for 2-4 h and ipheral nerve, and for all fixed tissues, which are more difficult to digest ,,
. · rephcas can trs Y I
m1cro-organ1sms,
transferred to 75% H,SO, for up to 24 ,
h or the treatment can be rey ,,11,
Cleaning schedule for specimens with a high lipid content
using the acid first. . thod that ensures the reco replicas of these specimens, it is recommended that the c1eaning proce-
Pearce (1983) suggested a c~~nmg ~~es a preliminary dehydrati starts with 1-2 h in a 2: 1 mixture of chloroform/methanol before
replicas of fractured_ leaf cells. l~~~~vi~ methanol and then trans£ 'sing agents are used (Moor 1970; Pearce 1983). When indexed gold
The replicated specimens
chloroform for 15 mm. T , y
:;e
~re then transferred to sodium hyp
o is used to transfer the
are used as an aid to the observation of complementary replicas of sus-
'ims, the Vaseline which may be used to hold the two grids in register
for 1 h and finally rinsed with watehr. A !do , p then floated with the s ig, 6.2; type Bh) should be removed with chloroform before the speci-
t d ld grid and t is gn is , ,
to a carbon-coa e go . .d f three days, The gold gn , digested. This is done, after thawing, by drying the grids, with the
.d 50% chromic ac1 or
on the upper s1 e, on o d then be used to transfer tb ed replicas, and then washing them in chloroform, with frequent agi-
h bon layer an can
separates from t e car h f d' t'lled water The cleaned , for a few minutes (Hess and Blair 1972),
d }" t a rinsing bat 0 lS I . ·
layer an rep ica o h 11 cted on the original gold gn 'd or Vaseline contamination can be also removed from mounted rep-
adherent carbon layer are t .e~ ~o e h . hloroforrn the cuti y placing the grid over boiling chloroform and allowing the chloro-
W"thout the initial was inc '
other support. I , , f n extensive layer that obsc apour to condense on the replica surface (Vail and Stollery 1978).
nised by the chromic acid and orms a '_:<:':-----
t al. (1971) suggest that the replicated specimen is treated with 25%
of the rest of the replica. . d handling replicas of cutinise~, lformamide in distilled water overnight, to extract lipid, before spe-
Procedures for d1ssolv1ng an ,-:; igestion and replica cleaning procedures.
402 Robards and Sleytr Low ten1perature methods in hiological electron microscopy Freeze-fracture replication
403
6.7 Mounting of replicas on grids a
b
c

After cleaning, the replicas are carefully rinsed in several changes of distilled
(preferably double-distilled) water. During this final step, it is particularly
important to avoid secondary contamination of the replica from dirty
water, dishes and handling tools (tweezers, loops, etc.). Floating replicas are
manipulated with a platinum wire loop while sunken ones are gently trans-
ferred with a pipette (e.g. a shortened pasteur pipette). After the final wash,
the replica is picked up on a coated or uncoated grid. This step is omitted Fig. 6.35. Techniques for tr' ·f. .
.
.
ans ernng a rephca from the fi1 l
when replicas remain attached to the grids which are used in the preparation. .·.a gnd down on top of a floatin . , . . na wash onto a grid. (a) Brin ·
. g rep 1ica. the fluid surface 5h Id gtng
should be lifted off horizontally. (b) H 0 Id" . _ ou not be broken and the grid
of complementary fracture faces (§6.2.2d). ·h'I . . Ingthereplicaag· tth .
.away h 1 1 st the gnd 1s removed from the fluid . ains e gnd, to prevent it floating
Replicas of suspensions of isolated cells or cell organelles have considet' underneath t11e replica parallel to th fl 'd 'with a mounted eyelash. (c) Raising a grid from
' e lll surface The ·d h
able redundancy of information and are picked up on uncoated low trans+'. a convex surface to the replica in order to h ··l gn . as been bent slightly to present
mission grids with wide bars. With more complex specimens, such as anima{"' e p centralise the replica on the grid.

or plant tissues, much useful information would be lost beneath the g '
bars. Replicas from this type of specimen are picked up on high transrnf 6.35b). When an uncoated rid. .
sion grids. Grids with a single, large central hole are especially useful. T g or sunken replica the chanc f . kg is raised from beneath a float-
' es 0 pie mg up th r .
grids are coated with a support film, preferably of Formvar which is rri' n are much better if the grid.i's b t l' h e rep ica ma central posi-
en s 1g tly conv · 1 ·
stable in the electron microscope than Collodion. Carbon reinforceme and then raised parallel to th ex m re at10n to the rep- ;!'·
. e water surface (Fig 6 35 ) .
not necessary, since the heavy metal/carbon replicas provide sufficient ramed away by placing the .d fil · · c . Excess hquid
. gr1 on 1 ter paper
lily themselves. ornet1mes replicas have a tendency to curl u .. . .
Jn practice, there are many different ways of transferring the replica process. To unroll a repl1.ca th p and smk durmg the clean-
. on e water surfa "t · ·
the final wash in distilled water to the centre of a grid. It is often replica (with a pipette) first to 100% h ce I is possible to transfer
carry out S1:1Ch manipulations under a dissecting microscope. If it is p 'n a drop of the organic solvent to e:h a:ol or acetone, and then back,
to keep the replica floating during the cleaning procedure, then it in surface tension the repl1·c , ol e tslllled water. Due to differ-
' a uncur s on touch. th
picked up from the surface by bringing a coated (or uncoated) gri breaks into smaller pi·e H mg e water but it also
..- . ces. owever the ·
in clean tweezers, down on it (Fig. 6.35a). The grid need not be gh Information to be useful 1· 1 i' . se pieces usually contain
, par ICU ar y If SU .
down into the trough and the fluid surface should only be depress her words, it is better t . spens1ons are examined.
. o save a small piece of hi
broken. The grid is subsequently lifted to a horizontal position an piece that cannot be stud· db . . somet ng useful than
Ie ecause it is curled up!
plus water is blotted off with filter paper, either by touching the ·
filter paper or by placing the grid with the replica uppermost flat
ter paper. It is important to be sure that there is no water left bet Interpretation ofJreeze:fracture replicas
tips of the tweezers before releasing the dry grid. Floating repli
be picked up with a wire loop (about 4 mm diameter) and thus >:.General considerations
to coated or uncoated grids. Finally, either floating or sunken.
be picked up by placing the grid underneath the replica and therr etal/carbon replicas obtained b £ .
ing it. Tbis is generally easier using uncoated grids. With coa, transn1ission electron . y reeze-fracture rephcation are stu-
m1croscopy Th fi 1 .
floating replica will often drift away. Working under a disse screen is determined b th . e ina Image seen on the mic-
scope provides the possibility of using an eyelash fixed to a rod'. ts a two-dimensional y. e_ topography of the fracture face and
PfDJect10n of the mass
. distribution ,·n th e rep-
404 Robards and Sleytr Low te1nperature methods in biological electron microscopy
Freeze-fracture replication
p ~ ~
lica. Thus, to interpret the latent three-dimensional information encoded in rovi ed that the fracturing process has ex
the electron micrographs, it is important to know the direction and angle quality freeze-fracture replica b posed areas of interest, a high
of shadowing. The direction of shadowing is generally indicated on pub- . can e charactensed by the following criteria:
lished micrographs by an arrow. (1) No major structural changes have b .
The correct appearance of the replica relief (surface sculpturing) is easily and pretreatment (see §2.3.2). een mtroduced during sampling
obtained by orientating electron micrographs with the shadows extending The freezing procedure has limited ice er
away from the observer or with the shadows pointing from the bottom to does not mterfere with mor h l . ystal growth to a size which
the top of the picture. Other orientations may give an unpractised observer remembered that disruptive i p o ogical analysis (§2.3. l ). It must be
t ce crystals may al d l
the optical illusion of reversed relief (i.e. depressions will appear as protru· men emperature has risen ab so eve op if the speci-
berances and vice versa). Since no heavy metal is deposited in the lee ofproc. preparation (§2.1 ). ove the recrystallisation point durin
minent structures, shadows appear as white areas on prints when electron Shadow· g
mg must have reproduced th
microscope negatives are directly used for photographic printing. Some . ture face as faithfully as possible S e surface geometry of the frac-
authors publish reversed images in order to obtain dark shadows. This giv~s/' shadows with good contrast R ·1· tructural details must show sharp
i _ · ep 1cas with too
pictures which provide the illusion that the freeze-fracture relief is ill11111ii•' ,ess s1ructural detail althou h th . great a contrast reveal
nated by light. A procedure has also been described which allows prinf ··. sional illusion of the top g h ey may provide a better three-dimen-
. ograp Y at low mag 'fi ·
of freeze.fracture pictures with black shadows without the necessity of p resistance evaporation techni ues m icat10n (§6.4.la). With
paring an intermediate negative (Schulz and Reynolds 1977). In fact, li quently due to insufficient e q (§6.4. lc), lack of contrast is fre·
b S h vaporat1on of platin .
is gained by this procedure and the danger of misinterpretation of fine st on. uc replicas have a soft um in relation to car-
turn! detail is obvious. As pointed out earlier (§6.4. la), the faithful re of structural details. appearance and show poor resolution
duction of surface structure information by a shadowing method is li The replica must be free of incom letel .
by the condensation process of heavy metal atoms. Reversed con which is seen in the micro ~ y dissolved specimen material
areas in which the accumulation of heavy metal does not accurately outlines. scope as Irregular dark patches with diffuse
the surface geometry (decoration phenomena) is particularly misleadi
The fracturing process frequently splits biological specimens with' e criteria are g ll
. enera y easy to establish but
branes, so leading to views of extended membrane faces. With the rmatwn of specific artefacts which 'll b . do not take into account
exceptions of isolated organelles and procaryotic cells, the fract t all areas of a rep!' . w1 e discussed in §6.8.3.
ica necessanly show h. h .
produced may have little similarity with images from thin sections. .of the replica may show th t ig quality fracture faces
e s ructure of th f ·
larly for the interpretation of fracture planes through tissue or en;{?r the surface of the . e un ractured edges of the
. specimen support F h
eucaryotic cells, it is important to have previous knowledge oft specunens smoothed down .th . . urt ennore, most repli-
fl wi a micro tom hib ·
of the specimen obtained from other electron (or light) microsco · .or at areas generated by local f . !' b e ex it areas with knife
ods. Unless replica reinforcement methods are used (§6.4.2), it i surface. These areas are dev .dnc f10n etween the knife and the spe-
o1 o any lllD t'
to maintain coherent replicas of fracture faces through a heterog s me 1led regions (§6 3 lb) s· ·1 onna wn and are inter·
''t · · · lilli arly eve d .
sue. It is commonly necessary for several pieces of replica to I y of shadowing may vary in diffe;e nun er optlmal conditions,
to obtain a complete view of structures exposed by a single f _the case if fracturing has created nt parts of the replica. This is
best aid in the understanding of replica fragments is the examin nentated at an extremely low l a coarse topography with some
sections of the same material, since the gross morphological fea mages of fine structures that ~ng e ~o the evaporation (shadowing)
sections and freeze-etching images are very similar. However,. ly low angle may be considera:~e . een shadowed at such an un-
differences in the preservation of fine structural details maj scope equipped with a go . y improved by tiltmg the replica
since chemical fixation and freeze-fixation act very differe ison of etched mometer stage.
and unetched fracture faces of the s .
§2.3.l). ame specimen
406 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze~fi·acture replication
407

can be of great value in the interpretation of the structure of different cell


components. Steere and Erbe (1979) have even developed a method that al-
lows the evaluation of etched and unetched images of complementary frac-
ture faces. Under condensation-free conditions, the structural differences
between etched and unetched images are entirely due to the loss of volatile
material which, for biological specimens, is predominantly water. Since
moderate segregation phenomena (§2.1.2) occur with most freezing techni-
ques, etching roughens hydrated regions by exposing the pseudo-eutectic
network (see Fig. 6.33a in §6.5.1 and Fig. 6.36). Thus spatial differences in
ice crystal size can reflect differences in water content (degree of hydration).
An obvious example of this is the contents of vacuoles which frequently
how a much rougher etched structure than the surrounding cytoplasm.
· ilarly, in procaryotic cells which have been chemically fJXed before freez-
' the nucleoplasm is easily distinguished from the surrounding matrix by
rougher structure after etching (Lickfeld 1968; Nanninga 1973). By com-
ring both etched and unetched complementary areas of freeze-fractured
us crystals, Steere and Erbe (1979) were able to show that etching may
ea1 useful information even in specimens which have been treated with
concentrations of cryoprotectant (25% glycerol plus 25% sucrose).
11 structures are not always enhanced by etclting, however. Fine struc-
detail in regions which fracture but are not in the main fracture plane
become obscured by lowering the surface of exposed ice crystals. For
pie, structural components of the cytoplasm or nucleoplasm which
been pulled up or deformed by the fracturing process are only clearly
guished if the background is smooth, as in an unetched fracture face
_ant et al. 1972). However, deep etching in combination with rotary
ing may help to reveal, for example, the fine structural detail of the
letal network of a cell (Heuser 198 I).
brane face views are generally not affected by etching but become
o distinguish from the hydrated surroundings which change from
to rough as etching progresses (Fig. 6.36). Nevertheless, it has been

Frozen-fractured yeast cells before and after etching. Yeast cells (Saccharomyces
I
11
were rapidly frozen (as a 30 µm thick sandwich) by dipping in liquid Freon 22 and
fractured and replicated without etching (a) or with 20 s etching at 173 K (b). The
fracture faces of the etched and unetched specimens reveal the same structural dc-
.'..frozen specimens, etching changes the appearance of the cross-fractured cytoplasm
,.
,.'
'
!
ii11
),'
oath to rough. Etching has also changed the dense coat of radially arranged fibrous '1
_erial (F) on the cell surface. The arrows indicate the direction of shadowing. 'I 1f1

'
:l,1 ' i,
:1,i

,,1; '"
' I..
408 Robards an d •)01eytr Lo"'
" tem'nerature
1' met h ods in biological electron niicroscopy
Freeze-fracture replication
409
lt . the formation of small holes in
shown that prolonged etching can resu19;~) which are possibly the sites of membranes strongly depends on the thickness of the evaporated layer, angle
membranes
CM eyer and Wmkelmann . (Tranum-Jensen and Bh a kd"I of shadowing and orientation of the membrane in relation to the evapo-
. f t ansmembrane proteins . ration source (Lickfeld et al. 1972).
concentrations o r h. h pear in unetched specimens as
f t d membranes w tc ap . Surfaces of single cells or isolated cell components (e.g. membranes),
1983). Cross- rac ure . ' ft lowering the ice matnx. It
. h d 0 uble-line structure a er .
single !mes, s ow a d" ensional image of a replicated which are not favoured by the fracture plane, can be exposed by deep etch-
1 b remembered that a two- im 1 ing (§6.5.2) provided the specimen can be suspended in distilled water. For
must a ways e . . the microscope screen. Consequent y,
three-dimensional structure i_s seen o~ ross-fractured single or multi-1ayered many bacteria it has been shown that extended areas of regular arrays of
as with other structures, the image o c macromolecules on the cell surface are clearly visible after deep etching (Fig.
6.37), even when cells have been frozen while suspended in the growth
medium (Sleytr 1978; Sleytr and Glauert 1975, 1982).
The study of complementary areas on both fracture faces created by a
single fracture (§6.3.3a(ii)) has been shown to be particularly valuable in the
interpretation of freeze-fracture images and has provided considerab]e in-
sight into the generation of artefacts (§6.8.3). As an aid to localising comple-
mentary areas of fractured suspensions, replicas are mounted on gold finder
grids. For finding complementary areas on replicas which were not origi-
ally attached to finder grids, the pattern of distribution of marker cells
uch as yeast), or easily identifiable features of tissue fracture faces, may
e examined in a conventional or, preferably, a projection light microscope
leytr and Umrath l974a). The replicas on grids are placed on a micros-
e slide and drawings or micrographs of the replica are made. If a projec-
microscope is available, the drawings can be made by tracing on trans-
ent paper. As another aid to finding identical areas more easily, it has
n suggested that one of the complementary replicas should be mounted
"de down so that the two images in the microscope are comparable and
mirror images of each other (Bullivant 1973).

Interpretation of membrane structures and nomenclature

dominating feature of freeze-fracture replicas of biological specimens


en membrane face views (Fig, 6.38), It is now well established, with
ception of the cytoplasmic membranes of those Archaebacteria which
. . ial cell suspension (Desu(fotomaculum
bnonnal lipid composition (Langworthy et al. 1982), that membranes
Fig. 6.37. Fracture faces of a b_acter (b) "th tching. In non-etched c long an interior plane and that true membrane surfaces are only
·
Pended in growth medium a w ) (· ) 1thout and WI e
. b cm and cm pr can be o
f h t 0 plasmic mem rane 1;;r d by etching (Fig. 6.38) (for reviews see Branton 1971; Verkleij and
only the fracture faces o t e cy . . mbrane· p, cross-fract
. . §6 8 2 ) cm cytoplasmic me
menclature see Fig. 6.41 in · .· · ' II
' .
pace· p surfaces facing cytop_
gaert 1975, 1978; Zingsheim and Plattner 1976; Bullivant 1977). With
f f ctngextrace 1
uars , ' . (e.g.< eptions, both membrane fracture faces are studded with particles
F fracture faces; E, sur aces a . I sed by deep etching
' 11 r~ , structure (S) is on Y expo . . --<
pression (pits) of various sizes (Figs. 6.38 and 6.39). Particles seen
crystalline cell wa sur ace . . d" t the direction of shadowing.
Arrows 1n 1ca e
·<:>·
branes are described as membrane-intercalated or membrane-asso-
410 Robards and Sleytr Low temperature inet h ods in biological electron microscopy
Freeze~/Yacture replication
411
.1 . 1 Convincing evidence has accu-
. st membrane part ices.
ciated part1c es, or JU . . . . . the bimolecular leaflet of the is a fixed characteristic of a given membrane type (Branton 1971). The com-
h th eflect d1scontinu1ties in position of most particles is not known but their distribution between the
mulated t at ey r . oprotein or lipoprotein structur-
1. "d enerated by protem, g1ye .. two complementary fracture faces reflects an asymmetric organization of
membrane 1p1 · s, g ted lipid . les (z·mgs h eim and Plattner 1976; VerkleiJ
. m1cel .
b biological membranes. Besides the two faces of the interior of the mem-
es, or even Y mvert 1975 1978· Ververgaert a nd Verkleij 1978 ,· Verkle1j et al.
and Ververgaer • ' b t pes it has been shown that the brane, both true surfaces of a membrane can be exposed by etching provid- ' I

1980). By comparing different me;' rane y both membrane fracture faces ed that they are in intimate contact with ice (Fig. 6.38). Thus, extended areas
number and distribution of part1c es seen on of membrane surface are only revealed by deep etching (§6.5.2) when iso-
lated membranes have been suspended in water, etchable organic solvents
{§6.5.3), or a volatile buffer. Otherwise, as a result of segregation pheno-
mena, the true membrane surfaces may become masked by non-etchab1e
aterial which is pushed against the membrane during ice crystal growth
2.3.1).
Several systems of nomenclature have been used in the literature to desig-
ate the two internal membrane faces exposed by fracturing and the two
urfaces only revealed by etching. The most common, among others, were
, B, C and D faces (Bullivant 1973), ( +) and ( - ) faces (Meyer and Win-
.lmann 1969) and convex and concave faces (Pinto da Silva and Branton

.CI)
,
_______ . '
70). As a logical consequence of this confusion of descriptions, a unifying
was suggested for plasma membranes by a group of scientists active

,',
. ,',
.
.

No etching

;I
,,,·I

Etching

. h . ,· how two frac t ur e faces are produced by spli.tt


al
Fig. 6.38. Diagram~ ov.1.ng , lane Non-etched specimens only rcve ,
, ;'I
rotoplasn1ic fracture face (PF; see Fig. 6.40) of a yeast vacuolar membrane. Mern-
membrane along a unique intern~! p . . f. embrane surface views beco, alated particles, and holes (pits; arrowed), from which similar particles have been
ture faces whereas, on the sublimation o ice, m
removed, can be seen.

I,

I
412 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-fracture replication
. 413
in the field (Branton et al. 1975). It was proposed that, for any membrane Tlm method for the d . .
. d es1gnat10n of frozen f
that can be split (Fig. 6.40), the half closest to the cytoplasm, nucleoplasm, m ependent of the number of m b - ractured bacterial envelopes is
em ranes or non b
chloroplast stroma or mitochondrial matrix be designated the protoplasmic spaces an d h elps to overcome the d" . -mem raneous layers and
d · · 1vers1ty f ·
half, abbreviated P, and that the complementary half, closest to the extracel- es1gnat10ns previously used in th l" o mcoherent topographical
lular space, or endoplasmic space, be designated accordingly the extracellu- ft . l G e iterature As ca b
o yp1ca ram-positive (Fig. 6.41a) and . n e seen from examples
lar, exoplasmic or endoplasmic half, and abbreviated E. The exoplasmic and "_lopes, membranes or non-memb Gram-negative (Fig. 6.41b) enve-
. d raneous layers fi
endoplasmic space would include the interior of endocytic vacuoles, phago" ·_o
t th.elf escription from th"in sections
. are Irst1y labelled accord"
usi . 1ng
somes, p'rimary and secondary lysosomes, food vacuoles, tonoplasts, cisteri_'_::-,, tud1ed. In general the ~ ll . , . ng specific data on the organi
, ) . ' o owing abbrev1atio sm
nae and vesicles of dictyosomes and the endoplasmic reticulum and the ciSc om, pept1doglycan (pg), cell wall (cw c tons am used: outer membrane
ternae formed between the inner and outer nuclear membranes. The abbre:..i'. present, surface layer (s) S di ), . y plasm1c membrane (cm) and
. t .h . econ y, an mdex . dd d . ,
viation E is also used to designate the half-membranes lying closest to the · es w ether a face is direct d t IS a e which clearly ind. -
e awards the cell t l I
space between the inner and outer membranes of mitochondria and chlqt, '. extracellular space (E). Th. dl cy op asm (P) or toward
· IT Y a supplem t · s
oplasts, as well as the intrathylakoid space. Views of the fracture face ofth~j.'. arly mdicates whether an f , en ary mdex is added which
un ractured surface (S) or a f ractured face (F) I I
half of the membrane associated with the protoplasm are referred to as t
protoplasmic face (PF) and that associated with the extracellular, exop1
mic or endoplasmic space as the extracellular face (EF). The true membr.lli a
surfaces, which may be revealed by etching, are then referred to as t
surface (PS) and E surface (ES), the former representing the membrane· Cytoplasmic membrane

face originally in contact with the protoplasm, the latter with the exo
Cell wall
Since the designation E or P is not applicable to procaryotic cells b , .. ,
,1,
of their interrnembraneous spaces, it was suggested by Schmid et al..
that the terms extracellular space (E) and protoplasm (P) should cw PS ',1'

used at the reference level. In this way, the fracture faces (F) and.
(S) of the membranes and non-mernbraneous layers of Gram-pos{
'•s
Gram-negative bacterial envelopes can be clearly designated (ff

ES Cytoplasmic membrane

l PeptidogJycan

,,,
, Outer membrane

I ' ' ' ,


Surface fayer
I
'
'
Protoplasm 1
EF
~~·-~J_
PF
PS om

PS '•s

Fig. 6.40. Nomenclature for the freeze-fracture faces of membranes e N


osmencl~ture system for the fracture faces an
and for the membrane surfaces exposed by etching. PF, Fracture face of{ha Ea typical Gram-positive bacterial env 1 ,d surfaces of procaryotic cell envc-
brane associated with the protoplasm. EF, Fracture face of that half oLt~'-- . 'Surfaces facing the extracellular s : o~e and (b) shows a typical Gram-nega-
ciated with the extra-cellular space. PS and ES, Surfaces of the membrart ·,· S, etched surfaces; cm cytopl . p ce, P, surfaces facing the cytoplasm· F
out ' asm1c membrane· JI ' '
with the protoplasmic or extra-cellular sides, respectively. (After Bra(i} ' er n1embrane; s, surface layer (R d. 'cw, ce wall; pg, peptidoglycan·
. e tawn after Schmid ct al. 1980.) '
414 Robards and Sleytr Low temperatw.e ' nethods in biological electron microscopy Freeze-fracture replication
415
. . th. olume have been labelled according to
is shown. The illustrat10ns m ts v 1 (1975) and Schmid et al. material onto the fracture face of the freshly cleaved specimen in the period
the nomenclature suggested by Branton et a . between fracturing and replication is one of the main sources of artefact.
(1980). Condensation contamination is frequently not easy to detect unless it has
led to obvious structural alterations by masking or enhancing fine structural
6.8.3 Artefacts details or has created prominent new structures which have no complemen-
tary partners on the opposite fracture face. Warts or flat plaques seen in
. . f th. Chapter the early hope that freeze4
As pointed o.ut at the be~1:nmg. oht ~;ve artefa~t-free images of biologic~<: areas which are normally devoid of such structures (e.g. smooth lipid or
membrane fracture faces) are strong indications of heavy contamination of
fracture rephcat10n met o s mI1g
s over optim1st1c t was soo
n realised that the individual prep.·. '.r.··. ·•.
the fracture face with water or hydrocarbon vapour (Deamer et al. 1970; i' I
'I
structures
· 1 wa - lead to ·th e pro duction of artefacts and that,
each . for. co. ··'.':'.·.<:.·.·
arat1on s eps can . . coded in the final two-d1mens1ona:lt-: Staehelin 1973; Bullivant 1973; Walzthiiny et al. 1981). It has been suggested
. tation of the mformat10n en . ,. that high quality replicas can be judged, for instance, from the number of
rect
. 1nterpre
f , freeze-fracture rep1tea . as seen on the microscope . . screen, i.t..)§/ · mall holes that correspond to the pattern of membrane particles on oppo-
image o a h [ t rs involved in deterrn1n1ng structu7_
essential to understand fully~ e ac_od Plattner 1976· Bullivant 1977; S • ·te fracture faces (Staehelin 1973). Although the existence of these mem-
changes (Branton 1973; Zmgs e1m abn d Kistler 1979· Lepault ne pits is an excellent indication of a good (high resolution) preparation,
b d 1977a 1982· Kellen erger an , ir absence is not necessarily due to specific condensation contamination.
and Ro ar s ' 'dz· h 1.m 1983 . Verkleij 1984). Since cryo
h t 1980· Plattner an mgs e ' ·
D~boc e ' . h minimise segregation phenomena are no~ a~a brnne pits may also be destroyed by deformation or dislocation during
e-fracturing or during shadowing (Sleytr and Robards 1982). In con-
at10n methods wh1c . d these consequently retain the d1stn
for most biological spe:1~e~;t~:ing in the native state, artefacts are m t, specific decoration phenomena (§6.9. 7) may easily mimic membrane
tides (Bullivant 1973; Zingsheim and Plattner 1976). ,,
of molecules close to t a p f on steps particularly fracturing and re '

caused by subsequent prepara 1 ' l. f om both halves o onsidering that condensation-free conditions can be maintained in_ the ·11'
··,I;'
bT duce and evaluate rep icas r imen area by using either cold shrouds or ultra-high vacuum techno,
tion. The a 11ty to_ pro (§ a(ii)) has proved a valuable asset in . 1'1,1
633 , the possibility of artefact formation should be restricted to the fractur- ,,
I'
zen-fractured specimen . . h ,. f artefact formation. Whil
. t th , ture and mec an1sm o '.8.nd replication processes. Nevertheless, it is also necessary to be aware
1:,i1'
to mterpre e na f l· nes match the finer details :-,1,'
features of complementary racture p a . h. , The most likely dehydration during etching can cause the aggregation of particles and r,.11'

tions ior t e 1ac 0


exhib~ theh an ticki p a/~~:P ~;:~~7a:f~r~~~~~:~ ~~~inally opposite
1 se phenomena (Kistler and Kellenberger I 977; Lepault and Dubochet
f,' I
1,11_
11'.111
Sleytr and Robards 1982). i.' I
?ii',)
11;11
important to remember that, using the right technical methods, speci- 1
(i) condensation of material on. the freshly fractured surface in freeze-fractured under liquid nitrogen (§6.3.3c(i)) or liquid helium :',_·, 11,',.i
:1·,,1)
between fracturing and rephcat10n (§6.5.2), ts du c(ii)) show the same structural details on fracture faces as specimens
(ii) deformation and dislocation of specimen componen d at similar temperatures under high vacuum (Bullivant 1973, 1'1
,,, 111
fracturing (§6.3.1); lfng from dehyd··· Sleytr and Messner 1978; Bullivant et al. 1979; Sleytr et al. 1981).
(iii) collapse and aggregation phenomena resu I ·.• '" 111
words, the formation of a disturbing, or resolution-limiting, non- r
iug etching; and . r ~ rmationX e adsorption layer as predicted by Moor (1973b) is not detectable 11

(iv) thermal load-induced alterations durmg rep ica o : hen specimens have been freeze-fractured under liquid helium and 11

.ed via liquid nitrogen into evaporation units for replication (Sleytr
. . th reatest variations in sp~,
In freeze-fracture rephcat10n, e g . 1 . §6 I) but l). Steere and Erbe (1979) reported that even a layer of contami- 11
· (see Fig. 6. m · ,.
tion are in the fracturmg process 1· t. n of the spei.·. e can be sublimed from fracture faces under etching conditions
h d
of which freeze-fracture me t o is
· used rep 1ca 10 ·
, nit (§6 3.2). Thus
·.:
t; old shroud device (§6.3.3b(i)) without detectable loss in resolution Ii
face must be carried out tn a vacuum u . with non-contaminated specimens. This correlates with the obser-
\'1
!,

111

·''
1:

i1,' X11
l1,1'il'lli
416 Robards and Sleytr Low temperature nw !ho ds in biological electron microscopy Freeze-fracture replication
417

. . (dee etching) of two-dimensional ar- form during freeze-fracturing even at the temperature of liquid helium (for
vation that, after par.tial freeze-dry)mg 1 pg the sublimation of a layer of
l les (proteins ' tnvo v1n general reviews see Sleytr and Robards l977a, 1982 and, with reference to
rays of macromo ecu
. t
. ~
thick betore exposm
. g the frozen specimens, the sha-
membranes, Zingsheim and Plattner 1976). Plastically deformed compo-
ice several m1crome res t ral details as vacuum-fractured
· 1 the same s rue1u nents frequently exhibit a fibrillar structure and show a clear lack of comple-
<lowed preparat10ns revea . ns (Crowther and Sleytr 1977;
and etched or negatively staineSd :ec1~\ 1980· Gross et al. 1982; Wild- mentarity on opposite fracture faces. Fig. 6.42 is a schematic illustration of
Sleytr 1978; Kistler et al. 1978; tu ere a. , the possible events during the fracturing and subsequent preparation steps
of a two-phase system (polymers in an ice matrix) (Sleytr and Robards
haber et al. 1982). . ts have shown that ductile and·'
brittle regions may alternate a ong
i'
As discussed earlier (§6.3. l ex~~:·~~~ture plane and polymers may
1982). The deformation products may either stick out from the fracture
plane or collapse onto the nearest surface. Strong evidence exists that, due
,I
I

to thermal load during evaporation, the survival chances of delicate fibrillar


deformation products sticking out from the fracture plane are very small.
Fracturing Etching and
Structure
replication his means that, generally, only fibrils which have collapsed onto a large
ominant structure and are in close thermal contact with it wi11 be seen on
e final replica. During specimen fracturing, deformed structures may have
e ability to recontract if freeze-fracturing is done at higher temperatures,
t may re-freeze in the extended position at lower fracturing temperatures.
his may give an apparent increase of plastically deformed structures when
A eze-fractnring is done at low temperatures (Steere et al. 1980; Sleytr and


bards 1982). Structural distortion and collapse phenomena may also oc-
during etching. For instance, in replicas of fractured and deep-etched
erial suspensions, most flage1la are seen in close contact with the cell
face (Sleytr and Robards 1977a; Sleytr 1978; Kellenberger and Kistler
9). Furthermore, the heat dissipation during local deformation processes
replication may be sufficient to facilitate elastic recontraction of the
nned structure.
conclusion, with regard to artefacts it can be stated that:
B
Freeze-fracture replication methods are available which have the
potential, even when freeze-fracturing is carried out at 4 K, to reduce
·the possibility of artefact formation to the fracturing and replication
stages.
'Structural deformation during fracturing may be expected at temper-
atures down to 4 K. This clearly shows that the freeze-fracture replica-
. "bl events during fracturing and subseq_
Fg
1 6 42. Schematic illustration of poss1 e. . . ) (A) The globular ,tion technique has a significant methodological limitation.
· · (· ol mer 111 an 1ce matnx · :
tion steps of a two-phase system a p y a· 1 ·x (") or may undergo comp_ ith so1ne structures the degree of deformation is temperature-de-
, ' f the surroun tng ma n
either separate clean1y rom 1 generally deform in a sym ..endent and may be reduced by lowering the fracturing temperature,
. B) F"b ·11 polymer struc ures ·.
tial (c) deformation. ( 1 n ar b ff t d during etching and/or repl ,-,'
(d). Exposed deformed structures ~an e ~ _ec e ) colla se (f) or total disinti
'y chemical fixation procedures (introducing covalent bounds), by
changes, such as elastic recontract1on (hfeahn~l (et~ and ;obards 1982.} ltration with glycerol, or by a combination of all of these.
occur. (Redrawn ram ey

I,!
418 Robard~ and Sleytr Low temperature methods in biological electron microscopy
Freeze--fracture replication
419
(iv) A number of fine structural details see_n on freeze-fracture replicas
TABLE 6.4
must be interpreted as the result of a rp__ulti-event process, involving Cytochemical methods for sp-'ecimcns
. .
examined
plastic deformation, reconstruction, collapse phenomena and thermal
---------~b"_y_'f~re:"e'.ze-fracture replication
load-induced alterations during replica .formation. All these factors ,
Pre--freezing methodr (/<lb.e1l"tng or specific
. alte , ( b
which are involved in producing the final topography of the fracture L Freeze-fracture autor· d' ra ions efore freezing)
a iography
face, may enhance or compensate for each other. 1.1. Freeze-fract ure autoradiography
. of bulk .
(v) Finally, in characterizing artefacts it must be remembered that it is 1.2. Freeze-fracture autorad· , specimens
a two-dimensional projection of a replica_ that is examined in the mic:-o 1.3. Freeze-fracture autorad'.ography ~f monolayers
L· b 11" . iography In combinatio .h h
roscope and that, during shadowing, the heavy metal atoms are not a e '.ng w'.th morphologically detectable m k n wit _t e sectioned replica technique i'I'
Labelling with electron-opa ar ers for rephca techniques
necessarily trapped at their first sites of impact (§6.4. la; also Aber, - que markers for the ( .
Spect 5ic pre-freezing alt ti sec roned replrca technique
. era ons of the molecular . . ;'I,
mann et al. 1972). components into recognisable patterns organisation of specimen (membrane)
4· 1· Thermal phase transitions
4.2. Changes in pH a d/ . . strength
. n or 1on1c ,·;I

6.9 Cytochemical and quantitative methods 4.3. Interactions with lipid perturbers ,·;',
,,1
,'
st-fracture
. methods (lab eII"ing or specific
. alte t" f
Freeze-fracture replication techniques produce inert replicas. This me Thm sectioning of post f t ra ions o fracture faces)
- rac ure labelled s .·
that, with the exception of autoradiography, histochemical methods are Replication of freeze- or .,. . pcc1mens
. en rca 1point-dried f
applicable for determining the chemical nature and function of struc. Thin sectioning of post-fr· t post- racture labelled specimens
. . . ac ure, post-shadow lab 11 d d
. Rep I!Catron of freeze- or critical po· t d . d e e an replicated specimens
features once the continuous replica has been deposited on the fta L" · in - ne post f
ipt~ ~xtraction of fractured frozen specimens - racturc. post-shadow labelled specimens
plane. As summarised in Table 6.4, the cytochemical techniques whi Spec1f1c decoration of fracture~ . .
be combined with freeze-fracture and/or replication and etching m aces tn vacuo with water vapour

can be classified into two major groups (see also Severs 1984). The pr
turing methods involve techniques where labels or specimen constitu
pendent alterations are introduced before freezing the specimen. 'f ers can be done directly after the fracture .
be done using radioactive labels which are then detected by freeze-~ .Sa and b) or after shado . d specimen has been thawed
wmg and subse h ·
autoradiography after specimen fracturing and replication (§6.9.J), t her method, the labelled sp . quent t awmg (§6.9.5c and d)
. . ecunens can b t d. .
sites on the specimen surface may also be labelled with morphologi ctiomng. Information about th h . e s u Ied by replication or
e c em1cal com ·t ·
cernible markers which later, after fracturing, become exposed du: on fracture faces may also b b . pos1 ion of structures
· e o tamed by ·fi 1
ing (§6.9.2). Similarly, specific sites on the specimen surface may b ton methods (§6.9 6) F" 11 . speci ic ow temperature
· · ma Y specific d ·
with electron-opaque markers (e.g. conjugates of antibody with fj ater v~pour in vacuo (§6.9. 7 ~a . ecoration ~f fracture faces
ochem1cal properties of exp~ed yt provide valuable Insights into the
,',
colloidal gold) which, after fracturing, replication and embedding.,
s ructures.
men in plastic resin, can be detected in thin sections cut parallel}
;!1'
lica/tissue interface (§6.9.3). Finally, selective pre-freezing alteraf .Freeze-fracture autoradiography 1!1'

organisation of specimen components (particularly membranes}-' '.!1'

'1 11'
nisable patterns by physicochemical means may provide valu .• 'fracture autoradiography, introduced . ,111

into the chemical nature of such structures (§6.9.4). •" et al. (!976), involves the a r . by Fisher and Branton (1976) 11
11

Post-fracture methods (Table 6.4) include techniques wher\l. ic em I . pp !cation of a mono] f 1


u s1on to replicated fr t f ayer o an autora- ,11

11
structural changes are carried out on the freshly exposed, radiographic principles ar:ct:;es aces and subsequent processing.
11
(§6.9.5). Labelling with morphologically discernable or ef (1977). The tecbniqu ame as those described fully by '11
1
e represents a promismg· · approach for the 1
:,11

,/
I
I'
' 'I
. b.10Iogi·cal electron microscopy
Low temperature me thods in
420 Robards and Sleytr Freeze-fracture replication
421
Shadowing and replication
sectioning before applying the emulsion (Fig. 6.43b). With tissue blocks, the

r~~ ·--·· , _ _
Labelled specimen c Pt/C

problem of the coarse topography of a fracture face making it difficult to

~~ ~~....
obtain uniform contact between the film and the replica is often encoun-

I@ tered. In addition, resolution may be limited by radiation from deeper levels


in the block. This is a particular drawback if labelled membrane fracture

\ /i =c,\ ! ~ : (/ ''' Ch•m,o•i '""' 0"


faces are to be studied but may be partially overcome by using short-range

~
125
radiation such as tritium or I (Schiller et al. 1978, 1979). Replicated frac-
ture faces of bulk specimens can be coated with dry monolayers of autora-
Coath1g (-143 K) : ,mO•ddmo mp<.,uo

I diographic emulsions in vacuo or ex vacuo. The practicability of ex vacuo


with emulsion '
• ' I .
\ a S•o••o"'"' """'' ' 0 """" :Coating was shown by Fisher and Branton (1976) and Rix et al. (1976).
With the Fisher and Branton technique (Fisher and Branton 1976; Fisher
978) specimens are fractured and replicated in a Balzers-unit (§6.3.2) and
bsequently transferred, without protecting the replica from hoar frost for-
j tion, to the surface of a cold pillar and then to a dark room. After
state (-173K)
Exposure in frozen nd develop rosting the replicas in flowing dry cold nitrogen gas, the autoradio-
at atm. pressure a
phic emulsion mounted on a wire loop is immediately applied to the rep-
I
Thawing and pushed against the replica contours using the pressure of a stream
I
Replica cleaning
ooled dry nitrogen. Nuclear emulsion monolayers made from flford L-4
ord; see Appendix 2) can be strengthened and stabilized using Collodion
~ s. Samples are stored for exposure at 193 K ( -80°C) in gas- and light-
containers and then thawed before developing the emulsion and dis-
·ng the specimen remnants adhering to the replica. Fisher and Branton
.
fracture autoradtograp hy of bulk.
. The two basic methods for freeze- d directly with an autoradiograp~i ) reported an attempt to apply emulsions in vacuo but their method
Fig. 6.43. d and replicated specimens are coate d for thin sectioning, after Wht ced inexplicably high background levels.
(a) Fracture . cimens are processe .
(b) Fractured and rephcated spr:diographic emulsion is applied. mpared with the method described above, the ex vacua coating proce-
developed independently by Rix et al. (1976) has the advantage that,
the Leybold EPA 100 freeze-fracture replication unit (§6.3.3c(i)) with
luble aud diffusible compounds and
a nalysis of labelled water-so 1 ·1hout encountering the pro sport device (e.g. Figs. 6.17d and 6.22b in §6.3.3), less condensation
d f n
brane concentrat10ns o md b ·hemical fixation, dehy ra '~ :t.
· f olecu es wi ination of the fracture face occurs during transfer of specimens to
fusion artefacts introduce ~· cgraphy can be applied to o, tat for coating with a monolayer of emulsion. With this method, the
d . g Freeze-fracture autora 10 ical stability of the emulsion-coated replica is increased by adding,
in. lers
bulk specimens or mono ay . ryostat, a layer of gelatin. After exposure in the dark at 253, 223
(-20, -50 or -100°C) and photographic processing of the emul-
6.9.la F reez e-fira cture autoradiography o1 bulk specimens specirnens
5 are transferred back to the vacuum unit. In a vacuum
7
- Pa (5.25 x 10- Torr) and at a specimen temperature of 188 K
d of bulk specimens by
The two basic methods for the s.tuFy 6 43 The fracture. d the monolayer is freeze-dried and, finally, strengthened by evapor-
·n t ted m ig. · ·
autoradiograpby are 1 dus r~ly (in vacuo or ex vacuo) wt.
't rbon layer on top of it.

specimens are. coated ~r:~a) or, alternatively, are first pr; hie and routine method for in vacuo emulsion coating of freeze-
graphic emuls10n (Fig. . ·.specimens has been developed by Schiller et al. (1978, 1979) using
422 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze~fracture replication
423

a special coating device designed for the Leybold Heraeus EPA l 00 freeze-
Transfer under LN Replication
fracture unit (§6.3.3c(i)). With this method (Fig. 6.44) tritium-labelled speci-

'~"'_ ~~:~·;~ - ~ ~
Vacuum
mens are frozen in type I holders (Fig. 6.2; §6.2.2) and subsequently frac- 5
1.33x10 Pa
tured under careful control with a Reichert (Austria) OM P microtome
C Pt/C
under liquid nitrogen (Fig. 6.22a) using the Geymayer (1966) technique
(§6.3.3c(i)). Samples with flat fracture faces are then transferred using the j/
transport device (Fig. 6.22b) under the protection of liquid nitrogen to the fll El
173K
--+1J -
stage of the vacuum unit. They are warmed up to 173 K ( - 100°C) for etch-
ing, shadowing and carbon replication at 1.33 x 10- 5 Pa (1x10- Torr).
1
j
Subsequently, specimens are cooled to 143 K (-130°C) and coated with a in vacuo coating

monolayer of autoradiographic emulsion in a vacuum of about 1.33 x 10-2 Vacuum


2
Pa (1 x l0- 4 Torr) using a specially designed coating device (Fig. 6.44) caf- Transfer
1.33x10 Pa

rying 3.0 mm plastic coating tubes. Before use, the coating tubes mnst b6 J Plastic
into
vacuum
freshly loaded with air-dried monolayers ofilford L-4 emulsion (10 g IlfotcC c,;iating
h;be-----il
L-4; 15 g distilled water; 1.6 g glycerol) on stainless steel loops. To aid adh\t
Emulsion-
sion of the emulsion, 1 µl of 1-butanol or a 1-butanol/2-butanol mix
(I: 1) is applied to each of the monolayers shortly before introducing 143 K__J@ mm
coating device into the vacuum chamber through the air-lock syste
means of the transport device. After applying the emulsion to the repli Autoradiographic exposure
(at 163 Kin dry N I I
specimen fracture face, samples are warmed from 143 K (- l30°C).t<j; atmosphere)
2 .'i i
K ( - 100°C) and kept at this temperature until the butanol is com~(
evaporated and the monolayer of emulsion is in close contact with. . .
j
Development at 20 oc
Jica. Subsequently, specimens are transferred from the freeze-fract
into a cold chamber and kept under atmospheric pressure
j I
,I,

( - 110°C) for exposure. Photographic processing was performed b . F•erng


et al. (1978) in tissue-isotonic solutions of standard photographic in vacuo carbon coating
(Amidol with Hostapon), diluted fixer, and 0.6% sodium chlorid
ing. Before dissolving and separating the tissue from the emulslo
replica in I N KOH, a carbon protection layer can be evaporat
of the specimen, as in the ex vacuo coating procedure described;
j
et al. 1976). Th Replica cleaning
The Schiller et al. (1978, 1979) technique for in vacuo coating•1 b' S e. routine method fo rm · vacuo emulsio .
strated that the autoradiographic background of the film e ,- y chiller et al. (1978). Tritiurn-labelled s n c?attng of freeze-fracture specimens de-
te a d~op of emulsion containing a small p~c1men~ are fractured, etched and repli-
significantly higher than in ex vacuo controls. The authors cl
he specimen is wanned to 173 Kat wh· h p oportton of 1-butanol is applied to the
tochemographic artefacts are observed and that the method n .as a inonolayer on the specimen ~~r:~:p~~ture t~e butanol evaporates leaving
comes potential ex vacuo coating hazards, such as ice conta_ nc pressure and the emulsion <level d . e specimens are exposed at 163 K
replicated fracture face before coating with the autoradiogt~ orated on t o the specimen before th ope at . 293. K ·. A protecting
· layer of carbon
. e specimen is digested aw·
and recrystallisation or thawing of the specimen ice durin · · Schiller et al. 1 _)
978
ay. (Redrawn frOm

cryostat for coating.


