Metallurgical Science and Technology, Vol. 19/1

METALLURGCAL
SCEHCE
AHD
TECHHOLOGY
Taxe Perue
Tassa Riscossa
Torino C.M.P.
8 '
A Iournol published by
P
o
s
t
e

l
t
a
l
i
a
n
e
S
p
e
d
i
z
i
o
n
e

i
n

A
.
p
.

7
0
%
D
.
C
.
/
D
.
C
C
.
l
.

T
o
r
i
n
o
N
.

2
/
2
0
0
0
Editorial Panel
aeIe AK1OKA
Materials Consultant
Ietre AK0lKO
Professore di Teconologia dei Materiali e Chimica Applicata, Facolt di
Ingegneria, Politecnico di Torino
0Iureppe CA0LlO1l
Professore di Fisica dello Stato Solido, Facolt di Ingegneria, Politecnico
di Milano
CarIe 0OKlILLl
Professore di Costruzione di Autoveicoli, Facolt di Ingegneria, Politecnico di
Milano
nrIte VAK0LlI1A
Professore di Metallurgia, Facolt di Ingegneria, Universit di Ancona
Herten C. FLHlK0I
Head Department of Materials Science and Engineering, M.I.T., Cambridge,
Mass.
IerIe 0ALLO
Libero docente di Chimica Applicata presso il Politecnico di Torino, Presidente
Onorario Teksid S.p.A.
CIaudIe H0I
Innovazione Centrale Teksid S.p.A.
HaIter KlCO0Hl
Professore di Siderurgia e Impianti Siderurgici, Facolt di Ingegneria, Politecnico
di Milano
Direttore Responsabile: Sergio GALLO
Registrazione presso il Tribunale di Torino n. 3298 del 12 maggio 1983
Associato alla Unione Stampa Periodica Italiana
Authors are asked to supply two typewritten copies of their papers. These
should be laid out in accordance with the Instructions for authors on the
inside back cover. All correspondence should be addressed to:
Pippo BONSIGNORE
Segreteria di redazione
Metallurgical Science and Technology
Teksid S.p.A. Via Pianezza, 123
10151 Torino (Italy) - Tel. +39.011.6854.111
Publication distributed free of charge
Copyright 1983 Teksid S.p.A. - All rights reserved.
Editorial coordination: Immagine e Comunicazione Esterna
Graphics and layout: EMMEDI pencil & mouse - Via S. Ambrogio, 23 - To
Printed in Italy: Graficat - Via Cuniberti, 47 - Torino
No part of the texts published in this journal may be reproduced, whether in
the original language or in translation, without permission in writing from Teksid.
Papers submitted for publication as a rule should illustrate unpublished original
research works, experiments, or critical reviews. Publication will depend on
the approval of the Editorial Panel.
1& ucUn~L e Lec
vLeL& c~
1&ece v&e e1&:
vvv.1&ec.ccM
z
Mc+=LLLncic=L
eciccc =o 1ccicLcc-
JcLn=L =LeLieico
e- 1ceio
+wicc = -c=n
vcL. = ~c. , uuc &cc
cnc~ cr 1& LUc c1 ece11c
=n&ee~c =ncc&ee
=c&
0. ChandIey
e Mcnce1nUc1UnL ~c M&c~cL
&rr&c1e cr LUc c1 ece11c
=n&ee~c c~ zee eM=L&e. .
H. . 8ehI, 0. CaIIetI
I. 8arene, C. Hur
& Mcnce1nUc1UnL &rr&c1 cr 1&
L...=. =ncc&ee c~ 1& zee
LUM~UM LLcv
H. CabIbbe, . vaneIIrta
e n&ec~e 1c c&v&Lc= LUc c1
ece11c =n&ee~c
I. 0aIIe, C. Hur, 0. HertarI
& 1& &rr&c1 cr LUc c1 ece11c
=n&ee~c c~ r1cU& =nc=&n1&e cr
L ee&c ce1~ce
. kemane, H. kerre,
C. Hur
&e =nc=&n1&e M=ncv&M&~1 cr
1crcnM&c =n1e ev LUc c1
ece11c =n&ee~c =ncc&ee
H. kerre, . kemane,
I. 8arene
LUc c1 ece11c =n&ee~c I. 0aIIe
LUc c1 ece11c =n&ee~c
I primi brevetti del processo di HOT ISOSTATIC PRESSING HIP
risalgono al periodo 1913-15.
Quando negli anni 60 si svilupparono le ipotesi per impiegare il processo
come post-trattamento di getti, il mondo delle fonderie manifest un
notevole interesse, vista la possibilit di eliminare, grazie ad una elevata
pressione di gas inerti ad elevata temperatura, quei piccoli difetti interni
dei getti (microporosit, microrisucchi, pin-holes da idrogeno) che ne
riducevano le propriet meccaniche, in particolare la resistenza a fatica.
Numerosi impianti sono stati realizzati negli anni, con notevole successo,
purtroppo limitato ad impieghi per mercati di nicchia e di classe elevata
(es. aviazione, competizioni) che potevano accettare il costo non
indifferente del processo HIP, legato al costo elevato degli impianti ed a
cicli di trattamento molto lunghi (circa 10 ore).