424 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-fracture replication
425
6.9.lb Freeze-fracture autoradiography in combination with the sectioned developed a different approach kn
diography (MONOFARG) MONown as monolayer freeze-fracture autora-
replica technique · OFARG1sac b. ·
f reeze-.f racture (Fisher 1975 1976b· . om mat10n of monolayer
Nermut 1984) and electro, . , for review see Nermut 1982a b and
This technique, suggested by Rash and co-workers (1980), can help to over- n microscope t d· ' '
come the limits to resolution known to arise in autoradiography of bulk spe- Bachmann 1972; Fisher l976a 1978 19 au ora wgraphy (Salpeter and
nally conceived (Fisher 1976 )'th" ' 80, Salpeter et al. 1977). As origi-
cimens (§6.9.la) where isotopes located at sub-membrane levels may still . . a is techmque can p "d b
contribute to the exposure of silver grains. As described below (§6.9.3 and and quant1tal!ve information b h rov1 e oth qualitative
. a out t e transme b
§6.9.5a), the sectioned replica technique allows an unambiguous correlation ut10n of radioisotopic molecules(~ . m rnne or translayer distri-
of freeze-fracture and thin section images within a single electron micro;---: The procedure is particularly suit::1:e;:;~:e: Fisher _1982a).
graph. As illustrated in Fig. 6.43b, the sectioned replica technique, in combic'' oplasts, membrane vesicle ( tudy of mtact cel!s and pro-
. s e.g. erythrocyte gh t d .
nation with autoradiography, involves the following steps: the radioactive!)\; odel bilayers or non-vesicular b os s an hposomes) and
sis of the MONOFARG t hmem rane_ fragments (Fisher 1982c). A syn-
labelled tissue is first fractured and replicated by standard methods bnt, i ec n1que is given in p· 6 .
stead of digesting the tissue, the replica-tissue sandwich is post-fixed ill: the P. rocedure involves the .fi . . ig. .45. The mitial step
. spec1 IC bmdmg of th I b
solution of 30% buffered glycerol containing 1% osmium tetroxide; support. Smee most biomemb e a elled specimen to
ranes possess a n t ·
sandwich is then post-stained, embedded in a plastic and sectioned pa y can be attached to positively h d . e negative surface charge ,,,
. . c arge sohd supp t ( '
to the replica/tissue interface. Thin sections from the replica/tissue int eIectrostal!c mteractions (see §6 2 2) Al or s e.g. glass or mica)
ly simple, good binding (with t.h. . dtho_ugh these techniques are rela-
are then coated on both sides with !lford L-4 emulsion. After exposnr ' · e pro ucl!on of d
development by standard methods (see Williams 1977), specimens are be achieved.
when the cells a f c
re ree1romanyc t . .
ense monolayers) can
ined stereoscopically in the transmission electron microscope to det us protem or cell debris wh. h on ammat10n with exoge-
. Nennut (1982b) recomm ic d may~ompete for positive sites on the sup-
which of the two emulsion layers has been developed. Because aut<\5
nic buffer just before use ends twhas mg the cells three to five times with
graphic grains are produced and recorded on both sides of the sectii . an en resuspend· th .
least one-half of the low energy emission is recorded without having ntrat1on. It is useful to ce t ·r ing em at an Increased
n r1 uge some sp ·
the electron-opaque platinum layer. Rash et al. (1980) demonstr For n1.aking preparations of 1· t t ecnnens down onto a sup-
. n ac membrane (F" 6 .
2.5 nm diameter 1251-a-bungarotoxin molecules are separated fro sis and the unrolling of b . . s ig. .45a) mvolving
. mem ranes it is de · bl .
re w1delv separated. Slfa e to bmd cells when
class of intramembrane particles present in junctional fold memh
so it can be concluded that half-membrane resolution is attain an alternative to the attachment of cell .
tioned replica autoradiography. Despite these advantages, this.I·': ely charged surface, covalent bindi s or isolated membranes to a
o glutaraldehyde-coated su (ng of ceHsurface (membrane) pro-
not suitable for the analysis of the distribution oflabelled water'i(
diffusible compounds, since extraction and displacement artefa
ter thawing tbe specimen.
hi 1979· B.. hi d ..
rts coated with a specific r d p:
_pports see Robmson et al. 1971· B .. hi
, ac an Buechi 1981 ·A r , uec
and Hughes 1981) or binding
h_es 1981) is also possible. igan sue as concanavalin A (see Aplin
6.9.lc Freeze-fracture autoradiography of monolayers bmding the radioactively labelled cells
pport, the MONOFARG th d . or membrane preparations
· · me o involves tw
There are a number of problems involved in freeze-fractu · . trung intact membran - o separate processes.
· e preparat10ns (F" 6 45 )
phy of membranes in suspension or in bulk specimens, inc moval of excess unbound b 1g. . a ' cells are first
. mem ranes can b d ·
ing from the smallness of exposed membrane areas, the_"· --cS:oIution from a syringe fitted with e ~ne using a stream
ermut l 982b) Aft f . a hypodenmc needle (Lang et
phy of fracture faces, the disturbances by isotopes un · er reez1ng spe ·
d freeze-dried S b ' c1mens are transferred to a coat-
fracture faces, and the necessity for long exposure of the
n of split membra:essequent steps are identical to those for the
men temperatures close to 173 K (- 1OO"C). Conseque
426 Robards and Sleytr Low temperature m ell'o d~· in biological electron microscopy Freeze~fracture replication
427

MONOFARG
b for lO s, rinsed with distilled water and dried with clean nitrogen gas. Glass
a
coverslips are cleaned as for the poly-L-lysine coating method (§6.2.2d).
Bind
Freezing of the sandwich is done either in propane or in Freon (§2.4). After
fracturing, which can be done using either single or complementary meth-
.·-'a~··===·····=;;;a~··=··•.···
=
~--.·.
Polylysine-coated support
ods, the specimen is freeze-dried. Both intact and split membrane prepara-
"ons are then paired and processed together. After shadowing and warming
Lyse Sandwich mounting to room temperature, specimens are removed from the vacuum and
F<e°'e'=•.~"""'·
...'?'"'<a~·.
·3""<0'5="
. :::"'·
..···•·
ansferred to a dark room. A Parlodion-supported Ilford L-4 photographic
ulsion is applied to the replica at room temperature and the samples are
ared together with non-radioactive controls in light-tight containers in a
atmoshere at + 5"C. After exposure, photographic processing is carried
t using standard methods for electron microscope autoradiography (Sal-
Freeze-dry r and Bachmann 1972; Williams 1977; Salpeter et al. 1977, 1978). The
ica, with the emulsion, is removed from the glass support by floating
,,
·::?'$ r.s to dilute hydrofluoric acid (HF:water ratio of 1:1). It is then rinsed in I
c ?"·
"lled water, mounted and examined.
Shadow "sher (l982b) has also studied the quantitative aspects ofMONOFARG

,....< - ....... (-
7 t
, as a model system, human erythrocyte membranes treated with radia-
ted and fluoresceinated concanavalin A (' 251-FITC-Con-A). The
Apply emulsion ' round was regularly < 2.5 x 10 4 grains µm- 2 day- 1 . The highest over-
tciency was between 25% and 45% and grain density and efficiency
oooao
,....e--.-, (.-
0
0000000 0 Ot $ 7 t
ependent on radiation dose for 1251 and 019 emulsion (Eastman Ko-
o.; see Appendix 2) development. The corrected grain densities were
Expose Develop
y proportional to 1251 concentration. Analysis of autoradiographs
that silver grain densities over split extracellular half membranes
Replica cleaning
:45b) were identical to those over intact unsplit single membranes
rr......t.:~ .....; • ;Jl ~ ~ ""
Fig. 6.45.
.. .
.:m

h (MONOFARG)tech
The monolayer freeze-fracture ~uto_rat~~~g::m~ranes (a) after ly ,
f
:·;.,,
Sa). These results demonstrate that 1251-FITC-Con-A remains
ly attached to the extracellular half of the membrane following
cturing and can thus be used as a quantitative marker for the pro-
. be used to examine tn a f extracellular split membrane halves. As well as this excellent
1978). This technique can f . t In both cases, spedmen_s
d b nes (b) after freeze- rac ure.
ing; or fracture mcm ra . . . d , room temperature. A_ brane resolution of membrane constituents, a unique feature of
d d the emulston JS app 1te a 1 a·----./
supports are shad owe an . d f om the support by o~. RG lies in its potential to determine the position of the radio-
r d emulsion are remove r "·-
development the rep ica an . .d (R-drawn from Fisher 1978.) · . .·•.·.·•.(: the plane of the split membrane. Preliminary studies indicate that
hydrofluoric ac1 · e ---_)
e resolution of MONOFARG for 125 [ is about 150 nm (Fisher
us, for membranes, MONOFARG clearly overcomes the resolu-
Half-membrane fracture faces (F"tg. 6 .45b) are obtained.;
beet.• .lhatapply to bulk specimens where isotopes at level~ in the speci-
. h' (O 2 o 5 mm) copper s ..• than the membranes may contribute to the formation of silver
the cell monolayer with a t m . - . h t. treated wi.. . . ·
. Before use, the copper s ee ts
glass covershp.
428 Rohard~ and Sleytr Low temperature methods in biological electron microscopy Freeze:fracture replication 429

6.9.2 Labelling with morphologically detectable markers for replica gated phytohaemagglutinin, Tillack et al (1972
fie erythrocyte membrane antige d . ) were able to analyse speci-
techniques tion of surface markers with p tnsl an to correlate the pattern of distribu-
. . . ar ic es seen on the fracture 1
This cytochemical technique involves labelling specimens with marker mole- With per1od1c structures the to o ra hie . pane.
markers may be considerably ,·m p g dp . al resolut10n of specific surface
cules before freezing. The markers must have a recognisable structure and . prove since optical .
·must be large enough to be detectable in high resolution replicas of deep: techmques can be used on the final ima e F or computer filtenng
tungsten or platinum/carbon shadowe: o~~xample,. on.freeze-dried and
T4
sites can be localised by labellin them w· : y~eads, mdmdual antigenic
etched (freeze-dried) or critical point-dried specimens.
The most commonly used techniques involve specific antibodies or agglu- 1
et al. 1978). Unlabelled antibodig h ' Fab fragments alone (Kistler
tinins coupled to marker molecules such as ferritin, hemocyanin or colloidal . es ave a 1so been obs d c
bac!enal flagella (Munn et al 1980) F erve on •ast-frozen
cationic or polyanionic ferri~in ca .b ma dy, charged particles such as poly-
metals (Horisberger and Rossel 1977; Perkins and Koehler 1978; Horis; 11
berger 1979; Roth 1982). Among the most commonly used immunochemi; · . n e use for labellin h d .
structures on cell and membra .r g c arge specimen
cal markers for demonstrating cell surface antigens in electron microsco.11¥· ne sur.aces Polycatio · ( · ·
are IgG-ferritin conjugates (Morgan 1972). This method has success£
js commercially available (Miles Israel· A .
me cat10msed) ferritin
duced by attaching aliphatic amines ;o tr;::,end1x 2) but may also be pro-
been used to label antigens on different membrane surfaces exposed by
ing carbodiimides as condens. carboxyhc groups of ferritin
etching such as the surface of erythrocytes (Pinto da Silva et al. 1971; H . mg agents (Danon et al 1972) E .
th bacterial cells (Fig 6 46) h h · · xpenments
and Biichi 1973; Elgsaeter et al. 1976), lymphocytes (Karnovsky et al. · · ave s own that th k
negatively charged sites even du . e mar er molecules stick
Unanue et al. 1973), or chloroplasts (Miller and Staehelin 1976). Th.e:
. 78). Similarly other suitable m rkmg prollonged etching (Sleytr and Friers
of the IgG-ferritin conjugates (approximately 20 nm) allows a positive! . ' ar ermoecules(eg ha .
tification of membrane surface components which are more than 2~ !Sh peroxidase) can be specifically charged for l~b~ff emocyanm, horse-
logy with the use of polycationic macromolec mg expenments. By
apart, provided that no genuine membrane structures resemble the
al markers should be usable as robe ules, polyamomc topogra-
in shape and dimensions. The question of how small a topogr . ts (Rennke et al. 1978). p s for pos1t1vely charged cell consti-
marker can be and still be consistently recognisable on freeze-etche.
brane faces has been studied by Carter and Staehelin (1979). Usi
erythrocyte ghosts as a model system, they compared the labelling
of !gG molecules, their antigen-binding fragments, !gG-horseradis
dase conjugates and also varying amounts of associated diamino-,
precipitates. The authors concluded that an organic surface
freeze-etched membranes must have a diameter of more than l
to be seen consistently over extended areas and against the
granularity of red blood cell ghosts. Markers with a distinct.
inorganic crystals) coupled to Fab' fragments of IgG can give:.
resolution.
Specific labelling of cell surface glycoproteins may be done>
tor-specific proteins of non-immune origin, such as phytoha~
protectins and lectins (Gold and Balding 1975). Lectins, a
magglutinins derived from plants, have specific affinity for
·, Cell surface of the bacterium Clostridiu h
or oligosaccharide groups and are used for labelling the ca th patches of hexa I m t ermohydrosulfuricun1 (revealed by deep
gona arrays of glyco r t · Th. . .
dues of cell membrane constituents (for references see Cook! nly bound to areas where th . p o e1ns. e polycat1on1c ferritin molecules
e negatrvely charged fd 1
and Cohn 1976; Sutherland 1977; Roth 1978, 1982). Usin_g; Oprotcin nlolecules. S S-I· ~ep og ycan layer is not masked
1

, ayer._ (Reproduced, with permission, from Sleytr a d


Fnersl978.) n
430 Robards and Sleytr Low temperature n1ethods in biological electron microscopy
Freeze-ji·acture replication
431

6.9.3 Labelling with electron-opaque markers for the sectioned replica 'vvith the distribution of specific marker molecules on membrane surface
technique exposed by deep etching. s
Considering the variety of immunological or affinity markers that are
The initial step involves labelling the specimen with an electron-opaque available (see §6 ..9.2), it is obvious that these simple techniques have great
marker. Both irnrnunoferritin and colloidal gold have been used (Rash et potential in studies on membrane architecture and, particularly, for trans-
al. 1982). Labelled specimens are then frozen, fractured and replicated using membrane features.
standard freeze-fracture methods. The subsequent preparations steps follow Fi.nally, membrane components, particularly membrane lipids, may be
the schedule used for the sectioned replica technique described in §6.9.lb. specifically complexed into characteristic structures detectable in surface
In thin sections cut parallel to the replica tissue interface, the distribution fractures (protrnberances, ripples, pits, etc.). To reduce the redistribution of
of the electron-opaque markers can be correlated with the superimposed t?ese ~omplexed membrane structures, probes are often used in combina-
freeze-fracture and thin section images. Rash et al. (1980, 1982) have shown t10n with aldehyde fixatives.
that, particularly with the help of stereoscopic images, it is possible to deter" As a probe. to reveal fl-hydroxy sterols, the polyene antibiotic filipin
mine which half of the membrane retains the electron-opaque markers afte! (Ttllack and Kmsky 1973; Verkleij et al. 1973; De Kruijf and Demel 1974·
fracturing. This technique, together with post-fracture labelling methods Kitajima et al. 1976; Elias et aL 1979; Montesano et aL 1979; Robinson and
(§6.9.5), has proved to be a valuable tool for analysing individual intranaeni& Kamovsky 1.980; Garcia-Segura et aL 1982) or saponins (e.g. digitonin and
brane particles. omatm) (Elms et aL 1978, 1979; Robenek et al. 1982; Severs and Simons
983) have been used .(see Severs 1984 for a review of this area) .
6.9 .4 Specific pre-freezing alterations of the molecular organisation of- . The, anltb10tJc, bacitracin (which forms a complex with C 55 prenol pyro-
membranes hospnate) has been shown to induce the formation of rod-like elevations
the plasma membranes ofprocaryotic cells (Sleytr et al. 1976).
These methods involve a variety of pre-freezing treatments of spec;'
with the objective of inducing a specific change in the organisation of
9.5 Post,fracture labelling procedures
brane components into recognisable patterns.
The simplest procedures involve temperature changes (cooling),
through phase transition phenomena, cause phase separations in ti:·-- e cytochemical labelling techniques for membrane or cytoplasmic consti-
brane lipids (transition from the liquid crystalline to the gel state) a11> ts exposed by frneze-fracturing follow two principles (Fig. 6.47). One
gation of membrane-intercalated particles (Verkleij and Ververgae< roach mvolves direct labelling of fractured and thawed specimens
1978; Maul 1979). Although such segregation is a common undes1 .Sa, b) whereas, with the other technique, fracture faces are first stabi-
fact caused by inadequate cryofixation (see §6.8.3), it can also pro by the deposition of heavy metal shadows before the specimens are
able information about the physicochemical properties of membr ed for labelling (§6.9.Sc, d).
As well as temperature shifts, a variety of other treatments (e~
tional forces, hydrostatic pressure, changes in ionic strengt~, --~:'.<'.
Thin sectioning ofpost-fracture labelled specimens
composition, infiltration with cryoprotectants, chemical fixat~o-~p
2.12 in §2.3.2) can induce particle aggregation. These changesm
Jar organisation of membrane constituents seen on fracture f ethod is illustrated. schematically in Fig. 6.47a and involves the grind-
be used for studying the topography of membrane surfaces ( ..c• f:azen samples while immersed in liquid nitrogen, followed by thaw-
et al. 1971; Shotton et al. 1978; Maurer and Miihlethaler 19!1~ . chemical labellmg of fracture faces and processing for thin section-
ing the pattern of aggregat10n of membrane partJcles seen o ~o da Silva et al. 1981b,c; Pinto da Silva 1982; Torrisi and Pinto da
4). Tissues or cell suspensions that have been fixed in glutaralde-
432 Robards and Sleytr Low temperature methods in biological electron 1nicroscopy Freeze-fracture replication
433

Post-fracture labelling methods


hyde (e.g. 1% glutaraldehyde for 15-20 min at 4°C) before embedding in
a protein matrix (e.g. 30% bovine serum albumin) cross-linked with glutar-
Cryoflxation of native or chemical fixed specimens
aldehyde (I% for 30 min at 25°C) are often used. The tissues and gels are
then sliced into pieces approximately I x 2 x 2 mm, gradually impregnated a b
with 30% glycerol and frozen by immersion in partially solidified Freon 22
Thawing In fixative
cooled with liquid nitrogen. The frozen pieces are then transferred to a
polished glass container (e.g. the base of a tissue homogeniser) filled with
liquid nitrogen and immersed in a slush of liquid nitrogen and solid carbon Post fixation
dioxide. Drops of a 30% glycerol:!% glutaraldehyde solution in 310 mOsm
Dehydration and
Dehydration and
phosphate buffer, pH 7.5, frozen in liquid nitrogen, are added in approxi- embedding
in resin critical point-
mately equal amounts to the gels or tissues. After the pieces have all sedi- drying

mented to the bottom of the container, they are repeatedly fractured with Pt shadowing
and C-coatrng
a glass pestle at 77 K ( -196°C) until pulverised into fine fragments. Tlie ·
glass container is then removed from the liquid nitrogen, placed at ambient Replica cleaning

temperature until the volume of the liquid nitrogen is reduced to approXi'{ lnlormatlon: thin section Information: replica of
image with electron labelled fracture face
mately one tenth, and 2-3 ml of the glycerol/glutaraldehyde buffer is added' opaque marker {marker electron opaque
and/ or shadowed)
in liquid form. The glass container is immersed briefly in water at 30°C Pt shadowing
on thawing of the glycerol, is transferred to an ice bucket for 15 min. {no carbon)

fragments of the gels and tissues are then gradually deglycerinated by dr c Thawing
d
wise addition of 1.0 mM glycyl-glycine in 310 mOsm phosphate buffer· Labelling

pH 7.5 and are then washed twice in 310 mOsm phosphate buffer, als\'!; Chemical fixation
Chemical fixation
pH 7.5. Centrifugation and pelleting of the gel or tissues pieces is not n ·
Dehydration
Freeze- or
ary because they settle rapidly. critical point-
in resin drying
Conventional cytochemical labelling techniques can then be used t
the membrane faces exposed during freeze-fracturing. Later, the sam Thin section Carbon coating
parallel to
fixed in glutaraldehyde and osmium tetroxide using standard pr replica
Replica cleaning

(Glauert 1974), dehydrated and embedded for thin sectioning. Bet lnlormaiion: superimposed
Image of thin section Information: replica of
labelled fracture face
trimming of the blocks, thick sections are used to localise suitable and replica of labelled
fraclt1re face (marker electron opaque

the specimen for fine sectioning. Most cytochernical techniques -r· not shadowed)

electron-opaque markers can be used and good results have been:,


in labe11ing anionic sites with cationised ferritin and concanaval\'
wheat germ receptor sites using ferritin conjugates or colloidal gol<l.
Q Label " ~;:~~:::;~~:•In
labels (Pinto da Silva et al. 1981 b,c; Pinto da Silva 1982). 4 7. Post-fracture labelling procedures (a) Th"
Frozen samples are grou d. 1· .d ·.
. .
in sectioning of post-fracture labelled spe-
In summary, although a considerable rearrangement ofmemb( n in iqu1 nitrogen followed b th · ·
belling and processing fo th" . . Y awing 1n fixative, cytoche-
tuents can occur during thawing and labelling (fractured merrt r in sectioning (b) Replica( f ..
labelled specinlens F . · ion o cnhcai point-dried, post-
-. · rozen specimens are fnictured th d d
reveal an image of unit membrane-like segments), the technique< heal point-dried The f t ,. ' awe an labelled as in (a) and
· rac ure 1aces are replicated and th .
rect, high resolution, chemical and immunological charac .samples are fractured and then th. I
a in ayer of platinum is
e specimen digested away. (c)
t d
plasma and intracellular membranes. he specimen is then thawed I b II d . d e~apora e onto the fracture
· ' a e e an processed for thi 1· ·
e as m (b) except that a thin l f . . n sec iorung. (d) The same
aye_r o platinum is evaporated onto the fractured surface
pnor to thawing.
434 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-fracture replication
435

6.9.5b Replicating critical point-dried, post-fracture labelled specimens membrane particles, exposed by the fractnrin . .
chenucal mformation to per ·1 b h . g process, retam sufficient bio-
. m1 ot 1mmuno 8 ·fi d
This technique (Fig. 6.47b) was developed by Pinto da Silva and Torrisi c1fic labelling, despite initial formaldeh - .pec1 ic an neurotoxin-spe-
(1982) and resembles the previous method except that the frozen tissues and are fIXed in I% osmium tetroxide I 0 yde fixat10n. After labelling, samples
cells within gels are not fragmented in a cold tissue homogeniser. Instead, aqueous uranyl acetate (Silva et ~i'. l~~i' nnsed and post-fixed with 0.5%
slices of gels and tissues are immersed in a Petri dish filled with liquid nitro- stabhsed, either by critical point-dr in ). The labelled surfaces are now
gen and fractured with a liquid nitrogen-cooled scalpel. After thawing and mens are then rotary carbon coated ;o g or by freeze-drymg. Dried speci-
labelling as described above (§6.9.5a), the tissue and gel fragments are fixed rehydration and specimen dig t' cprotect the labels dunng subsequent
es wn. arbon films of 15 20
in buffered glutaraldehyde/osmium, dehydrated in ethanol and critical found to be sufficient to stab·!· f . - nm have been Ii I I
. . I 1se reeze-dr1ed sp · H
point-dried. After drying, the specimens are orientated on a specimen car- the cnt1cal point-dried specim h ec1mens. owever, because
ment during rehydration ·1 . ens ave a strong tendency to swell and frag-
rier with the fracture faces pointing upwards, shadowed with heavy metal , I IS necessary to mcreas th th. k
(2.0 nm Pt/C) and subsequently replicated with carbon. The replicas are bon stabilising coat to abo t 50-80 . e e 1c ness of the car- i' I

less, both techniques allowu th nm, which reduces resolution. Neverthe- 1i ! :


separated and cleaned by digesting the specimen in a 5.0% solution of . e recovery of Jarg · t . . 'I
sodium hypochlorite, rinsed in water and collected on grids. The fine struc- underlymg tissue remnants a d .th I . e m act replicas without , I
n w1 c early v1sbl I
ture of the membrane fracture faces revealed by this technique differs from directly above the individu I b . I e e ectron-opaque labels I I
C . a mem rane-mtercalated particles 'I
that of conventional freeze-fracture replicas. Nevertheless, the electron-opa~ ompanng all the post-fracture labellin t h . . . I

que markers (e.g. colloidal gold) can clearly be distinguished from the r011gb < ata currently available that th h d g ec mques, it appears from the
e post-s a ow lab Jr h ·
membrane background structure. By analogy with the interpretation §!ability to the fracture face a d th . . . e mg tee mques give more
embrane components durin ~h~ . us, 1nh~b1t exc~ssive rearrangement of
post-fractured sectioned material, it is necessary to take into account
er, the shadowing step masgk wmg.dand mcubat10n with markers. How-
fact that, on thawing in an aqueous environment, significant rearrange . s a cons1 erable pr f f
of membrane components can occur. In conclusion, post-fracture label·- ndmg sites for the label Th' . . opor 10n o the potential
. is is a partlcular pr0 bl .f h
techniques allow structural and cytochemical dissection of membrane wing is not favourable. em 1 t e angle of sha- 1!'·,'
i'1
face sites and, in particular, the identification of those sites associated 1:.. 1

transmembrane proteins (Pinto da Silva et al. 198la; Pinto da Silva 1982 Thin t' · if 11.',,
,,
speci~::;onzng o post-fracture, post-shadow, labelled and replicated 1·/'
1,,.!i'

6.9.5c Replicating freeze- or critical point-dried, post-fracture, post-sha ,,


',,','11'·
,,,d1'1
labelled specimens lead of critical point- or freeze-dr in
be post-fixed, stained and embed~eli:ost~shadow. labelled specimens
This post-fracture labelling technique, together with the one that ; Rash et al. 1982) Th' . resm for thm sectioning (Rash
· m sect10ns are cut p 11 1
(§6.9.5d), was devised by Rash and collaborators (Rash et al. 1978;· face and are best examined as . ara e to the replica/tissue
method permits the s1·m It stereo-pmrs. The sectioned, labelled rep-
1982). They basically resemble the techniques described in the previcf · u aneous 1magmg of f f
lion except that the labelling reagents must penetrate a thin but disc subadjacent tissue th1'n t. reeze- racture replicas
sec 10ns and thei tt h d 1 '
ous layer of platinum superimposed on the molecules of inte.re Fig. 6.47c) It can b d r a ac e e ectron-opaque la-
. · e emonstrated that alth h h
6.47d). The methods follow the standard freeze-fracture replicatio braneous particles can be labell d .' oug t e shadowed intra-
binding to th .I e specifically, the relative amount of
Jes until heavy metal (Pt/C) has been evaporated. The specim•• .• ese partlc es exposed b b
thawed and labelled by one of the standard methods involving fe~ ally reduced from th t b y mem rane fracturing is sub-
a o served on sub-adj. a t f
body complexes or colloidal gold-affinity ligands. In a study of P~~: (Rash et al 1982) D . . cen non- ractured mem-
thod can b;expect~d ;sph1te such mtrinsic difficulties, this cytochemi-
membrane particles corresponding to the acetyl-choline receptQ , o ave a w1de application.
Rash and colleagues have shown that the unshadowed side of
436 Robards and Sleytr Low temperature met hod s m
. biological electron microscopy Freeze-ji·acture replication
437

6.9.6 Lipid extraction from fracture faces generally not a problem since diethyl-ether is a very poor solvent for ice at
this temperature; see §7.2), the specimens are transferred with cooled forceps
. 1 ( 1971 ) that lipids may be extracted under the protection of dry nitrogen gas back to liquid nitrogen. Any rem-
It has been shown by Richter and S eytlr t peratures before replication.
t: s at very ow em nants of the extraction liquid solidify on the specimen holder (the freezing
specifically from fracture ace les (Rhodotorula, Candida) and oil droplets
yeast species with hp1d vacuo . ere used as a model system. point of diethyl-ether is 157 K; - l l6°C) and are removed from the holder
l/ l t' /water matnx w . base with a cold dissection needle before inserting the holders into the
suspended in a glycero ge a m ted and frozen in holders smt-
d ·mens are best moun . . transfer device. The subsequent replication is done by any method which
For this proce ure spec! . (F'g 6 2) Freeze-fracturmg is
. 1 mentary rep1teas 1 . . . allows controlled etching and replication of fracture faces obtained under
able for prepanng
. . comp
. e 'th two palfs . o f cooled tweezers applying ten-. liquid nitrogen (§6.3.3c(i)). Etching at approximately 163 K ( -100°C) is im-
done under hqmd mtrogen WI h t ansferred in a small con tamer
f d pecimens are t en r . portant to remove, by sublimation, the solidified cliethyl-ether from the frac-
sile stress. The racture s . f Ii uid nitrogen, to a receptacle filled with
0 ture face. With this method, controlled extraction oflipids below the recrys-
(< 2 ml), under the protectwn. n ; uids (see § .2), maintained at 195 K
1 7 tallisation point of the specimen and a combination of morphological exa-
diethyl-ether or other extract10 qh (F' 6 48) or at lower temperatures
· / thanol bat 1g. · mination of fracture faces and chemical analysis of the extracted lipids by
( - 78"C) in a dry ice me . . cooled cold box. During extraction, gas chromatography is possible.
(173 K; -100."C) using a hqmd m~~o;e~~ hours or up to several days, the..
which may mvolve treatment f. h rd to avoid condensation of water~
diethyl-ether vessel is closed wit \ i_d traction (freeze-substitution. Specific decoration offracture faces in vacuo with water vapour
vapour from the atmosphere. After 1p1 ex
rom the analysis of condensation artefacts on lipid/water and glycerol/
ater 1nixtures, it was realised that residual gases in the vacuum chamber
condense on fracture faces in a non-random distribution (Deamer et al.
O; Staehelin and Bertaud 1971). As in the condensation of shadowing ',I,
I,
terial (§6.4.la), impinging atoms or gas molecules can migrate on a frac-
e face before being immobilised. By this lateral surface diffusion, stable
.1,,
lei at sites with high binding energy will be formed and eventually lead I

discrete structures. Experiments by Gross and Moor (1978) and Gross :.',.1,

l. (1978b) have shown that the paracrystalline regions on E faces of yeast


malernma exposed by fracturing at 77 K (-196°C) in a vacuum of
7 9
x J0- Pa (I x J0- Torr) (§6.3.3b(ii)) can be decorated specifically with
r vapour. In their experiments, pure water vapour generated by heating
Diethyl ether
er sulphate-pentahydrate was stored in a vessel and released through
admittance valve into the ultra-high vacuum system. The pattern of
condensation, seen as small cubic ice crystals, depended on the inci-
Fracturing Extraction flux of water molecules let into the system and the introduction of oth-
ultaneously impinging (nitrogen) molecules. The fact that water
. . of li ids from fracture f.ac es ·at low.·t ,·.
F. 6 48. Method for specific extraction p f d nder the protectiog_._.o_.,·. condensation on freshly exposed specimen fracture faces is highly
ig.
· . 'd ·t
Specimens fractured under hqu1 ~1 roge
n are trans erre , u
1 d extraction liquid (e.g. dietb? 1.· ate-specific was later confirmed by condensation experiments on che-
r1!
gen to a small container filled with a dpcb~-~kootoc liquid nitroe:en and processed..·~ well-defined specimens. Using multi-lamellar cardiolipin liposomes II
' ·
the extraction, specimens are
transferre ac
. . .
~ d d
f specimens fracture un.·./: J1
t" n and exam1natton o ) o/ ic acid crystals which, after fracturing, reveal hydrophobic faces 11:
dules suitable for the prepara to . df Richter and Sleytr 1971. .·f_.·
gen (see §6.3.3c). (Denve rom rophilic cross-fractured steps, as well as multi-lamellar octatriacon- I
\,I
11
I;
1,
438 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-fracfure replication
439
tane crystals which, after fracturing, expose hydrophobic faces and geome- nv =number of particles per unit volume
trically similar but hydrophobic steps, it was demonstrated that hydrophilic NA =numberofmoleculesperm l (A
surface features in hydrophobic surroundings can be specifically decorated midt's number=6.02 x 1023~ e vogadro's constant or Losch-
by particle-like ice crystals (Walzth6ny et al. 1981). This clearly shows that,
under ultra-high vacuum conditions, the controlled deposition of suitable Since relatively large errors can be expected as . .
condensing gases (specific decoration) may be a valuable tool for labelling accurate subhmation rates (§6 5 l) . a result of msuffic1ently
regions on fracture faces according to their physicochemical properties. It pie and the other contai . . . ' t':o solutions, one containing the sam-
mng a protem of known l ul .
also became obvious from these experiments that condensing water vapour dard) are spray-frozen (§2 4 le) f mo ec ar weight (stan-
. · , one a ter the oth · th
may lead to the formation of small ice crystals which resemble particles seen seL As standards molecules such .c . . er In e same propane ves-
. bl ' as 'err1tin or bovi
on membrane fracture faces. smta e (lunger and Bachmann 1977) Af ne serum albumin are
the frozen mixture the re Ii h . . ter freeze-fractunng and etching
' p cas s ow Side-by side r
6.9.8 Molecular and particle weight determination the sample and the standa d b th - spray •rozen droplets of
r ' o etched under th ..
unknown molecular or parti 1 . h e same cond1t10ns. The
c e we1g t of the sample is then:
A method for determining weights of macromolecules or colloid particle$
by particle counting on replicas of spray-frozen, fractured and etched sam"
ples was developed by Bachmann et al. (1974). The technique was tested (6.5)
the authors for particle molecular weights ranging from 104 to 108 daltons:,
Theoretically, the only requirements for accurate measurement are: ·-where
M, =molecular or particle weight of the sam l
(i) that the random distribution of particles or macromolecules in M" =molecular weight of standard pe
sample is preserved during freezing; nx : number of part~cles per unit area of sample
(ii) that the particles are suspended in a volatile solvent; number of particles per unit area of standard
(iii) that the particles do not sublime during controlled etching ofthef =concentration of the sample
ture face and remain on the frozen surface; =concentration of the standard
(iv) that no lateral movement of the particles occurs during etching;
(v) that the particles have a detectable size. optimal concentrations increase .
es have been recommend d b ;1t~e particle size and the following
The particle distribution seen on the replica will then represent ll'.. = 12,500) 0.001 mg/ml,e fer~tina~M;n~ et al. (1974): cytochrome c
dimensional projection of their original three-dimensional distribulf nm) 2 mg/ml. Under the cond1't1· d -. 750,000) 0.2 mg/ml, Latex
ons escnbed ab h
the solvent layer which has been removed by controlled etching(§ fr met hod depends mainly on th ove, t e accuracy of
e accuracy of th dry ·
observed number of particles per area (particle density) is then reji er and Bachmann 1977). e weight protein assay
tive for the sample and the molecular weight can be derived from
tion: Freeze-fracturing and freeze-fracture r 1· . .
negative staining ep zcatzon in combination with

live. staining h as b een used 1n


. com bin at· ·
where nmarily to facilitate th I 'fi . Ion with freeze-fracture techni-
. e c ass1 1cation of · t b
M =molecular weight ng to their size and sha e 6 8 In ramem raneous particles

S =dry weight concentration of solution of[


s of initial replication 21.
(§ t . . The method of Fujikawa (1979)
rac ure membrane faces with a uniform,
440 Robards and Sleytr Low temperature me til Ods in biological electron niicroscopy
Freeze-fracture replication
441

tment of the cleaned replica with a negative-


20 nm, carbon layer and trea . . . the replicas mounted on a shadowed object, the shadow itself causes an increase in size (§6.4. Ia). To
. 1f Before negative sta1n1ng, '
staining so _u IO~. . en side to the surface, are made wettable by
minimise such deviations, the diameters of uni-directionally shadowed par-
uncoated gnds with the. spec1ml . (2% hosphotungstic acid adjusted to ticles are measured at right angles to the shadowing direction (Gulik et al.
glow discharge. The stammg so utI~n ot:P f the replica and the excess 1982; Schmidt and Buchheim 1982). Willison and Moir (1983) suggested the
pH 6.9 with KOH) is appli.ed to t ~ sur a~~ ~ piece of filter paper. After use of the classical technique for particle size analysis involving a 180° inter-
liquid is removed by touchmgfthe e ge :;ion in the transmission electron shadow rotation of the specimen (bi-directional shadowing according to
. h ·men is ready or exam1n Williams and Wyckoff 1944; Kahler and Lloyd 1950). A method of recon-
drying, t e speci d" tribution of intramembraneous
. The method shows a c1ear is . structing the surface topography of isolated heavy metal-shadowed particles I·' I
m1croscope. . i hbourin structures and makes 1t , ,

particles without mterference from neg 1 ment~ by means of density dif-


'
by computer techniques has been described by Smith and Kistler (1977) and , ,

d . . . ate between contiguous e e Smith and Ivanov (1980) (see also Guckenberger 1984). Finally, for an auto-
easy to 1scnmm t. s and the flat lipid mono-
h . t membraneous par tc1e mated system for the quantitative analysis of membrane-intercalated parti-
ferences between t e m ra. d h t densit differences between individual
layer. It has even been claime t a y cles on membrane fracture faces, attempts have been made to create a refer-
particles can be seen. . . . of freeze-fractured membranes has been ence library of computer-generated synthetic electron-optical images of
A method for negat1veWst·a111mng (1976) Nermut (1977) and Nermut et al. morphologically defined particles prepared using a variety of different sha-
t d by Nennut and I rnms , f . dowing conditions (Weinstein et al. 1979).
repor e . ndwich between two sheets o mica, or __ ,_,
( 1978). Specimens are frozen as a sa . I §6 2 2)' '.·.' For the accurate determination of particle size and the dimensions of
. . . h 0 ther materials (e g. filter paper, see a so ... ,, membrane fracture face edges and cross-fractured bilayer membrane edges,
mica in combmat10n wit d r 'd ni;rogen by pulling the sandwich.;.},
and freeze-fractured by hand un er iqm 'th the adhering fractured mem,.Y; Ruben and Telford (1980) recommend the use of colloidal gold as a calib-
F f e staining the mica w1 .· "'' ration standard. A method for calculating the diameters of rotary-shadowed
apart or nega iv ' . .d .trogen and a drop of negative st ·
brane halves is removed from hquh1 nl1 d material is adsorbed onto ca. particles which, under ideal conditions, cast annular shadows, has been de-
dd d th melting surface T e re ease . scribed by Simpson (1979).
is a e to e . 1. to avoid possible precipitation of negat;
hon-coated gnds. Alternative y, b dded upon thawing and, sub··.· Characterisation of particle populations is best done by means of particle
. dro let of distilled water may e a .
d id which is transferred after a few mm~
ize histograms (Staehelin 1973). So far, counts of particle densities per unit
stam, a p .
quently, covered with.a coate ~h I e of the method has been dem. ea have been done without corrections for local tilting angles, although
onto a drop of negative sta~n. e va u ur le membrane of n::· eeze-fracture replicas are generally copies of non-planar specimens. Meth-
strated with Esc.herichia colz el~~~o)p e~:~d p~~: ~it~ a diameter of 4-6 s are available (see below) for determining the local angle of the replica
bacterium halohlum (Nermut a . , relation to to the electron beam.
and partlc. Ies o f 6-8 nm diameter were demonstrated. · Critical testing of topographical distribution by statistical methods (Ger-
n et al. 1979; de Laat et al. 1981; Van Winkle and Entman 1981; Duniec
6.9.10 Quantrifizcation of structural details and interpretation aids at 1982; Niedermayer and Wilke 1982) is often required in order to iden-
!:·I
significant changes in topography which are unconvincing on casual in- 1'·
h't ture is obtainable by analys
!:~r:~:e~r;b:n~ :~:~;:;:~t~~~t::~h~ :::~~;n;;:~t~~;:~t:'e
ction of electron micrographs. A method based on a statistic, the so-
ed 'coefficient of dispersion', has been used by Weinstein et al. (1976)
(§6.8.2). As discussed earlier (§6.4.lb), . lanes Robards et al. (I 981) to detennine if the two-dimensional distribution
geometry of fine structural details are depicted on fra~~~~;: betw embrane-intercalated particles approximates to a statistically random
on the shadowing procedure. Furthermore, ~ny corree membr ibution. In instances of non-randomness, the method indicates whether
pbologically distinguishable membrane constituen~:s\;~onsidert particies tend to be in a regular array or, alternatively, if the individual
calated particles) and the in vrvo or m slfu stlrut~;;e §6.8.3). Furthe s are statistically aggregated (clumped).
b1hty of preparation artefacts (Pncam et a . , stereological method, allowing an estimation of the relative surface
442 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze:fracture replication
443
area of a certain membrane type in a freeze-fracture preparation of subcellu- that the left and right hand .
lar fractions has been used by Weibel and collaborators (Weibel et al. 1976). 1978). mICrographs are not transposed (Allen et al.
The method exploits the fact that each membrane type has a characteristic
The examination of stereo-pairs of com
range of numerical density of membrane-intercalated particles. It is necess- can be a valuable approach for 1 . plementary areas (Steere 1973)
ary that the membrane vesicles are spherical and that the size distribution lion (§6.8.3) or condensation co:~a ysmg artefacts, such as plastic deforma-
of the various types of vesicle is known. to an apparent lack of co 1 ammat10n (§6.5.1), which often contribute
. mp ementanty A
Numerous methods are available to facilitate the interpretation of the compar1son of complementary . d' . more accurate and objective
topographical undulations recorded in the replica. Using a goniometer . per10 1c structur · ·
age processmg methods as h , es is possible by digital im-
stage, many replica areas that are not at first well resolved can be set to give , s own 'or the regul
lemma fracture faces by Ku""bl 1 ar arrays on yeast plasma
er et a . (1978). -
best resolution by tilting in the appropriate direction (Steere et al. 1974;
H6randner 1976; Steere and Rash 1979; Chalcroft 1984). This is of particu-
lar value if complementary replicas (§6.3.3a(ii)) are examined. One replica
Rej"erences
may look excellent while the complementary part, due to mechanical distor-
tion or displacement during attachment to the grid, may deviate considera-
Abermann, R., 1:1".M. Salpeter and L. Bachmann (1972 . .
bly from an ideal orientation. The best definition of a specimen is generally pies and techniques of elect.- . ), High resolution shadowing in· Pn· .
,on llilcroscopy MA H ' · nct-
obtained when the areas to be examined are normal to the electron beam pany, New York), p. 197. ' · · ayat, ed. (Van Nostrand Reinhold Com-
(Steere 1971). With coarse undulations in the replica, well-shadowed area~ Akahori, H., S. Okamura M Nishiu
p. 294. ' . ra and K. Uehira (1972), Proc. 30th Ann. Meeting EMSA
can even be hidden by protruding parts of a replica. By tilting replicas,.a;,"·
coincidence of the viewing angle with the shadowing angle is recognised ~¥·-'
Akahori, H. and M. Nishiura (1972) . ,
Aldrian, A.F. and HR F H 'Rev. Mtcrosc. elect. 1, 124.
the minimisation of the contrast (and by concentric shadows of memhrat\¢.ri d · · · om (1974), Trennung von D f ..
es Aufdampfens beim thermisch V d amp - und Warmestrahlung wahrend
en er ampfen em fi di" h .
intercalated particles). From this position the optimum viewing angl<(,;• Congr.
. Electron Microscopy , C anerra
b P in IC er ObJekte ' Proc· 8th I n.
t
1 406
more easily obtained by tilting the replica in the appropriate direction (Jiil llam,
h J M. and J. Ardonceau (1983) M d"fi• :
, o 11cation of the at' . .
c ert- ung freeze-fracture apparatus J M' n tcontanunation device of a Rei-
andner 1976). len iv JS G , . tcroscopyl34 217.
A method which allows a topographical map to be produced from. '·· ., · · ardner and W.M. Hess ' . . .
f~eeze-etch replicas, Protoplasma 93, 473 0978), Pseudoscop1c illusions in stereo views of
tised micrographs of a heavy metal-shadowed periodic specimeu;'
hn, J.D. and R.C. Hughes (1981), Prot~in derivatised .
reported by Smith and Kistler (1977). The calculations are based q o substratum adhesion Anal t B" h
, y. iocem.1131 44
glass coverslips for the study of cell
theoretical model that the thickness of metal deposited along the line. hi , I ·and M B .. h. ' ·
. uec 1 ( 1981 ), Internal structural difti . .
incident metal deposits itself uniformly onto the specimen surface w~ the Sendai virus maturation in· Th
nd R W , .
1· . erenhahon of the plasma membrane
e rep ication of negat· tr d .
. . Compans, eds. (Elsevier/North-Holl d A ive s an viruses, D.H.L. Bishop
diffusing or aggregating. Another method, which allows the three-di)Il. ann, L. (1962) Shadow c ,. . an , msterdam), p. 553.
nal reconstruction of surface profiles from microdensitometric data-<:: ' as 1ng using very high m 1 .
ectron Microscopy, Philadelphia J, FF3. e ting metals, Proc. 5th Int. Congr.
tron micrographs, has been described by Krbecek et al. (1979). mann
. L and
· . w w s
· · c hmi tt-Fumian (1973) s . .
The evaluation of stereo-pairs represents a simple procedure for~. hing, techniques and applications EL B ,edpra~ freezing and freeze-etching, in: Freeze-
r.'. , . . enethandPF·dd
terpretation of the spatial relationships of regular structures (Waf ,, oscop1e Electron1que, Paris), p. 73. . avar ' e s. (Soc. Franyaise de
1978). Steere (1971, 1973) reports that good stereo-pairs can be/ ann,· L., R. Abermann and Hp z· h .
· · tngs e1m (1969) H h fl ..
.chem1e 20, 133. , oc au osende gefrieriitzung, His-
with 15° tilt between the two micrographs but, generally, a total . ann L HF"
. t d' ., · · ntzmann and WW Sh · .
· · c m1tt-Funuan (1974) M
is sufficient. For best results he recommends tilting the specim etenn1nation by counting p th I ' olecular and particle
, roe. 8 nt. Congr. Electron Microscopy, Canberra 2,
angles to the direction of the shadow. To avoid the reversed app
<- annr' ....... , GL
'
depressions and elevations, special care must be taken to ens, · eupold, M. Barth and W B · '
,Jilase crystals, Proc. 8th Eur Re C nf .E aume1ste~ (1984), Decoration of freeze-etched
prints are correctly orientated (shadows pointing towards the b;. ", . g. o . lectron Microscopy, Budapest 2, 1289.
444 Robards and Sleytr L ow tempera t u,•e method~· in biological electron microscopy
Freeze-fracture replication
445
W I-I o T N Rhod1n . and B. M · s·iegel (l960)' Determination of surface
Bachmann, L .., · · _rr, · · lication, J. appl. Phys. 31, 1458. Bullivant, S., D.G. Rayns, W.S. Bertaud, J.P. Chalcraft and G.F. Grayston (1972), Freeze-
structure using ultrahigh vacuum rep . I t d ~ r metallic replication, J. Microscopy fractured myosin filaments, J. Cell Biol. 55, 520.
Bank, H. and J.D. Robertson (1976), A simple e ec ro e o
Bullivant, S., P. Metcalf and K.P. Warne (1979), Fine structure of yeast plasma membrane
106, 343. I
after freeze-fracturing in a simple shielded <;ievice, in: Freeze-fracture: methods, artefacts and
Barnett, J.R. (1969), Physical studies of cellulose biosynt h es1s,
· PhD . . Thesis, University of I
interpretations, J.E. Rash and C.S. Hudson, eds. (Raven Press, New York), p. 141. I

Leeds. . . ( 1979), Intercellular junctions in FANFT Carter, D.P. and L.A. Staehelin (1979), Evaluation of IgG molecules, Fab fragments and IgG-
. p KE K ettner and R.S. Wemste1n . f
Bend1cht, U . . , · · u . 1issue
. ulture·· in situ thin-section, freeze- rac- horseradish peroxidase conjugates as surface labels for freeze-etched membranes, J. Micros-
f
induced carcinomas o ra unn1 · ary bladder in
. .c 115 271 copy 117, 363.
. 1 tr icroscopy studies, J. Microscopy , .
tu re and scanning e ec on m ) F t h. ng techniques and applications (Soc. Chalcroft, J.P. (1984), Towards rational terminology and improved techniques in the quantita-
Benedetti, E.L. and P. Favar,d, eds. (1973 , ~eeze e c I ,
tive analysis of surface replicas, Proc. 8th Eur. Reg. Conf. Electron Microscopy, Budapest
. d M' ·e Electronique Pans). . 2, 1297.1
Franyaise e icrosc~pi . hi ' 1 fra ·ie freeze-fracture replicas, Micron JO, 139.
Bordi, C. (1979), Parlod1on coating of gh y . gt f latinum and carbon for possible use in Chalcraft, J.P. and S. Bullivant (1970), An interpretation of liver cell membrane and junction
Bradley, D.E. (1958), Simultaneous evapolrattton :U~roscope Nature Lond. 181, 875. structure based on observation of freeze-fracture replicas of both sides of the fracture, J.
l . h d wcastingfortheeecron ' ' . Cell Biol. 47, 49.
high-reso utlon s a o - . t' technique for the electron microscope
Bradley, D.E. (1959), High resolu~1on sha~oi:u:sa:: carbon, Br. J. appl. Phys. 10, 198. Collins, T.R., J.C. Bartholemew and M. Calvin (1975), A simple method for freeze-fracture
using the simultaneous evaporation of plat b Proc natl Acad. Sci. 55, 1048. of mono layer cultures, J. Cell Biol. 67, 904.
D (1966) Fracture faces of frozen mem ranes, . . S S
Branton, · , . d' f brane structure, Phil. Trans. R. oc. er: Cook, G.M.W. (1976), Techniques for the analysis of membrane carbohydrates, in: Biochemi-
Branton, D. (1971), Freeze-etching stu ies o mem cal analysis of membranes, A.H. Maddy, ed. (Chapman and Hall, London), p. 283.
Costello, M.J. and J.M. Corless (1978), The direct measurement of temperature changes within
D ( 1973) The fracture process offreeze-e tching, in·. Freeze-etching, techniques
B 261, 133. , and'
.
E.L., Benedetti and P. Favard , ed s. (S o c . Franyaise de Microscopie Electron1~,.
Branton, · freeze-fracture specimens during rapid quenching in liquid coolants, J. Microscopy 112, l 7.
applications, Crowther, R.A. and U.B. Sieytr (1977), An analysis of the fine structure of the surface layers
que, Paris), p. 107. . h ansk.I (1977) , Interpreting the results of freeze-etching, J. from two strains of Clostridia, including correction of distorted images, J. Ultrastruct. Res.
Branton, D. and S. Ktrc 58, 41.

Ill, 117. M.J Kamovsky, H Moor, K. Muhlethaler, D anon, D., L. Goldstein, Y. Marikovsky and E. Skutelsky (1972), Use of cationized ferritin
Branton, D., S. Bullivant, N.~. Gilul~, V S th LA Staehehn R L Steere and RS W as a label of negative charges on cell surfaces, J. Ultrastruct. Res. 38, 500.
Northcote L Packer, B. Sahr, P. Sahr, . pe '
. ' · h' omenclature, Science 190, 54.
vy, J.G. (1970), Vaporization mechanism of single ice crystals, Ph.D. Thesis, Univert;ity of
California, Berkeley.
I
stein (1975), Freeze-etc ingn 'th the Bullivant-Ames freeze-£ I !1
Brown, J.N. (1981), Replication of cell mono 1ayers WI
vy, J.G. and D. Branton (1970), Subliming ice surfaces: freeze-etch electron microscopy,
apparatus, Stain Technol. 56, 2h~7. f d hydrated biological material, Proc. 6th Etif~:
cience 168, 1216.
Bucheim, W. (1976), Freeze-etc 1ng o e
mer, D.W., R. Leonard, A. Tardieu and D. Branton (1970), Lamellar and hexagonal lipid
Congr. Electron Microscopy, Jerusalem 2, 122~· for freeze-fracture/freeze-etch sti{ es visualized by freeze-etching, Biochim. Biophys. Acta 219, 47.
Bucheim W. (1982), Aspects of sample prepara ion . , I 189 Kruijff, B. and R.A. Demel (1974), Polyene antibiotic-sterol interactions in membranes of
' . A view Food M1crostruct. , ·
proteins and lipids in food systems. re , and electron microscopy of the c_ choleplasma laidlawii celis and lecithin liposomes. III. Molecular structure of the polyene
d T B" h'1(1979) Immunofluorescence . tibiotic-cholesterol complexes, Biochim. Biophys. Acta 339, 57.
Biiechi, M. an · ac ' b d its interaction with Sendai/
mic surface of the human erythrocyte mem rane an
Laat, S.W., L.G.J. Tertoolen and J.G. Bluemink (1981), Quantitative analysis of the nu-
Cell Biol. 83, 338. . f t · in: Advanced techniques.~ 'ca! and lateral distribution of intramembrane particles in freeze-fractured biological
B Ir , t S (1973) Freeze-etching and freeze- rac unng'. -<:.> branes, Eur. J. Cell Biol. 23, 273.
u 1van, · ' d (S · r Berhn) p. 67.
cal electron microscopy, J· Koehler' e · pnnge ' f; 't and artefacts prod'U<\ y, G.P. and R.E. Beever (1979), Electron microscopy of the rodlet layer of Neurospora
Bullivant, S. (1977), Evaluation of membrane structure ac s .: conidia, J. Bacterial. 140, 1050.
freeze-fracturing, J. Microscopy lll, IOI. b e structure Micron 1ki.•.· y, G.P. and S. Bullivant (1976a), A copper block method for freezing non-cryopro-
t ti f yeast plasma mem ran '
Bullivant, S. (1980a), Interpre a on o . . . f 'mple cold block freeze-fr .. " tissue to produce ice-crystal-free regions for electron microscopy, J. -Microscopy 106,
Bullivant, S. (1980b), Advantages and lnrutatlons o a SI
Proc 38th Ann. Meeting EMSA, p. 748. . . thod for
q~e, S.. and A · Am es (1966) ' A simple freeze-fracture rephcatlon me
Bullivant,
Sey, G.P. and S. Bullivant (l976b), A copper block method for freezing non-cryopro-
d tissue to produce ice-crystal-free regions for electron microscopy. II. Evaluation using
roscopy, J. Cell Biol: 29, ~35. S d (1968) The type II simple e~fracturing with a cryoultramicrotome, J. Microscopy 106, 261.
Bullivant, S., R.S. We1nste1n and K. ome a ,
J. Cell Biol. 39, 19a.
446 Robards and Sleytr Low temperature me 'h od~· in biological electron microscopy
Freeze~(racture replication
447
H F k H Schwartz and E. Henne (1974) ' Complementary freeze-etch
. replicas
Demsey, A., . ran , · . I" the Balzers unit, J. Microscopy 100, Fisher, K.A. (1982b), Monolayer freeze-fracture autoradiography: quantitative analysis of
by adopting the Denton apposed specimen too mg 1o
transrnernbrane distribution of radioiodinated Concanavalin A, J. Cell Biol. 93, 155.
169. h J M Ducommun M Zol11nger . Fisher, K.A. (1982c), Preparation of planar membrane monolayers for spectroscopy and elec-
and E . Kellenberger (1971), A new preparation
34 14 tron microscopy, Meth. EnzymoL 88, 230.
Duboc et, ., · ' : fbiomacromolecules,J. Ultrastruct. Res. , 7.
method of dark-field ele~tron microscopy
0
( ) A method of estimating the radial distri- Fisher, K.A. and D. Branton (1976), Freeze-fracture autoradiography: feasibility, J. Cell Biol.
. J T DJ Goodch1ld and S.W. Thome 1982 , . h C 70, 453.
Dun1ec,
bution ·function
., · · of protein
. part1c . mem branes from freeze-fracture m1crograp s, om-
. 1es m
Fisher, K.A., K. Yanagimoto and W. Stoeckenius (1978), Oriented adsorption of purple mem-
put. Biol. Med.12, 319. S t facts of the freeze-etching technique, J, brane to cationic surfaces, J. Cell Biol. 77, 611.
Dunlop, W. and A.W. Robards (1972), ome are
Fujikawa, S. (1979), Negative staining-carbon replica methods, a new procedure to reveal
Ultrastruct. Res. 40, 391. d ( ) Temperature vacuum and fracturing in ultrastructure of freeze-fractured membrane, J. Microscopy 117, 261.
GR p · h and AW Robar s 1972 ' '
Dunlop,
the freeze-etch technique, Proc. 5th Eur. eg. onf. Electron Microscopy, Manchester, p,
W., · · ans · · R C Garcia-Segura, L.M., D. Baetens and L. Orci (1982), Freeze-fracture cytochemistry of neuro-
nal membranes: inhomogenous distribution offilipin-sterol complexes in perikarya, dendrit-
es and axons, Brain Res. 234, 494.
248. (1974) Th role of spectnn . 1n. d et enmn · ing the lateral distribution of the particles.
Elgsaeter,
exposedA. ' e
by freeze-fracure .
of ISolated h uman ery throcyte plasma membranes, Ph.D. Thesis, Gershon, N.D., A. Dempsey and C.W. Stackpole (1979), Analysis of local order in the spatial
distribution of cell surface molecular assemblies, Exp. Cell Res. 122, 115.
· c lif · B keley
University of a omta, er · h . f f eezc etch rotary shadowing, low temperature Geymayer, W. (1966), Die gleichzeitige elektronenmikroskopische Erfassung von Oberfliiche
78) A freeze-etc unit or r -
Elgsaeter, A. (l~ ' new tional freeze-etching, J. Microscopy 113, 83. . . und Querschnitt gefriergesschokter kolloidaler Systeme, Proc. 6th Int. Congr. Electron Mic-
roscopy, Tokyo 1, 577.
freeze-fractunng and conven . ( ) 1 tramembrane particle aggregatton in erythr. ·<·.
Elgsaeter, A., D. Shotton and D. Branton 1976 , n '::-:::.~-:<:
. Geymayer, W. (1967), Die elektronenmikroskopische Untersuchung temperaturempfindlicher
B. h. Biophys Acta 426, 101. . . ,
rocyte ghosts, ioc tm. .· d 0978) Freeze-fracture identification of sterol-d1gitonvf·-l;.:_ Kolloide, Staub-Reinhaltung Luft 27, 237.
Elias, P.M., J. Goerke and D.S. Fnen b , J C ll Biol 78 577. .,_,,. ··':·_;::;:- Glauert, A.M. (1974), Fixation, dehydration and embedding of biological specimens, in: Prac-
. 11 • d rposome
1 mem ranes, . e · '
complexes m ce ~n
r PM DS FnendandJ.Goer e 979 , k (! ) Membrane sterol heterogeneity. Freeze-fract tical methods in electron microscopy, Vol. 3, A.M. Glauert, ed. (North-Holland, Amster-
dam).
E ias, . ., . . . . . J Histochem. Cytochem. 27, 1247.
detection with sapon1ns and fi~1p1ln, .1980) Rapid freeze-fracturing at down to 12 K;.·· !itch, S. (1969), Verdampfung von Platin aus Kohletiegeln zur elektronenmikroskopischen
· J D G che and G Nico as ( ' ·
Esca1g, ., · under
a · vacuum, p roe. ?th Eur . Reg . Conf. Electron Microscopy: Schrfigbedampfung, Naturwissenschaften 56, 559.
freeze-drying ultrahigh
old, E.R. and P. Balding (1975), Receptor-specific proteins - Plant and animal lectins,
(Excerpta Media, Amsterdam).
Hague 2, 654. f t s de materiel biologique realis6es a
Escaig, J. and G. Nicolas (1976), Cryo- rac.ure . 283 1245
oldstein, J.I., D.E. Newbury, P. Echlin, D.C. Joy, Ch. Fiori and E. Lifshin (1981), Scanning
I , 'd c r Acad. Set., Pans , · .
temperatures en u tra-vi e, · · . . t freezing and freeze-etch1ng.o~ electron microscopy and X-ray microanalysis (Plenum Press, New York, London), p. 646.
E evik T. arid A. Elgsaeter (1981), In situ propane 1e - :.:.:
sp ' · 123 105 ·pp, E.B. and M.R. Bernfield (1978), Acquisition of synchronous beating between embryo-
layer cell cultures, J. Microscopy . , . . . t f•eezing freeze-drying and r .cheart cell aggregates and layers, Exp. Cell Res. 113, 263.
1 (1984) L1qu1d propane JC - i ,

Espevik, T. and A. E gsaeter d } embrane associated structures in human pp, E.B., J.H. Peacock, M.R. Bernfield and J.P. Revel (1978), Morphological and func-
Iication of the cytoskeleton an P asma m
!ional correlates of synchronous beating between embryonic heart cell aggregates and layers,
es, J. Microscopy 134, 203. h' .. El t on microscopy and cytochemistry of.pl.. xp. Cell Res.113, 273.
Fineran, B.A. (1978), Freeze-etc ing, in. ec r d· 279
H. ( 1977), Gefrieriitzung im Ultrahochvakuum, UHV, bei -196°C, PhD. Dissertation,
J.L. Hall, ed. (Elsevier/North-Holland,~= ~~~~cati~n by electron micr en6ssiche Technische Hochschule, Ziirich.
.
Fisher, KA· · (1975) , "Half' membrane ennc en .
H. and H. Moor (1978), Decoration of specific sites on freeze-fractured membranes,
ence 190, 983. b "halves'" JH-cholesterol l . 9th Int. Congr. Electron Microscopy, Toronto 2, 140.
Fisher, K.A. (1976a), Autoradiography ofmem rane .
, H., E. Bas and H. Moor (1978a), Freeze-fracturing in ultra high vacuum (UHV), at
rocytes, J. Cell Biol. 70, 2l~a. b hal es· cholesterol, Proc. natl. Acacf< 96°C, J. Cell Biol. 76, 712.
. her, KA
F is · · (1976b) ' Analysis ofmem rane v .
, H., 0. Kuebler, E. Bas and H. Moor (1978b), Decoration of specific sites on freeze-
73, 173. . . tides 9th Int. Congr. Eli tured membranes, J. Cell Biol. 79, 646.
Fis . . (1978) , Split membrane hp1ds and po1ypep
. h er, KA '
', H., D. Studer and I. Wildhaber (1982), Information contained in metal replicas, Proc .
copy, Toronto 3, 521. . Rev Physiol. 42, 261. Int. Congr. Electron Microscopy, Hamburg 3, 17.
Fisher, K.A. (1980), Split membrane analyfs1s,tAnn~utor~diography: origins berger, R. (1984), Reconstruction of surface profiles from shadowing data, Proc. 8th
. h er, K .A · (1982a) , Monolayer freeze- rac ure
F is , Reg. Conf. Electron Microscopy, Budapest 2, 1281.
Microscopy 126, 1.
448 Robards and Sleytr Low temperature metiio d~. in biological electron microscopy
Freeze-fracture replication
449
bkJCDedieuandT.Gu1-rzw1 l.k K 'cki 0982) ' Freeze-fracture
7 electron
Golik, A., L.P. Agger ec ' . ·. . 1 . I molecules J. Microscopy 125, 20 . Howe S.C. and T. B3chi (1973), Localization of erythrocyte membrane antigens by immune
microscopic analysis of solutions of bio ogica f t ~perature X-ray d,iffraction to eva- electron microscopy, Exp. Cell Res. 76, 321.
1
Gulik-Krzwicki, T. and M.J. Costello (1978)1, Utse ormocwrosecopy J Microscopy 112, 103.
· f f acture e ec ran ' · . Hubbard, A.L. and Z.A. Cohn (1976), Specific labels for cell surfaces, in: Biochemical analysis
luate freezing methods in reeze- r 'k (G undlagen und Anwendungen), (Spnnger-Ver- of membranes, A.H. Maddy, ed. (Chapman and Hall, London), p. 427.
c RA
Hae1er, . · (l 9 Sl) , Kryo-Vakuumtechnt r
JakopiC, E., A. Brunegger, R. Essl and G. Windisch (1978), A sputter source for electron mic-
lag, Berlin, Heidelberg, New York). (1977) A simple electron beam gun for plati- roscopic preparation, Proc. 9th Int. Congr. Electron Microscopy, Toronto 1, 150.
H 1 HK. W.W. Schulz and R.C. Reynolds ,
ager, · ' · 110 149 Jensen, T.E., M.A. Smith and W.M. Hess (1973), Washing apparatus for freeze-etch replica
num evaporation, J. Microscopy , r . thod for electron microscopy, J. appl. Physics fragments, Stain Technol. 48, 43.
Hall, C.E. (1950), A low temperature rep ica me
Johnson, R.P.C. (1968), Microfilaments in pores between frozen-etched sieve elements, Planta
81, 314.
21, 61. . . .. li technique for study of biological membranes
Heinmets, F. (1949), Mod1ficatton of s~hca_ rep! cat on microscopy, J. appl. Physics 20, 384.
. f t condensation tn e ec r .. d. R t Johnson, R.P.C., A. Freundlich and G.F. Barclay (1976), Transcellular strands in sieve tubes;
and application o ro ary .. . b. lo ·scher Gewebeproben fur te as er- what are they? J. exp. Bot. 27, 1117.
. U (1974) Verfeinerte Praparatton 10 gt
Heinzmann, · ' · A ta 76 145
Junger, S. and L. Bachmann (1977), Structure ofpyruvate dehydrogenase complex: compari-
Elektronenmikroskopie, Microscopica ~ 'd · . g of freeze-etch replicas which show
L Bl . (1972) Production an c1ean1n son between freeze-etching and negative staining, Biochim. Biophys. Acta 481, 364.
Hess W.M. and R. · air ' d h phae Stain Technol. 47, 249. Kahler, H., and B.J. Lloyd, Jr. (1950), Metallic evaporation and the diameter of tobacco
' f ff tured fungus spores an Y '
complementary sur aces o rac . . . f coated vesicle formation in fibroblasts.~ mosaic virus with the electron microscope, J. appl. Phys. 21, 699.
Heuser, J .E · (198 0) ' Three-dimensional VIsuahsatJ.on o
Kamovsky, M.J., E.R. Unanue and M. Leventhal (1972), Ligand-induced movements of lym-
J · Cell Biol. 84, 560. . b" 1 · 1 samp1es ,ior stereomicroscopy by the quick-freez.e~<
. <: phocyte membrane macromolecules, J. exp. Med. 136, 907.
Heuser, J.E. (1981), Prepanng 10 o-?ca .. Methods in cell biology, Vol. 22, J.N. Turn . Kellenberger, E. and J. Kistler (1979), The physics of specimen preparation, in: Advances in
d eep-etchi,
rotary
.
replication technique,. tn.
97 structure research, W. Hoppe and R. Mason, eds. (Friedrich Vieweg and Sohn Verlagsge-
ed. (Academic Press, New York)., p. . ss bridge in insect flight muscle, J. sellsch., Wiesbaden) 7, 49.
H euser, J .E · ( 1983a) ' Structure of the myos1n cro
istier, J. and E. Kellenberger (1977), Collapse phenomena in freeze-drying, J. Uitrastruct.
169, 123. f d ying molecules adsorbed to mica flakes, J, Res. 59, 70.
Heuser, J .E · (1983 b) ' Procedure for reeze- r
istler, J., U. Aebi, L. Onorato, B. ten Heggeler and M.K. Showe (1978), Structural changes
Biol. 169;
Heuser and R. Cooke (1983 ), A ctin-my osin interactions visualised by the
J E 155. quick-~! during the transformation of bacteriophage T4 polyheads: charaterization of the initial and
' · · . h · e J mol Biol. 169, 97. . final states by freeze-drying and shadowing Fab-labelled preparations, J. mo!. Biol. 126, 571.
deep-etch replica tee niqu , · · . t r anisation revealed in platinum
J E and M.W. Kirschner (1980), Fdamen o g
Heuser, · · C B. 1 86 212 ·.
'.<· ajima, Y., T. Sekiya and Y. Nozawa (1976), Freeze-fracture and ultrastructural alterations
11 10 nduced by filipin, pimaracin, nystatin and amphotericin B, Biochem. Biophys. Acta 445,
of freeze-dried cytoskeletons, 1. e · . ' . · f tyicholine receptors in quicK. 52.
H euser J.E. and S.R. S alpeter (l979) , Organ1sahon o ace · 1<
11·cmembrane J. Cell Biol. 8 f,·
' r t d Torpedo postsynap ' ·"" inschmidt, A., D. Lang and R.K. Zahn (1960), Darstellung molekularer van DNS, Natur-
deep-etched, and rotary rep tea e
SR ndDMD Lan ts
a· (1976) , Preservation of synaptic structuti. '. issenschaften 47, 16.
Heuser,J.E.,T. · e~sea · · ~. uant.Biol.40, 17. . . "<" G., G. Oebel and H. Plattner (1982), A simple sandwich-cryogen-jet procedure with
freezing, Cold Spnng Harbour ~-y p Q E H er (1982), Organisation of a;ctl
G T'ln
1 K Fujiwara and J. · eus J cooling rates for cryofixation of biological materials in the native state, Protoplasma
Hirokawa, N., L. · ey, ·. h b der of intestinal epithelial ce 11s, ' I, 161.
and intermediate filaments in the brus or
!er, J.K. (1966), Fine structure observations in frozen-etched bovine spermatozoa, J.
425. ·1 1980) Semi-automatic washing device for.: astruct. Res. 16, 359.
H6henberg, H. and K.C. Mannwe1 e:d( ,. 1' onditions Mikroskopie36, 145. ::
f l" as under i en tea c ' ·th .c r, J.K. (1972a), The freeze-etching technique, in: Principles and techniques of electron
cleaning of sur ace rep IC . d. freeze-etch replicas WI "11!.•
(19 76) E aluation of correspon ing , y scopy, Vo!. 2, M.A. Hayat, ed. (Van Nostrand Reinhold Company, New York), p.
H6randner, H. , v Re Conf. Electron M1croscop.· ,~
stage and by densitometry, Proc. 6th Eur. g. ,,
r, 1.K. (l 972b), Human sperm head ultrastructure: a freeze-etching study, J. Ultrastruct.
125. . Id s a cytochemical mark~. 39, 520
. M (1979) Evaluationofcol101dalgo a ..... :
Honsberger, · ' . Biol Colloids 36, 253. ., L.A. Staehelin and E. Nettersheim (1973), A new type of storage container for freeze~
and scanning electron rrucroscopy, . 11 .dal Id a useful marker for.·.·. ·"specimens, J. Microscopy 99, 349.
(1977) Co 01 · go '
Horisberger, M. and J. Rasset J H.'t hem Cytochem. 25, 295.
h e Licht- und w··armes
, R., C, Gebhardt, H. Gruler and E. Sackmann (1979), Three dimensional microscopic
scanning eIec t ron m1"crOSCOpy ' · IS OC .
Horn H.R.F. (1962), Thennische BedampfunMg~ n opy Philadelphial, profiles of membranes reconstructed from freeze-etching electron micro graphs, Bio-
;.]iophys. Acta 554, l.
des' Objektes, Proc. 5th I nt. C ongr· Electron tcrosc ,
450 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-fracture replication
451
er o um freeze-etch device, J. Cell Biol. 47, 11 lA.
Kreutziger G.O. (1970), The Y P P t structures of membrane fracture Moor, H. (1959), Platin-Kohle-Abdruck-Technik angewandt auf den Feinban der MilchrOh-
, d H Moor (1978) Comp1emen ary
Kiibler, 0., H. Gross an · ' ' . _1 96oc and digital image process- ren, J. Ultrastruct. Res. 2, 393.
faces obtained by ultra high vacuum freeze-fractunng a 1
Moor, H. (1970), Die Gefrieriitztechnik, in: Methodensammlung der Elektronenmikroskopie,
ing, Ultramicroscopy 3, 161. . . (l Sl) Ultrastructure of sheep erythrocyte G. Schimmel and W. Vogell, eds. (Wissenchaftliche Verlagsanstallt, Stuttgart), p. 1.
A M V Nermut and L.D. W111tams 9 , b
Lang, R.D . . , · · . ~
plasma membranes bound to sohd supports - e ec
tsofselectiveremovalofmem ranecom- Moor, H. (1971), Recent progress in the freeze-etching technique, Phil. Trans. R. Soc. Ser. B.
261, 121.
ponents, J. Cell Sci. 49, 383. ( ) Lipids of archaebacteria, Zbl. Bakt. Moor, H. (1973a), Evaporation and electron guns, in: Freeze-etching, techniques and applica-
Langwor thy, T ·A ., T ·G · Tornabene and G. Ho1zer 1982 ,
tions, E.L. Benedetti and P. Favard, eds. (Soc. Franyaise de Microscopie Electronique,
Hyg., I. Abt. Orig. C. 3, 228 . f . g and etching artefacts in particulate Paris), p. 27.
Lepault, J. and J. Dubochet (1980), Freezing, rac1unn
Moor, H. (1973b), Etching and related problems, in: Freeze-etching, techniques and applica-
72 223
suspensions, J. Ultrastruct. Res. • · . Ein Beitrag zur Kliirung seiner Tertiiir- tions, E.L. Benedetti and P. Favard, eds. (Soc. Fran¥tise de Microscopie Electronique,
Lickfield, K.G. (1968), Der frostgeiitzte Baktenenkern. Paris), p. 21.
560
struktur, Z. Zellforsch. 88, · . fbacteria in· Methods in microbio- Moor, H., K. Miihlethaler, H. Waldner and A. Frey-Wyssling(l961), Anew freezingultramic-
L. kfield K.G. (1977), Transmission electron nncroscopy o ' . rotome, J. biophys. biochem. Cytol. JO, 1.
IC , . d (Academic Press, London), p. 160.
logy, Vol. 9, J.R. Norns, e · . h( ) The interpretation of images of cross- Morgan, C. (1972), The use of ferritin-conjugated antibodies in electron microscopy, Int. Rev.
K G M Achterrath and F. Hentnc 1972 ' CytoL 32, 291.
Lickfield, · ., · d b es J Ultrastruct. Res. 38, 279 ·
df t hed and shadowe mem ran , · . . t "
fracture rozen-e c (l 977) Rotary rephcation for freeze-e C1ung, Miihlethaler, K. (1973), History of freeze-etching, in: Freeze-etching, techniques and applica-
. . L .H ., A · Elgsaeter and D. Branton
Margant1s, ,
tions, E.L. Benedetti and P. Favard, eds. (Soc. Franyaise de Microscopie Electronique,
Paris), p. l.
J. Cell Biol. 72, 47. han es in intramembrane particle distribution,;_- ..
Maul, G.G. (1979), Temperature dependen~ ~ t g tation J.E. Rash and C.S. Hudson, ed~~: Miihlethaler, K., E. Wehrli and H. Moor (1970), Double fracturing methods for freeze-etch-
in: Freeze-fracture: methods, artefacts an in erpre , -------- ing, Proc. 7th Int. Congr. Electron Microscopy, Grenoble 1, 449.
37
(Raven Press, New York), P· · .fi b ll"ng of glycoproteins in yeast plas~:~~< Miihlelhaler, K., W. Hauenstein and H. Moor (1973), Double fracturing method for freeze-
A d K Miihlethaler (1981), Spec11c 1a e l ->':;cY<-
Maurer, · an · . E J C ll Biol 25 58. '.''/ etching, in: Freeze-etching, techniques and applications, E.L. Benedetti and P. Favard, eds.
membrane with Concanavahn A, ur(.1~74~ Polyiysi~e as an adhesive for electron mig!- (Soc. Fram;:aise de Microscopie Electronique, Paris), p. 101. -
Mazia, D., W.S. Sale and G. Schatten ' ,_>
iiller, T., H. Gross and H. Moor (1984), A double replica device for freeze-fracturing and
copy, J. Cell Biol. 63, 212a. . (1967) Freeze-etching with radiant energy in a simple:_:, freeze-etching at defined specimen temperatures, Proc. 8th Eur. Reg. Conf. Electron Micros-
McAlear, J.H. and G.O. Kreutziger , ,- copy, Budapest], 1735.
. p 25th Ann Meeting EMSA, p. 11 6 .
block device, roe. · d . (Academic Press, London). lier, M., T. Marti and S. Kriz (l980a), Improved structural preservation by freeze-substitu-
(197 8) Fundamentals of freeze- ry1ng .k
on, Proc. 7th Eur. Reg. Conf. Electron Microscopy, The Hague 2, 720.
Mellor, J.D. , . S h 'dt (l 972) Gefrierbruchtechru zur
Menold, R., B. Liittge, W. Katser an~ A. cc::m
In . T~chn. 44, 1226. Her, M., N. Meister and H. Moor (1980b), Freezing in a propane jet and its application
chung von Sus~_ensione~ : d:a::~s~~·;;:), Freez.e-fr~cturing, a new m~thod for freeze-fracturing, Mikroskopie 36, 129.
Menold, R., B. Luttge an · . Ad Colloid Interface Set. 5, 281. , ,_.;_: , E.A., L. Bachmann and A. Feinstein (1980), Structure-of hydrated immunoglobulins
. a· . by electron
gatlon of 1spers1ons
trucroscopy, v. . J p l
. .ds by vacuum evaporation, . a_ Jl antigen-antibody complexes. Electron microscopy of spray-freeze-etched· specimens,
Meryman, HT · ·( 1950) ' Replication of frozen 11qu1 chim. Biophys. Acta 625, I.