Gli studi di Dixon Chandley, presidente della Metal Casting Technology
Inc., una joint venture tra la Hitchiner Mfg.Co.Inc. e la General Motors,
hanno portato alla realizzazione di un piccolo impianto dimostrativo di
laboratorio ed alla successiva richiesta di brevetto, di un nuovo sviluppo
del processo HIP in cui il mezzo che genera la pressione isostatica non pi un gas,
ma un liquido, una miscela opportuna di sali fusi, con un ciclo di trattamento di
pochi minuti : il processo chiamato LIQUID HIPPING LHIP.
Alla fine del 1998 la Teksid e la Idra-Presse hanno ottenuto da GM e
Hitchiner lautorizzazione allo sviluppo industriale di un impianto LHIP
destinato, per il prevedibile basso costo di esercizio, al trattamento di
particolari destinati alle produzioni di grande serie, quali quelle per
lindustria automobilistica.
Il presente numero della rivista Teksid Metallurgical Science and
Technology dedicato alla presentazione dei lavori effettuati per la
realizzazione del primo impianto
i ndustri al e di LHIP, del l e
problematiche tecniche superate,
dei ri sul tati ottenuti vi a vi a
sui getti trattati , sul l e
considerazioni economiche rela-
ti ve al l abbi namento del
trattamento LHIP ai vari processi
di colata dei getti, in particolare
a quelli di alluminio.
La possi bi l i t di i nseri re i l
processo di LHIP direttamente
nel ciclo di trattamento termico,
immediatamente prima della
tempra in acqua dei getti, aggiunge
una significativa riduzione dei
costi eliminando in pratica la
richiesta energetica per il riscaldo
dei getti da sottoporre al
trattamento di pressatura
isostatica.
I ri sul tati ottenuti e l e
problematiche tecniche risolte
hanno cos consenti to l a
definizione degli impianti, di varie dimensioni,
destinati alle produzioni industriali.
La presente pubblicazione vuole essere anche
l occasi one per ri ngrazi are quanti hanno
contribuito a far si che il progetto LHIP diventasse
una realt industriale a disposizione della comunit
metallurgica mondiale:
in primo luogo Dixon Chandley che ne ha in-
tuito e dimostrato la fattibilit
John Morrison e Paul Mikkola rispettivamente
C. E. O. ed Executi ve V. P. Operati ons
della Hitchiner che hanno avuto fiducia nella
Teksid e Idra affidando loro lo sviluppo del
processo LHIP
Paolo Filomeni C.E.O. della Teksid e Aldo Palma
C.E.O. della Idra-Presse che hanno creduto
nel potenziale del nuovo processo ed hanno
accettato di investire quanto necessario per
questo sviluppo.
Francesco Rangoni, Senior Vice President della
Teksi d, Di vi si one Al l umi ni o, che ha fi n
dallinizio sostenuto liniziativa seguendone
continuamente i progressi.
Graziano Mortari, vice-presidente della Idra-
Presse ed i suoi collaboratori, per il continuo
e costante contributo di idee nella difficile lotta
nel superare le difficolt tecniche che via via si
sono presentate.
Il Prof. Merton Fleming, MIT di
Boston, che con i suoi illuminati
consigli ci ha confortati nella
ricerca delle migliori soluzioni.
I professori: Caglioti, Doniselli,
Evangelista, Nicodemi e Rosso
dei Pol i tecni ci e Uni versi t
Italiane per le prove sui materiali
trattati e per le fattive discussioni
sui risultati ottenuti.
Il Ministero dellUniversit e
del l a Ri cerca Sci enti fi ca e
Tecnol ogi ca (MURST) che
partecipa al finanziamento del
programma.
Ed infine tutti i componenti del
Teksid Technical Center che
sotto la guida di Claudio Mus,
direttore del centro, con umilt
e passione hanno visto nascere
e crescere il processo LHIP fino
a raggiungere la realt industriale
qual oggi.
IerIe 0aIIe
LUc c1 ece11c =n&ee~c
IerIe 0aIIe
The very first patent on HOT ISOSTATIC PRESS-
ING HIP dates back to 1913. When, the po-
tential improvements of the quality of the cast
components, offered by the HIP process were
recognized, in the sixties, the foundry commu-
nity showed strong interest through the applica-
tions of a high inert gas pressure at high tem-
perature: it seemed possible to eliminate defects
like shrinkage, porosity and hydrogen pin-holes
increasing the mechanical properties of the final
products.
Many HIP units have been put in place since then
and some of those are successfully treating aero-
space and high performances components for rac-
ing engines and others niche markets. Cost is still
the main barrier to the HIP process to be used
for high volume productions: high investments and
long cycle times (up to 10 hours) being the main
cost issues.
Studies performed by Dixon Chandley, president
of Metal Casting Technology Inc., a Hitchiner
Mfg.Co.Inc. and General Motors joint venture,
introduced an innovative concept to the proc-
ess, today patented as Liquid Hot Isostatic Press-
ing, where the fluid used to apply the isostatic
pressure on the components to be treated is a
molten mixture of salts.
The process capability was demonstrated on a
small scale pilot plant: few minutes cycle time was
enough to achieve results similar to those ob-
tained by the much longer HIP cycle time.