M;:~!n, HT and E Kafig (1955), The st:~y;::'~~:;::;;mi:,~~~~;;'~'.;:~ 'nga, N. (1973), Freeze-fracturing of microorganisms. Physical and chemical fixation of
illus subtilis, in: Freeze-etching, techniques and applications, E.L. Benedetti and P.
growth by electron trucroscopy, Naval M . G f tzung und dte Struktur rd, eds. (Soc. Fram;aise de Microscopie Electronique, Paris), p. 151.
M eyer, H .W . and H Wmkelmann (1969), Die e nera
, M. V. (1973), Freeze~drying and freeze-etching of viruses, in: Freeze-etching, techniqu-
Membranen, Protopl~sma 68, 253. Die Darstellung von Lipiden bei der applications, E.L. Benedetti and P. Favard, eds. (Soc. Fran~aise de Microscopie Blee::·.·
Meyer, H.W. and H. W1nkelmann (1970k), I se biolooischer Membranen, ue, Paris), p. 135.
· hung zur Stru turana Y
paration un d l.hre B ezte t:r
, M.V. (1977), Negative staining in freeze-drying and freeze-fracturing, ·Micron 8, 211.
t, M.V. (l982a), The "cell monolayer technique" in membrane· research, Eur. J. Cell
Mill~r K.R. and L.A. Staehelin (1976), An~lysis of th el B'yla
47 th l koid outer surface:-
68 30. 28, 160.
' . . k d brane regions, J. Cel 10 . ' - .
tor is hm1ted to unstac e mem . 0 . (1979) Absence of filiptn-.s.t.. 1:, M. V. (l 982b), Freeze-fracture of red blood cells attached to solid supports: structural
R A p let p Vassalh and L. rc1 ' S . tr--
Montesano, ., · erre ' · f lt e cells Proc. natl. Acad. Ct._>:: ,Chemical analysis of the outer membrane leaflet, Proc. 10th Int. Congr. Electron Micros-
from large coated pits on the surface o cu ur , , :_:':- ·;r-Hamburg 3, 203.
N rmut M.V. (1984), Ultrastructural an d ch ermc · al analysis
3 of intramembranous particles;
1737
e ' C f Electron Microscopy' Budapest , . Pinto da Silva, P., C. Parkison and N. Dwyer (l98lc), Fracture-label: cytochemistry offreeze- ,,
Proc. 8thMVEur.
and L.D. on
Reg. ·.
Williams (197 6) , F reeze- f ra cturing of monolayers of cells, membranes fracture faces in the erythrocyte membrane, Proc. natl. Acad. Sci. U.S.A. 78, 343.
, I
Nermut, . . R C f Electron Microscopy, Jerusalem 2, 131. PJatt-Alloia, K.A. and W.W. Thomson (1982), Freeze-fracture of intact plant tissues, Stain
N and viruses, Proc. 6th Eur:lli eg. (~~?7)Freeze-fracturing of monolayers (capillary layers) Technol. 57, 327. 'I
I
t M V and L D Wi ams ' · 110 121
ermu
0f 11, membranes
. . . and
. viruses.
. . some t echnical considerations, J. Microscopy
. . f ,f . Plattner, H. and H.P. Zingsheim (1983), Electron microscopic methods in cellular and molecu- I

Nermut,ce s,M.V., l.D.J. Burdett and L .D . w·ir . l iam,s (1978) ' Negative staining of reeze- ractured lar biology, in: Snbcellular biochemistry, Vol. 9, D.B. Roodyn, ed. (Plenum Press, New
'York), p. 1.
vela es of Escherichia coli K12, J. Microscopy I 14, 229. . .
N en b Per D Ch and H.P. Zings . h e1m· (I 9?9) , Apparent holes 1n rotary shadowed proteins: Porter, K. and J. Blum (1953), A study of microtomy for the electron microscope, Anat. Rec.
117, 685.
cuge au , . . gl f hadowing and replica thickness, J. Microscopy 117, 313.
dependenceonan eo s · ·I · 1 t ·I "th
Nickel, E., G. Oebel and P. Psc h e1·d (1978), Freeze-fractunng of b10 og1ca ma ena w1 a new Prescott, L. and M.W. Brightman (1978), A technique for the freeze-fracture of tissue culture,
J. Cell Sci. 30, 37.
. table J Microscopy 113, 101. . .
specunen W. '(1982a),
Niedermeyer, · Freeze- f rac t unng. at low temperatures.
. I. A device for fractunng b10- Pricam, C., K.A. Fisher and D.S. Friend (1977), Intramembranous particle distribution in hu-
. . t 77-10 K under high vacuum, J. Microscopy 125, 307. man e~ythrocytes: effects oflysis, glutaraldehyde and poly-L-lysine, Anat. Rec. 189, 595.
logical specimens
Niedermeyer, a
W. (1982b), . 1e atr
A s1mp . 1oc k system for hoar frost free specimen exchange in Pschcid, P. and H. Plattner (1980), Cryofixation of cell monolayers: Avoidance of unusual sub-
h m Jants, J. Microscopy 125, 299. . strate r..ateria!s, detachment of cells, chemical fixation and antifreeze impregnation, Proc.
p W1"lke (1982) , Quantitative analysis of intramembrane particle (IMP}-_.'>'.-'-' 7th Int. Congr. Electron Microscopy, Grenoble 2, 716.
N .freeze-etc
d r vacuu
W and H
tedistribution
enneye ' on· b1omembranes
. · a ft er f re eze-fracture preparation by a computer-based techni->/<>- ," Pscheid, P., C. Schudt and H. Plattner (1981), Cryofixation of monolayer cell cultures for
freeze-fracturing without chemical pretreatments, J. Microscopy 121, 149.
J Microscopy 126 259 · I
que, G.R.
Parish, · (1973), The app . · o f the freeze-etch technique to the study of plant Vascu 11r--
' 11cat1on Rash, }.E. (1979), The sectioned-replica technique: direct correlation of freeze-fracture replicas
D Ph"!
1 Th sis University of York. ___. and conventional thin section images, in: Freeze-fracture: methods, artefacts and interpreta-
cells,B.U.,
· R.S.
· We1nste1n,
e. ' . L. W · Sobl e an d J · Alroy (1977), Freeze-fracture ofmonolayer.cu,1,t.' tions, J.E. Rash and C.S. Hudson, eds. (Raven Press, New York), p. 153.
Pauli, --:i\
ash, J.E. and C.S. Hudson (eds.) (1979), Freeze-fracture: methods, artefacts and interpreta-
tures, J. Cell Biol. 72, 7163.. thod ensuring recovery of delicate freeze-fracture repli_ ;_· tions (Raven Press, New York).
Pearce, R.S. (1983), A c ean1ng me -: ;;.>
11 f lls J Microscopy 129, 229. ash, J.E., C.S. Hudson and W.F. Graham (1978), Direct and indirect labelling of particles
of cerea . ffe (1976) , A simple selective method for freeze-fracturing sphef and pits in freeze-etch replicas, J. Cell Biol. 79, 232a.
Peng, H.B. eaandceL.F.' Ja
_ash, J.E., T.J.A. Johnson, C.S. Hudson, D.S. Copio, W.F. Graham, M.E. Eldefraw and F.D.
cells, J. Cell Biol. 71, ~4.
ht (1978) Antibody-labelling techniques, in: Advanced tee · Giddings (1980), Identification of intramembrane particles by pre- and post-fracture labell-
Perkins, W.D. and J.K. oe _er ' V J K Koehler ed. (Springer Verlag, Be ing techniques: a progress report, Proc. 38th Ann. Meeting EMSA, p. 692.
es in biological electron microscopy, o1· 2, · · '
sh, J.E., T.J.A. Johnson, C.S. Hudson, F.D. Giddings, W.F. Graham and M.E. Eldefraw
39. K.R. (1980), Penning
Peters, . sputtenng . o f u ltrathin metal films for high resolution 982), Labelled-replica techniques: post-shadow labelling of intramembrane particles in i.",11
eeze-fracture replicas, J. Microscopy 128, 121.
. S nning Electron Microscopy 1980, 1, 143. .
microscopy, ca . d (1975) Methods for the freeze-fracturing ofne~: er, L. (1967), Elektronenmikroskopische Untersuchung- und Priiparationsmethoden l'.I
Pfenninger K.H. and E.R. R1n erer ' ,_'-·· ringer-Verlag, Berlin), p. 303.
' 11 la ers J Cell Biol. 65, 15. ·
cultures and ce mono Y ' · . t hemistry labelling of plasma and in r, 0. {1961), Ein einfaches Verfahren zur Dickenbestimmung von Kohlenstoff-Auf-
Pinto da Silva, P. (1982), Freeze-fracturing cy oMc . , Hamburg 3 195. pfschichten, Naturwissenschaft 48, 519.
p lOth Int Congr Electron 1croscopy, '
membranes, roe. · ( 1970) Membrane splitting in freeze-etching.\, e, H.G., Y. Patel and M.A. Venkatachalam (1978), Glomerular filtration of proteins:
Pinto da Silva, P. and D. Branton 'c II B. I 45 598. -'
arance o[ anionic, neutral and cationic horseradish peroxidase in the rat, Kidney Int. 13,
.. brane marker J e 10 · , .
bound fernttn as a mem T .. (1982) Freeze-fracture cytochemistry: partitta_
Pinto da Silva, P. and M.R. ornst 'b s J Cell Biol 93 463.
. . f f ture erythrocyte mem rane , . . ' , ., 1i J.P., A.G. Lee and A.J. Hudspeth (1971), Gap junctions between electronically coupled
phonn in reeze- rac (l ?I) Localization of antigen sit in tissue culture and brown fat, Proc. natl. Acad. Sci. U.S.A. 68, 2924.
. t 0 d a Si.Iva' p ., Sp
P Ill · · Douglas and D. Branton 9 '
144 r, H, and U.B. Sleytr (1971), Fettextraktion bei - 78°C: Nachweis im Gefrieriitzbild,
erythrocyte ghosts, Nature, Land. 232_,. C. B . nd C Parkison (1981a).F aturf. 26, 470.
S ·1 p B Kachar M R Torns1, . rown a .
Pinto da 1 va, ., · ' . ·.
cytochemistry: replicas of cnttcal point ne · . d . d cells and tissues after fracture~_ ., A. Schiller and R. Taugner (1976), Freeze-fracture autoradiography, Histochemistry

213, 230. 9 b) Freeze-fracture cytoch , A.W, (I 974), Ultrastructural methods for looking at frozen cells, Sci. Prag. (Oxford)
Pinto da Silva, P., C. Parkison and N. Dwyer 0 81 , J Histochem. Cyt
tions of cells and tissues after labelling of fracture faces, .
, A.W. (1978a), Freeze-etching of plant cells, Mikroskopie 34, 12.
454 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-fracture replication
455

Robards, A.W. (1978b), An introduction to techniques for scanning electron microscopy of Sc~ille~, A., E..Rix and R. Taugner (1978), Freeze-fracture . . .
plant cells, in: Electron microscopy and cytochemistry of plant cells, J.L. Hall, ed. (Elsevier/ c~attng technique, Histochemistry 59, 9. autorad1ography. The m vacuo
North-Holland, Amsterdam), p. 343. Schiller, A., R. Taugner and E. Rix (1979) F f
. ' reeze- racture autoradio h p
Robards, A.W. and M.H. Cooper (1971 ), An apparatus for preparing frozen-etched specimens a routine technique J Histoche C t h grap y: rogress towards
Sch ·d ' · m. Y oc em. 27, 1514
for electron microscopy, Vacuum 21, 465. mi' E.N., U.B. S!eytrand K.G. Lickfield (1980 .
Robards, A.W. and G.R. Parish (1971), Preparation and mechanical requirements for freeze~ envelopes, J. Ultrastruct. Res. 70 86 ), Nomenclature of frozen-etched bacterial
etching thick-walled plant tissue, J. Microscopy 93, 61. Schmidt, D.G. and W. Bucheim (1~82). 0 th . f
l t . , n e size o small protein part. I d
Robards, A.W. and W. Umrath (1978), Improved freeze-etching of difficult specimens, Proc. e ec ran microscopy of unidirectionall shad IC es etermined by
126, 347 . y owed freeze-etched preparations , J . M.tcroscopy
9th Int. Congr. Electron Microscopy, Toronto 2, 138.
Robards, A.W., W.R. Austin and G.R. Parish (1970), The role of the microtome assembly in Schulz, W.W. and R.C. Reynolds (1977) E h
' n ancement of three dime · l
the freeze-etching technique, Proc. 7th Int. Congr. Electron Microscopy, Grenoble I, 447. f reeze- f racture images by reve 1 . ns1ona appearance of
rsa processing of electron ·
Robards, A.W., U.B. Sleytr, G.R. Bullock, P. Sibbons and H. Quine (1980), A thin plate clea- copy 112, 249. nncroscopy sheet film, J. Micros-
vage device for the Leybold-Heraeus Bioetch 2005 freeze-etching unit, Proc. 7th Eur. Congr. Severs, N.J. (1984), Freeze-fracture cytochemistry Proc 8th E R
Electron Microscopy, The Hague 2, 740. copy, Budapest J, 1747. ' · ur. eg. Conf. Electron Micros-
Robards, A.W., G.R. Bullock, M.A. Goodall and P.D. Sibbons (1981), Computer assisted ana- Severs, N.J. and H.L. Simons (1983), Failure of fili i t d
lysis of freeze-fractured membranes following exposure to different temperatures, in: Effects smooth muscle plasma membrane N t L p n o etect cholesterol rich domains in
. , a ure, ond. 303 637
of low temperatures on biological membranes, G.J. Morris and A. Clarke, eds. (AcademiC Shibata, Y., T. Arima and T. Yamamoto (1984) D bl' . . .
ing, J. Microscopy 136, 121. ' ou e-axis rotary rephcation for deep-etch-
Press, London)p. 219.
Robenek, H., R. Gebhardt and W. Jung (1982), Freeze-fracture identification of complexes: Shotton, D., Thompson K. Wofsy L dD B
' ' ' · an · ranton (1978) App d .
between cholesterol and filipin, digitonin or tomatin in hepatocyte plasma membranes of surface proteins of the huma yth ' earance an distribution
Silva, M.T. F.C G , d n er rocyte membrane, J. Cell Biol. 76, 512.
ing formation of the biliary pole, Proc. 10th Int. Congr. Electron Microscopy, HEcmlbm·1j"1,' . ' . uerra an M.M. Magalhaes (1968) The fi . .
457. . m electron microscopy, Experientia 24, 1024. , ixahve action of uranyl aceta.ie
Robinson, J.M. and M.J. Karnovsky (1980), Evaluation of the polyene antibiotic filipin imps~n, D.J. (1979), Freeze-fracture studies on barle 1' .
cytochemical probe for membrane cholesterol, J. Histochem. Cytochem. 28, 161. the light-harvesting chlorophyll-protein Carlsb. y p astid membranes. III. Location of
Robinson, P.J., P. Dunnill and M.D. Lilly (1971), Porous glass as a solid support for irnii1 leytr, U.B. (l970a), Die Gefrieriitzung korresp:~~e~:~e~o:mun ...44, 305.'
sation of affinity chromatography of enzymes, Biochim. Biophys. Acta 242, 659. Aufkliirung van Membranenstrukt p ruchhiilften ein neuer Weg zur
uren, rotoplasma 70 101
Roth, J. (1978), The lectins, molecular probes in cell biology and membrane researcl:ic- ytr U.B. (1970b), Fracture faces in intact cells a d ' .
Gustav Fischer, Verlag Jena). illus. A study by conventional f h. n protoplasts of Bacillus stearothermoph-
. . reeze-etc 1ng and freeze-etchin 0 f .
Roth, J. (1982), New approaches for in situ localisation of antigens and glycoconj 01et1es;Protoplasma 71, 295 . g corresponding fracture
thin sections: the protein A-gold technique and the lectin colloidal gold marker sys ytr, U.B. (1974), Freeze-fracturin at Ji uid h 1"
10th Int. Congr. Electron Microscopy, Hamburg 3, 245. t. Con gr. Electron Microscopy, ~anb~rra 2~ ;~m temperature for freeze-etching, Proc. 8th
Ruben, G.C. and J.N. Telford (1980), Dimensions of active cytochrome C oxidasein''' tr, U.B. (1975), A method for adaptin the Le. b 1 . . .
tuted liposomes using a gold ball shadow width standard: a freeze-etch electron licas to the Balzers unit J M1'c g 103 y o d device for obta1n1ng complementary
, · roscopy , 377.
study, J. Microscopy 118, 191. , ~-B. (1978), Regular arrays of macromolecules on b .
Salpeter, M.M. and L. Bachmann (1972), Autoradiography, in: Prin~iples and tee: emistry, assembly and flinction I t R C actenal cell walls: structure
T U B and . , n . ev. ytol. 53, I. '
electron microscopy, Vol. 2, M.A. Hayat, ed. (Van Nostrand Reinhold Com, ' .. . G.P. Fners (1978), Surface charge and mar h .
York), p. 219. ems on bacteria! cell walls Proc 9th I t C p ogenes1s of regular arrays and
UB ' · n · ongr. Electron Microscopy T
Salpeter, M.M., H.C. Fertuck and E.E. Salpeter (1977), Resolution in electron,_-- ' · . and A.M. Glauert (19 75) An 1 . f , oronto 2 , 346.
. ' a ys1s o regular arrays of s b ·1 b .
autoradiography. III. Iodine-125, the effect of heavy metal staining, and a rea : evidence for a dynamic proc f, bl u uni s on actenal sur-
i; U B dA ess o assem y, J. Ultrastruct. Res. 50 103
critical parameters, J. Cell Biol. 72, 161. ' , . an .M. Glauert (1982), Bacterial cell walls and me b .' .
Salpeter, M.M., F.A. McHenry and E.E. Salpeter (1978), Resolution in ele:tr_~ of proteins, Vo!. 3 J.R H, . d ( , . m ranes,in: Electronmicros-
U.B. and PM , . arns,e. AcadelUlcPress,NewYork),p.41.
autoradiography. IV. Application to analysis of autoradiographs, J. Cell Biol, . essner (1978), Freeze-fracturing in norm I
Sawada, H. and E. Yamada (1981), A freeze-fracture deep-etching replica met alemma structures, J. Cell Biol. 79, 276. a vacuum reveals ring-like yeast
cryoprotectant, J. Electron Microscopy 30, 34 L _U.B. and A.W. Robards (1977a) Pl · .
Schiller, A. and:R. Taugner (1980), Freeze-fracturing and deep-etching will\ 'croscopy 110, I. , asttc deformation during freeze-cleaving: a review,
protectant ethanol reveals true membrane surfaces of kidney structures, C~_ f:J,B.andAW· ·
Rb0 ar d s (1977b), Freeze-fracturing·
oscopy Ill, 77. ·a review of methods and results,
57.
456 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-fracture replication
457

Sleytr, U.B. and A.W. Robards (1982), Understanding the artefact problem in freeze-fracture Steere, R.L. and J.M. Moseley (1969) N . . .
Meeting EMSA, p. 202. , ew dimensions In freeze-etching, Proc. 27th Ann.
replication: a review, J. Microscopy 126, 101.
Sleytr, U.B. and W. Umrath (1974a), A simple fracturing device for obtaining complementary Steere, R.L.. and J.M. Moseley (1970), M o difi1ed f reeze-etch equipment ·t .
replicas of freeze-fractured and freeze-etched suspensions and tissue fragments, J. Micros- preparation of 2-10 double replicas 7 th I C penru s simultaneous
Steere, R.L. and J.E. Rash (1979) U, f dnt.blon~r. Ele_ctron Microscopy, Grenoble 1, 451.
copy JOI, 177. ' se 0 ou e-tdt device (goniomet ) t b
Sleytr, U.B. and W. Umrath (1974b), A simple device for obtaining complementary fracture contrast in freeze-fracture replica .. F fi er o o tain optimum
. s, In. reeze- racture· methods t fi
planes at liquid helium temperature in the freeze-etching technique, J. Microscopy 101, 187. tlon, J.E. Rash and C.S. Hudson d (R · , ar e acts and interpreta-
Sleytr, U.B. and W. Umrath (1976), Freeze-etching: technical developments and general inter- Steere, R.L., E.F. Erbe and J.M. M:s:ie al~e;14Press, New York), p. 161.
study of biological specimens prep dy b( ), Impo~ance of stereoscopic methods in the
pretation problems, Proc. 6th Eur. Reg. Conf. Electron Microscopy, Jerusalem 2, 50. . are Y cryotechn1ques p 8th I C
Sleytr, U.B., D. Kerjaschki and L. Stockinger (1972), Gefrieratzpraparation von Gewebes- Microscopy, Canberra 2, 36. ' roe. nt. ongr. Electron
tiicken in der EPA 100 am Beispiel des Nasenseptums der Maus, Mikroskopie 28, 257. Steere, R.L., E.F. Erbe and J.M. Moseley (1977) A re . . .
automatic regulation of shadow d , fi s1stance morutor with power cut-off for
Sleytr, U.B., R.C. Oliver and K.J.L Thome (1976), Bacitracin-induced changes in bacterial an support dm thickness in f: h'
techniques, J. Microscopy I Ii, . reeze-etc 1ng and related
plasma membrane structure, Biochim. Biophys. Acta 419, 570. 313
Sleytr, U.B., H. Groesz and W. Umrath (1981), Interpretation of morphological data obtained Steere, t R.L., E.F. Erbe and J.M. Moseley (1980) p f
, re- racture and cold fract · ·
by freeze-fracturing at very low temperatures, Acta Histochemica, Suppl. 23, 29. yeas p1asma membranes, J. Cell Biol. 86 - unng images of
113
Smith, P.R. and I.E. Ivanov (1980), Surface reliefs computed from micrographs of isolated Stolinski, C. (1975), A freeze-fracture repli~atio.n a a . .
roscopy 104, 235. pp ratus for biological specimens, J. Mic-
heavy metal shadowed particles, J. Ultrastruct. Res. 71, 25.
Smith, P.R. and J. Kistler (1977), Surface reliefs computed from micrographs of heavy m<ital·>' Stolinski, C. (1977), Freeze-fracture replication in biolo .
shadowed specimens, J. Ultrastruct. Res. 61, 124. p~acti~e and future prospects, Micron 8, 87. g1cal research: Development, current
Sommer, J.R. (1977), To cationize glass, J. Cell Biol. 75, 245a. , Stohnski, C., G. Gabriel and B. Martin (1983) R , '
ff 'e1n1orcementandprotect" "th
Southworth, D., K.A. Fisher and D. Branton (1975), Principles of freeze-fracturing and -eti?hf.i o reeze-fracture replicas during thawin a d d. . . ton WI po1ystyrene
Stolinski, C., L. Landmann and H. Abt (~ ; ) ~::~ion ::;issue'. J. ~icroscopy 132, 149.
ing, in: Techniques of biochemical and biophysical morphology, Vol. 2, D. Glick and 94
stage with a cleaver device for Balzers 'freeze-: h' ~ applicatto_n of a freeze-fracture
Rosenbaum, eds. (Wiley Interscience, New York, London), p. 247.
Staehelin, L.A. (1973), Analysis and critical evaluation on the information contained int Studer, D., H. Moor, V. Vogt and H. Gross l98~c in~ apparat~s, J. Microscopy 135, 317.
purple membrane, Proc. 7th Eur Re Co~f El), High r~soluhon shadowing of freeze-dried
etch micrographs, Freeze-etching, techniques and applications, E.L. Benedetti a
utherland, I. (ed.) (1977), Surfac~ ca!boh d. ectron Microscopy, !he Hague 2, 616.
Favard, eds. (Soc. Frant;aise de Microscopie :E:Iectronique, Paris), p. 113.
London). y rates of the prokaryottc cell (Academic Press,
Staehelin, L.A. and W.S. Bertaud (1971), Temperature and contamination deper1deJ1fi
etch images of frozen water and glycerol solutions, J. Ultrastruct. Res. 37, 146. esche, B. (1975), Ein Elektronenstoss-Verdam fer m"t . .
stellung diinner Schichten aus hochschmel p I ge?n~em Le1stungsbedarf fiir die Her-
Steere, R.L. (1957), Electron microscopy of structural detail in frozen biological
llack, T.W. and S.C. Kinsky (1973) A f zenden Matenalien, Vakuum Technik24, 104.
J. biophys. biochem. Cytol. 3, 45. , reeze-etch study of the effects f fil" .
Steere, R.L. (1969), Freeze-etching simplified, Cryobiology 5, 306. -and h un1an erythrocyte membranes B. him B' o 11p1n on liposomes
k TW ' ioc · iophys.Acta323,43.
Steere, R.L. (1971), Retention of 3-dimensional contours by replicas of freeze-fr ' . ., R.E. Scott and V.T. Marchesi (1972) Th
udied by freeze-etching II L al' t' f , e structure of erythrocyte membranes
mens, Proc. 29th Ann. Meeting EMSA, p. 242. . · · oc 1za ion o receptors f. hyt h , ,
1rus to the intramembranou t' J or P 0 emagglutln1n and influenza
Steere, R.L. (1973), Preparation of high resolution freeze-etch, freeze-fracture, fr . spar IC1es, . exp. Med.135 1209
and freeze-dried replicas in a single freeze-etch module, and the use of stereo ele ._osaki, A. and T.Y. Yamamoto (1974) Doub] . '. .
etina, J. Ultrastruct Res. 47 86. , e replicating method for the freeze-fractured
copy to obtain maximum infonnation from them, in: Freeze-etching, techniqU
cations, E.L. Benedetti and P. Favard, eds. (Soc. Frans:aise de Microscopie. 'si, M.R. and P. Pinto da Silva (1984) Co . .
Paris), p. 223. cocomponents is revealed by fracture label~i.a~~;~~~~~tt~~ of intracellular membrane
Steere, R.L. (1982), Reliable use of resistance evaporation of Pt and C for-,h·-· um-Jensen, J. and S. Bhakdi (1983), Freeze-fracture ana , . .
freeze-fracturing and a crystal surface image complementary to the E-face-- an complement, J. Cell Biol. 97, 618 lys1s of the membrane lesion of
ath, w c1977) P , , , , .
membranes, J. Microscopy 128, 157. rierat " M'k, nnz~p1en der Kryopni.parationsmethoden: Gefriertrocknung und ii
Steere, R.L. (1983), Preparation of freeze-fracture, freeze-etch, freeze-dry Zl.:,lg, I roskop1e 33 I 1. !;
th w ' ,
replicas for electron microscopy in the Denton DFE-2 and DFE-3 freeze:,e_ , W. (l978), Automattsche Gefrieratzung, Mikroskopie 34 6 11
rent trends in morphological techniques, Vol. 2, J.E. Johnson Jr., ed, ~ ' . (1981), An improved technique for the freeze f ' . '
s, J. Microskopie 1 , . - racture of cell cultures and mono-
Raton FL). 21 229
Steere, R.L. and E.F. Erbe (1979), Complementary freeze-fracture, f ' W. (1983) Berech
p .. . ' nung von Gefriertrocknungszeiten fiir die elektronenmikrosko-
Microscopy 117, 211. raparation, Mikroskopie 40, 9.
I I

458 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-fracture replication
459
Unanue, E.R., M.J. Karnovsky and H.D. Engers (1973), Ligand-induced movements of lym- ''Nildhaber, I., H. Gross and H Moor (1984) H" .
phocyte membrane macromolecules. III. Relationship between the formation and fate of electron beam evaporation ~d . b ' igh _resolution shadowing fihns produced by
anti-lg-surface complexes and cell metabolism, J. exp. Med.137, 675. ion earn sputtenng Proc 8th E R
Microscopy, Budapest 2, 128 5 . ' · ur. eg. Conf. Electron
Vail, W.J. and J.G. Stollery (1978), A convenient method to remove lipid contamination from
Williams, M.~. (1_977), Autoradiography and immunoc oc . . . .
freeze-fracture replicas, J. Microscopy 113, 107. electron microscopy Vol 6 AM GI _yt hetnistry, in. Practical methods in
Van Harreveld, A. and J. Crowell (1964), Electron microscopy after rapid freezing on a metal . . ' · , · · auert, ed. (Elsevier/North H 11 d A
WJ!hams, R.C. and K.M. Smith ( 1958) Th - o an , msterdam).
surface and substitution fixation, Anat. Rec. 149, 381. . . ' e polyhedral form of the r· l . •.J •
ch1m. B1ophys. Acta 28, 4 64 _ lpu a tnuesent vrms, Bio-
Van Winkle, W.B. and M.L. Entman (1981), Accuratequantitation of surface area and particle Williams, R.C. and R.W.G. Wyckoff (1944) Th .
density in freeze-fracture replicas, Micron 12, 259. appl. Phys. 15, 712. , e thickness of electron microscope objects, J.
Verkleij, A.J. (1984), Limitations and perpectives of freeze fracturing and freeze etching, Proc. Willison, J.H.M. and R.M. Brown (1977) A . .
8th Eur. Reg. Conf. Electron Microscopy, Budapest 3, 1729. , n exatnination of the develop· u fib
_ar~ d p Iasmalemma, Protoplasma 92 , 1. ing co on 1 re: wall
2
Verkleij, A.J. and P.H.J.Th. Ververgaert (1975), The architecture of biological and artificial Wtlhson, J.H.M. and R.D. Moir (1983) Bid" . . .
membranes as visualised by freeze-etching, Ann. Rev. Phys. Chem. 26, 101. copy 132, 171. ' rrectional shadowmg In freeze-etching, J. Micros-
Verkleij, A.J. and P.H.J.Th. Ververgaert (1978), Freeze-fracture morphology of biological Willison, J.H.M. and A.J. Rowe 0980) R r . ' I
membranes, Biochim. Biophys. Acta 515, 303. Practical methods in electron m· , epVica, shadowing and freeze-etch _techniques, in· ' I

Verkleij, A.J., B. de Kruijff, W.F. Gerritsen, R.A. Demel, L.L.M. Van Deenen and P.H.J.Th. icroscopy ol 8 A M Gl t d (El . · ·
land, Amsterdam). ' · ' · · auer, e · sevierJNorth-Hol-
Ververgaert (1973), Freeze-etch electron microscopy of erythrocytes, Acholeplas1na /aidlawii Winkelmann, H. and H.W. Meyer (I968) A . f .
cells and liposomal membranes after the action offilipin and amphotericin B, Biochirn. Bi(l~­ .
bYsimple , routine reeze-etchingtechn' fh' h ~
technical means I Th . . E ique o ig ·euectivity
. . . e pnnc1p1e, xp. Pathol. 2 277
phys. Acta 291, 577. ·;WmkeJmann, H. and S. Wammetsberger (1968) Eine . m1t. '.
Verkleij, A.J., C.J.A. Van Echteld, W.J. Gerritsen, P.R. Collis and B. de Kruijff(l980), Thy~ e ti . h M.
Routinegefrieratztechnik hoher Effektivit'"t Ir' D" . Ill ac en 1tteJn durchfiihrbare
lipidic particle as an intermediate structure in membrane fusion processes and bilayer to he~~:: 113. a. . ie techn1sche Anordnung, E:;<.p.~Pathol. 3,
agonal HI! transitions, Biochim. Biophys. Acta 600, 620. · ingsheim, H.P., R. Abermann and L. Bachmann 197 . .
Ververgaert, P.H.J.Th. and A.J. Verkleij (1978), Lipid intermembranous particles, Proc. 9 freeze-etching technique J Phys E· , ( Oa), An ultra-shadowing. unit for the
. . , · · . sc1ent. 1nstrum. 31 39
Int. Congr. Electron Microscopy, Toronto 2, 154. gshe1m, H.P. and H. Plattner (1976) Electron . ' .· ,
Waldhausl, P. (1978), An approximate solution for the restitution of stereo electron , M h , m1croscop1c methods in b .. b.
m: et ods in membrane biology, Vol. 7 ED· m~m- r_an,e 101ogy,
graphs, Photogrammetr. Eng. Remote Sensing 44, 1005. New York), p. I. ' · · Korn, ed. (Plenum Pubhshrng Company,
WalzthOny, D., H. Moor and H. Gross (1981), Ice crystals specificalJy decorate hydr, gsheim, H.P., R. Abermann and L Bachm (I
sites on freeze-fractured model membranes, Ultramicroscopy 6, 259. Proc. 7th Int. Congr El · ann 970b), Shadow casting and heat damage,
ectron Microscopy, Grenoble l, 411
WalzthOny, D., H. Gross and H. Moor (1982), Contamination sources in freeze-fr'!:~-<--·
Proc. 10th Int. Congr. Electron Microscopy, Hamburg 3, 213. ,I
Washburn, E.W. (1924), The vapour pressure of ice and of water below the freezi!lg,
Monthly Weather Rev. 57, 488.
Wehrli, E., K. Miihlethaler and H. Moor (1970), Membrane structure as seen with
replica method for freeze-fracturing, Exp. Cell Res. 59, 336.
Weibel, E.R., G. Losa and R.P. Bolender (1976), Stereological methods for estimati
membrane surface area in freeze-fracture preparations of subcellular fractions,
copy 107, 255.
Weinstein, R.S., W.D. Selles and I.T. Young (1976), Quantitative analysis of ttt~ll!.-
graphy in normal and malignant epithelium, Proc. 6th Eur. Reg. Conf. Bl
copy, Jerusalem 2, 104.
Weinstein, R.S., D.J. Benefiel and B.U. Pauli (1979), Use of computers in the a,
membrane particles, in: Freeze-fracture: methods, artefacts and interpret_a_
and C.S. Hudson, eds. (Raven Press, .New York), p. 175.. , , --<:,
Westmeyer, H. and E. Lorenz (1960), Die Herstellung van Aufdampfschi
kleiner elektronenstossgeheitzer Dampfquellen, Optik 17, 244. :''--' ,,
Wildhaber, I., H. Gross, and H. Moor (1982), Surface structure preservatio_r(~­
dried RPI-layer, Proc. 10th Int. Congr. Electron Microscopy, Hamburg-:3;_-
'I
Chapter 7
:I

Freeze-substitution and
low temperature embedding

7.1 Introduction

Freeze-substitution as an electron microscopical preparation technique is


based on rapid freezing of specimens followed by snbstitution (solution) of
the specimen ice by an organic solvent at temperatures which inhibit damag-
ing recrystallisation of the frozen water in the system. Generally, the substi-
tuted specimens are subsequently infiltrated with resin monomers and, after
polymerisation, sectioned. Alternatively, substituted specimens are dried by
critical point-drying or freeze- or evaporation-drying methods (Chapter 5).
As with all other low temperature techniques, it is the ultimate goal of
freeze-substitution to be able to examine specimen structures and specimen
constituents trapped by freezing. The basic technique was first described by
Simpson in 1941 for preparing tissue for light microscopical studies and
several modifications of the technique for light microscopy followed (see re-
iew by Feder and Sidman 1958).
Fernandez-Moran (1957, 1959a,b) was the first to test freeze-substitution
a technique for electron microscopical studies. He primarily used helium
at about -271°C (2.0 K) as a quenching fluid for pre-glycerinated tissue
mples. The frozen specimens were substituted in alcohol, or in alcohol/
etone/ethyl chloride mixtures, at temperatures varying from -130°C to
. 80°C (143 to 193 K). Stains and/or fixative substances (e.g. platinum
oride, osmium tetroxide, gold chloride) were sometimes added to these
bstitution media. Substituted specimens were then infiltrated with mixtur-
of methacrylate at temperatures between - 100 and - 75°C (173 and 198
and, finally, the polymerisation was initiated by UV light using benzoin
a catalyst at -80 to -20'C (193-253 K) (Fernandez-Moran 1960,
la,b).
461
462 Robard~ and Sleytr Low temperature methods in biological electron microscopy
Freeze-substitution and low temperature embedding
463

Subsequently, Bullivant used Fernandez-Moran's technique but modified Specimen - Pretreatment: e.g.
the procedure by introducing methanol and ethanol as substitution media chemical fixation,
cryoprotection
Freezing
and also by quenching in liquid propane at -175°C (98 K) since this gave ,---_-----:-------1
1 Optional additives: e.g. 1
much better cooling rates than helium II (§2.2.5). Bullivant also abandoned I fixatives, I
- - - ---j staining components, I
the low temperature polymerisation of the methacrylate embedding medium I activated molecular sieve, 1
:_ .!l:'~~p~t~t~o~ ~e~~~t~ _:
(Bullivant 1960, 1965). At the same· time, others became involved in devel- Infiltration with embedding
medium (at room temp. or change to other
oping freeze-substitution procedures for electron microscopical studies, at 243K to 213K using organic solvents
low temp. resins)
usually using acetone or ethanol/acetone mixtures for substitution (Van
Critical point-drying Evaporation drying
Harreveld and Crowell 1964; Van Harreveld et al. 1965; Rebhun 1972). An Polymerisation (al room
tamp. or low tamp.)
{see Chapter 5)

extensive review of these early developments is given by Pease (1973), who


srn
had himself been exploring the potential use of small glycols as substitution Standard thin
sectioning procedure
CTEM

media for 'inert' dehydration at room temperature as well as for freeze-sub- Sectioning under anhydrous _ - - For: localisation ol water soluble
conditions {dry sectlor:ing) >1ubstances by analytical EM
stitution (Pease 1966, 1967, 1968). As already stressed by Pease (1973), the STEM or CTEM
{po•nibillty for (X-rny m!croanalysin, or
autoradiography) or LAMMA
reason why these early attempts (and in many respects this remains applica- cytochemical studies)

ble to the present state of the art) did not culminate in a single methodology Fig. 7 .I. Sequences of electron microscopical procedures involving freeze-substitution.
lies in the fact that three distinct, but intertwined, decisions are involved in
the freeze-substitution technique:
,,,",.:_._;
biological specimens with an average water content, this critical temperature
(i) the selection of the best method of specimen freezing, including pre~ . •
can be expected to lie within the range -100 to -80cc (173 to 193 K)
treatment (although there remain difficulties, recent developments:,,,,
(§2. l.3) remembering that the recrystallisation rate is itself a temperature-
have produced good techniques for many systems, see Chapter 2); . ,'r,'
dependent phenomenon. For many biological specimens much higher tem-
(ii) the question of which substitution liquid should be used, for wh19~F
peratures might be safe for a short period of time (Steinbrecht 1982). Several
period of time, and at what temperature and, also, whether fixati~,~
substitution liquids and dissolved additives (fixatives and staining and preci-
and/or staining components should be added. Cons1dermg alL;
pitating agents) have been used. As discussed below, the choice strongly
potential applications of freeze-substitution (Fig. 7.1 ), it is likely,;':
depends on the scientific questions posed (e.g. ultrastructural or analytical
different processing schedules will be required for optimal results
_electron microscopical studies). After completion of freeze-substitution, the
different specimens; . '1ubstitution liquid may be exchanged against another solvent which is more
(iii) the choice of the best resins and infiltration procedures to provide; suitable for the subsequent stage of the preparation procedure.
sections for ultrastructural, analytical and cytochemical studies·
The procedures for subsequent handling of freeze-substituted specimens
recent .development of the polar and non-polar methacrylate,,
re outlined in Fig. 7.1. For straightforward ultrastructural studies, substi-
low temperature resins (§7.5) has led to considerable progress
ted specimens are infiltrated with embedding medium at room tempera-
area. re or at -30 to -60°C (243 to 213 K), and, after polymerisation of the
in, are thin-sectioned using standard ultramicrotome techniques (see
The different electron microscopical procedures involving a freez
eid 1974). Alternatively, sectioning under anhydrous conditions (dry sec-
tu ti on step are shown in Fig. 7 .1. Following specimen pretreatm
pning) (§7.4.2) can be used if water-soluble substances are to be localised
and freezing (§2.4) the specimen ice is substituted by an organic sol
means of analytical electron microscopy. Freeze-substituted specimens
ble 7.1). As a basic requirement, the substitution temperature must
n also be critical point-, evaporation- or freeze-dried directly for exa1nina-
tained_~below the recrystallisation temperature of the specimen ice._,_
_n in the scanning or transmission electron microscope (Chapter 5).
464 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-substitution and low ten1perature embedding
465

TABLE7.l 7.2 Substitution media


Liquids used in freeze-substitution"
· - - - - - - - --· - - · - -
Formula Melting Point Only those liquids that can dissolve (substitute) the specimen ice at tempera-
Solvent (+synonyms)
tures below the recrystallisation temperature can be used for freeze-substitu-
·c K tion. To minimise the extraction or migration of soluble specimen consti-
tuents, it is desirable to have the specimen in contact with the substitution
- 117.J 155.7
liquid for the shortest possible time at the lowest practical temperature. In
Ethanol C 2H50H
addition, time becomes particularly important if freeze-substitution is to be
(alcohol, ethyl alcohol)
a practical alternative to the standard chemical fixation/thin-sectioning pro-
Methanol CH,OH -93.9 179.l cedure. As a more specific requirement, it must be possible to dissolve in
(methyl alcohol) the substitution liquid components which stabilise the overall specimen
-95.4
structure (fixatives) or help to 'stain' specifically, or otherwise localise (e.g.
Acetone CH3 COCH3 177.6
by precipitation - see §7.4.la), certain specimen constituents.
(2-propanone, dimethyl ketone)
To our knowledge, very few attempts have so far been made to determine
Diethyl ether CH 3CH 20Cll2CH3 -116.2 156.8 the temperature dependence of the rate at which ice dissolves in different
(ether, ethyl ether) substitution media. Zalokar (1966) used a model system consisting of pieces
- 126.5
of Millipore filter soaked in an aqueous solution of methylene blue. These
1-Propanol ClI 3CH 2CH20H
pieces were frozen and then placed in various substitution liquids. The com-
(n-propyl alcohol)
pletion of the substitution process was determined from the decolouration
1-Butanol CH 3CH2 CH2CH20H - 89.5 of the filter. According to Zalokar's observations, acetone is unable to dis-
solve ice during several days at the temperature of solid CO, (dry ice)
-46.7
(- 78°C; 195 K). Alcohols are much better substitution liquids, and metha-
1-Hexanol CH 3(CH2)sOH
nol is preferable to ethanol. A mixture of methanol and acrolein is even
(n-hexyl alcohol)
more efficient. Based on these results, Zalokar used for his work three parts
Hexane CH 3(CH2)4CH3 -95.0 of absolute methanol containing 20% uranyl acetate and one part of acro-
jein. Acrolein has to be added to the pre-cooled methanol to avoid the rapid
- 112.1
rmation of acetal. Thus, Zalokar's results from his model system argue
Propylene oxide C 3 H"O
rongly against the use of acetone as a suitable substitution liquid. Neverth-
(1,2 epoxy propane)
ess, excellent results have been obtained from its use (e.g. Steinbrecht
Tetrahydrofuran CJfeO - 108.5 80, 1982). The reason for this discrepancy and for similar controversial
servations in the literature relating to other substitution fluids, can be
-87.0
lained as resulting from variations in the following parameters govern-
Acrolein CH 2:CHCHO the substitution process:
(Propenal, acrylaldehyde)

Ethylene glyco\b HOCH 2CH20H the water content of the substitution liquid at the beginning, during,
and at the end of the substitution process;
the temperature and water content of the substitution liquid at which
. tr and Physics' 60th edition (1980).
a Data taken from 'Handbook of Chem1s y .. ' . h. h 'reeze< ~' an equilibrium with the specimen ice is obtained (no substitution
b 70% Ethylene glycol + 30 % water gives 88) w tc •·
a .eutectic m1xture
0 .

should occur below this temperature);


(223 K) (Pease 1967a; see also Barlow and Sleigh 1979, p. .
466 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-substitution and low temperature embedding
467
(iii) the presence of a molecular sieve which constantly removes water
from the substitution liquid (agitation of the substitution liquid to ... 100 ·50%
243±2K (-30"C)
compensate for the reduced diffusion velocity of the water molecules 243:!:2K 1-ao"cJ

.--·&a;:
in cooled (cold) substitution liquids may be of great importance); and
---·
50 ___

(iv) the rate at which the specimen, immersed in the substitution liquid,
is warmed up (as discussed below, the actual substitution may only,
.:
~

:
.~--o,;------;--~~-~-
4 8 t {h)
. ,·" ~·

: 0 2
or mainly, have occurred during this final stage). ~ 100 ! 8 I (h)

..~.. l
'/
,{
213:!:2K {-BO"C) ~ 100 1ij/•ao%
5 10%

213:!:2K (-60'C)

For methanol, acetone and diethyl ether, the most accurate data on sub- 0 50
.,£ l/
0
Ci 50 /
stitution of ice as a function of temperature and water content have been 0
0 0

~
obtained by Muller (personal communication), Humbel et al. (1983) and
Humbel and Muller (1984) (Fig. 7.2a-c). The purest chemicals (Merck) were
0


•""""'=;c====,!====~====t::::
100 /
8 t fh)

.:•
>
0 ~·~·-,,,·-=-·_,,-~·-=--=--=--=;~~----"""'-'""-_·~'_:':'
O'i> 5 '1>
8 t {h)

used. Acetone and methanol were dried with activated molecular sieve £ I.~ :
:; 100
;,:;:~J-10%

,. {1· /
183:!:2K (-90"C) 1B3:!:2K (-90'C)

(Linde 0.3 nm; Merck) and diethyl ether containing a piece of sodium. Piec- 0
20%

es of filter paper (approximately 3 mm in diameter) were soaked in l.O µl


l ,. JI/ "---·
of tritiated water (l.O mCi ml- 1), frozen in liquid nitrogen and substituted .~~~~~.~~~,~~~.=.=!"
at -30, -60 and -90°C (243, 213 and 183 K) in 1 ml of substitution liquid.
0 Methanol
Time ol substitution
• Time of substitution
The substitution curves shown in Fig. 7.2a-c were determined by analysing •Acetone o Ether

the release of tritiated water with time. In another experiment using frozen:>;
aqueous solutions of methylene blue, Miiller confirmed the basic observ~:::'.--' ... / 15'1)
l'·.1
'

tion of Zalokar (1966) that methanol substitutes better than acrolein, whiCh. 1CIO '//20%
! -~·
- - - · 25 %
'I
1:· I
is better than acetone, and that diethyl ether is the worst (Muller et al. 19 i!
50 ' . / Fig. 7.2. Substitution of ice by methanol, ace-
Muller, personal communication). Thus, the most important information< 243±2K {-30"C) I ,I
'

tone and ether as a function of temperature and


be derived from Fig. 7.2a is that methanol is the fastest substitution liq 1'1 .I
water content of the substitution media. The
at a temperature safely below any possibility of recrystallisation ( -9 • ' 8 t (h)
d~t.a were obtained by analysing the release of
1''!.1
il·,1
183 K). Nevertheless, both acetone and diethyl ether also substitute ice( tntiat~d ~ater :rom
the frozen specimen. (a)
i) ,1
II
slowly) at this low temperature provided that they are absolutely dry. Substitution of ice as a function of organic sol-
bel et al. (1983) and Humbel and Muller (1984) also examined the su vent and substitution temperature. Methanol is
the :astest substitution liquid at 'crystallisation-
tion properties of methanol and acetone in relation to their water
safe temperatures. (b) Substitution of ice b
.~--;------o,--7----C,-,-IM
for the temperatures -30, -60 and -90°C (243, 213 and 183 K methanol as a function of water content an~
found that methanol (Fig. 7.2b) substitutes ice at -90°C (183 K) temperature. Methanol substitutes ice at
it contains 10% water. Using the tritiated water model system, it was ~90~C even ifit contains IO% water. (c) Sub-
that a solution of methanol which contains 10% water is capable.o stitution of ice by acetone as a function of
water content and temperature. In comparison
tuting water completely within 3 hat -90°C (183 K) (Fig. 7.2b). In
t~ methanol (b), acetone is a much poorer sub-
ison, acetone containing 1% water substitutes the same test samp_ . 4 6 8 I (h) stitution liquid. (Redrawn from Humbel and
75% after 8 h at the same temperature (Fig. 7.2c). The reason w T11ne of substitution
MUiler 1984.)
is a poor substitution fluid at low temperatures is evident fro
which shows that the amount of water in acetone in equilibrj,
at -85°C (188 K) is only about 2%. Thus, if this water is no
Il
if,
I
££%

468 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-substitution and /ow temperature embedding
469

Temperature (K)
220 210 200 190
1S

~
c 10
'·~. ~
!c
0
0 ~. ;; so
.!!
! s ~ 0
:c
N


3:

0
·-----·-----·--·-· "
• Small ice crystals
o Large ice crystals
-4S -SS -6S -75 -as
Temperature ("C)

Fig. 7.3. The water content (volume percent) of acetone in equilibrium with ice as a function Hours of substitution
of temperature. (Redrawn from Van I-larreveld et al. 1965.)
Fig. 7.4 . Efflux of tritiated water fron1 quick-frozen (helium-cooled metal block· small .
crysta 1s1 and quench-frozen (Freon 22 . 1 . . . ' ice
. ' arge ice crystals) ferntJn solution at -83"C ( 190 K)
Jn acetone. (Redrawn from Ornberg and Reese 198L)
removed, or if in relation to the amount of specimen ice the volume of sub-
stitution liquid lias not been properly calculated, the substitution liquid cal'!
soon reach saturation level. It is frequently suggested that molecular a. fact which he explained by assuming that the rate-limiting step is the diffu-
be added to bind the water liberated from tlie frozen sample (Harvey et sion ~fwater.through the solvent and not the rate of dissolution of ice. To
198lb; Humbel and Muller 1984). Humbel and Muller (1984) have report<: explain the difference shown in Fig. 7.4, he would suggest that the lower
that a 0.3 nm molecular sieve is very efficient in removing water from diethi' freez1ng ~ate _gives more time during the cooling of any specific volume for
;recrystalhsat10n, or grain growth which is a , l' ff
ether saturated with tritiated water at 213 and 183 K (-60 and -90~ 'b h. · ' sea mg-o process (the
To our knowledge no systematic studies have yet been done which W< ran~ ing ice 'necks-off and produces separate co1npartments) while faster
confirm how active sieves are in other substitution liquids at low subs freezmg leaves a continuous skeleton through the spaces of whi~h the sub f-
ution proceeds. sI
tion temperatures, but it can be expected that the addition of molecular.
es supports the substitution capacity of other organic solvents. ,. In combination "".'ith_procedures for the localisation of water-soluble sub-
According to Ornberg and Reese (1981), the substitution rat~ tances,·dother subsl!tut10n media have also been used (e ..
g n-h exane propy-
depends on the ice crystal size within the specimen. These authors_ -_- ne ox1 e and ethanol; see review by Harvey 1982). However, to ou~ know-
mined the substitution rate for acetone at - 83°C (190 K) by measur· ge, no accurate data, such as those available for methanol acetone a d
thyl ether, have been published. ' n
efflux of tritiated water from ferritin samples that had been frozen i
22 or on a helium-cooled metal surface (§2.4.2) (Heuser et al. J.9
shown in Fig. 7.4, the ferritin samples frozen on a helium-cooled; Freeze-substitution apparatus
(which have small ice crystals that extend much deeper into the pr
than Freon-frozen specimens) substitute much faster than the Fre ipmcnt for freeze-substitution must satisfy the follow1'ng requ1re1nents:
.
samples. However, the freeze-substitution rates measured by Ma :~ ~ecimen must be kept immersed in the substitution medium at
Appendix to Ornberg and Reese 1981) using a light microscope e1;ned, controlled temperature (between 193 and 178 K· -80 d
cated that the substitution rate is independent of gross ice cryst, - 95 C) for up to several days; ' an
470 Robards and Sleytr Low temperature methods in hiological electron 1nicroscopy Freeze-substitution and low temperature embedding
471

(ii) during the final stages of substitution it should .be possible to control ~---- Temperature measurement
the rate of temperature rise (up to about ambient temperature) and I r - - - - Supply for heater plate
also to maintain certain temperature levels constant over defined peri-
- 1 - - - L i d of Dewar vessel
ods of time; . .
(iii) the substitution medium must.be protected from atmospheric mois-
ture; and
(iv) it should be possible to manipulate both small (e.g. isolated cells) and
tissue specimens with relative ease.

Either compressor or liquid nitrogen cooling systems can meet the tem-
perature requirements in (i) and (ii). .
· ·
The simplest type of freeze-subsl!tul!on eqmpmen t , which at the .same
Dewar vessel
time meets the most important parameters mentioned. above, was des1gn~d
by Sitte (personal communication). As illustrated in Fig. 7.5, a small plastic Fig. 7.6. Simple freeze-substitution in a Dewar vessel over liquid nitrogen (cooled by liquid
·
container fit · t o a 1'oam
1 sin ' - insulated metal container.
. After
. the· container has
l nitrogen vapour). The substitution vessel is connected to the lid of the liquid nitrogen con-
been filled about three-quarters full with the subst1tut10n medmm, the who e tainer. A temperature sensor-controlled heating plate ensures that the substitution liquid ls
assembly is cooled in liquid nitrogen. The frozen spec1m.en 1s then placed maintained at the desired temperature. (Redrawn after Sitte 1981.)

on the surface of the frozen substitution medmm. If freezmg took place


a cold block (e.g. a metal mirror; §2.4.2), the best frozen part should termines the rate at which the whole assembly gradually warms up, thus
placed face-downwards. After closing the contamer, the .metal allowing the specimen to slowly sink into the molten substitution liquid.
covered with the insulated lid. The insulation charactenst1cs of the foam d With this assembly, it is not possible to maintain the specimen at defined
·temperatures during the warm-up period unless the container is placed in
Plastic container
a refrigerator kept aHhe desired temperature. It can be anticipated that, for
''Inany systems, the well-frozen border zones are substituted in a very gentle
ay. As the. substitution temperature increases, additives, such as fixatives
Ild/or staining components, interact with the specimen.
Specimen Another simple freeze-substitution apparatus proposed by Edelmann (see
Metal block
itte 1981), illustrated in Fig. 7.6, consists ofa substitution vessel which can
Grid
connected to the lid of a liquid nitrogen container. The substitution vessel
not in contact with the liquid nitrogen surface but is constantly cooled
the liberated nitrogen vapour. A temperature sensor ensures_ that the
ting cartridge at the bottom of the vessel maintains the substitution
Frozen cryofixed object Melting 1ub1tltut1on
id at the desired temperature. Sitte has stated (personal communication)
medium the liquid nitrogen consumption of this freeze-substitution unit is not
h higher than the evaporation loss of the standard vessel. Both of the
Fig. 7.5. Simple freeze-substitution metho d using· a ~o am -insulated epre-cooled
of the rr· >
-substitution units shown in Fig. 7.5 and 7.6 can easily be made using
al workshop facilities.
(cooled by liquid nitrogen). The fr.ozen specimen is placed on thed:u:i~ac warms uPi
tion medium in the plastic container. The whol~ as~em~ly .gra d/awn after·Sif( ,, r some freeze-substitution procedures, a dry ice/acetone bath with a
specimen to sink slowly into the molten substitution hqu1d. (Re :·:·.:;'.· ::· erature of about -78°C (195 K) may be adequate, particularly if the
472 Robards and Sleytr Low temperature methods in biological electron 1nicroscopy
Freeze~substitution and low temperature embedding
473
Plaatic container

Control unit
for cryostat i====
••••
Cryostat

-warming ---Exchange of
subotitution medium
{a.g. to methanol)

I
Critical point-drying
a
Dry ice/acetone in C0 2

Substitution
medium
195
bath

I( (-78 °C) I
304.1 K 8.3 X 1015 Pa
(1200 psi)

Substitution

Fig. 7.7. Simple freeze-substitution procedure in a dry ice/acetone bath for both single ce!!s
and small pieces of tissue. Frozen specimens are placed in modified srnall centrifuge tubes
which are placed in small glass tubes containing the pre-cooled substitution medium. (Redrawn
after Barlow and Sleigh 1979.)

substituted specimen is to be further processed (e.g. by critical point-drying)


for examination in a scanning electron microscope. The flow diagram of a·
preparative procedure recommended by Barlow and Sleigh (1979) for both,
single cells and small pieces of tissue is illustrated in Fig. 7. 7. The substitu;,
tion is carried out in a 9 mm (outer diameter) plastic tube (e.g. an embed; Fig. 7.8. Freeze-substitution apparatus developed by Miiller (Millier et al. 1980). (a) Overall
dmg capsule, as suggested by Newell and Roath (I 975), or a plastic cen -~iew of the apparatus. (b) Rack used to support the substitution pots at different temperatures.
fuge tube) with the conical end cut off and the other opening covered Illustration compiled from photographs kindly provided by M. MUiler and redrawn after
a 20 µm nylon mesh filter. The modified containers are placed in narr. Miiller et al. 1980.)
glass specimen tubes containing 3.0 ml substitution medium cooled.<
immersion into a dry ice/acetone bath. The frozen specimens are then t~:. If, during the course of freeze-substitution, defined temperatures are to
ferred to the substitution liquid. Freeze-substitution and any subseq maintained according to specific time schedules, an automatically con-
fluid exchange is carried out by draining and filling the tube outsideth: lled apparatus is most desirable. One such apparatus, which automati-
tainer. Since the tube containing the substitution liquid is not clo ly controls substitution time and temperature, has been developed by
will dissolve in the substitution fluid. If, in the course of prepar ··11er (Miiller et al. 1980). As illustrated in the overall view shown in Fig.
substitution liquid is warmed too rapidly to room temperature, t , the complete device consists of a Jiquid nitrogen Dewar vessel, the
men may be damaged by liberation of CO, gas bubbles. AlthougJ;i stat unit and a control unit. The solid state control unit is designed to
gested by the authors, this problem could easily be overcome b constant three presettable temperatures (between 0 and -200"C; 273
closed container instead of a glass specimen tube. The techn1qu~ 73 K) for three presettable times (between 1.0 min and 100 days). The
cally designed for freeze-substitution followed by critical point erature is measured using a 100 ohm platinum resistance element and
analysing delicate structures (ciliated surfaces) in the scanning e. ulated within ± 1.0 K by balancing the supply of liquid nitrogen from
roscope (§7.4.lb). ·war vessel with heat from an electric element. The actual te1nperature
474 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-substitution and low temperature embedding
475
. . . . .cular temperature level (i.e. within a con-
of the substitution hqmd, the part1 ) d the time elapsed since the start 7.4.la Freeze-substitution for thin sectioning
. . f d.f' r nt temperatures an
secut1ve series o I ie e I ·t The cryostat
ll d. I ed on the contro um ·
of this temperature level are a isp ay . h. h ck (Fig 7 Sb) sup- A variety of schedules has been suggested for freeze-substitution of biologi-
. 1 d ta·ner into w ic a ra . .
itself consists of an msu ate ~;sn T~e plastic centrifuge pots will hold 3.0 cal specimens prior to plastic embedding and thin sectioning (see, for exam-
porting the subst1tut10n pots . d .th a Teflon lid during the substi- ple, reviews of Rebhun 1972; Pease 1973; Harvey 1982). The most common
. . d' and are covere w1
ml substitution me ium . . the frozen samples are transferred criteria used to assess the quality of freeze-substitution results are:
. - For freeze-subst1tutlon,
tut10n process. f . ) .nto the Teflon pots which each
from liquid nitrogen (storage af~~O;~_z;~; ~)substitution liquid.
(i) a dense appearance of the cytoplasm, indicating that little extraction
of specimen components has occurred;
contam 1.5 ml pre-cooled (e.g. 'h I gical studies developed by Miill-
The substitution proced~re for mo! rpl9o80o Humbel et al. 1983) using this
(iz} that the volume of different cytoplasmic compartments remains
unchanged;
d h. . orkers (Muller et a. '
er an is co-w . .b d in detail below. It is anticipated that this type of (iii) that cell membranes are sharply distinguished and continuous; and
apparatus IS desert e f bstitution becomes a reproducible, and that there is no 'healing' of the ice crystal damage caused by specimen
apparatus will ensure that ~eeze;s~ morphological studies of cryofixed bio-
(iv)
freezing (Dempsey and Bullivant 1976; Steinbrecht 1980).
co~sequently _routine, proce u~t ~here is little commercial apparatus avail- Fixation
log1cal matenaL At the mome
able for freeze-substJtut10;;h:s f~~~:~~e.
relatively mfrequent use
0
d
No:
·ch has robably contributed to the
that the criteria for optimal
d e believe that more commercial
One of the major problems in ultrastructural work is to stabilise (fix) the
cellular (specimen) components after the ice has been removed, i.e. to reduce
b · t" e better un erstoo , w ,
freeze-su stJtu.ll 10n the extraction of low and high molecular weight cell constituents to a mini-
b .art duced and t h at I.t w1.11 ' wi.th the possible exception
equipment w1 _e mtro ling systems instead of liquid nitrogen, be simil_ar mum. While gross structures may frequently remain wel1 preserved, organic
of usmg closed c1Icm coo . b M ·· 11 t al (1980). solvents remove considerable amounts of lipid and affect protein conforma-
d. tly in a low temperature cabme~.
. · th aratus descnbed Y u er e · .• ·
m design to e app d tion (Glauert 1974; Hayat 1981). Thus methanol is a preferable substitution
Freeze--substitution can also be one i~ec d n to -900_. fluid to acetone because it is a less effective lipid solvent (Glauert 1974).
. f each1ng temperatures ow .
(deep freeze) with _the capacityh or ~d·f nal advantage that the substituf; Consequently, for morphological studies it becomes obligatory to add both
(183_ K). Such eq~pmen~~;sbte ::ry ~~~ily processed for low tempera! lipid and protein fixatives to the substitution medimn. Aldehydes have been
:::~~~~~:gc~s::~ 7 s~~~~cluding polymerisation at low temperatures, shown to be the most suitable protein fixatives for standard chemical fixa-
tion (Glauert 1974) and so acrolein or glutaraldehyde are often added to
UV light.
the substitution media, the former in quite high concentrations (e.g. 30%
/v; Zalokar 1966). In addition to aldehydes, heavy metal salts (such as ura-
yl acetate or osmium tetroxide) can be added to stabilise and stain lipids
7.4 Freeze-substitution procedures d other specimen components.
In order to optimise freeze-substitution schedules it is important to know
7.4 .1 Ultrastructural studies
teinperature-dependent rate of interaction of fixatives with specimen
nstituents. The dissolved fixatives and staining media penetrate the speci-
· · tm:d'a
~f and
fine. addifr
For morphological studies, different subst1tut10n
have to be used to obtain the opt11~1u~ preserva electron I:
strl\qt

cytological details for either transm1ss1on or sfcann fg acture rep,····.


niigf;
en along with the substitution front. Nevertheless, they cannot be
pected to be chemically active at the commonly used ('recrystallisation-
fe') substitution temperatures (,,:; 80°C; 193 K). For this reason, most sub-
It obtained from reeze- r J
For many systems, resu s . ( Chapter 6) can serve• ution schedules include a stepwise or gradual temperature increase after
d oprotected specimens see . 11.
unfixe , non-cry . b ht l976- Browning aif initial, low temperature, substitution period. So far, few systematic stu-
references· and comparisons (e.g. Stein rec ' ::..~ ,I ,;!
have been carried out to determine the 'activity' of fixatives at different 1:1 .i
1

1979; Hunziker et al. 1984).


1,
1·.I 1
1
i
1
1:1 ;1
f,1
1
!1
476 Robards an d SIey tr L ow temnerature
1' methods in biological electron microscopy Freeze-substitution and low temperature embedding
477

temperatures. Miiller and his colleagues (perso~al co~m~:~~:~~:ha:~ pective of the fact that the number of cross-linking reactions to prevent
Humbel et al. 1983) have investigated the proper\Jes of 3% g . y structura1ly observable extraction of protein from an 'average' biological
in methanol The fixating capacity of this common cross-hnkmg relagtent specimen during freeze-substitution remains to be established, Fig. 7.9
was determined· by mJectmg
. . . 10 .0 µ 1 o f a 2°1
lo bovine serum
. albumin sou90'C 10n clearly shows that there are characteristic temperature-dependent fixation
(BSA· Merck) in a 3%0 solution of glutaraldehyde m methanol at - d . properties for glutaraldehyde in methanol and that, even at -50°C (223 K),
(183 K).
' The fixative was then warme d up t o the temperatures 1nd1cate in about 50% of the protein can be expected to be cross-linked within 24 h.
Fi . 7 .9 and maintained there. The reaction of the glutaraldehyde was No similar systematic studies have been carried out on Os0 as a fixative,
st:pped by removing the substitution medium and subsequently m w~hmg 4
but blackening of tissue does not occur below -25°C (248 K) (Van Harre-
re methanol at the temperature of the experiment. The cross-hn ed pro- veld et al. 1965) and significant reactions with unsaturated lipids can only
~e~n was pelleted in a Beckman Microfuge B (which can be operated at tem- be expected above -30°C (243 K) (Muller et al. 1980 and Muller personal
P eratures down to - 9o "C·' 183 K) · The specimen was then warmed toII room I communication). However, both uranyl ions and Os04 can be expected to
temperature an d was h ed WI"th double-distilled water to ensure
· that a · sol u-
N react with lipids to some extent even at 183 K ( -90°C) (Humbel et al. 1983).
ble protein. was removed . The cross-linked protein was dissolved . rn Based on his systematic studies both on different substitution media (Fig.
NaOH and the protein content determined by the Lowry techmque. 1rres- 7.2) and on different additives (Fig. 7.9), Muller and his co-workers (Miiller
et al. 1980, Humbel et al. 1983) concluded that 1% Os0 4 , 0.5% uranyl ace-
tate and 3% glutaraldehyde in methanol (containing 3% water) provide the
273K
100 263K best freeze-substitution 'cocktail' for routine ultrastructural studies on a
range of biological specimens. It must be remembered that, as opposed to
~ acetone, methanol substitution occurs at 183 K (-90°C) even in the pres-
c ence of 10% water (Fig. 7.2b). Thus, glutaraldehyde (available as 50-70%
••0 aqueous solutions) can be added to the substitution medium. (Note: high
~
~
concentrations of glutaraldehyde have an increased tendency to polymerise
•c
~
during storage.)
-;
••0
::; 50 '_Miiller'sfreeze-substitution schedule I I

233K
0
c,
0
E The frozen sample is transferred from liquid nitrogen (storage) into
• pots containing the substitution liquid (methanol containing fixatives)
•>
••" 223K pre-cooled to 183 K (-90°C). If the specimens have been frozen by
the sandwich (propane jet) technique (§2.4.1 b), the two holders are
separated in the substitution liquid or in liquid nitrogen using pre-
cooled forceps and the pots are subsequently slightly shaken. 1.5 ml
2 4 6 t (h)
substitution fluid and a substitution time of about 8.0 h is sufficient
Time of fixation
for a few tens of micrometres thickness of sandwich-frozen specimen.
F . 7 9 Fixation properties . of glutaraldehyde 10
. methanol at different
. substituti_o
lb .
ig. · · . db · ·ectin a 2% bovine serum a Variations in the substitution time depend on the dimensions of the
tures. The fixation capacity was determtne y lilJ . g . do t 9ooc (183 !().
into a 3% solution of glutaraldehyde in methanol ma1nta1ne a(R-edrawn from fr',_
well-frozen specimen area. Unless high pressure freezing has been
· was then warmed up to th e 10
so1ution · d"catcd
1 temperatures.
.) used (§2.4. ld), there is no sense in substituting deeper, but poorly fro-
1983 zen, specimen areas at a 'recrystallisation-safe' temperature.
478 Rohards and Sleytr Low temperature methods in biological electron microscopy
Freeze-substitution and low temperature embedding
479
(2) The temperature is subsequently raised to 213 K ( -60°C) and main-
tained here for 8.0 h.
(3) After a further temperature rise to 243 K ( - 30°C) for a period of 8.0
h, the specimens are removed from the methanol and washed twice
in cold (243 K; - 30°C) acetone. Suspensions are pelleted using a
'"••
small centrifuge (e.g. Beckman Microfuge B; Appendix 2) operating
in a deep freeze at 243 K ( - 30°C). (The Beckman Microfuge B can
be operated at temperatures down to 183 K; -90°C).
(4) The substituted pellets or tissue samples are further processed fo r th in
sectioning using standard, or low temperature (§7 .5), embedding tech-
mques.

Some typical results obtained with the M iiller freeze-substitution techni-


que are illustrated in Figs. 7.10 and 7.11. The Euglena gracifis cells (Fig.
7.10) were propane jet-frozen and the frozen suspension was then fract ured
under liquid nitrogen and subsequently freeze-substituted and embedded in
Araldite/Epon at 4°C (277 K) for thin sectioning. Fig. 7.1 1 shows part of
a high pressure-frozen mouse growth cartilage substituted and embedded in
the same manner as the Euglena cell. Muller (personal communication and
Muller 1981) bas also found that pure phospholipid structures remain well
preserved even if freeze-substituted in methanol without any dissolved fixa-
tive (stabilizing additives). Propane jet-frozen lecithin liposomes (Fig. 7.12)
were freeze-substituted in methanol for 8 h at temperatures of - 90, -60
and -30°C (183, 213 and 243 K) and embedded in Lowicryl HM20 (§7.5).
The image contrast arises from post-staining the cut sections with lead ci-
trate and uranyl acetate.
It has also been reported that freeze-substitution in pure methanol, fol-
lowed by low temperature embedding (§7 .5), can give good structural pres-
ervation of muscle tissue (Humbel et al. 1983). These authors fo und that,
if freeze-substitution is followed by conventional embedding and heat poly-
merisation (e.g. in Epon/Araldite), a fixative must be added to the substitu-
tion fluid, while, if followed by low temperature embedding techniques, no
additional fixative seems to be necessary as long as the temperature is kepi
below the collapse temperatures of the specimen structures. These tempera-
tures, which are different fo r different solvents and for different cell ular
components (MacKenzie 1972), are within the range 21 3 to 243 K (-60 to
- 30°C) . Because methanol is more polar than acetone the collapse temper-
I
1.o· I'm
I
f I

ature in methanol is lower than in acetone. The collapse temperature in pol- Fig. 7.10 Freeze-substituted Eug/ena racilis II .
fractured under [' 'd . . g ce . A suspens10n was propane jet-frozen
ar Lowicryl K4M (§7.5) is probably between that of acetone and methanol at 277 ~qui " nitrogen and then freeze-substituted and embedded in Araldite/Epo~
(4 C). (Unpublished micrograph kindly provided by M. Miiller.)
480 Robardr and Sleytr Low temperature methods in biological electron microscopy
Freeze-substitution and low temperature embedding
481

Fig. 7. 12. Propane jet-fro zen lecithin liposomes freeze-substituted in methanol a nd embedded
in Lowicryl HM20 at 243 K ( - JO"C). (Unpublished micrograph kindly provided by
M . M uller.)

. · h. h-prcssure fro/en Jnd Fig. 7.1 3. Chicken muscle fibre bundle p repared in relaxation buffer containing 50% glycerol,
Fig. 7 .11. F reeze-substituted mouse growth carlllage. The tissue wa~ ig U n u bl1shed m1.:r<>·
then freeze-substituted and embedded in Araldite/ Epon at 277 K (4_C) ( . t1984) frozen in liquid propane and freeze-substituted in methanol before embedding a t 243 K
graph kindly provided by M . M ul.ler. For details, see Hunziker el a . . (-30"C) in Lowicryl K4M. The M-line protein was dem onstrated by immunola bclling using
lhe protein A-gold complex method. (Reproduced, with permission, from Humbel et al. 1983.)
482 Robards and Sleytr Low temperature method~ in biological electron microscopy
Freeze-substitution and low temperature embedding 483
It was found experimentally that biological material freeze-substituted in
(1982) also investigated primary freezing damage caused during cooling the
the absence of any fixatives is more strongly influenced by the polar K4M
specimen and secondary freezing damage caused by recrystallisation when
than b y th e apo I ar HM20 (Humbel et al · 1983) · Obtaining thm
. sect10ns of
the specimen was warmed up in the substitution medium. The relevant re-
completely unfixed biological specimens with good preservat10n of structure sults can be summarised as follows:
and biological activity is of particular importance for 1mmunolabellmg tech-
.
n1ques. A n exam pie (from Humbel et al. 1983) is illustrated m Fig. 7.13
(i) The duration of substitution at 194 K ( - 79"C) could be reduced from
which shows the fine structure of a chicken muscle fibre bundle prepar~d
seven days to 5 min (the shortest time tested) without generating any
in relaxation buffer containing 50% glycerol. The specimen was frozen m
recrystallisation artefacts provided that the rate of warming up was
liquid propane and freeze-substituted in pure methanol before bemg embed- low (half-time = 3.0 h).
ded in K4M at 243 K (-30°C). The M-line protein was demonstrated by
(ii) Severe recrystallisation occurred when the specimen was substituted
immunolabelling techniques (using a protein A-gold complex method; Roth
for 5.0 min at 194 K ( - 79°C) and subsequently warmed up quickly
et al. 1978). (half-time = 30 s).
(iii) No detectable recrystallisation artifacts were seen when the specimen
Van Harreveld's freeze-substitution method
was substituted at 194 K ( - 79°C) for seven days and subsequently
warmed up rapidly (half-time = 30 s).
Using silk moth antennae as a model system and freeze-fracture replicat~on
. Stein
to assess quaI1ty, · b recht (1980) obtained good structural preserval!on
.
These data confirm that acetone dissolves water (ice) slowly at 194 K
with freeze-substitution by the method of van Harreveld et al. (1965), which
(- 79°C) (Fig. 7.2). Unless long substitution times are used at such a tem-
is as follows:
perature, most of the ice will dissolve during the final warming up period.
Thus recrystallisation appears to be a minor danger during freeze-substitu-
(!) The frozen sample is transferred from liquid nitrogen (stora~e) to ace· tion provided that the specimens are warmed slowly from low temperatures
tone containing 2% (w/v) Os0 4 maintained at 193 K (:--80 CJ m t~e (allowing the ice to dissolve before any recrystallisation can occur). Conse-
presence of molecular sieve (Linde, Type 4A; AppendIX 2) m closl)~ quently, the cold metal block freeze-substitution device (Fig. 7.5; §7.3),
vials for three to ten days. With silk moth anten?ae (drnmet~~ which warms up slowly after loading at low temperature, can be expected
< I 00 µm) as a test specimen, no significant structural differences 'Se[:) to give good results. Steinbrecht (1982) also showed that, if water was added
detectable between the longest and shortest durat10ns of subsl!tull~p)> to the substitution acetone (5% v/v, without molecular sieve) while other-
(2) The specimens are warmed up to ambient t~mperature over ape~; wise following the standard procedures described above, significant freezing
Of4 h (during which the OsO,acts as a fixal!ve). . i<t damage occurred in all parts of the tissue. As shown in Fig. 7.3 (§7.2), ·at
(3) After two changes of acetone and propylene ox1·d e, spec1mens. /;c
a 194 K ( - 79"C), acetone containing 5% water is in equilibrium with ice, and
embedded in Epon 812 or another resin at 293 K (20°CJ· _so the actual substitution of specimen ice must have occurred at much
igher temperatures during the slow warming up period, allowing recrystal-
isation before substitution. As discussed in §2.1.1 and §2.1.2, great differ-
By comparing freeze-substitution with standard chemical f:lti;,
nces in the recrystallisation temperature of different specimens, or even of
freeze-etching results, it was shown that, if antennae wer~ fixefi _--:
Iferent compartments within an individual cell, can be expected and a
(Zetterqvist buffer pH 7.4; 300 mOsm) prior to being rapidly r~~
~:~':
neralisation from Steinbrecht's findings is not recommended. Neverthe-
freeze-substituted, ,the images no longer resembled_ cryofixed
s. as Steinbrecht-points out, his results are in agreement with the data of
were almost identical to specimens that were chem1cally fixed ---::
acKenzie (1977) who measured glass transitions and other second order
ventionally dehydrated. .
se behaviour in model systems and a variety of animal tissues and
Using the same mo d eI System a nd substitution medium, Ste.
erved no significant recrystallisation below the 'ante-melting' transition

'I
'
484 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-substitution and low temperature embedding
485
at about 223 K (-50°C). Again, Steinbrecht (1982) emphasises that'. with
ing a variety of procedures, Barlow and Sleigh (1979) suggested the follow-
tissue in an organic solvent, the ante-melting transition may occur ~t shghtly
ing medium for preservation of the metachronal wave pattern of ciliated
lower temperatures (MacKenzie, personal communication to Ste1nbrecht, surfaces of protozoa:
see Steinbrecht 1982). . . .
Methanol for the substitution medium is prepared by mixing, at room
Summarising the state of the art of freeze-substitution for thi~ se~hon1~g,
temperature, one volume of methanol saturated with CaC1 with one
it appears that, for most specimens, freeze-substitution in c~mb1nahon with 2
volume of methanol saturated with HgC12 . To 50 parts of this mixture are
low temperature embedding (§7.5) will be the method of chmce. Further stu-
added 40 parts of pure ethylene glycol and 10 parts of acrolein. The resul-
dies to optimise methods have the ultimate goal of findmg cond1t10ns where
tant mixture has a freezing point below 193 K (-so·q and, although vis-
the only water removed is that which is not necessary for the maintenance
cous, is smtable for substitution temperatures as low as 197 K (- 76"C). The
of structural and functional integrity. Thus, it is particularly important to
content of the two chlorides in the final mixture is approximately 8% H Cl
learn more about the behaviour of different substitution media with differ-
ent water contents at different temperatures (Humbel et al. 1983).
and 3% CaCl2; an increase in concentration causes precipitation at
0
~ 19;
( - 76°C). Ethylene glycol is added to the methanol to reduce the extraction
of phospholipids from the membrane since, with pure solvents, it was
7.4.lb Freeze-substitution for scanning electron microscopy
observed that there was insufficient 'stiffness' of structures to resist distor-
tion during critical point-drying. The inclusion of acrolein HgCl and CaCl
As illustrated in Fig. 7.1 (§7.1), freeze-substitution can be used for scanning 'fi . ' 2 2
as 1xatives' proved to be essential to stabilise the delicate ciliated surfaces
electron microscopy (SEM) as an initial preparation step in combination
In compa~ison,. osmium-fixed material always gave unsatisfactory result~
with critical point-, evaporation- or freeze-drying. As discuss~d in ~ha~ter
bec~use distortion due to shrinkage always occurred during critical point-
5 once the water has been replaced by an organic solvent, air-drying in a -_/
drymg. The end of the substitution process at dry ice temperature can be
s;ream of dry air or inert gas (e.g. N 2) is the simplest way of removing t_he• • •
judged by observing the disappearance of the ice drop. The specimens are
solvent. However, severe distortion (e.g. flattening) can be expected wit~
then sJowly warmed up to room temperature. With soft tissue a warming
this technique even if the water has been replaced with solvents with a low~f
surface tension (Revel and Wolken 1973; Albrecht et al. 1976). Altern~~
up period between 203 and 293 K ( - 70 and 20°C) of at least his recom- 6
mended and warming up overnight is convenient. After reaching ambient
vely, the substituted specimen can be frozen and subsequently freeze-dn~~
temperature, the substitution medium is gradually exchanged with pure ace-
Nevertheless, the specimen surface structure is usually the ~a1n intere_~~-;F
.tone or alcohol. Critical point-drying from the pure acetone or alcohol is
SEM studies and, with properly frozen, fully hydrated specimens, the~~-~ then carried out, using C02 as the transition fluid.
little point in not performing the freeze-drying procedure d_irectly. For·~·
SEM, the most common procedure for removing the subst1tut10n hq • _7.4.2 Localisation l!f water-soluble substances
critical point-drying (Hayat and Zirkin 1973). If suitabk subst1tu.t10n ll I I
are used it is possible to transfer the specimen directly mto the cn!Ical
addition to its uses for structural studies, freeze-substitution can also
drying apparatus. This can be done at low temperature and most
rve as a method for preparing biological specimens for the localisation of
CO,/intermediate fluid exchanges can be made at low tempe~atnre.
!er-soluble substances. Thin sections of freeze-substituted, embedded
warming the CO, to make the transition through the cn!Ical pomt
CJmens can be evaluated using analytical electron microscopy, autoradio-
1978).
phy or LAMMA (Fig. 7.1) (§7.1). As with freeze-dried cryosections, ide-
It has been shown that the substitution mixtures used for TE•
standards for judging the results of freeze-substitution are frozen-hyd-
contain low concentrations of fixatives in pure solvents, are-· not n_
ed sections (§3.4.l) from the same material. A basic requirement for the
suitable for SEM. Particularly for the preservation of delicate su~,
ess of analytical studies is to minimise the movement and dissolution
tures, special substitution mixtures may have to be. devel~p~:;-_:-
cimen c~mponents during substitution and the subsequent embedding
shrinking and distortion during subsequent critical pomt-drymg.
ess. Substitution in the presence of precipitating reagents may provide
=Z""?P/5,!!."'ill"'f0 "'C?/&?:%<_-'-·"'/ ?!?!i'./0TCJ;23;;;,7q-. "'"" w,, ,,,,, ..;'- " ' ( \ " / , <---"
r

486 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-substitution and low temperature embedding
487

a means for detecting water-soluble substances by conventional electron Carlemalm and colleagues (Carlemalm et al. 1982; Armbruster et al. 1982;
microscopy. Unfortunately substitution 'cocktails' which minimise the Carlemalm 1984) for this purpose and are commercially available (see
movement of soluble specimen components often give poor structural pres- Appendix 2). (At the time of writing, an intermittent newsletter, Lowicryl
ervation. An important difference in sectioning material to be used for letters, is produced by Ch.emische Werke Lowi [see Appendix 2] which pro-
ultrastructural studies and for localisation of water-soluble components is vides up-to-date mformat10n on progress with the use of these resins.) Other
that while the former may be sectiohed onto a liquid bath, the latter must low t~mperat~re resins, such as London L.R. Gold, are also becoming com-
be dry-sectioned (even excluding non-polar flotation media) to avoid the mercially available (Appendix 2).
loss of soluble materials.
Typical examples of the application of freeze-substitution to the localisa- 7.5.1 Lo1v temperature resins
tion of water-soluble substances in multicellular plants are: Fisher and
Housley (1972), Van Steveninck and Van Steveninck (1978), Harvey (1980), Two resins are available: Lowicryl K4M and Lowicryl HM20 (subsequently
Hall et al. (1978), Browning and Gunning (1979) and Harvey and Kent referred to as simply K4M and HM20). These are highly cross-linked acry-
(1981). . . late-. and methacrylate-based embedding media, specially formulated to
In comparison with studies on plant material, relatively little freeze-sub- provide low viscosity at low temperatures. K4M is a polar (hydrophilic) res-
stitution work has been carried out on animal tissues (e.g. Chandler 1980; m usable down to -35°C (238 K), while HM20 is non-polar (hydrophobic)
Marshall 1980; Ornberg and Reese 1980, 1981). . and can be used as low as - 70°C (203 K). Both resins can be polymerised
A comprehensive review of freeze-substitution methods for the Jocahsa~ by long wavelength (360 nm) UV light at low temperatures or can be chemi-
lion of water-soluble substances is provided by Harvey (1982). For relevant cally polymerised at 60°C (333 K).
discussions of related problems, see also the book on Microprobe Analysis The hydrophilic properties of K4M provide some advantages; in particu-
of Biological Systems edited by Hutchinson and Somlyo (1981). lar the specimen can be kept in a partially hydrated state during dehydration
and mfiltrat10n since K4M can be polymerised with up to 5% (w/w) water
m .the block. Furthermore, K4M is particularly useful for labelling sections
7.5 Low temperature embedding techniques ~sing spec1fic antisera or lectins, since it gives better structural preservation,
improved retention of antigenicity and a significantly lower background
Implicit in almost everything that has been said. earlier in this book is labelling than HM20.
belief that maintenance of low temperatures durmg the processmg of·. There .are e~sentially two different processing routes for embedding in
frozen specimens is a valuable aid to the preservation of u_lt_rastructureff these resins: v~a normal (ambient temperature) fixation with subsequent,
retention of soluble components close to their in vivo pos1t1ons. Beca~> gradu~l, lowermg. of. the temperature during the dehydration stages, and
is often necessary to cut thin sections of biological specimens for o_~'­ ps1ng rreeze-subst1tut1on so that the frozen specimen is never warmed until
tion in an electron microscope the material must be infiltrated witl1 has.been fully infiltrated with resin and polymerised at low temperature.
form of matrix that can be hardened. This is usually a resin (mosf; he first pathway 1s the most frequently used although, in principle, the
manly an epoxy resin but sometimes an acrylic or polyester). Suc_e-:_· cond has much to recommend it.
can themselves be effective lipid solvents, the more so as they are o
merised at relatively high temperatures. The problems are compou .2 Fixation and dehydration
cause the resin viscosity is usually quite high and becomes wor~e ~~:
perature falls, so that infiltration can be difficult. The product10 Y ~ormal fixation schedule can be used although fixatives which also
that can be infiltrated into biological specimens at sub-zero te _ ntnbute to contrast (e.g. osmium tetroxide) should be avoided because
and then polymerised, also at low temperature, is clearly an.": Y mterfere with photopolymerisation (by blocking penetration of the UV
some importance. The Lowicryl resins were developed in S-wrt:- t) and may also (e.g. OsO,) react with unsaturated double bonds in the
488 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-substitution and low temperature embedding
489

273 TABLE 7.2

-~~-::~~.-- ......
Dehydration at low temperature
-10 ............ 263 -::-:----:-~~~~~__:~~~---~
Ethanol (vol. %in H 20)
E ,..
'":.;,., ......
....
.Acetone 253
g Temperature ("C) Tin1e (min)

,:! -20 "-t. .. ...... ~

• Methanol " 0

-- -- EGOH:
I
B
~
30
50 -20
0 30

-' ' I
'\·\·.. .. ........
60
~ -30 \ ·• •. Ethanol I
243 •0. 70 - 35 (- 50)' 60
E ' E
,_• \~··· ... '' I ,_• 95
100
- 35 (- 50 to -70) 60
233 -35 (- 50 to - 70)
'' ,.I
~::'-~-'-~-'-~-'~--,"-~~E~G-0-H~·~··_··_··_··.~··~·~~~~~-:-:
60
100
" - 35 (- 50 to -70) 60
·''' . 223
a Suggested steps for work at lower temperatures-· only possible with HM20.
0 20 40 60 80 100% Solvent
100 80 60 40 20 0%Water
Solvent/water (volume %)
7.5.2a Maintaining low temperatures
Fig. 7.14. Freezing points of commonly used dehydrating agents (solvents) as a function of
concentration. Note the rise in the freezing point of ethylene glycol (EGOH) at higher tempera-
tures. (Redrawn from Carlemalm, unpublished. Available in details supplied with Lowicryl res- When _working at a number of different low temperatures, the problem of
ins; see Appendix 2.) achievmg and maintaining the relevant temperature(s) has often deterred
potential users from applying low temperature methods (something that is
resin. Naturally occurring pigments, present in reasonable amounts, do not not only a problem in low temperature embedding). In the present context
normally interfere with polymerisation although, if specimens are heavily the following methods for achieving low temperatures during dehydration,
pigmented with strong absorbers at 360 nm, polymerisation may be less mfiltrat10n and embedding can be used:
than optimal. It will be clear that photopolymerisation necessitates
of the UV light throughout the specimen block to ensure even polynaer:isa,,,, (i) p . .
ropnety umts are available, such as the Balzers Low Temperature
lion. Blocks shonld thus be small ( ~0.5 mm 3). Emb_eddmg Apparatus (Appendix 2), which provides four sample-
If it is decided to pass the specimen from a fixative at ambient (or holdmg blocks, each of which can be independently controlled at tem-
0°C; 273 K) temperature into the low temperature resin, then this invo peratures between 0 and -50°C (273 and 223 K). This unit also al-
a stepwise reduction in temperature concomitant with increasing conce Jows continuous agitation of the samples. ·
tion of dehydrating agent. A temperature is selected at each step wh A temperature of -20°C (253 K) is obtained by using a 3:1 (w/w) ice/
just above the freezing point for the concentration used in the pre NaCl mixture. The temperature should be carefully monitored and
stage. This corresponds to the concentration of the dehydrating the mixture periodically replenished to ensure that the temperatnre is
actually in the tissue block as it is introduced into the next higher con maintained. As with all such low temperature baths, they will main-
tion in the series. Freezing points in relation to concentrations of-V: tam their low temperature longer if they are placed inside a glass-lined
dehydrating agents are shown in Fig. 7.14. It is important that the0 Dewar vessel which is itself kept in a refrigerator or cold box.
mens are periodically agitated during dehydration and infiltration; eit For temperatnres in the range down to about -40°C (233 K) ordin-
gently stirring or by swirling the sample vials. Most polar or no~; ary domestic deep freezers are suitable.
dehydrating agents can be used with either resin although, due to it. For temperatnres below -35 to -40°C (238 to 233 K) either special
phobic nature, HM20 is immiscible with ethylene glycol or dimetll low temperature cabinets (Appendix 2) or dry ice/o- and m-xylene
mide. A typical dehydrating schedule, suitable for either K4M baths can be used. Obviously, the low temperature cabinets provide
is given in Table 7.2. greater convenience but are expensive. If dry ice/xylene baths are
490 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-substitution and low temperature embedding
491

used then a range of temperatures between about -26 and -63"C 7.5.3 lrifiltration and polymerisation
(247,and 210 K) can be obtained depending on the proportions (v/v)
of o- and m-xylene (Fig. 7.15). The xylene mixture appropriate to the 7.5 .3a Preparation of resins and infiltration
temperature required is made up and then crushed dry ice is added
to form a thick slurry. This bath will provide a constant temperature Methacrylates, like many other embedding resins, may cause skin reactions
for 8-10 h if used in a Dewar vessel. (Caution: xylene fumes are toxic in sensitive individuals. Therefore protective rubber gloves should be used.
- only use in an effective fume hood). Furthermore, the resins should be mixed in a fume hood and the vapour
should not be inhaled. As the resins are of very low viscosity they require
As the samples used in low temperature embedding are almost inevitably little mixing. Indeed, excessive stirring results in the incorporation of too
small, they are likely to show sharp temperature changes during transfer un- much oxygen into the mixture and this may interfere with polymerisation
less they are suitably protected. For this reason, it is advisable to accommo- (especially if thermal polymerisation is to be used).
date the sample vials (or other containers) in holes drilled in an aluminium The monomer and cross-linker should be weighed out into a preweighed
block (aluminium is a convenient metal to use and has good thermal capa- vial and then mixed gently for 3-5 min by one of the following methods
city; see Table 2. 7 in §2.2.4). The block can be allowed to equilibrate at a (note that this represents the 'published' formulation; in practice, such
particular temperature before the sample vials are placed in it and, during accurate weighings are probably unnecessary and, in any case, the mixtures
transfer, its high thermal capacity will prevent any steep temperature gradi- are more easily measured by volume):
ents within the specimen.
(I) Bubble a stream of dry nitrogen gas through the mixture via a Pasteur
pipette (this both mixes the resin and prevents incorporation of oxy-
gen).
(2) Mix gently with a glass rod.
-20 253
(3) If the vial is sealable, rock slowly from side to side (avoiding the for-
mation of air bubbles or foaming).
-30 243

ii g When the monomer and cross-linker have been mixed, the initiator is
! added and mixing is continued until it has been completely dissolved in the
,~ .'l resin .