In 1998 Teksid and Idra Presse have been author-
ized by GM and Hitchiner to develop an indus-
trial LHIP unit with the goal of having the process cost competitive for
the automotive industry.
This Metallurgical Science and Technology issue is dedicated to the LHIP
project: tackled technical problems and achieved results on aluminum al-
loys casting are described.
The possibility to integrate the LHIP process in the heat treatment cycle,
in between solubilization and quenching, helps in having a significant cost
energy and cost reduction.
This give the opportunity to thank everybody involved in this activity making
today available to the metallurgical world the LHIP technology.
First of all Dixon Chandley inventing the LHIP process and proving its
feasibility
John Morrison and Paul Mikkola, Hitchiner C.E.O. and Executive VP
Operations trusting in Teksid and Idra commitment and competencies.
Paolo Filomeni, Teksid C.E.O. and Aldo Palma Idra Presse C.E.O. believ-
ing in the new technology potential and financially supporting the
project.
Francesco Rangoni, Senior Vice President of Teksid, Aluminum Division,
who has supported this project since the very beginning and in show-
ing continuous interest in its development.
Graziano Mortari, Idra Presse Vice President, and his team contribut-
ing with continous efforts in the technical development.
Prof. Merton Flemings, MIT Boston, bringing bright suggestions to the
team in finding smart solutions.
Professors: Caglioti, Doniselli, Evangelista, Nicodemi and Rosso, of
Italian Politechnics and Universities testing, evaluating and discussing
the achieved results.
The Italian Government that, through MURST, is financially supporting
this program.
Last but not least, all the Teksid Innovation Technical Center staff, guided
by Claudio Mus, innovation manager, contributing in the day by day
development and problem solving.
3 - Metallurgical Science and Technology
~1nccUc1c~
In the late 1980s, when the author was president of Metal Casting Technol-
ogy, Inc. (MCT), a joint venture proccess development company between
Hitchiner Mfg. Co., Inc. and General Motors, a new low cost casting proc-
ess for titanium alloys was invented and later patented, US 5,299,619. All
titanium castings required Hot Isostatic Pressing (HIP) to achieve good
mechanical properties. The conventional HIP process, which uses argon
gas to apply high pressure for closing internal porosity, is very effective, but
very dangerous and costly - too much so for the commercial applications
cnici c= +ic
LicLio c+ ece+=+ic =nceeic =ncccee
0Ixen ChandIey
VIte rerIdent 1ethneIey, KItthIner HanufatturIn Cerp., lnt., ketIred
==ncc ~c n&eUL1e
Most of the time involved in gas HIP cycle was, and is, involved in heating
up the charge and raising the pressure of the argon gas, resulting in very
long, 16-24 hour, machine cycle times. Reference one is an excellent article
with some other references for anyone desiring more information about
gas HIPping. Since it takes so long to get the gas HIP units up to tempera-
ture and pressure (and to cool them down), there was little information
about exactly how long it actually takes to close porosity in castings at a
given pressure and temperature. It was surmised that, if the actual time at
temperature and pressure was on the order of 15 to 30 seconds, an ap-
proach like that in Figure 1 might be feasible. First, a pressure vessel con-
taining molten salt, first view in the upper left of the figure, would be main-
tained at a temperature, say 550 C, for which a number of alloys have
anticipated for the new low cost casting process.
As a result of this need, the author devised an
approach using mechanical means to pressurize a
non-reactive molten salt, which in turn, pressu-
rized the casting surface, closing interna1 poros-
ity. This brief article describes some of the early
work which led up to the patenting of the Liquid
Hot Isostatic Pressing (LHIP) process as disclosed
in US Patent 5,340,419.
adequate strength to sustain up to1650bar (25,000
psi). Second, a container with a higher tempera-
ture salt and the parts to be LHIPped, second from
upper left view, would be heated separately to
the desired LHIP temperature. Third, the hot con-
tainer with parts and salt, third from upper left
view, would be placed in the lower temperature
salt. Subsequent steps to the right upper and left
to right on the lower views as described would
have to be done rapidly, before the hot salt sur-
rounding the parts was cooled significantly by the
lower temperature salt.
Armed with the optimism required in all devel-
opment, we set up an old hydraulic press, and cast
a small pressure vessel in INCO 713. It had inside
dimensions of only about 100mm diameter and
250mm deep. The press had a 50mm ram to apply
pressure. Many small castings of plain steel, stain-
less steel, titanium, nickel alloys, aluminum alloys,
and malleable iron were run through the small
vesse1. When aluminum was done, it was not nec-
essary to use a separate container as shown in
Figure 1, as the salt in the pressure vessel was
kept at the desired LHIP temperature. The cool-
ing curves when doing high melting point alloys
were measured and found to be fairly quick, but
even in these small containers, there was adequate
time to completely close the porosity in the al-
loys shown. This, of course, was great news since
scaling the vessel up to larger sizes would make
the cooling of the hot salt less of a problem.
Fig. 1: Steps in the LHIP process
!o|t, vesse| ot
1000 |
Potts ia hottet
so|t, ot hip
tempetotote
hot potts/coatoiaet p|oced ia
vesse| so|t Ii||ed to top
!heet p|oced ovet
vesse|
Assemh| iato
ptess
Ptess c|osed
kom octivoted
to ochieve
desited
ptessote
Fig. 2: Some fatigue properties of aluminum castings
The vessel was scaled up to the largest size we
could reasonably handle, inside dimensions of
250mm diameter and 400mm deep. In scale up,
a variety of mechanical problems were encoun-
tered with the old press, but only two relating to
the process itself will be discussed. First, although
most of us had remembered from early physics
that liquids were incompressible - in the rea1
world, it turned out that liquids are indeed com-
pressible at high pressure. Our salts compressed
some 4%, with the larger pressure vessel, this used
up too much of the displacement of our 50mm
ram, so the press was modified to provide a
100mm ram. Using copper sheets for the top pres-
sure vessel seal enabled us to LHP many more
and larger castings of all of the above alloys. Me-
chanical testing gave results in fatigue of aluminum
and malleable iron that were very impressive, as
shown in Figures 2 and 3. Of most interest was
cast aluminum, for which the endurance limit was
improved almost 100%. These data were from
actual castings. For shaped aluminum castings, it
was necessary to add hydrogen to the melt be-
fore casting to assure the absence of surface con-
nected porosity in the casting. It was found in cast-
ing aluminum that almost any isolated boss would
have some surface connected shrinkage and would
be impregnated with salt during LHIP. Adding the
hydrogen before cast to counteract shrinkage also
enabled significant reduction in gating, especially
when countergravity casting. Parts such as sus-
pension arms and steering knuckles were satis-
factorily countergravity cast and LHIPped with
only 1/o gating weight when the hydrogen was
properly added to the melt. Semi-permanent
molded aluminum castings always contained sur-
face connected porosity on the sand core sur-
face unless hydrogen was added. It was also found
that adding hydrogen to semi-permanent molded
castings did not always result in the uniform dis-
tribution of hydmgen bubbles as was found in sand
castings (Reference 2). Also, many aluminum die
castings were LHIPped; it was a pleasant surprise
to us that this solved major problems with after
machining leakage and low thermal conductivity
due to internal porosity as cast. No problem was
found with blistering of the die castings due to
LHIP.
Penetrant inspection of the sample parts was al-
ways required when setting up foundry practice
to assure there was no surface connected poros-
ity in the approved practice and to yield optimum
parts after LHIP. This practice not only yielded
parts with high fatigue properties for alloy 356 as
shown above, but also very good tensile proper-
ties from sand cast alloy 356 parts in 18 mm thick
Fig. 3: Some fatigue properties of malleable iron castings
sections as follows (average and range ()[ ] for five tests):
Yield Strength,ksi,[MPa] Tensile Strength,ksi,[MPa] % Elongation
39.4(39.1-39.9)[269-275] 29.1(27.2-30.5)[187-210] 7.7 (6.2-9.1)
Note: parts were quenched from LHIP temperature and aged.
It is interesting to note that some premium quality specifications for alloy
356 require higher yield and tensile strengths than those in the table. In
most casting thicknesses, those higher strengths require the use of struc-
ture modification and/or rapid cooling (chilling) to round the eutectic par-
ticles and refine the structure to give improved response to heat treat-
ment. In the above test results, peak strengths were not reached as peak
aging could not be achieved with the segregated structure in the slow
cooled sand castings. Other test bars that were solution treated and aged
after LHIP did not show tensile properties better than the above. How-
ever, the soundness created by the LHIP did give fine tensile properties
overall with both methods.
The other mechanical problem related to the process was that the copper
sheets did not always seal the ram from the salt. Also, the cost of the
sheets was significant. Thus, the press was modified so the grease around
the ram in the top platten of the press could be pressurized to the same
level as the salt pressure. This eliminated the need for the copper sheets.
Other liquids have been used successfully to overcome this difficulty as
well.
rU1Un& cr L=
Teksid and IDRA have taken LHIP far beyond the early work at MCT. They
have overcome many problems of scaling up to sizes that accommodate
engine heads and blocks. Certainly the future af LHIP to do aluminum
castings should be bright, as it seems to be the only known, potentially
economic way which can upgrade aluminum castings to the degree needed
for safety and other high quality parts for automotive applications.
n&r&n&~c&e
1. Robert Widmer, The Current Status of HIP Technology and Its Pros-
pects for the 1990s, Fourth Annual Conference on Isostatic Pressing,
Stratford-Upon-Avon, U.K., November 5,6,7 1990.
2. P.H. Mikkola, J.R. Redemske, Optimizing the Properties of Cylinder
Heads Cast of Aluminum, SAE Paper 940153, February, 1994.
c~cwL&cc&M&~1
It gives me great pleasure to acknowledge the
enthusiasm and dedication of MCT, Teksid and
IDRA to making this new process work. My best
wishes to all for a great success!
Mc+=LLLncic=L
eciccc =o 1ccicLcc-
JcLn=L =LeLisico
e- 1csio
+wicc = -c=n
Mcnce1nUc1UnL ~c M&c~cL
=n&ee~c c~ ee eM=L&e. .