~
0.
-40 233
•••
a
E
E ...
...• -so 223
TABLE 7.3
Mixtures for ultra-violet polymerisation

K4M
HM20
-60 213
Amount (g)
10
0
9 8
2
7
3

4
5 o-xylene
5 m-xylene
Component Amounr.(g)

2.70 Cross-linker D
Volume parts 3.00
17.30 MonomerE 17.00
0.10 Initiator ca 0.10
Fig. 7.15. Temperatures of crushed dry ice/xylene slurries, as a function of the
tom-xylene. (Redrawn from Carlemalm, unpublished. Available in details supplied
or polymerisation between - 50°C and 0°C (223 and 273 K). Above 0°C, the· illitiator
icryl resins; see Appendix 2.) Ould be replaced by the same amount ofbenzoin ethylether.
492 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-substitution and low temperature embedding 493

Mixtures for ultra-violet polymerisation 7.5.3b Polymerisation


The mixtures indicated in Table 7.3 produce blocks of average hardness.
The hardness can be varied by incorporating more or less cross-linker in the
Samples may be conveniently held for polymerisation in either BEEM-type
mixtures (more cross-linker gives more hardness). For HM20, the cross-
or gelatin capsules. A suitable holder is required so that capsules receive UV
linker concentration can be varied from 5 to 17% (w/w) (1.0 to 3.4 g/20 g
radiation from all sides (e.g. Fig. 7.16a). The polymerisation chamber is
resin). For K4M, the cross-linker concentration can be varied from 4 to 18%
made so that it will fit into a chest-type deep freeze or a cold room (Fig.
(w/w) (0.8 to 3.6 g/20 g resin).
7.16b). The light source must provide long (360 nm) UV, preferably from
two 15 W fluorescent tubes (e.g. Philips TLD 15W/05; see Appendix 2) simi-
Mixtures for thermal (chemical) polymerisation at + 60°C
lar to those used for thin-layer chromatography. To provide diffuse illumi-
Although HM20 and K4M are primarily designed for use at low tempera-
nation, a right-angle deflector is suspended below the fluorescent tubes. All
tures and UV polymerisation, it is also possible to polymerise them by the
six inner surfaces, as well as the deflector, should be constructed from UV-
more conventional thermal method. If this is to be done, then the cross-
reflective material or lined with aluminium foil. The capsules should be
linker and monomer are first mixed as described in the previous section
positioned about 30-40 cm below the UV source. The whole box should not
(preferably using bubbling nitrogen) and then, instead of initiator C, 0.5%
be made too air-tight: ventilation from top and bottom allows air circula-
w/w (for HM20) or 0.3% w/w (for K4M) dibenzoyl peroxide is used. tion and help to minimise temperature gradients.
Dibenzoyl peroxide is usually supplied as a paste with dibutyl-phthalate
Other UV sources can be used, including some hand-held lamps. How-
or as a powder moistened with water. The extra ingredients should be com-
ever, they must emit at 360 nm and shorter wavelengths should be filtered
pensated for so that the resins receive the above amounts of the peroxide
out. Smaller sources should be positioned correspondingly nearer to the
exclusive of additives.
samples (10-15 cm) and the overall dimensions of the polymerisation box
should also be reduced. If weak UV sources are used, then radiation from
Infiltration at low temperatures
the bottom of the cabinet may help because the light will have less resin to
Infiltration with K4M or HM20 at low temperature is essentially similar to
infiltration with other resins at room temperature. The exact procedure will
depend on the dehydrating agent used and the temperatures chosen in rela- a b
tion to the viscosities of the dehydrating agent and resin at those tempera-
tures. The schedule in Table 7.4 illustrates a typical route, through eth1an•Dl{
which is applicable to either K4M or HM20.

TABLE7.4
Typical infiltration schedule

3-~
Resin: Ethanol (v/v) Temperature(° C) Time

1:1 _ 35a 60min


2:1 - 35 60min
Pure resin -35 60min
Fig. 7.16. Apparatus for low temperature polymerisation. (a) A wire capsule holder for UV
Pure resin - 35
polymerisation. (b) A polymerisation chamber for indirect UV irradiation. The UV source (1)
is diffused by a right-angle reflector made, for example, from aluminium foil (2). The capsule
a L ower t emperat ures ( - 50 to - 70 ° C , 223 to 203 K) are possible for infiltration witfr holder (3) is placed 30-40 cm below the UV source. (Redrawn from Carlema\m, unpublished.
and higher temperatures can be used for both resins.
Available in details supplied with Lowicryl resins; see Appendix 2.)
494 Robards and Sleytr Low temperature methods in biological electron microscopy
Freeze-substitution and low temperature embedding 495

traverse before polymerising the resin in the sample. Whatever the arrange- 7.5.4 Sectioning and staining
ments for polymerisation, trials without samples should first be attempted.
If there is shrinkage and deformation of the blocks, then polymerisation has For best results with HM20 or K4M, the blocks should be trimmed to pyra-
been too rapid and the distance between the lamps and the capsules should midal faces on a microtome or trimming apparatus. The sides and face
be increased. should be clean and transparent with the pyramid having an angle of about
28-30° from the cutting face. (See Reid 1974 for details ofultramicrotomy
Ultra-violet polymerisation at low temperatures procedures.)
HM20 is a highly cross-linked methacrylate which, when of the correct
(1) Fill capsules to the top (to minimise dead space filled with air) with hardness, sections easily with either glass or diamond knives.
fresh, pre-cooled resin. K4M is a hydrophilic resin. Therefore, as with other polar (water-misc-
(2) Transfer samples to capsules using Pasteur pipettes; close the capsules ible) resins, precautions should be taken to ensure that the block face does
and allow the temperature to equilibrate for 10-15 min. not wet during sectioning. This is best accomplished by sectioning with the
level of liquid in the trough slightly below normal. In these circumstances
(To minimise condensation of water and crystallisation of ice on the sample the reflection from the trough fluid along the knife edge is slightly darker
vials and capsules, all apparatus should be pre-cooled and steps (1) and (2) than the normal bright silver colour. It is important that the fluid level is
should be performed in the cold.) not reduced so much that the edge becomes dry. This is particularly impor-
tant with diamond knives because their edges are normally hydrophobic.
(3) Polymerise for at least 24 h under UV light at -30 to -50°C (the The best procedure with a diamond knife is to orientate the trimmed block
lowest recommended temperature for polymerisation is -50'C). with the knife edge before the trough is filled. The specimen arm of the mic-
(4) Remove capsules from the cold and continue 'curing' under UV for rotome is placed in its lowest position and the trough is overfilled to form
2-3 days at room temperature. a 'reverse meniscus' along the knife edge. The knife is left in this positi_on
for 10-15 min. Immediately before sectioning, the level of the trough fluid
Chemical polymerisation at +60'C is lowered to produce a dark silver reflection along the knife edge. The final
advance of the knife block should then be made and sectioning commenced.
(1) Place fresh resin in gelatin capsules; transfer samples to capsules and Since K4M is a hydrophilic resin, the sections should be collected as soon
fill capsules approximately 3/4 full. as possible after they have been cut. Cutting speeds of 2-5 mm s-' are
(2) Close capsules and polymerise at 60°C for 2-3 days. recommended for both resins.
There are significant differences between the _staining properties of the
The use of gelatin capsules is recommended for chemical polymerisation at, _ , two resins but, as general guidelines, secti.ons may be stained with either
+ 60°C since the plasticiser in BEEM capsules may interfere with polymeric. aqueous or alcoholic solutions o_f uranyl acetate, with Reynold's lead ci-
sation at the edge of the block. The chemical polymerisation of HM20 an~: trate, or with Millonig's lead acetate. Completely unstained sections of
K4M with peroxides is an exothermic reaction. To prevent an uncontrol HM20 have been imaged in STEM instruments using so-called 'Z~contrast'
rise in temperature, the capsules sholild be in contact with a heat sink s , (Carlemalm and Kellenberger 1982; see also Kellenberger 1984). A particu-
as an aluminium block with predrilled holes into which the capsules· . larly useful staining combination consists of saturated aqueous uranyl ace-
firmly. tate (for 35 min for HM20 or 5-10 min for K4M) followed by Millonig's
lead acetate (for 1-3 min for both resins).
Since K4M is hydrophilic, the sections should be incubated on drops of
:the stains for short periods Of time. ·Prolonged staining may cause contami-
·nation and distortion of the sectio'ns. As with all staining Procedures, wash
496 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-substitution and low temperature embedding 497

well between the uranyl and lead stains and take precautions against CO, Femindez-Morin, H. (1959b), Cryofixation and supplementary low temperature preparation
dnring the lead staining and the subsequent rinsing. techniques applied to the study of tissue ultrastructure, J. appl. Physiol. ·iO, 2038.
Femindez-Morin, H. (1960), Low temperature preparation technique for eiectron microscopy
K4M has been successfully used in cytochemical and immunocytochemi-
of biological specimens based on rapid freezing with liquid helium II, Ann. N.Y. Acad. Sci.
cal studies (e.g. Bendayan 1981, 1983; Roth, 1982a, 1983; Roth et al. 1981; 85, 689.
Roth and Berger 1982). Femindez-Monin, H. (196la), Lamellar systems in myelin and photoreceptors as re'vealed by
high resolution electron microscopy, in: Macromolecular complexes, M.V. Edds, Jr., ed.
(Ronald Press, New York), p. 113.
References Fernandez-Moran, H. (196lb), The fine structure of vertebrate and invertebrate phcitorecep-
tors as revealed by low temperature electron microscopy, in: The structure of the eye, G.K.
Albrecht, R.M., D.H. Rasmussen, C.S. Keller and R.D. Hindsill (1976), Preparation of cul- Smelser, ed. (Academic Press, New York), p. 521.
tured cells for electron microscopy - air drying from organic solvents, J. Microscopy 108, Fisher, D.B. and T.L. Housley (1972), The retention of water soluble compounds during
21. freeze-substitution and microautoradiography, Plant Physiol. 49, 166.
Armbruster, B.L., E. Carlemalm, R. Chiovetti, R.M. Garavito, J.A. Hobot, E. Kellenberger Glauert, A.M. (1974), Fixation, dehydration and embedding of biological specimens, in: Prac-
and W. Villiger (1982), Specimen preparation for electron microscopy using low temperature tical methods in electron microscopy, Vol. 3, A.M. Glauert, ed. (North-Holland, Amster-
embedding resins, J. Microscopy 126, 77. dam).
Barlow, D.I. and M.A. Sleigh (1979), Freeze-substitution for preservation of ciliated surfaces Hall, J.L., D.M.R. Harvey and T.J. Flowers (1978), Evidence for the cytoplasmic localisation
for scanning electron microscopy, J. Microscopy 115, 81. of Betaine in leaf cell of Suaeda maritiJna, Plan ta 140, 59.
Bendayan, M. (1981), Ultrastructural localization of actin in insulin-containing granules, Biol. Harvey, D.M.R. (1980), The preparation of botanical samples for ion localisation studies at
Cell41, 157. subcellular level, Scanning Electron Microscopy, 1980, 2, 409.
Bendayan, M. (1983), Ultrastructural localization of actin in muscle, epithelial and secretory Harvey, D.M.R. (19_82), Freeze-substitution, J. Microscopy 127, 209.
cells by applying the protein A-go!d immunocytochemical method, Histochem. J. 15, 39. Harvey, D.M.R. and B. Kent (1981), Sodium localisation in Suaeda maritima leaf cells using
Browning, A.J. and B.E.S. Gunning (1979), Structure and function of transfer cells in the Spor- zinc uranyl acetate precipitation, J. Microscopy 121, 179.
ophyte haustorium of Funaria hygrometrica Hedw. IL Kinetics of uptake of labelled sugars Harvey, D.M.R., J.L. Hall and T.J. Flowers (l98lb), The use of freeze-substitution and freeze
and localisation of absorbed products by freeze-substitution and autoradiography, J. exp, fracture study of bacterial spore structures, J. Ultrastruct. Res. 76, 71.
Bot. 119, 1247. Hayat, M.A. (1981), Fixation for electron microscopy, (Academic Press, New York).
Bullivant, S. (1960), The staining of thin sections of mouse pancreas prepared by the Fernan- Hayat, M.A. and B.R. Zirkin (1973), Critical point-drying method, in: Principles and techniqu-
dez-Moran helium II freeze-substitution method, J. biophys. biochem. Cytol. 8, 639. es of electron microscopy, Vol. 3, M.A. Hayat, ed. (Van Nostrand Reinhold Company, New
Bullivant, S. (1965), Freeze-substitution and supporting techniques, Lab. Invest. 14, 440. Yock), p. 297.
Carlemalm, E. (1984), Low temperature embedding resins, Proc. 8th Eur. Reg. Conf. Electron Heuser, J.E., T.S. Reese, M.J. Dennis, Y. Jan, L. Jan and L. Evans (1979), Synaptic vesicle
Microscopy, Budapest 3, 1786. exocytosis captured by quick-freezing and correlated with quanta! transmitter release, J. Cell
Carlemalm E. and E. Kellenberger (1982), The reproducible observation of unstained embed- Bio!. 81, 275.
ded cellular material in thin sections; visualization of an integral membrane protein by a Humbel, B. and M. Miiller (1984), Freeze-substitution and low-temperature embedding, Proc.
new mode of imaging for STEM, EMBO J. 1, 63. 8th Eur. Reg. Conf. Electron Microscopy, Budapest 3; 1789.
Carlemalm, E., R.M. Garavito and W. Villiger (1982), Resin development for electron micros-.: Humbel, B., T. Marti and M. Millier (1983), Improved structural preservation by Combining
copy and an analysis of embedding at low temperature, J. Microscopy 126, 123. freeze-substitution and low temperature embedding, Beitr. elektronenmikroskop. Direk-
Chandler, J .A. (1980), Application of X-ray microanalysis in reproductive physiology, tabb. Oberfl. 16, 585.
ning Electron Microscopy 1980, 2, 475. Hunziker, E.B., W. Herrmann, R.K. Schenk, M. Millier and H. Moor (1984), Cartilage ultras-
Dempsey, G.P. and S. Bullivant (1976), A copper block method for freezing non-cryoprote~ tructure after high pressure freezing, freeze substitution, and low temperature embedding.
tissue to produce ice-crystal-free regions for electron microscopy. I. Evaluation using fre I. Chondrocyte ultrastructure - implications for the theories of mineralisation and vascular
substitution, J. Microscopy 106, 251. />' invasion, J. Cell Biol. 98, 267.
Feder, N. and R.L. Sidman (1958), Methods and principles of fixation by freeze-substi~~i,·' Hutchinson, T.E. and A.P. Somlyo (eds.) (1981), Microprobe analysis of biological Systems
J. biophys. biochem. Cytol. 4, 593. · .:. .;: (Academic Press, London).
Fernandez-Morin, H. (1957), Electron microscopy of nervous tissue, in: Metabolism.,.O~: Kellenberger, E. (1984), Thin sections with low temperature embedding: successes and remain-
nervous system, D. Richter, ed. (Pergamon Press, Oxford), p. 1. , , ._.<< ing limitations, .Proc. 8th Eur. Reg. Conf. Electron Microscopy, Budapest 3, 1781.
Fernindez-Morin, H. (1959a), Electron microscopy of retinal rods in relation to local Mackenzie, A.P. (1972), Freezing, freeze-drying and freeze-substitution, Scanning Electron
of rhodopsin, Science 129, 1284. Microscopy 1972, 2, 274.
498 Robards and Sleytr Low temperature methods in biological electron microscopy Freeze-substitution and low temperature embedding 499

Mackenzie, A.P. (1977), Non equilibrium freezing behaviour of aqueous systems, Phil. Trans. Roth, J., M. Bendayan and L. Orci (1978), Ultrastructure localisation of intracellular antigens
R. Soc. Ser. B 278, 167. by the use of protein A-gold complex, J. Histochem. Cytochem. 26, 1074.
Marshall, A.T. (1980), Freeze-substitution as a preparation technique for biological X-ray mic- Roth, J., M. Bendayan, E. Carlemaim, W. Villiger and M. Garavito (1981), Enhancement of
roanalysis, Scanning Electron Microscopy 1980, 2, 335. structural preservation and immunocytochemical staining in low temperature embedded
Milller, M. (1981), Demonstration of liposomes by electron microscopy, in: Membrane pro- pancreatic tissue, J. Histochem. Cytochem. 29, 663.
teins, A. Azzi, U. Brodbeck and P. Zahler, eds. (Springer-Verlag, Berlin), p. 252. Simpson, W.L. (1941), An experimental analysis of the Altmann technique of freeze-drying,
Milller, M., T.H. Marti and S. Kriz (1980), Improved structural preservation by freeze-substi- Anat. Rec. 80, 173.
tution, Proc. 7th Eur. Reg. Conf. Electron Microscopy, The Hague 2, 270. Sitte, H. (1981), Ultramikrotomie - Aktuelle Trends und Neuentwicklungen im biologisch-
Newell, D. and S. Roath (1975), A container for processing of small volumes of cell suspen- medizinschen Bereich, Git. Labor-medizin 4, 317.
sions for critical point drying, J. Microscopy 104, 321. Steinbrecht, R.A. (1976), Freeze-substitution and freeze-fracturing of insect sensilla without
Ornberg, R.L. and T.S. Reese (1980), A freeze-substitution method for localising divalent cryoprotectants, Proc. 6th Eur. Reg. Conf. Electron Microscopy, Jerusalem 2, 111.
cations- examples from secretory systems, Fed. Proc. 39, 2802. Steinbrecht, R.A. (1980), Cryofixation without cryoprotectants - freeze-substitution and
Omberg, R.L. and T.S. Reese (1981), Quick freezing and freeze-substitution for X-ray microa- freeze-etching of an insect olfactory receptor tissue, Tissue Cell 12, 73.
nalysis of calcium, in: Microprobe analysis of biological systems, T.H. Hutchinson and A.P. Steinbrecht, R.A. (1982), Experiments on freezing damage with freeze-substitution using moth
Somlyo, eds. (Academic Press, London), p. 213. antennae as test objects, J. Microscopy 125, 187.
Pease, D.C. (1966), The preservation of unfixed cytological detail by dehydration with ''inert'' Van Harreveld H. and J. Crowell (1964), Electron microscopy after rapid freezing on a metal
agents, J. Ultrastruct. Res. 14, 356. surface and substitution fixation, Anat. Rec. 149, 381.
Pease, D.C. (1967a), Eutectic ethylene glycol and pure propylene glycol as substituting media Van Harreveld, H., J. Crowell and S.K. Malhotra (1965), A study of extracellular space in cen-
for the dehydration of frozen tissue, J. Ultrastruct. Res. 21, 75. tral nervous tissue by freeze-substitution, J. Cell Biol. 25, 117.
Pease, D.C. (1967b), The preservation of tissue for fine structure during rapid freezing, J. Van Steveninck, R.P.M. and M.E. Van Steveninck (1978), Ion localisation, in: Electron mic-
Ultrastruct. Res. 21, 98. roscopy and cytochemistry of plant cells, J.L. Hall, ed. (Elsevier/North-Holland, Amster-
Pease, D.C. (1968), Unfixed tissue prepared for electron microscopy by glycolic dehydration, dam), p. 187.
Proc. 4th Eur. Reg. Conf. Electron Microscopy, Rome 2, 11. Zalokar, M. (1966), A simple freeze-substitution method for electron microscopy, J. Ultras-
Pease, D.C. (1973), Substitution techniques, in: Advanced techniques in biological electron truct. Res. 15, 469.
microscopy, J.K. Koehler, ed. (Springer-Verlag, Berlin), p. 35.
Rebhun, L.L (1972), Freeze-substitution and freeze-drying, in: Principles and techniques of
electron microscopy, Vol. 2, M.A. Hayat, ed. (Van Nostrand Reinhold Company, New
York), p. 3.
Reid, N. (1974), Ultrarnicrotomy, in: Practical methods in electron microscopy, Vol. 3, A.M.
Glauert, ed. (North-Holland, Amsterdam).
Revel, J.P. and K. Wolken (1973), Electron microscope investigations of the underside of cells
in culture, Exp. Cell Res. 78, I.
Robards, A.W. (1978), An introduction to techniques for scanning electron microscopy of
plant cells, in: Electron microscopy and cytochemistry of plant cells, J.L. Hall, ed. (Elsevier/
North-Holland, Amsterdam), p. 343.
Roth, J. (I982a), New approaches for in situ localization of antigens and glycoconjugates on
thin sections - the protein A-gold technique and the lectin colloidal gold marker system,
Proc. 10th Int. Congr. Electron Microscopy, Hamburg 3, 245.
Roth, J. (1982b), The protein A-gold (pAg) technique - a qualitative and quantitative ap-..·.
proach for antigen localization on thin sections, in: Techniques in immunocytochemistryy:
Vol. I, G.R. Bullock and P. Petrusz, eds. (Academic Press, London), p. 107.
Roth, J. (1983), The colloidal gold marker system for light and electron microscopic cyt_~:
chemistry, in: Techniques in immunocytochemistry, Vol. 2, G.R. Bullock and P. Petrus_~~:
eds. (Academic Press, London, New York), p. 217.
Roth, J. and E.G. Berger (1982), Immunocytochemical localisation of galactaryltranferase·
HeLa cells: codistribution with thiamine pyrophosphatase in trans-Golgi cisterna, J,
Biol. 83, 223.

11:
Chapter 8

Future outlook

Although this is a book dealing with practical techniques in low temperature


electron microscopy, it would not be serving its full purpose if it did not
end by pointing towards some of the more probable future developments.
Many low temperature techniques are still relatively new and will continue
to develop beyond the level described here. Thus, it is important that the
reader should be aware of the probable direction of developments so that
he is in a better position to make the right choices in both techniques and
equipment.
Much of the apparatus currently available has either been produced as
'one-off equipment within a single laboratory or is only in the first gene-
ration of commercial production. It is usua1, in the evolution of techniques,
for methodology and apparatus to become simpler for the operator to use
and for the equipment to become more fully automated. Such a trend is seen
at its best in electron microscopes themselves but can also be discerned in
equipment relevant to the topic of this book, such as cryoultramicrotomes,
freeze-driers and freeze-etching units. The availability of simple micropro-
cessor control systems will undoubtedly play an important part in making
work faster and easier. For example, the availability of a simple micropro-
cessor-based cooling rate meter has resulted in a great improvement over
conventional methods for determining rapid cooling rates (§2.2.2). Indeed,
it is difficult to see apparatus in the low temperature field where micropro-
cessors could not fulfil a useful role! Many methods currently use liquid ni-
trogen 'by the bucketful'. Not only is this expensive; it also generally exacer-
bates problems of condensation on cooled surfaces. Excessive use of liquid
nitrogen for cooling cold stages and other components is usually a good
indication of poor design. Refrigerated cryopumps (§6.3.2; Haefer 1981)

501
502 Robards and Sleytr Low temperature methods in biological electron 1nicroscopy Future outlook 503

have, as yet, made hardly any impact in the area of low temperature electron the first detailed theoretical assessment of the relative merits of different
microscopy and yet, for many purposes, they should be ideal. They can freezing methods. Such work not only suggests ways of optimising present
achieve very low temperatures using a closed system and exploration of methods but also points to new techniques that should be tried (e.g. freezing
their merits in applications such as routine freeze-etching is surely long over- in super-critical nitrogen, §2.2.3a).
due. In general, it can be expected that techniques will become more conve- We will always have to fight against the inflexible limits of the physical
nient to apply to different materials and that the choice of the most appro- properties of water. It may well be that future work will involve further
priate technique to solve a particular problem will become easier. In all low experiments into deep-cooling specimens without allowing them to freeze:
temperature work it would be useful to have a suitable test-specimen that if ice formation were to be prevented, no crystal da1nage could occur.
could be used to compare techniques from different laboratories: such a spe-
cimen would provide great advantages in evaluating the comparative merits
of different methods. 8.2 Direct viewing methods

While low temperature SEM techniques are now fairly routine and the
8.1 Freezing apparatus is commercially available, it cannot yet be said that low tempera-
ture TEM is as widely used. The instrumental problems are more severe,
Although our current state of knowledge about the mechanisms of freezing as are the difficulties of preparing specimens. Furthermore, there are a
biological specimens is very much greater than just a few years ago, there number of demands made of low temperature stages in EMs: in addition
still remains much scope for improvement in techniques to ensure that the to the simple requirement of keeping a specimen cold while it is viewed and/
best method is used in solving a particular problem. Moreover, more atten- or analysed, far more specialised (TEM) cold stage will be needed where
tion will have to be given to sampling from large specimens, such as by ·the high resolution and protection from radiation damage are the major objec-
use of improved 'cryoguns' which minimise the otherwise inevitable prob- tives. While it is to be expected that many laboratories will acquire cold
lems of mechanical damage during sampling. The range of commercially stages for ordinary TEM and SEM use, the more sophisticated systems will
available apparatus for freezing specimens is very limited. It is possible to only be available in relatively few laboratories with advanced ancillary faci-
buy spray-freezing, propane jet-freezing, cold block-freezing and plunge- lities. In such circumstances, as in other fields, it is important for the 'nor-
freezing equipment, although the number of suppliers is restricted and it is mal' user to be aware of the results from the more specialised equipment
to be hoped that additional, competitive versions will become available so since these very often have a direct bearing on the use of the simpler systems.
that a greater number of research workers can use these techniques and dif- We will also have to learn more about the physical and structural behav-
ferent approaches can be compared. The commercial realisation of a high iour of ice in the electron microscope. Under the conditions of cryomicro-
pressure freezing device (Appendix 2) permits wider investigation of this scopy (frozen-hydrated specimens) only 4 of the 12 known types of solid
technique; the advantages of being able to freeze specimens while water need to be considered. These are two forms which are crystalline (hex-
them to high pressure are so great that this avenue of research should agonal and cubic ice) and two forms of amorphous water (generally called
tainly not be abandoned. vitreous ice because it retains the order of liquid water). Since temperature-
We can thus expect to see the introduction of new commercial apparat~~;.~ and electron dose-dependent transitions between the four forms of solid
which will make freezing both simpler and more reliable. At the same ti water occur (Fig. 8.1), it will be worth studying the relevance of these transi-
the joint contributions of low temperature physicists, cryogenic engine. tions to changes in the specimen structure in the course of high resolution
and biologists will ensure that there is a greater awareness of the prob! studies of frozen-hydrated specimens.
involved in freezing and in the selection of appropriate methods for partl:
Jar problems. We can expect significant improvements in the 'quality~,;
freezing to result from work such as that of Bald (in press), who has n\
504 Robards and Sleytr Low temperature methods in biological electron microscopy
Future outlook 505

Ice in four temperature ranges


more fundamental information about the relationship between the nature
8- 30 30- 70 70 - 100 >100 K
of a frozen specimen and the sectioning/fracturing process. As this becomes

·-
available, new cryoultramicrotomes will doubtless be produced to take ad-
as, h
• '
vantage of the latest knowledge.

••''
as, I

le

• • :) .)
• 8.4 Freeze-drying

a Ih • • • • Freeze-drying has been with us for a longer time than most low temperature
methods but it still remains far from perfected in most apparatus as far as
e - irradiated

..
()
ultrastructura1 research is concerned. The primary requirements to ensure
as, h , reproducible and effective freeze-drying are that the specimen should be
I held at a precise, known temperature and that the ambient vacuum shonld
as, I '
• .)
~;_ ~:
also be well controlled. Once these conditions can be met, then there ap-

b
le

Ih • •


·-•".
'
pears every reason to think that freeze-drying will show an increased appli-
cation to biological specimens, not least because of criticisms of critical
point-drying as a preparation method for SEM (see §5.4L
Fig. 8.1. (a) Temperature transition ranges for the 4 types of solid water relevant to cryomic-
roscopy. as,h, high density amorphous ice (specific density 1.1 g cm- 3); as,l, low density amq.r-
phous ice (specific density 0.94 g cm- 3); le, cubic ice; Ih, hexagonal ice.(~) Possible.transitions 8.5 Freeze-etching
induced by electron bombardment. (Diagram constructed from data kindly provided by H.
Heide and E. Zeitler (personal communication) and from Heide and Zeitler (1984).)
Freeze-fracture and freeze-etching have now become extremely well estab-
lished techniques in numerous laboratories and have consequently made a
8.3 Cryoultramicrotomy
major contribution to our understanding of biological ultrastructure.
Nevertheless, they remain methods that require a high level of technical
Among the whole gamut of low temperature microscopical techniquesi
expertise. Furthermore, they are time-consuming and the interpretation of
there cannot be one that remains fraught with more technical problems than
results is a feature requiring the most detailed consideration. The cost of
cryoultramicrotomy. There is now some very fine commercial apparatus
good freeze-etching eqnipment is high and, partly for this reason, there is
available for cutting ultra-thin sections of frozen specimens (§4.2). However,
little choice of apparatus. This is a pity because further competition would
much remains to be done to determine how best to section different tissues .
undoubtedly stimulate still further improvements in both the applications
with different water contents. The influence of the frozen specimen itselfo.~,i~L:.
of the technique and in the ease of use of the eqnipment. With modern tech-
the results obtained is nowhere seen more clearly than with this techniqu~. ;\;:
nology it should be possible to produce freeze-fracture/etching equipment
The (relatively unsupported) school of thought that believes that sectio~mi;t
that would be less expensive to run (more economical in use of liquid nitro-
can take place at about 240 K without adverse effects from icnecrystalh~~
gen), would process many specimens within a short period of time and
tion is countered by the protagonists of the belief that the specimen temper
would not require 'high-level' operator expertise. Using modern techniques
ature must always be maintained below about 200 K. Furthermore, the ..
of microprocessor control, the latest information on fracturing devices,
ture of the ice crystals within the specimen seem to have an important effi
refrigerator cryopumps and vacnum/cryotechnology, the 'ultimate' freeze-
on the sectioning properties of the specimen. Thus, while very goo_d app~
etching device is waiting to be constructed. By fracturing and replicating at
tus is now available, it cannot be used to its fullest advantage unttl we h~.
4 K it should be possible to minimise artefacts.
506 Robards and Sleytr Low temperature methods in biological electron microscopy

8.6 Freeze-substitution Appendix l


As with freeze-drying, freeze-substitution has been undergoing something
of a revival. This is partly because the latest substitution methods do allow
complete solvation of ice at very low temperatures and partly because, when
combined with fow temperature embedding methods, a highly advanta-
geous processing pathway of sections for TEM is available. Unfortunately,
the very superiority of fixation during freeze-substitution means that infilt-
Safety
ration with resin can be difficult or even impossible because membranes are
so well preserved. Nevertheless, this is a potentially important preparation
route and, therefore, one worthy of much further work. Many are fright-
ened away from freeze-substitution because of the lack of appropriate
apparatus. Reference to Chapter 7 should dispel any notion that freeze-sub-
stitution is particularly difficult to use. Performed carefully, it should pro- Low temperatures are inherently dangerous. In addition to direct hazards
vide results from mauy specimens that challenge the very best that can be fro~ lo'": temperatures, workers in laboratories involving cryogenic meth-
produced by conventional techniques - with the added advantage that ods m microscopy will often be concerned with extremes of pressure (high
extraction of soluble components is very much reduced. vacuu~ apparatus such as electron microscopes or freeze-fracture appara-
tus; high pressure_ equipment such as gas cylinders and gas/liquid supply
Imes); with very high voltages; with toxic chemicals; with highly inflamma-
8.7 Conclusion ble materials; and with ionising radiation. The EM laboratory is potentially
a very unhealthy place! Nevertheless, sensible laboratory procedures ensure
It is inevitable that low temperature techniques will play an increasingly im- that nsks are minimised. It is particularly important that, when a scientist
portant role in the elucidation and understanding of biological ultrastruc- ~mbarks on a new field of study (for example, low temperature methods),
ture. Not only this, but they will also open up new possibilities for analytical ne or she attempts to become fully acquainted with the particular hazards
and cytochemical techniques so that new types of information can be associated with the work. By understanding potential difficulties they are
obtained. In turn, the improvement in preparation methods, with the conco- usually easily avoided. This Appendix cannot cover all aspects of safety in
mitant increase in 'quality' of data, means that our attention must also turn an EM laboratory although the references at the end will serve as an intro-
towards the greater use of quantitative methods in the analysis of purely duction for those who are concerned to enquire further into this matter.
structural data. Only in this way is it possible to have any long-term confi- The specific hazards oflow temperature techniques will now be discussed
dence in the results that are obtained from any micoscopical technique. in more detail. Among the useful references specifically dealing with hazards
at _low temperatures is the Cryogenics Safety Manual published by the
Bnt1sh Cryogenics Council and a pamphlet produced by Edwards High
References Vacuum (see references).

Bald, W.B. (In press), Heat transfer processes in cryofixation (Adam Hilger, Bristol). Categories of hazard
Haefer, R.A. (1981), Kryo-Vakuumtechnik (Grundlagen und Anwendungen) (Springer-Vet:- It is possible to separate the potential hazards that can be encountered while
lag, Berlin). using low temperature methods into different categories:
Heide, H.G. and E. Zeitler (1984), Water in Cryomicroscopy, Proc. 8th Eur. Reg. Con[ Bled:-.····:
tron Micoscopy, Budapest 2, 1388. (i) Direct contact with cold gas, liquid o; solid.
(ii) Problems of inhalation and asphyxiation with cryogenic gases.

507
508 Robards and Sleytr Low temperature methods in biological electron microscopy Safety
509

(iii) Expansion associated with decompression of liquefied gases. Insufficient protection against cold (for example when working in a cold-
(iv) Inflammable nature of some cryogenic gases. room) can lead to hypothermia. If the whole body temperature has become
(v) Toxic nature of some cryogenic gases. depressed (massive cold exposure) then the patient should be immersed in
a bath at _the temperature given above (between 42 and 45°C). Beware of
the poss1b1hty of shock as the patient re-establishes thermal equilibrium.
Al .1 Direct contact with cold gas, liquid or solid . The pal!ent should not be given alcohol or allowed to smoke after receiv-
mg a cold burn because these reduce blood flow to the frozen tissue.
Extreme cold destroys cells and tissues to a degree related to the duration While tissues that are still frozen look waxy and pallid and are usually
and temperature of exposure. Moreover, cold burns can be insidious in th8.t pamless, they become extremely painful, swollen and susceptible to infec-
frozen tissue does not necessarily feel painful or look very different from l!on when thawed. If a casualty occurs away from the possibility for medical
normal tissue. It is only on thawing that pain and damage become evident, attent10n, then it is best not to make any special attempt to thaw the frozen
a fact that will be well known to anyone who has suffered frost-bite. Because l!ssue but to transport the patient to the nearest source of medical help. Se-
heat transfer to a high thermal conductivity solid, such as a piece of metal, vere_ cold burns may require the administration of pain-killers and/or tran-
is far more efficient than heat transfer to a cold liquid, it is particularly im- qmhsers durmg thawing. If a severe cold burn has thawed before receiving
portant not to allow unprotected parts of the body to come into contact medical attent10n, then cover with a large sterile dressing to reduce the like-
with such cold solids. Protection against cold 'burns' thus takes the form lihood of infection. It may be advisable to administer an anti-tetanus boos-
of using appropriate protective clothing. ter injection after a severe burn; medical advice should be sought.
When handling cold liquids care should be taken that any spillage will
not accumulate in clothing, such as pockets, insides of boots or insides of
gauntlet-type gloves. Therefore protective clothing should be close-fitting so Al.2 Problems of inhalation and asphyxiation with cryogenic
that liquids will run straight off it. Similarly, it should be non-porous and gases
be kept dry. Gloves should be relatively loose-fitting (ease of removal) but
without large gauutlets into which liquids might fall. Types of glove vary: Cold _gas is not likely to be inhaled voluntarily over long periods but long-
if objects at low temperature are to be handled, then good insulation is at term mhalat10n can seriously damage the lung.
a premium; if liquefied gases are being used then the hands should not come Asphyxia can occur under any conditions where atmospheric oxygen con-
into direct contact with the liquid and leather or lined rubber gloves are ade: ,\"<.·.;; tent falls below _its normal level of 21% (v/v). Thus all cryogenic gases
quate. Overalls should not have open pockets or trouser turn-ups. Trouse(S ;<'"'· should be used m well-ventilated surroundings. Liquid gases release an
should be worn outside boots. The face, and particularly the eyes, should · · enormous equivalent volume of gas as they change from the liquid to the
be protected with a face mask. Use tongs or other insulated handling tool•~<;;, gaseous state. Thus, nitrogen expands by a factor of about 680 as it changes
whenever dealing with objects under a liquid cryogen. . .'.!:i• from hqmd to gas. Precautions must be taken to avoid asphyxiation which
If a person has suffered cold 'burns', first attempt to establish the severi))f;' . may start quite suddenly and, in some situations, can be at an advanced
For a small, localised injury treat the affected area with cold water and th stage before the patient is even aware of it. The remedy is to ensure that
with a cold compress. Do not use dry heat. If the skin is blistered, if high gas concentrations are never allowed to build up. Use small quantities
eyes are affected or if there are any other untoward symptoms, then me . of hqmd gas m a large, well-ventilated room or fume cupboard. Handling
advice should be obtained as soon as possible. For larger areas of darn. oflarge quantities ofliquid gas should be undertaken only in specially desig-
ensure that clothing is not restricting the blood supply and remov~. nated areas where there is ample ventilation. Do not travel in a lift with a
casualty to a sick room or hospital as soon as possible. Place the bur large ( > 1.0 litre) container of cryogenic fluid; if the lift becomes stuck you
water with a temperature between 42 and 45°C (higher temperatures o wiU.be in a confined space with a constantly increasing atmospheric concen-
heat will further 'burn: the already damaged tissue). · tratwn of the cryogen gas. In the event that asphyxiation does occur, the
-- - - - - - - - - - - - - - -
--
510 Robards and Sleytr Low temperature methods in hiological electron microscopy
Safety 511

victim should as quickly as possible be transferred to a well-ventilated envir- gen it is most important that the correct procedures for handling it are fully
onment (even outdoors) and have clothing around their neck and chest loos- appreciated.
ened. Medical advice should be sought as quickly as possible. Liquid oxygen, with a boiling point of - l 83°C (90 K) will condense into
any liquid bath colder than this (for example, into liquid nitrogen at
- l 96"C (77 K) or liquid propane at - l 89°C (84 K)). Although oxygen
Al .3 Expansion associated with decompression of liquefied itself will not burn, it will fuel the co.mbustion of other substances to burn
gases more fiercely and, under some circumstances, explosively. It will also allow
materials to ignite at temperatures lower than their normal ignition temper-
As mentioned above, liquefied gases expand greatly when vapourising, a ature. As liquid oxygen is seldom, if ever, used in low temperature micros-
I
volume increase of about 800-fold being quite typical. If this expansion is copy the problem is merely one of avoiding the condensation oflarge quan-
not under controlled conditions then it can be both dramatic and danger- tities of liquid oxygen from the atmosphere. Therefore, low temperature
ous. For example, if a cryogenic liquid such as liquid nitrogen is placed in cryogens < - l 83cC (90 K) should be covered as far as possible to avoid
a sealed container, any heat gain by the liquid will serve to boil off a little contact with the air (remembering, however, to allow venting to avoid pres-
more of the nitrogen as gas. Ultimately this will lead to an explosion. Thus sure build-up) and, for the same reason, the container should not be left
all liquid cryogen containers 1nust be provided with a means of venting to standing around for prolonged periods. Liquid oxygen is usually easily iden-
the atmosphere, whether valved or not. Similarly, storage Dewar vessels tifiable as, of all the common low temperature liquids, it has a characteristic
must never be completely sealed; this also applies to vacuum containers of light blue colour (the others are colourless).
the type used for insulating hot drinks (it is very easy to forget that the screw If there is a fire in which oxygen is involved, then shut off the oxygen sup-
seals on such containers must not be fully tightened). Large storage contain- ply (if possible); use water to extinguish the fire unless electrical equipment
ers, obtainable commercially, are built to stringent specifications and, are is involved, in which case use C0 2 , dry che1nical extinguishant or vapouris-
suitable for their specific purpose. Even here, however, care must be taken ing liquid. In attempting to extinguish fires involving cryogenic equipment,
to ensure that pressure-release valves do not become iced-up and inopera- do not spray water onto pipes, valves, etc. at low temperatures: the water
tive. Mistakes can be made as liquids are decanted from one container to will freeze and may immobilise valves and cause pressure to build up.
another, smaller, one. Unprotected vacuum flasks should not be used for Liquid propane, and similar combustible gases require very careful han- I
fear of implosion. At the very least they should be wrapped around with dling. The major problems arise during condensation of the gas (usually '

thick adhesive tape to prevent sharp pieces of glass from flying about if an !1I
from a pressurised cylinder) and during disposal of the gas because, at these
implosion occurs. In many situations it is simpler, cheaper and safer to use times, the temperature of the gas can rise above the flash point (for liquid d
!!
containers made from expanded polystyrene, polyurethane or similar mate- propane -104°C; 169 K). It is, therefore, important to duct propane gas ,I
rial· stainless steel containers are also satisfactory. Containers designed for
use,with C02 are not generally suitable for use with N,.
through coils of thin copper pipe immersed in liquid nitrogen before con- !I
densing it into a container (see §2.4.la). In this way, the propane is always
Always label containers clearly as to what they contain. Ensure that con- at a temperature « -100°C (173 K) when exposed to the atmosphere. In 11
tainers are clean and dry before filling with a cryogenic liquid. disposing of the propane (and again remembering the 800-fold volume in-
crease) care must be taken to lead the gas away through a spark-protected i
fume-hood or, alternatively, to tip the liquid onto the ground in the open
Al .4 Inflammable nature of some cryogenic gases air well away from all buildings. It goes without saying that there should i'I
be no smoking and no ignition of naked flames while working with combus-
Some cryogenic fluids, such as liquid oxygen and liquid propane, are tible gases and clear warning signs should be placed on the doors of rooms
I
inflammable in the gaseous state. With increasing use of propane as a in which such work is taking place.
512 Robards and Sleytr Low temperature methods in biological electron microscopy

Al.5 Toxic nature of some cryogenic gases


Appendix 2
Apart from the asphyxiation hazard (§Al.2), most cryogenic liquids/gases
are of relatively low toxicity. Some of the hydrocarbons are mild narcotics
and may cause dizziness, narcosis and nausea at low concentrations and
unconsciousness on continued exposure. Removal from exposure to vapour
will cause the symptoms to disappear. There have been reports of the car-
diac toxicity of some of the halocarbons used as the propellants in aerosol List of suppliers
cans and as cryogens (Freons, Arctons, etc.; Taylor and Harris 1970) but
little further information has been subsequently published. Ethane is an
anaesthetic gas and requires particular care in use.

Al.6 Conclusions
N.B. The following list includes the names of suppliers known to the auth-
By understanding the simple physical principles of what is involved and ors at th~ presen~ time. For reasons of space it has not been pos"Sible to in-
applying common sense and usual laboratory safety measures, most day-to- clu~e a111nternat1onal addresses or agencies for large companies; there will
day operations with low temperature equipment and chemicals are accom- obviously also be unintentional omissions and we will be pleased to hear
plished in absolute safety. However, if apparatus for cryogenic use is to.be of names and addresses of other suppliers for inclusion in later editions.
'home-made' then it is absolutely essential that a person who is fully The first list gives the names of suppliers of the many routine items
acquainted with cryogenic engineering is consulted about the use of appro- reqmred_for all methods in electron microscopy. Many of these suppliers
priate materials, tolerances and insulation. provide items relevant to low temperatures. Subsequent lists provide the
~ames and addresses for suppliers of specific items. If you cannot find a par-
References ticular product, then it is often useful to contact one of the major suppliers
(§!),who may well be able to help.
Alderson, R.H. (1975), Design of the electron microscope laboratory, in: Practical methods
in electron microscopy Vol. 4, A.M. Glauert, ed. (North-Holland, Amsterdam), p. 114.
Coward, H.F. and G.W. Jones (1952), Limits of flammability of gases and vapours, Bulletin l. General suppliers of materials and equipment for electron microscopy
503, U.S. Bureau of Mines.
Cryogenics Safety Manual - A Guide to Good Practice, Safety Panel, British Cryogenics (a) Agar Aids for Electron Microscopy (Agar also
Council, c/o Institution of Chemical Engineers, 165-171 Railway Terrace, Rugby CV21 Aids) Balzers High Vacuum Ltd.
3HQ, England. (U.K. agents for Ladd) Northbridge Road
Humphreys, W.J. (1970), Health and safety hazards in the SEM laboratory. Scanning Electron 66a Cambridge Road Berkhamsted
Microscopy I, 537. Stansted Hertfordshire HP4 JEN
Recommended safety precautions for handling cryogenic liquids (1979), Edwards Essex CM24 8DA England
Vacuum Publication 59-Kl00-00-880, B.O.C. Cryoproducts, Manor Royal, Crawley, England
Sussex RHlO 2LW, England. also
Taylor G.J. and W.S. Harris (1970), Cardiac toxicity of aerosol propellants, J. Amer. M .·. (b) Balzers Union AG (Balzers) Balzers
Ass. 214, 81. P.O. Box 75 FL-9496 Balzers 8 Sagamore Park Road
Fiirstentum Hudson
Liechtenstein New Hampshire 03051
U.S.A.

513
II
514 Robards and Sleytr Low temperature methods in biological electron microscopy
List of suppliers
515

(c) BEEM Graticules Ltd. (n) Polysciences Inc. (Polysciences) Extech International Corp. (Extech)
P.O. Box 132 (U.K. agents for EFFA) (U.S. agents for Polaron) (U.S. agents for Taab)
Jerome Avenue Station Morley Road Paul Valley 1-ndustrial Park 177 State Street ·
Bronx Botany Trading Estate Warrington Boston i
New York 10468 Tonbridge Pennsylvania 18976 Massachusetts 02109
U.S.A. Kent TN91RN U.S.A.
I
U.S.A.
England
(d) CW Frtinch Inc. (French) Polysciences Ltd.
58 Bittersweet Lane (r) Tousimis Research Corp. (Tousimis)
Touzart & Matignon 24 Low Farm Place
Weston 6000 Executive
(French agents for EFF A) Moulton Park Rockville
Massachusetts 02193 8 Rue Eugene Henaff Northampton NN3 IHY
U.S.A. Maryland 20852
94400 Vitry sur Seine England U.S.A.
France
(e) Denton Vacuum Inc. (Denton) (o) Soquelec Ltd.
Cherry Hill Industrial Center (s) Vaughn Electron Microscope Supplies Inc.
(j) Ladd Research Industries Inc. (Ladd) 5757 Cavendish, Suite 407
Cherry Hill 2176 Dunn Road
(U.S. agents for Agar Aids) Montreal
New Jersey 08003 Memphis
P.O. Box 1005 Quebec H4W 2W8
U.S.A. Tennessee 38114
Burlington Canada U.S.A.
Vermont 05402
(f) Electron Microscope Aids U.S.A. (p) SPI Supplies (SP!)
6 Lime Trees (t) Walter Mccrone Associates Inc.
Division of Structure Probe Inc. (Mccrone)
Malford (k) Marivac Ltd. P.O. Box-342
Chippenham 2820 South Michigan Avenue
1872 Garden Street West Chester Chicago
Wiltshire Halifax Pennsylvania 19380 lllinois 60616
England Nova Scotia B3H 3R6 U.S.A. U.S.A.
Canada
(g) Electron Microscope Sciences (EMS)
(q) Taab Laboratories Equipment Ltd. (Taab)
P.O. Box 251 (I) Nisshin EM Co. Ltd. 40 Grovelands Road
Fort Washington Espowarl Ichigaya 40-10 Reading
Pennsylvania 19034 Tomihisa-cho Shinjuku-ku Berkshire RG3 IBR
U.S.A. Tokyo 160 England
Japan
(h) EMscope Laboratories Ltd. (EMscope)
Kingsnorth Technology Park (m)Polaron Equipment Ltd. (Polaron) 2. Electron microscope accessories (special stages, mechanisms, etc.)
Wotton Road (U.K. agents for Polysciences)
(your electron microscope manufacturer should also be contacted)
Ashford 53-63 Greenhill Crescent
Kent TN23 2LN Watford Business Park (a) Advanced Metal Research (AMR) (d) EMscope Laboratories (EM scope)
England Watford Bedford (see lh)
Hertfordshire WDJ 8XG
Massachusetts 01730
(i) Ernest F. Fullam Inc. (EFF A) England U.S.A. (e) Ga tan Inc.
P.O. Box444
Schenectady 780 Commomwealth Drive
Ted Pella Inc. (Pelco) (b) Balzers (see lb) Warrendale
New York 12301 (U.S. agents for Polaron)
Pittsburgh
U.S.A. P.O. Box 510
(c) Cambridge Instruments Ltd. Pennsylvania 15086
Tustin
Rustat Road U.S.A.
California 92681
Cambridge CBI 3QH
U.S.A.
England
I
!
516 Robards and Sleytr Low temperature methods in biological electron microscopy List of suppliers 517

also (i) Philips Scientific and Industrial Division (b) FC 150 freezing attachment for Reichert also
Electron Optics Department OMU3 and FC4 and FC4D for Ultracut E DuPont
6678 Owens Drive TQ 1114 Eindhoven
Biomedical Products Division
Pleasanton The Netherlands C. Reichert AG Wedgwood Way
California 94566 Hernalser Hauptstrasse 219 Stevenage
I

also .I
U.S.A. A-l170Wien Hertfordshire SG 1 4QN

also Pye Unicam Ltd.


Austria England ·'I
(U.K. agents for Philips) Reichert Jung UK also
Ingolstiidter Str. York Street 820 Yeovil Road
Du Pont de Nemours France S.A.
40 D-8000 Miinchen 40 Cambridge CB l 2PX Slough 9 Rue de Vienne
West Germany England Buckinghamshire SLl 4JB Paris 8"
England France
(f) Hexland Ltd. (j) Carl Zeiss (Oberkochen) Ltd. 1,
'

W & G Estate P.O. Box 78 (c) Frozen thin seclioner attachment FSJOOO (d) SLEE TUL cryoultramicrotome
Faringdon Road Woodfield Road for MT6000 Sorvall ultramicrotome '
I
East Challow Welwyn Garden City
Slee Medical Equipment Ltd.
NrWantage Hertfordshire AL7 I LU Du Pont Company Lanier Works
Oxon OX12 9TF England Instrument Products Division
1'

Hither Green Lane


England Sorvall Operations London SE13 6QD
also Newton England
(g) Oxford Instruments Connecticut 06470
Osney Mead Carl Zeiss U.S.A.
Oxford OX2 ODX P.O. Box 1369
England D-7082 Oberkochen
I•
West Germany 4. Freeze-etch apparatus ii
(h) Anton Paar K.G.
(a) Balzers Union (see lb)
P.O. Box 58 (f) Leybold-Heraeus GmbH and Co. KG
Kaerntnerstrasse 322 Bonner Strasse 504
(b) Cressiugton Scientific Instruments D-5000 K6ln 51
A8054 Graz
34 Chalk Hill West Germany
Austria
Oxhey
Watford WDl 4BX (g) Polaron
3. Cryoultramicrotomes England
(see Im)

(a) Cryokit for LKB Ultramicrotome I, Ill (h) C. Reichert Ltd. (see 3b)
{c) Denton Vacuum Inc. (see le)
and IV also Cryonova
(i) Technics Electron Microscopy Systems
(d) Edwards High Vacuum Inc.
LKB Instruments Ltd. LKB Instruments Inc.
Manor Royal 7950 Cluny Court
232 Addington Road 12221 Park.lawn Drive
Crawley Springfield
South Croydon Rockville
West Sussex RHIO 2LW Virginia 22153
Surrey CR2 SYD Maryland 20852
England U.S.A.
England U.S.A.
(e) Hitachi Ltd.
LKB Instrument GmbH LKB-Produkter AB
4, I-chrome
Lochhamer Schlag 5 P.O. Box 305
Marunouchi Chiyoda-ku
P.O. Box 1396 S-161 26 Bromma
Tokyo
D-8032 Grafelfing Sweden
Japan
West Germany
518 Robards and Sleytr Low temperature methods in biological electron microscopy List of suppliers 519

5. Freeze-drying apparatus (f) L' Air Liquide (h) Thor Cryogenics


(Refer also to General Suppliers; see 1) Division Materiel Cryogenique Henley Road
57 Av. Carnot Berinsfield
(a) Freeze-Drying Device FD U 001/FDU0 JO Polaron (see lm) 94500 Champigny-s/Marne Oxford
France England
Balzers Union (see 1b) Chemlab Instruments Ltd.
602 High Road (g) Statebourne (Cryogenics) Ltd. (i) Union Carbide
Ilford 18 Parsons Road Union Carbide House
(b) Freeze-drying apparatus
Essex Parsons Industrial Estate P.O. Box 72
(see 4d) England Washington Rickmansworth
Edwards
Tyne & Wear Hertfordshire WD3 lAS
(see lh) (c) VirTisfreeze-dryers England England
EMscope

Meditrionics Inc. Techmation Ltd.


P.O. Box 11209 58 Edgware Way 8. Thermocouples
Dallas Edgware
Texas Middlesex HA8 8JP (a) Centemp (c) TC Ltd.
U.S.A. England 62 'curtis Road P.O. Box 130
Houndslow Uxbridge UB8 2YG
Middlesex TW4 5PT England
6. Low temperature freezers England

(a) Gesellschaft flir Labortechnik GmbH (c) Borolabs Ltd. (b) Labfacility Ltd.
Schulze-Delitzsch-Str. 4 Paices Hill 26 Tudor Road
D-3006 Burgwedel Aldermaston llampton
West Germany Berkshire RG7 4QU Middlesex TW12 2NQ
England England
(b) Scientemp Corp. (LABCOLD low temperature cabinets)
Adrian
Michigan 9. Temperature monitors and controllers
U.S.A.
(Scien Temp Lo-Cold freezer) (a) Bright Instrument Co. Ltd. (d) E. Leitz (Instruments) Ltd.
Clifton Road 48 Park Street
Huntingdon PE18 7EU Luton
7. Storage vessels England Bedfordshire LU 1 3HP
England
(a) Air Products Ltd. (c) Day-lmpex Ltd. (b) Euro therm Ltd.
Coombe House Earls Colne Faraday Close (e) Roach Associates Ltd.
St. George's Square Colchester C06 2ER Durrington IA Sidmouth Road
New Malden England Worthing Orpington
Surrey KT3 4HH West Sussex BN13 3PL Kent BR5 2EG
England (d) Edwards (see 4d) England England

(b) Air Reel (Cryoproducts) Ltd. (e) Jencons (Scientific) Ltd. (c) Hird-Brown Ltd. (Divn. ofMTE Ltd.) (f) Rosemount Engineering Co. Ltd.
Riverside Industrial Estate Cherrycourt Way Industrial Estate Hartford Works Heath Place
Stanbridge Road Weston Street Bognar Regis
Bridge Road
Littlehampton Leighton Buzzard Bolton West Sussex P022 9SH
West Sussex BN17 SDF Bedfordshire LU7 8UA England England
England England
520 Robards and Sleytr Low temperature methods in biological electron microscopy List of suppliers 521

(g) SLEE (see 3d) (h) Stanton Redcroft Ltd. 13. Cooling rate meters
Copper Mill Lane
London SWI 7 OBN (a) Agar (see la) (c) Hexland (see 2f)
England
(b) EM,.,ope (see lh)

10. Joule-Thomson cryotips


14. Liquid N, worktable (cryoprep stand)
(a) Air Products Ltd. (see 7a) (b) The Hymatic Engineering Co.
Glover Street Polaron (see Im)
Redditch
Worcestershire B98 BBQ
England 15. Cryogens; quenching media

(a) Liquid nitrogen DuPont (see 3c)


11. Thermoelectric coolers Air products (see 7a) (sold under the name 'Freon')

(a) Bailey Instruments Inc. (d) MelcorfMaterials Electronic Products BOC Ltd.
Dean and Wood (London) Ltd.
515 Victor Street 992 Spruce Street Great West Road
83/85 Mansell Street
Saddle Brook Trenton Brentford London El
New Jersey 07662 New Jersey 08648 Middlesex TW8 9DQ England
U.S.A. U.S.A. England
(Sold under the name 'Isceon')

(b) Cambion (e) Mullard Ltd. CryoService Ltd. IC! Ltd.


Electronic Products Ltd. Mullard House Units G/H 1-3 Mond Division
Castleton Torrington Place Blackpole Trading Estate East Runcorn
Sheffield S30 2WR London WClE YHD Worcester Cheshire
England England West Midlands WR3 8SJ England
England
(Sold under the name 'Arcton')
(c) Eberbach Inc.
P.O. Box 1024 L'Air Liquide (see 7e) J.T. Baker Chemical Co.
Ann Arbor
Phillipsburg
Michigan 48106 Matheson New Jersey
U.S.A. 30 Seaview Dr. U.S.A.
P.O. Box 1587 (Sold under the name 'Ucon')
Secaucus
12. Freezing devices New Jersey 07094 (c) Propane; isopentane
U.S.A.
(a) Balzers (see lb) (c) Reichert (see 3b)
(KF80 Immersion Cryofixation System, British Drug Houses Ltd. (BDH)
(Spray-freezing, jet-freezing, {b) Halocarbons
Escaig cold-block device) Broom Road
high-pressure freezing HPM 010 Sold under various trade names by most Poole
local refrigeration supply companies. Dorset BH12 4NN
(b) Polaron (see Im) Available from a wide variety of sources England
(Slammer (cold block)) and in different containers ranging from (see also companies under 15a and b)
small pressurised cans to large cylinders.
Contact local suppliers or general EM
supply companies (see also 1).
522 Robards and Sleytr Low temperature methods in hiological electron microscopy List of suppliers 523

16. Materials for shadowing and making replicas 19. Specimen grids

Most general suppliers (see 1) (c) Metals.for evaporation (a) General suppliers (see 1) (e) Smethurst High-Light Ltd. (see la)

(a) Carbon rods for evaporation Johnson Matthey Metals Ltd. (b) Gilder Grids (f) VECO Zeefplatenfabriek NV
100 High Street 23 Macfarlane Road P.O. Box 10
Johnson Matthey Chemicals Ltd. Southgate London Wl2 7JA Karel van Gelreweg 22
Orchard Road London N 14 6ET England Eerbeek (GLD)
Royston England The Netherlands
Hertfordshire SG8 SHE (c) Graticules Ltd. (see li)
England (d) Platinum-carbon pellets
(d)' LKB-Produkter AB (see 3a)
Morganite Special Carbons Ltd. Engelhard Industries Inc.
Northfields Delancey Street
Wandsworth Newark
20. Chemic~ls, resins etc.
London SW18 ING New Jersey
England U.S.A.
Most general_ suppliers (see 1) Fisons Scientific Apparatus (Fisons)
Bishop Meadow Road
(b) Formvar (e) Tungsten wire
G.T. Gurr (see 16b) Loughborough
Leicestershire LEl 1 ORO
G.T. Gurr Lamp Metals Ltd.
Polysciences Inc. (PolySciences) (see ln) England
Searle Scientific Services East Lane
Coronation Road Wembley
Taab Laboratories (Taab) (see lq) Polaron (see Im)
Cressex Industrial Estate Middlesex
High Wycombe England
(a) Alcian blue Polysciences (see ln)
Buckinghamshire
The Tungsten Manufacturing Co. Ltd. Alcian blue 8 GX p.a.
England
C.I. 74240 Taab (see lq)
Fishergate Works
Portslade Serva
Shawinigan Ltd.
Heidelberg (e) Mounting media.for cryosectioning
118 Southwark Street Brighton
G.D.R. Raymond A. Lamb
London SE! Sussex
6 Sunbeam Road
England England
(b) CA-15 London NWlO 6JL
DeloGmbH. England
Munich (Tissue-Tek II)
17. Quartz crystal thin-film monitor West Germany
Polaron (see lm)
Most larger general suppliers (e.g. Agar (c) Chromerge (Cryoembed)
Aids, see la; Polaron, see Im; EMscope, Fisher Scientific Co.
see lh) Fair Lawn (f) Ethylene glycol
New Jersey
U.S.A. Fisons Scientific Apparatus __
18. Nebulizers Bishop Meadow Road_
(d) Dimethyl sulphoxid~_ (DMSO) Loughborough
General suppliers (see 1) Leicestershire LEI I ORO
EMscope (see lh) England
Ted Pella Inc. (see lm)
(Pelco nebulizer)
524 Robards and Sleytr Low temperature methods in biological electron microscopy
List of suppliers 525
Fisher Scientific Co. (k) Polycationic ( cationised) ferritin
633 Greenwich Street also
New York Miles Scientific Monsanto Co.
New York 10014 Division of Miles Laboratories Ltd. 800 N. Lindbergh Boulevard
U.S.A. P.O. Box 37 St. Louis
Stoke Pages Missouri 63167
Kodak Ltd. (U .K. agent) Slough SL2 4L Y U.S.A.
Chemical Division England
Kirby
Liverpool also 21. Miscellaneous
England Miles Laboratories Inc.
Division of Research Products (a) Antistatic pi~tols (e) Cats-Neparofa (heat-sealing plastic lami-
(g) Hystoacryl blue P.O. Box2000 Agar (see la) nate)
1127 Myrtle Street Apparently no longer available; alter-
Braun-Dexon GmbH. Elkahart (b) Abrasive unit SBU 010 natively:
Melsungen Indiana 46515 Balzers (see lb)
West Germany U.S.A. Camvac Ltd.
(c) Airbrushes Burrell Way
(h) London L.R. Gold (1) Poly-L-lysine Simair Graphic Equipment Thetford
Sigma Chemical Co. 16 Woodsley Road Norfolk IT24 3QY
London Resin Company Ltd. P.O. Box 14508 Leeds LS3 IDT England
P.O. Box 34 St. Louis England (Melinex 850)
Basingstoke Missouri 63178 (Olympos HP 100 Series)
Hampshire RG22 5AS U.S.A. (f) Diamond knives
England The De Vilbiss Company Ltd. Diatome
47 Holborn Viaduct P.O. Box 551
Also available from general suppliers and also LondonECI CH-2501
agents (see l). England Switzerland
Sigma Chemical Co. (Aerograph 'Super 63')
(i) Lowicryl Fancy Road (g) Elvanol 51-05
Poole (d) Beckman Microfuge Centnfuge B Dupont de Nemour (see also 3c)
Chemische Werke Lowi Dorset BH17 7NH Beckman Instruments Inc. Div. Analytical and Biomedical Products
Gesellschaft mit Beschrankter Haftung England 250 Harbour Boulevard 50-52 Route des Acacias
P.O. Box 1660 P.O. Box 3100 CH-1211 Geneva24
D-8264 Waldkraiburg Miles (see 20k) Fullerton Switzerland
West Germany California 92634
(m)Polyvinyl alcohol Vinol 205 S U.S.A. (h) Flammable gas sensors
Also available from general suppliers and Air Products and Chemicals Inc. Draeger Safety
agents (see 1). Allentown also Draeger House
Pennsylvania Div. Beckman-Riic Ltd. Sunnyside Road
(j) Pl3N (Polyimideresin) U.S.A. Turnpike Road Chesham
Cressex Industrial Estate Buckinghamshire HP5 2AR
Ciba (n) Skybond 705 ( Polyimide resin) High Wycombe England
Monsanto PLC Buckinghamshire HP4 3NR
Edison Road England
Dorcan
Swindon SN3 5HN
England
526 Robards and Sleytr Low temperature methods in biological electron microscopy List of suppliers 527

Sabre Gas Detection Ltd. (m)Indium sheet Also widely available from general sup- (u) Ultra Turrax On1nimixer
Ash Road Johnson Matthey (see l 6a) pliers and agents (see 1) Sartorius Ltd.
Aldershot 18 Avenue Road
Hantshire GU12 4DD Goodfellow Metals (r) SIMCO-Single-Spike Belmont SM2 6JD
England Cambridge Science Park The SIMCO Company Inc. England
Milton Road 2257 North Penn Road
Sieger Gas Detection Cambridge CB4 4DJ Hatfield Janke & Kunkle
Fulwood Close England Pennsylvania 19440 P.O. Box 447813
Fulwood Industrial Estate U.S.A. Staufen
Sutton-in-Ashfield (n) Melinex Sheet West Germany
Nottinghamshire NG 17 212 !CI Ltd. (s) Thermanox sheet
England Plastics Division Lux Scientific Corp. (v) Zeolite 1JX niolecular sieve
P.O. Box 6 Thousand Oaks Leybold-Heraeus (see 4f)
(i) Gloves for low temperature use Bessemer road California
General suppliers and agents (see 1) Welwyn Garden City U.S.A. Also widely available from general
Hertfordshire AL 7 1HD suppliers and agents (see 1)
Boro Labs Ltd. England (t) Thin nickel screen
Paicas Hill Perforated Products
Aldennaston (see also Camvac 16e) Brookline
Berkshire RG7 4QU Massachusetts
England (o) Petriperm U.S.A.
W.C. Heraeus GmbH.
Medalnorth Ltd. Hanau
Markyate West Germany
Hertfordshire
England (p) Photographic materials
Eastman Kodak Co.
Gl Hardened stainless steel knives 343 State St.
R. Jung AG Rochester
P.O. Box 1120 New York 14650
D-6901 Nussloch bei Heidelberg U.S.A.
West Germany
llford (see 22c)
(k) Hedin heaters (cartridges)
Hedin Ltd. Also widely available from general sup-
Raven Road pliers and agents (see 1)
South Woodford
London EIS !HJ (q) Polystyrene latex beads
England Particle Information Service
Los Altos
(I) Ilford L-4 emulsion California
Ilford Ltd. U.S.A.
Ilford
Essex
England
526 Robards and Sleytr Low temperature methods in biological electron microscopy List of suppliers 527

Sabre Gas Detection Ltd. (m)Indium sheet Also widely available from general sup- (u) Ultra Turrax Omnimixer
Ash Road Johnson Matthey (see l 6a) pliers and agents (see 1) Sartorius Ltd.
Aldershot 18 Avenue Road
Hantshire GU12 4DD Goodfellow Metals (r) SIMCO-Single-Spike Belmont SM2 6JD
England Cambridge Science Park The SIMCO Company Inc. England
Milton Road 2257 North Penn Road
Sieger Gas Detection Cambridge CB4 4DJ Hatfield Janke & Kunkle
Fulwood Close England Pennsylvania 19440 P.O. Box 447813
Fulwood Industrial Estate U.S.A. Staufen
Sutton-in-Ashfield (n) Melinex Sheet West Germany
Nottinghamshire NG17 212 ICI Ltd. (s) Thermanox sheet
England Plastics Division Lux Scientific Corp. (v) Zeolite 13X molecular sieve
P.O. Box 6 Thousand Oaks Leybold-Heraeus (see 4f)
(i) Gloves for low temperature use Bessen1er road California
General suppliers and agents (see 1) Welwyn Garden City U.S.A. Also widely available from general
Hertfordshire AL 7 1HD suppliers and agents (see 1)
Boro Labs Ltd. England (t) Thin nickel screen
Paicas Hill Perforated Products
Aldermaston (see also Camvac 16e) Brookline
Berkshire RG7 4QU Massachusetts
England (o) Petriperm U.S.A.
W.C. Heraeus GmbH.
Medalnorth Ltd. Hanau
Markyate West Germany
Hertfordshire
England (p) Photographic materials
Eastman Kodak Co.
G) Hardened stainless steel knives 343 State St.
R. Jung AG Rochester
P.O. Box 1120 New York 14650
D-6901 Nussloch bei Heidelberg U.S.A.
West Germany
Ilford (see 22c)
(k) Hedin heaters (cartridges)
Hedin Ltd. Also widely available from general sup-
Raven Road pliers and agents (see 1)
South Woodford
London El8 lHJ (q) Polystyrene latex heads
England Particle Information Service
Los Altos
(1) Ilford L-4 emulsion California
Ilford Ltd. U.S.A.
Ilford
Essex
England
Index for list of suppliers

Numbers referred to are the sections in Appendix 2.

Abrasive unit, 21 b EPA 101 vacuum unit, 2h


Airbrush, 21 c Electron beam guns, 1b, lj, 4b
Alcian blue, 20a Elvanol 51-05, 22g
Antistatic pistols, 2la Escaig cold block device, 12c
Arcton, 15b Ethylene glycol, 20f
Autoradiographic emulsions, 211, 21 p Evaporation materials, 16c

Beckman microfuge, 21 d FC J 50, FC4 and FC4D cryoultramicrotomy


attachments, 3b
CA-15, 20b Ferritin, 20k
Carbon rods, 16a Flammable gas sensors, 21h
Cats-Neparofa, 2le Formvar, 16b
Centrifuge, 2ld Freeze-etch apparatus, 4
Chromerge, 20c Freeze-drying apparatus, 5
Cold block-freezing device, 12b, 12c Freezing devices, 13
Cooling rate meter, 13 Freon, 15b
Cryoimmersion device, 12c Frozen thin sectioner attachment FSlOOO, 3c
Cryoultramicrotomes, 3
Cryoembed, 20e Gloves for low temperature use, 21i
Cryogens, 15 Grids, 19
Cryokit, 3a
Cryoprep stand, 14 Halocarbons, l 5b
Cyanoacrylate glue for low temperature use, Heat-sealing plastic, 2le
20b,20g Heating cartridges, 21k
Hedin heaters, 2lk
DMSO, 20d High pressure freezing device, l 2a
Diamond knives, 21 f Hystoacryl blue, 20g
Dimethyl sulphoxide, 20d
Ilford L-4 emulsion, 211

529
530 Robards and Sleytr Low temperature methods in biological electron microscopy

Indium sheet, 2lm Polyimide resin, 20n, 20j


Isceon, l5b Polystyrene latex beads, 2Iq
Isopentane, l5b Polyvinyl alcohol, 20m
Propane, J 5c
Jet-freezing device, l2a
Joule-Thomson cryotips, IO Quartz crystal thin-film monitor, 17
Quenching media, 15
KF80 Immersion Cryofixation System, l2c
Knives (steel), 2lj
Subject index
Resins, I, 20h, 20i
Knives (diamond, for low temperature use), 2
SIMCO-Single-Spike, 2lr
LABCOLD low temperature cabinet, 6c SLEE TUL cryoultramicrotome, 3d
Liquid N 2, l5a Scien Temp Lo-cold freezer, 6b
Liquid N 2 worktable, 14 Skybond 705 (polyimide resin), 20n
London L.R. Gold, 20h Slammer, l2b
Low temperature freezers, 6 Specimen grids, l 9
Lowicryl, 20i Spray-freezing apparatus, 12a
Accelerating voltage Ammonium acetate buffer
Stainless steel knives-hardened, 2lj
very low for viewing uncoated for freeze-drying 273
Melinex sheet, 2le, 2ln Storage vessels, 7
specimens 284 Amorphous ice l l
Metal screens (mesh), 2lt Accessory chamber AMR Bio-Chamber
Metals for evaporation, l 6c Temperature monitors, 9 for LTSEM 167 for LTSEM 169
Molecular sieve, 21 u Temperature controllers, 9 Acetone Amyl acetate
Mounting media, 20e Thermanox sheet, 21 s freeze-substitution solvent 466-467 in freeze-drying 270
Thermoelectric coolers, 11 freezing point vs. concentration 488 Amylamine
Nebulizers, 18 Thermocouples, 8 Acrolein for positively charging substrates 325
Thin nickel screen, 2 l t for freeze-substitution 465-466 Ancillary equipment
Olympos HP 100, 2lc Tissue-Tek II, 20e Adiabatic expansion for low temperature microscopy 156-158
of nitrogen gas 154 Animal tissue
Tungsten wire, 16e
Adsorption artefacts methods for cleaning replicas of 401
Pl3N (polyimide resin), 20j in freeze-drying 293 pretreatment for free;;:ing 86-88
Pelco nebulizer, 18 Ucon, l5b Adsorption technique Anode assemblies
Peltier stages, 11 Ultra Turrax Omnimixer, 2lu for small specimens prior to freeze- for electron-beam guns 380
Petriperm, 200 Ultra-violet fluorescent tubes, I drying 274 Anticontaminator
Photographic materials, 2lp (also 1) Air brush for LTSEM 155, 175
Plastic laminate for heat sealing, 2le Vino! 205 S (polyvinyl alcohol), 20m for spray-freezing 115-116 Anti-roll systems
Platinum-carbon pellets, 16d VirTis freeze-dryer, Sc Air-drying 244 for cryoultramicrotomy 205, 226-227
Poly-L-lysine, 20k artefacts 294 Antistatic pistols
Polycationic ferritin, 20k from different liquids 245 in cryoultramicrotomy 227
Zeolite, 21 v
Air-lock system Apparatus
for freeze-fracture replication units 333 for freeze-drying 261-268
Alcian blue for freeze-substitution 469-474
for positively charging substrates 273, for low temperature polymerisation 493
325 for vacuum embedding and vapour fixa-
practical method for coating tion 290
substrates 326 Arctons (see Halocarbons)
Algal cells Artefacts
pretreatment for freezing 81 adsorption during freeze-drying 293
Allium cepa air drying 294
pretreatment of roots for freezing 83 collapse during freeze-drying 295-299

531
532
533

critical point drying 269 of liquid cryogens 49 Carbon rods rates 62-65
dehydration during freeze-drying of liquids used for air-drying 245 for evaporation 376--377 Cold block devices
295-299 of some liquids 246 Cardiac myocytes Boyne type 125
drying 245 Boltzmann's constant 250 frozen at different rates 68 Dempsey and Bullivant type 125
fixation prior to freeze-drying 293 Bond breakage Cardiolipin liposomes 437 Escaig type 126
freeze-fracture replication 414--418 from beam damage 179 Cationisation of specimen supports 273 Heuser type 123-124
freezing prior to freeze-drying 295 Bound water 10 Cell aggregates hydraulic-pneumatic damping system
recrystallisation in freeze-drying 295 significance in freeze-drying 255 method of orientating for freeze- for 125
rehydration after freeze-drying 299 Bovine serum albumin fracture 321-322 magnetic anti-bounce mechanism
sampling and pretreatment 74-75 freeze-substitution test specimen 476 Cell components for 123-124
shadowing and replication 300 molecular weight standard 439 pretreatment for freezing 79-83 specimen assembly for 123
solvents in mounting media 283 Buffers Cell fragments Van Harreveld type 122
washing prior to freeze-drying 293-294 volatile, for freeze-drying 273 methods for cleaning replicas of 399 Cold block freezing
Asphyxiation Bulk specimens Cell structure comparison with other freezing
with cryogenic gases 509-510 examination in SEM 279-285 in relation to cooling rates 67---68 methods 93
Autoradiography freeze-drying, containers for 282 Charged particles comparison with plunge-freezing 222
of freeze-fractured specimens 419-427 freeze-drying of 278-293 deflection of during shadowing 374, 378 for cryoultramicrotomy 221-222
using frozen thin sections 235-236 freeze-fracture, autoradiography of Charging practical methods for 121-127
42()-423 uncoated specimens for LTSEM 161 Cold burns 508
Bacitracin 431 frozen 158-177 Chemical desiccants 260 Cold gases
as a wetting agent 273 frozen-fractured 163-164 Chemical polymerisation effects of direct contact with 508-509
Bacteria mounting for freeze-fracturing 316--318 at low temperature 494 Cold plates
freeze-fractured 408 Bullivant and Ames freeze-fracture of Lowicryl resins 492 anticontaminators in SEM 155
nomenclature for freeze-fractured device 338 Chicken muscle fibre bundle (see also Anticontaminators)
envelope 413 deep etching with 394 freeze-substituted 481 Cold shroud devices
Ballistic methods of sampling and description of 361-363 Chips for freeze-fracture replication 350--356
freezing 88-91 for freeze-drying 265 of ice in freeze-fracturing 331, 395 Cold shrouds
Balzers Bungarotoxin 424 Chloroform as heat shields 152
freeze-etching unit, description of Butanol in air-drying specimens 245 Cold stages
343-347 in freeze-fracture autoradiography 423 Christensen system different possibilities in SEM 153-155
freeze-etching unit, liquid helium~cooled Butyl benzene for cryoultramicrotomy 210--211 for high resolution 193-195
shroud for 354--356 dissolving from replicas 398 Chromel/Alumel 25 forLTSEM 176
UHV freeze-fracture device, description for mounting cryosections 231 Cleaning for transmission use in SEM 192
of 358-361 for mounting freeze-fracture specimens freeze-fractured specimens 397--403 for X-ray microanalysis 152, 194-195
Beam current meter 317-318 Clearance angle liquid helium-cooled 152
for X-ray analysis 193 in spray freezing 115 in cryosectioning 203 temperature sensing elements for
Beam heating low temperature glue 206 Clostridium thermohydrosulfuricum 153-155
of cold specimens in electron micro- cell surface of by deep etching 429 transfer module for JEOL JEM
scopes 179-183 CA-15 Coated pits !DOCX 189
Beckman Microfuge 476 for mounting monlayers for freezing 323 deep-etched, rotary shadowed 286 Cold surfaces
Beryllium Calibration Coating heat transfer at 60--65
for X-ray analytical cold stages 177 of thermocouples 28 frozen-dried bulk specimens 284 Cold traps 389
Bi-directional shadowing 385 Carbon possibilities for cold specimens 157-158 for freeze-fracture replication 311
for particle size analysis 441 for coating cold specimens 173 specimensforLTSEM 171-173 in freeze-drying apparatus 257-264
Binary system 13 physical properties of 370 thin frozen specimens 189 in LTSEM 161
simple phase diagram for 16 Carbon dioxide Coefficient of dispersion Collapse
Bioetch (see Leybold-Heraeus Bioetch 2005) critical point of 269 for freeze-fractured membrane during freeze-drying 295-299
Biological specimens Carbon fibre string particles 441 of structures during freeze-fracturing
treatment prior to freezing 66--91 for carbon films 158 Cold blocks 416-417
Biot modulus 41 for low temperature specimens 163 measurement of freezing rate on 35--37 Collapse temperature
Bismuth 17 Carbon films silver surface for 122 in low temperature embedding media
Boiling points interference colours from 381-382 theoretical assessment of cooling 478-479
534
535