*H. . 8ehI, *0. CaIIetI - **I. 8arene, **C. Hur
*0IpartImente dI lnenerIa KutIeare, eIItetnIte dI HIIane
lrtItute KazIenaIe FIrIta deIIa HaterIa, HIIane, ltaIy
**1ekrId I.p.A., 8erarette (1erIne}, ltaIy
Mcnce1nUc1UnL ~c M&c~cL
=n&ee~c c~ ee eM=L&e. .
*H. . 8ehI, *0. CaIIetI - **I. 8arene, **C. Hur
*0IpartImente dI lnenerIa KutIeare, eIItetnIte dI HIIane
lrtItute KazIenaIe FIrIta deIIa HaterIa, HIIane, ltaIy
**1ekrId I.p.A., 8erarette (1erIne}, ltaIy
Abstract
Comparative tensile tests and microscopic images are
reported of AA356 samples subjected to solubilization,
quenching in water and T6 aging (as cast), and on samples
of the same alloy subjected in addition, at the end of the
solubilization period, to Liquid Hot Isostatic Pressing
(LHIP). LHIPing is a recent technical variant of conventional
HIP-ing, where the pressure is transferred to the metallic
components through molten salts rather than through a
gas. Different solubilization times (1 8 hours) at the
temperature of 520 C have been explored. The results
confirm that, provided that the solubilization time exceeds
3 hours, LHIP-ing improves strongly the elongation to
fracture and sensibly improves the ultimate tensile
strength, UTS, practically without affecting yield.
In order to assess criteria for optimizing the casting
process, the mechani sms of HIP-i ng
and their roles in the densification and the improvement
of the ductility of the alloy are qualitatively analyzed for
the specific heat treatments and LHIP parameters adopted
in the present study.
Riassunto
Sono riportate prove di trazione uniassiale su campioni di lega AA356 sottoposti a
trattamenti di solubilizzazione, tempra in acqua calda e invecchiamento T6 e immagini al
microscopio di sezioni dei campioni stessi; i risultati sono confrontati con quelli
relativi a prove su campioni della stessa lega sottoposti in aggiunta, subito dopo la
solubilizzazione, a pressione isostatica a caldo (Liquid Hot Isostatic Pressing LHIP).
LHIP una recente variante tecnica dello HIP convenzionale, nella quale il trasferimento
di pressione al getto metallico assicurato da sali fusi anzich da un gas. Sono analizzati
e discussi i risultati relativi a diversi tempi di solubilizzazione (1 8 ore), alla temperatura
di 520 C. Le prove confermano che, se il tempo di solubilizzazione supera le 3 ore,
LHIP-atura aumenta fortemente lallungamento a rottura e migliora sensibilmente la
resistenza a rottura, UTS, senza peraltro influire in modo rilevante sullo snervamento.
Al fine di razionalizzare e, per quanto possibile, di quantificare i criteri correnti di
ottimizzazione del processo di produzione di getti, anche alla luce dei risultati
sperimentali sono analizzati i meccanismi di HIP-atura e il loro ruolo nella densificazione
e nel miglioramento della duttilit, con riferimento specifico ai trattamenti termici e ai
parametri di LHIP adottati nella sperimentazione.
. ceu&c1v&e
This contribution is part of a series of papers
devoted to Liquid Hot Isostatic Pressure (LHIP)
- a variant, developed by Teksid S.p.A. and Idra
Presse S.p.A., of the traditional HIP densification
process, where, perhaps for the first time, the
external pressure is applied to the metallic
components through molten salts rather than
through a gas [M. Rosso et al, 2000]. Its objective
is twofold:
To focus on the rol es of the LHIP-i ng
parameters in both the densification of the
porous structure of the as cast AA356 alloy
under study and the improvement of its
mechanical performance;
To report and to compare the results of preliminary tests on the
elongation to fracture and UTS of as cast vs. as cast + LHIP-ed samples,
aimed at identifying optimal thermal treatments and LHIP-ing proce-
dures.
&. 1& M&c~cL 1&e1e
64 heat treated sand cast samples have been submitted to standard tensile
uniaxial tests.
32 samples (8 sets of 4 samples each) have been submitted to the following
heat treatment: solubilization for 1, 2, 3 8 hours in a salt bath at 520 C,
quenching in water at 65 C, and T6 aging for 6 hours at 160C.
The other (8 4 =) 32 samples, in addition, have been submitted also to
LHIP: immediately after the solubilization treatment at 520 C, while the
sample temperature is still that selected for the solubilization
Fig. 1: Picture of the pore distribution in a longitudinal section of a) a sand
cast sample and b) an as cast + LHIP-ed sample, heat treated as specified in 2.
Magnification: 5X. (Courtesy of Mr. Olimpio Di Tollo). The effect of the LHIP-ing
process apparently is the collapse of the submillimetric pores
Fig. 2: Density of a set of sand cast AA356 Longitudinal Guided Arm (BLG)
components submitted/not submitted to the LHIP treatment as specified in 2
(corresponding to 0,85 T
m
, where T
m
= 933 K is the melting temperature
of the aluminum), isostatic pressures of the order of 115 MPa are applied
for 35 s through molten salts.
In paragraphs 3. and 4. below the mechanisms of LHIP are examined with
reference to optic macrographs and SEM micro-
graphs of samples submitted and not submitted
to LHIP. The results of the mechanical tests are
presented in 5.
. M&c~eMe cr LUc =-~c: 1&
v&Lc~c e1c& ~c 1& crrUec~
cc~1ncLL&c e1c&
The density of a sand cast Al alloy component turns out to be generally
smaller than the ideal one. The density reduction is mainly a consequence
of the 6,5% reduction of the volume per unit mass of the alloy in passing
from the liquid to the solid state: the tensile thermal stresses, created
during the non-homogeneous cooling of the sample while it solidifies,
produce an essentially random distribution of pores throughout the bulk
(Fig. 1). However the initial density never drops below 96% of the ideal
density, neither for the series of the as cast tensile samples submitted to
the mechanical tests described in the previous section (Fig. 2) nor for a
set of sand cast Longitudinal Guided Arm (BLG) 936 automotive compo-
nents (Fig. 3).
Density increase on suspension arm
We focus below on the densification produced
by LHIP and the underlying mechanisms, in the
neighborhood of the temperature and pressure
values indicated in 2. To this end, we refer to
the classic works of J. Weertman and J.R.
Weertman [1970], A.S. Argon [1966], M.F. Ashby
[1990], and H. Exner and E. Arzt [1996].
The first two papers deal essentially with creep.
The third and fourth papers, specifically dedicated
to HIP and sintering, correlate the mechanisms
of sintering promoted by high temperature and
pressure to the mechanisms of creep promoted
by high temperature and tensile stress; the
approach of these last two papers is essentially
that of interpreting HIP and creep in dual terms:
HIP is envisaged as a kind of inverted creep.
During a creep test, in the bulk of the specimen
under tensile stress (locall y under tensi l e
compressi on or shear), the initial vacancy
concentration increases toward values more and
more removed from those at the thermodynamic
equilibrium. As the tertiary, final creep stage is
reached, necking is formed; in the necked region,
vacancies coalesce lumping into submillimetric
voids, and these voids grow to millimetric
di mensi ons whi l e the densi ty di mi ni shes
catastrophically until fracture occurs.
As far as LHIP is concerned, the application of
isostatic pressure to the cast component after
solubilization produces two effects:
1. During the initial, yielding stage (stage I) -: an
increase of the initial density via a reduction
of the sizes of the pores.
(1)
2. During the final, diffusion controlled stage
(stage II) -: the junction of the facing portions
of the surfaces of the partially collapsed pores:
a kind of local welding occurs, assisted by a
drift diffusion ultimately promoted by the
applied pressure.
As indicated above, after the solubilization
treatment and prior LHIP-ing, the initial density
of the quenched as cast samples already exceeds
96% of the ideal density. (Figs. 2 and 3). But since
the void volume fraction due to the randomly
(1) In a fatigue or tensile test of an as cast solubilized and T6 aged sample
not submitted to HIP-ing, the pores created during solidification act as
cracks. The reduction of the pore sizes produced by (L)HIP during this
Stage II., qualitatively accounts both for the sensible increase of the UTS
reported in 4 and discussed in 5, and for the conspicuous increase of
the number of cycles to rupture observed in fatigue tests which will be
reported elsewhere (M. E. Beghi, G. Caglioti, S. Barone and C. Mus, to be
submitted for publication, Metallurgical Science and Technology).
distributed submillimetric pores doesnt exceed
the percolation threshold of 8% set by M.F. Ashby
[1990], the pores are sealed off. The macrographs
in Fig. 1 confirm that under these topological con-
ditions no percolation toward the external sur-
face could favor densification during (L)HIP-ing.
Furthermore the very presence of these
submillimetric pores in the as cast samples sug-
gests that the concentration of vacant sites in
the bulk of the untreated samples exceeds by
several orders of magnitude the equilibrium
concentration.
! 5J=CA 1
At the beginning of (L)HIP-ing, under a local
pressure P
lim
much larger than the yield stress at
the solubilization temperature, the pores shrink
rapidly. According to Ashby this local pressure,
high enough to cause yielding of the spherical
shells surrounding each pore is given by:
2

y
l
(1) P
lim
= ln [ ] ,
3 l - A
yield
where A
yield
is the level attained by the density of
the alloy at the end of the yielding stage (stage I).
The histograms in figures 2 and 3 show that in
the operating conditions of 2., the average value
of A
yield
falls in the range (0,985 0,996). According
to eq. (1), the lower limit of the local pressure,
P
lim
, required for the pore collapse, increases mildly
with the densification level attained by the alloy,
and doesnt exceed 10 MPa, a value roughly three
times larger than the yield stress of the alloy at
the solubilization temperature. This threshold of
the local pressure turns out to be about ten times
smaller than the applied external pressure, whose
value is selected in the 100 MPa range by the
press producer.
This comparatively large value of the applied
pressure is more than sufficient to promote
promptly the pore collapse observable when
Figures 1 a) and 1 b) are compared.
The densification induced during stage I. of LHIP
can be perceived also comparing the macrographs
of Figures 4 a) and 4 b).