Collodion criteria for citation of 60 asphyxiation with 509-510 LKB Cryokit attachment 214--215
for replica strengthening 386 effects on water withdrawal from cells 67 disposal of 511 merits and demerits of different types
Comparison in liquid cryogens 55-60 effects of inhalation 509-510 207-208
of cold block and plunge freezing 222 in relation to plunge depth 33 inflammable nature of 510--511 Reichert-Jung Cryochamber FC4 system
of rapid freezing methods 127-128 in relation to plunge velocity 34, 52 toxic nature of 512 for 214--218
Compensating wire in relation to specimen shape 52 Cryogens (see also Liquid cryogens) sublimation rate in 2~_3
for thermocouples 33 in relation to structure and viability Cryoguns 89-91 Cryoultramicrotomy 201-236
Complementary replicas 66-71 Cryomicroscopy 148-149, 503 cold block freezing for 221-222
device for Leybold-Heraeus Bioetch in relation to temperature of cryogen 95 Cryoprotectants 72-79 embedding and mounting specimens
2005 352-353 in relation to thermocouple size 24 artefacts from use of 74 for 219~220
devices for producing 340-343 influence on ice crystal size 10 natural 77 freezing specimens for 220--223
interpretation of 409 measurement of 25-38 non-penetrating 78 future outlook 504--505
specimen holders for 314 representative from model specimens 45 penetrating 78 knife holders for 213
specimen mounting methods for 325-327 required for vitrification 12 properties of 76 specimen preparation for 218-225
Concanavalin A Copper volatile 397 theory of_ 201-20_5
radio-iodinated and fluoresceinated 427 braided, for heat conduction 154 Cryopumps wet sectioning 233-235
use for binding to specimen supports 425 fine thread, for heat conduction 154 for freeze-drying apparatus 260 Crystal embryo 9
Condensation specimen support 103-104 Cryosectioning Crystallisation
physical conditions for during etching thermal conductivity of 62 anti-roll plate for 205 extracellular 13
388-392 Copper/Constantan 25 clearance angle for 203 Cubic ice
Condensation contamination CPD (see Critical point-drying) critical temperature during 204 conversion to hexagonal ice 21
artefacts arising from in freeze- Cracking knife angle 203-204 Cyanoacrylate glue
fracture 415 from thermal stress during freeze- knives for 205 mounting monql_ayers for freezing 323
during shadowing 371, 379 drying 295 melt-freeze theory 202 Cyclohexane
in freeze-drying 264 Criteria metal machining model 202-203 trough-liquid in cryoultramicrotomy 234
Condenser system for citation of cooling rates 60 plastic deformation in 202-203 Cytochemical methods.
for freeze-drying apparatus 260 for efficient cooling in liquid rake angle for 203-204 in freeze-fracture replication 418-443
Conduction losses cryogens 110 theory of 201-205 on frozen thin sections 235-236
in thermocouple lead wires 30 for establishing frozen-hydrated Cryosections 187 Cytochrome c
Connectors state 182-183 applications of 235-236 molecular ·weight determination of 439
for thermocouples 32 for high quality freeze-fracture collection using sucrose droplet 229-230 Cytoskeletal elements
Contamination replicas 405 contrast enhancement of 234 deep-etched, rotary shadowed 286
artefacts arising from in freeze- Critical cooling velocity 10 contrast in 233 Cytoskeleton
etching 415 ofwater 7 devices for flattening (pressing) 231-233 frozen-dried 285-287
avoidance in electron microscopes 156 Critical point freeze-drying of 233, 292-293
by condensation in freeze-drying 264 of carbon dioxide 269 mounting of 230--233 Damage
in freeze-fracture replication 311 of Freon 13 269 sublimation rate of ice from 292 to frozen-dried specimens__Qy_infiltration
phenomena during etching 392-396 ofwater 244 Cryosorption 287-288
problems during high resolution Critical point-drying 244, 269 during freeze~drying 267 Damping system
shadowing 369-374 comparison with freeze-drying 298 Cryosorption pumps for rapid injection devices 106
Contrast of mouse leukemia virus 298 for freeze-fracture replication 356 DC conductivity of ice 161
in cryosections 233 replication of specimens 434 C_ryostat ultramicrotomes 208-210 Decoration effects
Contrast enhancement Critical temperature Cryostats 205 during freeze-fracture replication 368
for 'wet' cryosections 234 during cryosectioning 204 Cryotransfer unit during shadowing 371 -
Convection Cross-linking for Philips 400 TEM 188-189 specific, of hydrophobic areas 396
natural or forced 38 from beam damage 179 Cryoultramicrotomes D"eep etching 392
Cooling methods Cryofixation basic requirements for 206--207 with Bullivant-type apparatus 394
comparison of 65-66 evaluation of 38 Christensen system 210--211 Deep freezes
Cooling rate meter 31 Cryofixatives 73 different types of 207-218 ultra-low temperature 130-
Cooling rates Cryogenic fluids Dollhopfsystem 211 Deflection systems
citation of 23 heat transfer in 38-53 for fracturing prior to freeze-etching for charged particles during _
comparative, in different cryogens 56 Cryogenic gases 365-366 shadowing 374
536
537
Dehydrating agents Disposal
freezing points of 488 of inflammable compressed cryogenic Elvanol 320 methods for low temperature specimens
Dehydration gases511 EM - electron microscope 157-158
artefacts during freeze-drying 295-299 DMP (see Dimethoxypropane) Embedding vs. sputter coating for LTSEM 162-163
for low temperature embedding 487-489 DMSO (see Dimethylsulphoxide) after freeze-drying 278, 287-292 Excitation volume
shrinkage artefacts 294 Dollhopf system at low temperature 486--494 influence of space charging on 162
Density for cryoultramicrotomy 211 specimens for cryoultramicrotomy Extracellular crystallisation 13
influence on cold block freezing 64 Double-axis rotary replication 385 219--220
Denton freeze-fracture replication unit Double-tilt specimen tip Emscope Sputter-Cryo system 166-167 Fab' fragments 429
complementary device for 342 for Philips EM 400 cold stage 190 Emulsion coating Faraday cage
description of 348-351 Drop-freeze-fracturing device for freeze-fracture autoradiography for X-ray analysis 193
Desiccants for complementary replicas 343 421-423 Ferritin
for freeze-drying 260 Droplet method Encapsulating media efflux of water in freeze-substitution 469
Desorption of freeze-fracturing 343 for cryoultramicrotomy 220 model system for rotary shadowing
artefacts during freeze-drying 294 of rapid freezing 120 Energy 383-384
during freeze-drying 255 Dry ice/xylene produced during sectioning 202 polycationic 429
Desulfotomaculum nigrificans low temperature mixtures 489--490 Entry velocities standard for molecular weight 439
fracture faces of 408 Dry sectioning for rapid plunging devices 108-109 Ferritin-antibody complexes 434
Deuterium oxide in cryoultramicrotomy 225 Environmental cell l, 243 Filipin 431
in determining rate of freeze-drying 254 Drying artefacts 245 Equilibrium temperature Film boiling 22-23, 48
Dextran 76 Drying methods ofwater 5 Film thickness
Dibenzoyl peroxide 492 different types for electron microscopy Erythrocytes monitoring 380--383
Dielectric constants 243 membrane structure vs. cooling rate 70 Finite elements method
of water and ice 35-36 Dupont-Sorvall cryoultramicrotome Escaig UHV freeze-fracture device for calculating cold block cooling
Diethyl ether 210-214 description of 356-358 rates 64
for lipid extraction from fracture Etching 388-397 Fire
faces 436 non-aqueous systems 396-397 extinguishing in cryogenics
freeze-substitution solvent 466--468 Earle's basic salt solution 321 (see also Freeze-drying, Sublimation of ice, laboratory 511
in air drying specimens 245 Elastic recontraction Freeze-etching) Fixation
Diffusion resistance of deformed structures in freeze- Ethane artefacts in freeze-drying 280, 293
for water molecules during freeze- fracture 416 as a cryogen 58, 186 for low temperature embedding 487
drying 259 Electrical capacitance method Ethanol in freeze-substitution 475-477
Digitonin 431 for determination of freezing rates 35-37 freezing point vs. concentration 488 of bulk specimens for freeze-drying 280
Dilauryl lecithin Electrode assemblies use in air-drying specimens 245 properties of glutaraldehyde in
in estimation of cooling rates 38 for resistance heating 376-377 virtues as a liquid cryogen 57 methanol 476
Dimethoxypropane 269 Electron beam evaporation volatile cryoprotectant 397 Flash point
Dimethyl sulphoxide anode materials for 380 Ethylene glycol of liquid propane 511
cryoprotectant 78, 84 methods of 378-380 cryoprotectant 78 - Flow diagram
properties of 76 Electron dose-dependent transitions freezing point vs. concentration 488 for freeze-fracture replication 312
trough liquid in cryoultramicrotomy of ice 503-504 properties of 76 Fluoroethylene-propylene foil
233-234 Electron guns (see also Electron beam evapo- Euglena gracilis for growth of monolayers 322
use in cryoultramicrotomy 206 ration) freeze-substituted 478--479 for mounting monolayers 319
Dimethylformamide design of 373-374, 378-379 survival of under different pressures 117 Forced convection 38, 57
as replica cleaning agent 387 Electron microscopes Eutectic temperatures Forceps
Diode sputter-coating cold stage requirements for 151-155 of some solutions 15 special, for holding paired rivets 324
for low temperature specimens 157-158 Electron transmission method Eutectics 11-20 Fourier rate equation 39
Dioxane for detecting loss of ice from Evaluation Fracture faces
volatile cryoprotectant 397 specimens 183 of cryofixation 38 lipid extraction from 436-437
Direct observation of cold specimens Electrostatic charging Evaporants specific decoration with water vapour
147-196 of sections in cryoultramicrotomes 227 physical properties of 370 437-438
different preparation possibilities for 150 Electrostatic interactions Evaporation Fracture plane
future outlook 503 during freeze-drying 293 for freeze-fracture replication 367-385 fine topography of 329-310 1!
for SEM coatings 162 gross topography of 330--332
Ir

::'II
538 539

Fracturing Freeze-fracture autoradiography Freeze-substitution 461--486 Frozen thin sections 178-183


for freeze-fracture replication 329-367 ofmonolayers 424-427 apparatus for 469--474 Frozen thin specimens 177-196
possibilities for L TSEM specimens 164 with sectioned replicas 424 characteristics of liquids used for 464 preparation of 184-185
specimens for LTSEM 165-170 Freeze~fracture de-Vices for SEM 484--485 Fungal mycelia
Fracturing/coating chamber Bullivant and Ames type 338 for thin-sectioning 475-484 pretreatment for freezing 81
for LTSEM 170 converging wedge type 339-340 future outlook 506 Fungal spores
Fragile specimens for producing complementary fracture localisation of water-soluble substances pretreatment for freezing 80
mounting method for freeze-fracturing faces 340--343 485-486
317-318 slidiri.g cold block type 338-339 practical methods for 474-486 Gallium 17
Free radicals Freeze-fracture replicas processing sequences for 463 Gas composition
produced from r,adiolysis of ice 179, 181 interpretation of 404-418 rate of dissolution of ice 465-467 artefacts from 74
Freeze-drying 243-300 labelling with morphological schedule ofMilller 477--478 Gas constant 250, 257, 389
artefacts from -293-300 markers 42-8-429 schedule of Van Harreveld 482 Gases
basic principles of 247-256 Freeze-fracture replication 309--443 Freezing expansion of liquefied on decompression
changes in linear dimensions during 297 (see also Freeze-etching) artefacts prior to freeze-drying 295 510
comparison with CPD of virus 298 artefacts of 414-418 comparison of methods 92-93, 127-128 inflammable nature of cryogenic
designs of apparatus for 261-268 autoradiographic methods for 419--427 for cryoultramicrotomy 220-223 gases 510--511
examination of bulk specimens in cytochemical methods for 418--443 future outlook 502 Gast carbon vane turbine pump
SEM 279-285 flow diagram for 312 hyperbaric 42 for freeze-fracture replication 356
flow diagram of possibilities for 247 future outlook 505 methods 91-128 Geymeyer freeze-fracture method
from non-aque-OUs solvents 268-272 · historical aspects of 310--311 pathways of specimen processing for 94 description of 363-365
future outlook 505 in combination with negative staining patterns of glycerol 14 Glitsch-type evaporation 375--377
in freeze-etching units 263 439-440 patterns of sodium chloride 14 Glues
mass and heat transfer during 252 interpretation aids for 440-443 speciinens for freeze-fracture replication cyanoacrylate for mounting mono-
modem vacuum systems for 262 principles of 309-313 327-328 layers 323
of bulk specimens -278-293 quantification of structural details time taken vs. heat transfer coefficient 47 for use at low temperature 206, 223
of cryosections 233, 292-293 441l-443 with liquid cryogens 91-120 Glutaraldehyde
of mouse leukemia virus 298 quantitative metliOds for 414-443 Freezing points artefacts from use of 75
of specimens mounted on EM grids 265 shadoWing aiid replication for 367-388 of commonly used dehydrating fixation properties in methanol 476
possible ice recrystallisation during 253 specimen mounting for 313-327 agents 488 for fixation of plant cells 85
practical aspects-· 256--268 specimen preparation for 313-328 Freezing rates (see also Cooling rates) Glycerol
procedure from amyl acetate 270 specimen supports for 313-:-327 determination of by electrical capacitance as a cryprotectant 78
prolongation faCtOi during 254 ultra-high vacuum systems for 356-361- 35-37 cryoprotective effect on yeast cells 69
rate in biological specimens 251-256 Freeze-fracturing 329-367 Freon 13 freezing patterns of 14
simple device fo,:t:_ use .in liquid nitrogen conditions for 332-333 critical point of 269 properties of 76
266-268 instruments for 336-367 for freezing in cryoultramicrotomy Glycerol mono~oleate
simple vacuum systems for 262 methods for producing single 220-221 use in liquid paraffin freezing
small biological specimens 272-278 fractures 337 Freons (see Halocarbons) method 119
specimen containers for bulk prior to freeze-dryi_ng 281 Frictional heating Glycerol/water system
specimeiiS 282 single fracture methods 336-340 contamination from during freeze- phase diagram for 18
specimen holders for 291 under improved vacuum conditions fracture 396 Glycols
specimen replication for TEM 285-287 347-361 Frozen bulk specimens 158-177 for freeze-substitution 462
times at different temperatures 248--249 under liquid gas 361-365 Frozen-fractured bulk specimens 163-164 Glycoproteins
Freeze-drying apparatus under liquid helium 363-365 Frozen-hydrated bulk specimens freeze-dried from bacterial cell walls 296
for vapour fixation and embedding 290 under liquid nitrogen 361-363 practical methods for 159--163 specific labelling of in freeze-fracture 428
schematic repri;:;;entati()n o_f.main under liquid nitrogen for freeze- Frozen-hydrated specimens Gold-affinity ligands 434
types 263 drying 270 criteria for assessing 182-183 Grain boundaries
simple, for use in vacuum systems with a microtome 343-347 Frozen specimens of ice crystals 13
264-266 Freeze-sectioning 201-236 alternative preparation methods for Grain structure
Freeze-etching (see also Cryoultramicrotomy, Cryo~ LTSEM 159 of coatings for SEM 158
(see also Freeze-fracture replication) sections) wanning of 20-21 Gravitation forces
future outlook 505 applications of 235-236 Frozen thin layers 183-184 artefacts from 74
540 541

Helium (see also Liquid helium) DC conductivity of 161 Inflammability


Grids
in cold block-freezing 122-127 dielectric constant of 36 ofsomecryogenicgases 510-511
mounting replicas on 402--403
Heptane heat capacity 7-8 Inhalation of cryogenic gases 509-510
Gross UHV freeze-fracture device
low temperature glue 206, 223 hexagonal 12 Injection devices for rapid freezing
description of 358-360
Heterogeneous nucleation interconversion in relation to tempera- (see also Plunging, Immersion)
of water 9 ture 21 entry velocities for 108-109
Halocarbon 11
Hexagonal ice 12 lattice structure of 13 solenoid-operated 109
in air-drying specimens 245
lattice structure of 13 mass loss in electron microscopes 181 spring-powered 105-108
Halocarbon 13
Hexland microcrystalline 12 Interference colours
in air-drying specimens 245
cold stage for LTSEM 176 nucleation of 5-11 from deposited carbon films 381-382
Halocarbons
High pressure freezing 18, 116--119 radiolysis of 178-181 Interpretation aids
condensation for use as cryogens 99
apparatusfor 118-119 rate of dissolution in freeze-substitution in freeze-fracture replication 440-443
contamination in cryoultramicrotomy
comparison with other freezing 465-467 Ion-beam sputtering 157-158
22()-221
methods 92 rate of sublimation 247-251 Iridium
definition of 48
of mouse cartilage 480 recrystallisation during freeze-drying 253 physical properties of 370
health hazards of 58
High resolution microscopy recrystallisation of 7, 19 Isentropic compression 42
practical use as cryogens 98-100
cold stages for 193-195 saturation vapour pressure of 390 Isolated cell components
problems when used for X-ray
High resolution replication specific volume in relation to temperature pretreatment for freezing 82
analysis 58
of frozen-dried specimens 285-287 20 Isolated cells
Hazards
High voltage electron microscope sublimation in low temperature pretreatment for freezing 79-83
categories at low temperatures 507-508
observation of frozen thin specimens microscopy 156--157, 170-171 Isopentane 206
(see also Safety)
in 184 sublimation of by radiant heating 171 for mounting cryosections 231
Heart tissue
Hoar frost sublimation of by raising bulk tempera- in air-drying specimens 245
sampling with ballistic device 88-90
removal of during freeze-fracture 336 ture 171 to lower melting point of propane 59
Heat capacity
Homogeneous nucleation 13 subliming from microtome knife Isopropyl alcohol
definition of 53
ofwater 6 edge 395 in high pressure freezing apparatus 118
ofice 7,8
Homogeneous nucleation point thermal conductivity of 46
of liquid water 7, 8
ofwater 7 thermal consequences of sublimation Jet-freezing
Heat damage
Horseradish peroxidase method 157 comparison with other freezing
during evaporation 372
as a marker for freeze-fracture vapourisation rate of vs. tempera- methods 93
Heat flow
replicas 428 ture 391 modification of spray-freezing 116
through specimen and cold stage in
Hydrophilic membranes Ice crystal-free zone Joule-Thomson refrigerator
SEM 155
method of producing 322 depth of without cryoprotection 73 for SEM cold stage 153-154
Heat flux
Hydrophilic polymers Ice crystals in AMR Bio-Chamber 169
during resistance heating for evapora-
as cryoprotectants 79 grain boundaries 13
tion 375
Hydrophilic support films 230-231, 322 size at different depths in tissues 72 Kamovsky's fixative 87
Heat of condensation
Hydrostatic pressure size in relation to cooling rate 10, 67-68 Kinetic energy
during resistance heating for evapora-
artefacts from 74 Immersion freezing during deposition of thin films 367
tion 375
ft-Hydroxy sterols 431 practical methods for 96--110 Kinetic gas theory 247-251
during shadowing 373--374
Hydroxyethyl starch 76 Immersion methods Kleinschmidt technique 274-276
Heat output
cryoprotectant for animal tissues 88 for plunging into liquid cryogens Knife angle
during sputter-coating 172
cryoultramicrotomy mountant 220 102-110 for cryosectioning 203-204
Heat removal
Hyperbaric freezing 42 Immunocytochemistry Knife holder
theoretical rates of, from water 6
Hypoeutectic point 15, 17 (see also Cytochemical methods) for cryoultramicrotomy 213
Heat transfer 22
Hystoacryl blue on frozen thin sections 235--236 for Reichert-Jung FC4 Cryosystem 224
across metal specimen supports 103-104
for mounting monolayers for Impinging flux Knife marks
at cold surfaces 6Q-65
freezing 323 of water molecules in vacuum 389 for cryosectioning 205, 208
during freeze-drying 252
Inert dehydration 462 in freeze-fracturing 331
effects of specimen surfaces and shapes
Ice
Infiltration Knudsen equation
50-53
absolute sublimation rate of 250 for low temperature embedding 491-492 for calculation of ice sublimation
in cryogenic fluids 38-53
amorphous 11 of Lowicryl resins 492 rate 250
Heat transfer coefficient
influence on time taken to freeze 47 cubic 21
542 543

LAAMA 485 for Balzers freeze-etching unit 354-356 Mass loss of ice pretreatment for freezing 82
Labelling Liquid nitrogen in electron microscopes 18 l Microanalysis
freeze-fracture replicas with comparison of cooling in 44 Mass spectrometer cold stages for 194--195
markers 428--429 expansion of on changing to gas 509 vacuum measurements with 393 Microcrystalline ice 12
Labelling procedures freeze-fracturing under 361-363 Mass transfer Microorganisms
for post-fractured specimens 433 sub-cooled 57 during freeze-drying 252 methods for cleaning replicas of 399
Laminar flow 39 super-critical for rapid cooling 50, 503 Maximum nucleate boiling point 50 Microtomes
Latent heat of evaporation Liquid nitrogen slush 43 Mean free path (see also Cryoultramicrotomes)
during sublimation of ice 252 (see also Sub-cooled liquid nitrogen) of gas molecules in vacuum 389 double-bladed, for freeze-fracturing 331
Latent heat of fusion Liquid oxygen 100 of nitrogen and water vapour vs. for freeze-fracturing 343-347
ofwater 6 safety hazard 511 pressure 259 Model specimens
Lattice structure Liquid paraffin method Measurement representative cooling rates for 45
of hexagonal ice 13 for rapid freezing of specimens 119-120 of cooling rates 25-38 Molecular sieve 260, 266
Lead-through Liquid propane Mechanical desiccants 260 Molecular weight determination
for thermocouples into vacuum 32 safety hazard 511 Mechanical stress from freeze-fracture replicas 438--440
~ccithin liposomes Liquids for freeze-substitution 464 artefacts from 74 Monitoring
freeze-substituted 481 LKB Cryokit Melt-freeze theory film thickness 380--383
Leidenfrost phenomenon 23 attachment for cryoultramicrotomy of cryosectioning 202 MONOFARG 424-427
Lenses 214-215 Melting points Monolayers
super-conducting 149 London LR Gold 487 of liquid cryogens 49 freeze-fracture autoradiography of
Leybold-Heraeus Bioetch 2005 Loschmidt's number 250 of some liquids 246 424-427
description of 351-353 Loss of ice Membrane structure method of mounting for freeze-fracturing
Leybold-Heraeus EPA 100 preparation methods for detecting 182-183 interpretation of freeze-fracture replicas 318-324
unit 363-365 Low temperature 4-09-414 Mountants
Light source maintaining for embedding 489--490 Membranes artefacts from with frozen-dried speci-
for low temperature polymerisation 493 Low temperature embedding 486--494 asymmetric organisation of 411 mens 283
Linear dimensions apparatus for 489--490 nomenclature of freeze-fractured for cryosections 230--233
changes in during freeze-drying 297-299 media for 478 409-414 for cryoultramicrotomy 220
Lipid extraction resins for frozen-dried-specimens pre-freezing alterations to 43D--431 for frozen-dried specimens 283
from fracture faces 436-437 291-292 splitting of during freeze-fracture 410 Mounting
Lipid micelles Low temperature glues 206, 221 Metal machining model for monolayer freeze-fracturing 319
in freeze-fractured membranes 410 Low temperature microscopy for cryosectioning 202-203 replicas on grids 402--403
Lipids (see also Cryornicroscopy) Metals single cells for freeze-fracturing 315-316
methods for cleaning from replicas 401 transfer devices for 155-156 evaporation of for SEM coatings 162 specimens for cryoultramicrotomy
Liposomes 43 7 Low temperature polymerisation Metastable supersaturation 17 219-220
Liquefied gases apparatus for 493 Methacrylates specimens for freeze-fracturing 313-327
expansion of during decompression 510 Low temperature resins 478, 291-292 low temperature embedding 491 specimens on cryoultramicrotomes
Liquid cryogens sectioning and staining 495--496 Methanol 223-225
characteristics of 47-50 Low temperature stage for thawing freeze-fractured replicas 286, Mouse growth cartilage
cooling rates in 55-60 for transmission use in SEM 192 398 freeze-substituted 480
criteria for rapid cooling in 110 Lowicryl HM20 481 freeze-substitution solvent 466--467 Mouse leukemia virus
freezing methods with 91-120 Lowicryl K4M 481 freezing point vs. concentration 488 comparison of CPD and freeze-
methods for immersion of specimen Lowicryl resins 291, 486--495 Methyl cellulose drying 298
into 102-110 LTSEM (low temperature scanning electron cryoultramicrotomy mountant 220 Miiller
Liquid helium 96 microscope) Mica freeze-substitution apparatus 473--474
as a coolant 48 alternative preparation methods for 159 for complementary replicas 325 freeze-substitution schedule 477--478
as a cryogen 102 transferdevicesfor 173-175 specimen substrate in freeze-drying
cold stages cooled by 152 273-276 Natural convection 38
device for freeze-fracturing under 366 M-line protein Mica flakes Negative staining
freeze-fracturing under 332, 363-365 demonstrated by immunolabelling 482 for adsorption of macromolecules during freeze-drying 272, 278
use prior to freeze-substitution 461 Masking material 275-277 in combination with freeze-fracture
Liquid helium-cooled shroud on frozen-dried bulk specimens 280 Micrasterias denticulata 439-440
544
545
Newton's rate equation 38 for polyvinylpyrrolidone 77
Nickel screen forwater 3,19,244 spacers for mounting monolayers 318 Propane
for mounting cells for freeze- Phase separation 11-20 Polytetrafluoroethylene addition ofisopentane to 59, 102
fracturing 315 in membranes 430 model material in cooling experiments 42 condensation for use as cryogen IOI
Nitrogen Phloem to enhance specimen wetting for flash point 1OI
(see also Liquid nitrogen) freezing of 43 cooling 51 hazardous nature 59, 100
adiabatic expansion of 154 Physical properties Polyvinyl alcohol lowering of melting point using isopentane
hyperbaric, for rapid cooling 42 of evaporants 370 for mounting monolayers 319-320 59
sub-cooled, for rapid cooling 42 Phytohaemagglutinin influence on fracture plane 320 practical use as a cryogen l 00-102
super-critical 50, 503 ferritin-conjugated 429 Polyvinyl pyrrolidone 76--79 Propane jet-freezing
triple point of 96 Pits cryoprotectant for animal tissues 88 (see also Jet-freezing)
Nomenclature in freeze-fractured membrane faces 409 cryoultramicrotomy mountant 220 practical methods of 110-114
of freeze-fracture membranes 409-414 Plant material phase diagram for 77 specimen holders for 314
Nucleate boiling 22, 24 methods for cleaning replicas of 399--401 support for LTSEM specimens 165 Propane jet-freezing devices
Nucleation pretreatment for freezing 83-86 Pool boiling 22 double-sided 111-112
homogeneous 13 Plastic deformation 363 Positive charging single-sided 113
ofice 5-11 during freeze-fracturing 330, 363, (see also Cationisation) Propane/isopentane
Nucleation centres 416-417 of substrates for mounting speci- boiling and melting points of mixtures 59
during formation of thin films 368 in cryosectioning 202-203 mens 325 Propylene
Nusselt's number 40 Platinum practical methods for 325--326 as a liquid cryogen 57
Nylon films 187 physical properties of 3_70 Post-fracture labelling 431-433 Protective clothing
Plunge depth Post-fracture methods for low temperatures 508
Orb-ion pumps influence on cooling rate 33 for freeze-fracture cytochemistry 419 Protein A-gold 482
for freeze-fracture replication 356 Plunge-freezing Potassium phosphotungstate Protozoa
Organic solvents (see also Immersion-freezing) for staining cryosections 234 pretreatment for freezing 81
to replace water prior to freeze-drying comparison with cold block freezing 222 Prandtl's number 40 Pseudo-eutectic 407
270-272 comparison with other freezing Preferential nucleation Pseudoreplication
vapour pressures at low tempera- methods 93 during freeze-fracture replication 368 after freeze-drying 278
tures 271 Plunge velocity Pre-freezing of frozen-dried specimens 285
Oscillograph influence on cooling rate 34 alterations to membrane structure PTFE (see Polytetrafluoroethylene)
ultra-violet, recording 31 Polycationic ferritin 429 430-431 PV A (see Polyvinyl alcohol)
Osmium tetroxide Polyglycerol ester methods for freeze-fracture cyto- PVP (see Polyvinylpyrrolidone)
artefacts from use of 75 in liquid paraffin freezing method 119 chemistry 419
Osmium/thiocarbohydrazide/osmium Polyimide Preparation Quantification
method general structure of 181 of bulk specimens for LTSEM 164-167 of structural details in freeze-
prior to freeze-drying 284 substrate for thin film thermocouples 28 Preparation possibilities fracture 440--443
Osmotic shock 87 Polyimide films for direct observation of cold speci- Quantitative methods
OTO (see Osmium/thiocarbohydrazide/ for frozen sections 181 mens 150 for freeze-fracture replication 418-443
osmium) for frozen thin layers 185 Press devices Quartz-crystal film-thickness monitors 382
Poly-L-lysine for cryosections 231-233
Palmeria palmata for coating thin nickel screen 316 Pressure Radiant heating
pretreatment for freezing 82 for positively charging substrates 273, inside electron microscope for ice sublimation in LTSEM 160-161
Partial pressure of water 325 Pretreatment Radiation damage
in freeze-fracture apparatus 335 practical method for coating substrates artefacts from 74-75 of L-valine at different temperatures 148
relationship with vapour pressure of 325-326 of specimens for cryoultramicrotomy of specimens in electron micro-
ice 258 Polymerisation 218 scopes 148-149
Particle weight determinations of low temperature embedding resins principles of 71-79 reduction at low temperatures 149
from freeze-fractured replicas 438-440 491--494 Procaryotic cell envelopes Radiation dose
Peltier effect 22 with ultraviolet radiation 291 freeze-fracture nomenclature 413 to specimens in electron micro-
Penning sputtering 157 Polystyrene Procaryotic cells scopes 148-149
Phase diagram for replica strengthening 386--387 pretreatment for freezing 80 Radiolysis of ice 178-181
for glycerol/water system 18 Polystyrene latex beads Prolongation factor Rake angle
during freeze-drying 254 for cryosectioning 203-204
546
547
Rapid freezing Reversal
(see also Freezing) of images of freeze-fracture replicas 404 labelling with electron-opaque mark- Space charging
comparison of practical methods Reynold's number 33--40 ers 430 in SEM 162
for 127-128 Roots Sectioning influence on excitation volume 162
droplet method 120 pretreatment for freezing 83 after freeze-drying 287-292 Specific decoration
liquid paraffin method 118-119 Rotary shadowing 383-385 energy produced during 202 of fracture faces with water vapour
using cold blocks 121-127 calculation of particle diameters 441 in cryoultramicrotomy 225-227 437--438
Razor blades double-axis method 385 low temperature resins 495--496 Specific heat
as knives for cryosectioning 208 of frozen-dried macromolecules 276 Seebeck effect 23 definition of 54
Recrystallisation Rotating apertures Seeding influence on cold block freezing 64
artefacts during freeze-drying 295 to shield radiant energy during of cooling water 9 of liquid cryogens 49
behaviour for water and different cells 9 shadowing 374 Self-assembly products of liquids and solids at low tempera-
during freeze-drying 253 freeze-dried, from bacterial cell walls 296 ture 54
during freeze-substitution 463, 475 Saccharomyces (see Yeast) Self-shadowing 383 of water as a function of temperature 53
of ice 7, 19 Safety Self-shrouding stub Specific volume
Red blood cells (see Erythrocytes) (see also Hazards) for LTSEM 175 of water and ice 20
Refrigerator cryopumps 50 I in low temperature electron microscopy SEM - scanning electron microscope Specimen holders
(see also Cryopumps) 507-512 Shadow-casting (see also Specimen supports)
for freeze-drying 262 ofhalocarbons 58 methods for 372 for double-sided propane jet device 112
for freeze-fracture replication 333, 335 when using propane 59 Shadowing for freeze-drying and embedding 291
Rehydration artefacts Salt (see Sodium chloride) - after freeze-drying 277 for high pressure freezing 119
after freeze-drying 299 Sampling automatic regulation of 376 selection for freeze-fracture replica-
Reichert-Jung Cryochamber FC4 artefacts from 74-75 bi-directional 385 tion 314
system for cryoultramicrotomy 214-218 principles of 71-79 for freeze-fracture replication 367-388 Specimen mounting
Reichert-Jung UHV freeze-fracture device Sampling and freezing high-resolution 369-374 for freeze-fracture replication 313-327
decription of 356-358 ballistic methods of 88-91 methods for 372 Specimen preparation
Replica relief Sampling and pretreatment Shape for cryoultramicrotomy 218-225
correct interpretion of 404 practical methods 79-91 of specimen in relation to cooling rate 52 for freeze-fracture replication 313-328
Replicas Sand-polishing Shear zone Specimen supports
after freeze~drying 277-278 specimen holders for freeze- in cryosectioning 203 (see also Specimen holders)
cleaning of 397--403 fracturing 315 Shrinkage for freeze-fracture replication 313-327
criteria for high quality 405 Sandwich freezing during dehydration 245, 294 for small frozen-dried specimens 283
for freeze-fracture replication 367-388 specimen holders for 314 Silicon monoxide thickness, in relation to heat
from freeze-fracture specimens 309--443 Saponins 431 as reinforcing material for replicas 367 transfer l 03-104
interpretation of 403--418 Sapphire support films 185 Specimen temperature
mounting on grids 402--403 use offor cold block-freezing 121 Silk moth antennae artefacts arising from changes in 74
of frozen-dried specimens for TEM Sartorius muscle freeze-substitution of 482 Specimens
285-287 of frog, freeze-dried and embedded 289 Silver technique different types for low temperature
reversed images of 404 Saturation vapour pressure for replica strengthening 386 observation 150
strengthening methods for 385-388 of ice vs. partial pressure of water 258 Single cells storage at low temperature 128-131
techniques after freeze-drying 285 of ice vs. temperature 390 mounting for freeze-fracture 315--316 Spherical cells
unrolling of 403 relationship with sublimation rate of Size method of mounting for freeze-fracture
Residual gases ice 251 of ice crystals at different depths 72 315-316
composition in vacuum system 258 Scanning electron microscopes Skin 86 Spotting plates
Residual water cold stages for 153-155 Slee cryoultramicrotome 208-210 for freeze-fracture replica cleaning 398
in plant tissue after freeze-drying 256 Scanning electron microscopy Sodium chloride Spray-freezing
Resins freeze-substitution for 484-485 freezing patterns of 14 comparison with other freezing
London LR Gold 487 Section collection simple phase diagram for binary system methods 92
low temperature 486-495 in cryoultramicrotomy 227-230 of 16 practical methods of use 114-116
Resistance heating Sectioned replicas Sodium silicotungstate to produce vitrified thin films 186
methods of 375-378 freeze-fracture autoradiography with for staining cryosections 234 Spraying technique
Response times of thermocouples 29 424 Solute concentration for small specimens prior to freeze~
artefacts from 74 drying 273-274
548
l 549

Spreading technique rate of 247-251 in freeze-drying specimens 261 Thin layers


for small specimens prior to freeze- theory of 247-256 Temperature sensing elements preparation for freezing 187
drying 274-277 thermal consequences of different for SEM cold stages 153-155 Thin sectioning
Sputter-coating methods 157 Temperature transitions after freeze-drying 287-292
after freeze-drying 278 to reveal frozen hydrated specimens 184 of ice in cryomicroscopes 504 freeze-substitution for 474--484
for low temperature specimens 157-158 Sublimation rate Ternary systems 15 Thin specimens
heating curves during 172 absolute, calculation of 250 Thawing coating of 189
vs. evaporative methods for LTSEM actual, calculation of 257 freeze-fractured specimens 397-403 Three-dimensional reconstruction
162-163 during etching 390-392 viability after 130-131 from freeze-fracture replicas 442
Sputter-ion pumps in cryoultramicrotomes 233 Thermal conductivity Tissue-Tek
for freeze-fracture replication 335 of bulk specimens vs. time 255 influence on cold block fi-eezing 64 for cryoultramicrotomy specimens 220
Staining of ice from cryosections 292 of copper 62 support for LTSEM specimens 165
cryosections 234 relationship with vapour pressure of of liquid cryogens 49 Titanium
tow temperature resins 495--496 ice 251 of liquids and solids at low tempera- as a specimen support 103-104
Static electricity Substituting liquids 464 ture 62 Titanium getter pumps
problems caused by in cryoultra- Substitution media 465-469 of polycrystalline ice 46 for freeze-fracture replication 356
microtomy 227 Substitution rate 468 of water as a function of temperature 53 Tokuyasu method
Statistical methods Sucrose droplet method Thermal inertia for collection of cryosections 229-230
for freeze-fracture replicas 441--443 for collection of cryosections 229-230 ofthermocouples 28 Toluene
Steere-Denton freeze-fracture unit Super-conducting lenses 149, 183 Thermal kinetic energy as low temperature glue 206
(see Denton freeze-fracture unit) Super-cooled water 19 during deposition of thin films 367 Tomatin 431
Stefan-Boltzmann law Super-cooling Thermal load Tonoplast
for determining intensity of radiant ofwater 6 during evaporation 371 fracture face of in yeast 411
heat 373-374 Super-critical nitrogen 50, 503 Thermal stress cracking 295 Topography
STEM - scanning transmission electron Supersaturation Thermanox plastic sheets of fracture plane 329-332
microscope metastable 17 for monolayers 314, 319, 324 Total pressure
Stereological methods Supporting films Thermocouples of water in freeze-fracture apparatus 335
for analysis of freeze-fracture for cryosections 230-231 and temperature measurement 25-35 Toxicity of cryogenic gases 512
replicas 442 making hydrophilic 230-231 British Standards specifications for 27 Transfer
Stereo-pairs Surface temperature calibration of 28 of frozen specimens for freeze-
from freeze-fracture replicas 442-443 maximum rise in thin and thick compensating wire for 33 fracture 328
Stereoscopic techniques sections I 80 conduction losses in lead wires 30 of thin frozen specimens 189-193
for deep-etched replicas 287 Surface tension connectors for 32 Transfer devices
Storage during drying 245 influence on cooling rate 23 cooled, for LTSEM 174
containers for specimens at low tempera- Survival lead-through into vacuum 32 for Leybold-Heraeus Bioetch 2005 352
ture 129 of cells at different cooling rates 70 materials for 26 for low temperature microscopy 155-156
of specimens at low temperature 128-131 of Euglena gracilis under pressure 117 preparation from thin wire 27 focLTSEM 173-175
specimens for freeze-fracture replication response times of 29 for Philips 400 TEM 188-189
327-328 Tantalum size in relation to cooling rate 24 simple, for LTSEM 173
Stratum corneum 86 physical properties of 370 thermal inertia of 28 Transition boiling 22, 24
Strengthening methods TEM - transmission electron microscope thin film 28 Transmission holder
for replicas 385-388 Temperature Thickness loss for Philips cryosystem 190
Stub arrangements for measuring 29 of frozen hydrated specimen vs. Trichlorethylene
self-shrouding for LTSEM 175 of cryogen in relation to cooling rate 95 temperature 180 for mounting freeze-fracture specimens
Sub-cooled liquid nitrogen 57 Temperature control Thin aqueous films 317-318
practical use as a cryogen 96-98 of specimen before propane jet-freezing vitrification of 185 Tritiated water
Sublimation of ice ll4 Thin film thermocouples 28 effiux from ferritin solution 469
foe LTSEM l 70-171 Temperature measurement Thin films Trough liquids
need for control of 160 using thermocouples 25-35 formation for freeze-fracture 367-369 for wet cryoultramicrotomy 233-234
physical conditions for 388-392 Temperature profiles Titin frozen layers Tungsten
principles in low temperature microscopy for cold specimens in electron micro- preparation of 185-187 physical properties of 370
156-157 scopes 179-180
550

Turbomolecular pumps Viability


1 Water-soluble substances Xylene/dry ice
551

for freeze-fracture replication 333 after thawing 130--13 l localisation by freeze-substitution low temperature mixtures 489-490
Turbulent flow 39 Vinol 320 485-486
Vitreous water 11 Wet sectioning Yeast
Ultra-high vacuum (UHV) conversion to cubic ice 21 in cryoultramicrotomy 233-235 fracture face ofvacuolar membrane 411
for freeze-fracture replication 335 Vitrification frozen at different rates 68
in freeze-fracture replication 322 cooling rates required for 12 X-ray diffraction frozen-fractured before and after etching
technology for freeze-fracture replication of thin aqueous films 185 to show ice crystal size 38 406-407
356-361 Vitrification temperature X-ray microanalysis lipid vacuoles in 436
Ultramicrotome cryoattachments 210-218 of water 7 LTSEM cold stages for 177 viability in relation to cooling rates 69
Ultra-violet polymerisation Volatile buffers of frozen-dried bulk specimens 283
at low temperature 494 for freeze-drying 273 of frozen-dried sections 292 Zeolite molecular sieves 260, 266
resin mixtures for 491 Volatile cryoprotectants 397 provision of low background environ-
Ultra-violet recording oscillograph 31 Volume (see Specific volume) ment 193
Umbelliferone 25
Warming
Vacuolar membrane of frozen specimens 20--21
fracture face of in yeast 411 Washing artefacts
Vacuum inside electron microscope during freeze-drying processing 293-294
Vacuum collection of sections Water
in cryoultramicrotomy 227-229 bound 10
Vacuum evaporator critical point of 244
for fracturing and coating LTSEM dielectric constant of 35
specimens 168 effect of pressure on 17
Vacuum pro be device equilibrium temperature of 5
for cryoultramicrotomes 210 heat capacity of 7, 8
Vacuum units heterogeneous nucleation of 9
for freeze-fracture replication 333-336 homogeneous nucleation of 6, 7
schematic diagram showing basic increase in volume during freezing 18
principles 334 latent heat of fusion 6
L-Valine lowering of freezing point under
radiation damage vs. temperature 148 pressure 116
Van Harreveld pressure-temperature phase diagram 3,
freeze-substitution method of 482 19, 244
Vapour fixation replacement of with organic solvents
for cryoultramicrotomy 219 268-272
of frozen-dried specimens 289-291 residual in tissue after freeze-drying 256
Vapour pressure specific heat as a function of tempera-
(see also Saturation vapour pressure) ture 53
during evaporation of shadowing specific volume in relation to tempera-
materials 370--374 ture 20
of liquids used for air drying 245 super-cooled 19
of organic solvents at low tempera- super-cooling of 6
tures 271 thermal conductivity of 53
of water at different temperatures vapour pressure at different tempera-
248-249 tures 248-249
Vaseline vapour pressure in freeze-drying 258
for complementary replication 327 vitreous 11
Very low accelerating voltage vitrification temperature 7
observation of frozen-dried speci- Water content
mens 284 of acetone in equilibrium with ice 468

Potrebbero piacerti anche