Finally, the same trend can be recognized once
more, examining the micrographs of the fracture
surfaces of not LHIP-ed and LHIP-ed samples in
Fig. 5 a) and 5 b) respectively. Although even for
the LHIP-ed sample the density of the pores
intersecting the fracture surface largely exceeds
the average pore density in the bulk (responsible
for the aforementioned void volume fraction of
the order of 1%), the sizes of the pores in the
Fig. 3: Density of a set of sand cast AA356 Longitudinal Guided Arm (BLG), 936
components submitted/not submitted to the LHIP treatment as specified in 2
Density increase on suspension arms
neighborhood of the fracture surface are much smaller for the LHIP-ed
samples than for the not LHIP-ed ones. The residual difference between
the actual density of the LHIP-ed samples, in the range (0,985 0,996 %
of the ideal density), and the ideal density (2,685 10
3
Kg m
-3
) could be
due to the excess of vacancies in non-equilibrium in the quenched sam-
ples and the hydrogen yet dissolved in the bulk.
Fig. 4: Picture of the pore distribution in a chemically etched longitudinal
section of a) a sand cast sample and b) an as cast + LHIP-ed sample, heat treated as
specified in 2. Magnification: 10X. (Courtesy of Mr. Olimpio Di Tollo). While it is
confirmed that LHIP-ing promotes the collapse of the submillimetric pores, the
chemical etching reveals the dendritic nature of the microstructure
Fig. 5: Micrographies of the structure of the pores intersecting a fracture surface for
a not-LHIP-ed sample a), and for a LHIP-ed sample b).
(Courtesy of Mr. Olimpio Di Tollo). A consequence of LHIP is the healing of the
dendritic branches, at the micrometric scale
! 5J=CA 11
Only few seconds after the application of pressure, LHIP-ing enters the
final stage. Although the densification achieved by local yielding is
immediately completed during the short time allotted to stage I, at this
point the facing portions of the folded surfaces of the collapsed pores
could yet be envisaged as thin cracks hampering the mechanical
performance of the alloy. However an important process should start
now: a sort of welding of the facing portions of the squashed pores brought
into contact during stage I., promoted at this point by thermally activated
and pressure-assisted local diffusion.
The mechanism underlying this process should be essentially the same as
the mechanism described by H. E Exner and E. Arzt [1996], responsible
for the nucleation and growing of necks at the junction between initially
loosely packed particles during sintering. According to these Authors, the
pores act as vacancy sources or atom sinks for the surrounding particles:
the atoms diffuse along and across the boundaries of the crystal grains;
and their diffusion is enhanced in hot-pressing because the additional
energy is gained on moving an atom with volume from the
interparticle boundary which is acted on by a compressive traction into
the pore space or to the surface. This compressive traction essentially
a shear stress should arise, even if the applied pressure originating it is
globally isostatic, as a consequence of the irregular shapes of the pores:
geometric irregularities lead to anisotropy in the local redistribution of
the external stress. A dual statement equivalent to that one quoted above
could be: the vacancies diffuse along and across the boundaries of the
crystal grains; and their diffusion is enhanced in hot-pressing because the
additional energy is gained on moving a vacancy with volume from
the surface of the pore space into the interparticle boundary which is
acted on by a tensile compression .
Therefore, as remarked above, although the densification promoted by
local yielding is completed, the collapsed pores could still hamper the
mechanical performance of the alloy if they were not healed up due to
the Al atom diffusion toward the pores and/or the vacancy diffusion toward
the bulk. This dual diffusion process can be envisaged as a drift-self-diffusion
enhanced by the shear stresses s originated locally by the applied isostatic
pressure; it leads to the healing of the pores, whose linear dimensions are
now reduced to few tens of micrometers.
According to Ashby, if the grain size G is smaller than the size of the
particles, the HIP-ing is controlled mainly by volume diffusion (Nabarro-
Herring creep). To a minor extent, at high temperature, grain boundary
diffusion (Coble creep) contributes also to HIP-ing. The following rate
equation for the diffusion flow is then proposed [Ashby 1990, eq. (36)]:
d 14

D
v
D
b
(2) = [ + ],
dt KT G
2
G
3
where K is the Boltzmann constant, D
v
and D
b
are the volume and the
grain boundary diffusion coefficients, and G are the grain boundary thick-
ness and grain size respectively. In the numerical estimates we will neglect
the contribution to the Coble creep. (Should power law creep conditions
dominate, the linear dependence would be substituted by a power law of
).
Since the self diffusion is a thermally activated process, the kinetics of the
densification is enhanced exponentially by the thermal energy and
controlled exponentially by the activation energy
for vacancy diffusion or atomic self diffusion. The
driving factor, 14

/(KT), depends linearly on
the work of the local shear force promoting
an atomic displacement of the order of the atom
size.
The atoms responsible for the strain rate of eq.
(1), diffusing from a grain contiguous to a par-
tially squashed pore toward the facing portions
of its folded surfaces, produce a kind of junction
of these surfaces ultimately promoted by the
applied pressure. Their drift velocities <v> is:
14

D
v
(3) <v> = .
KT G
<. &rr&c1e cr L= c~
1& c&r&c1v&
Mcnce1nUc1Un& cr 1&
eM=L&e U~c&n e1Ucv
In this paragraph equations (2) and (3) are utilized
in order to account for the evolution of the
microstructure promoted by the action of LHIP
on the AA356 samples analyzed in the present
investigation.
In Fig. 4 the dendritic nature of the structure of
our samples is evidenced. Dendrites grow during
any diffusion limited aggregation (DLA) process.
The DLA develops during solidification provided
the controlling parameters local temperature
gradients, surface tension and local curvature of
the solidifying front happen to lie in specific
intervals. DLA is characterized by a sequence of
elementary steps promoted at the atomic scale
by the systematic, recurring application of the
same set of rules. A natural consequence of the
systematically repeated application of the same
set of rules is that the pattern produced by the
process exhibits the feature of being statistically
invariant against a change of scale: each part of
the structure turns out to be a copy of a part of
the structure and eventually of the entire
structure: the structure exhibits the statistical
selfsimilarity typical of the fractals [see e.g. T. Viksek,
1986].
No wonder if the dendritic nature of the
structure remains evident not only in the
macrographs of Fig. 4, but also in the micrograph
of fig. 5 a), showing details of the internal surface
of the pores intersected by the fracture surface
of a sampl e whi ch was not LHIP-ed. The
appearance of the pore surface changes drastically
when the alloy is LHIP-ed: at the microscopic level
(~ 30 50 mm) the boundaries between
contiguous dendritic protuberances disappear.
Eqs. (2) and (3) are helpful to account for the
LHIP induced healing of the dendritic branches
at the micrometric scale, provided that we
interpret the grain size diameter G as the average
linear dimension of the microscopic dendritic
protuberances.
Setting:
~ 100 MPa,
= 12 10
-30
m
3
(the vacancy volume is
estimated equal to the volume of an aluminum
atom in the crystal),
D
v
= 6 10
-13
m
2
/s,
(2)
G ~ 5 10
-5
m,
one gets for the drift velocity of the atoms <v>
the following order of magnitude:
(4) <v> ~ 2 10
-8
m/s.
During the duration time t = 35 s of stage II, the
atoms migrate along and across the boundary
between adjacent protuberances by distances of
the order of
(5) <v> t ~ 0.7 m.
(3)
Since the thickness of the boundaries between
contiguous dendritic protuberances is indeed of
the order of 1 m, we believe that the fact that
the internal surfaces of the pores loose their
dendritic appearance should be attributed to the
LHIP assisted local diffusion promoting a kind of
local welding of contiguous protuberances.
In conclusion, the fracture surface of the LHIP-
ed sample exhibits fewer pores of very reduced
dimensions, whose internal surfaces dont exhibit
any more pronounced dendritic features, in
agreement with the above considerations.
(2) This value is derived from J.L. Bocquet et al. (1996), p. 575. It refers to
the alloy in thermodynamic equilibrium. However since selfdiffusion is
associated to a vacancy diffusion mechanism and since in the not LHIP-ed
samples the vacancy concentration largely exceeds the equilibrium
concentration, the numerical values of the strain rate (Eq. (3)), of the
average atomic drift velocity (Eq. (4)) and of the atomic displacements
along and across the interdendritic protuberances (Eq. (5), should be
sensibly underestimated.
(3)

Traces of gases dissolved in the ingot could slightly modify the above
picture. At high temperature the oxygen, in the presence of water vapor,
reacts with the aluminum producing alumina, and liberates hydrogen At
atmospheric pressure the hydrogen solubility in liquid aluminum does not
exceed 1 cm
3
/ 100 g Al. In practice the hydrogen concentration in the
solid is usually confined in the range (0,07 0,4) cm
3
/ 100 g . These values
are more than sufficient to produce supersaturation of this gas during the
solidification at ordinary temperatures, and could be partly responsible for
the lower than ideal density of the solubilized alloy. The hydrogen could
accumulate preferentially in the pores; but since the atomic size of the
hydrogen is much smaller than that of the aluminum atom, most likely,
during solubilization and LHIPing, hydrogen diffuses interstitially in the
bulk, distributing homogeneously in it.
e. n&eUL1e cr 1& M&c~cL 1&e1e
The experimental tests described in 2 lead to the following results:
1. The process of LHIP-ing promotes an increase of the elongation to
fracture, from 0.6 % to 3.0 %. Below a solubilization time of 3 hours
at 520 C, LHIP-ing is ineffective;
2. The process of LHIP-ing promotes also an increase of the ultimate
tensile strength, UTS, from 210 MPa to 250 MPa, again provided that
the duration of the solubilization time exceeds 3 hours.
3. The process of LHIP-ing seems to promote also a slight increase of
the yield stress,
0,2
, - from about 203 MPa to 213 MPa -, barely exceeding
the experimental error.
e. cecUeec~ ~c cc~cLUec~e
The experimental results exhibit a conspicuous improvement of the
elongation to fracture for the LHIP-ed samples in comparison with the as
cast ones.
This finding is coherent with the results obtained by other Laboratories
for the same alloy submitted to other casting processes characterized by
the application of a pressure while casting proceeds. For instance, P.A.
Karnezis, et al. [1996] report an increase by a factor eight of the elongation
to fracture of the squeeze cast and the roll cast alloy with respect to the
gravity cast (Table 1).
0.2% prooI sIress, MPu UTS, MPu ElonguIion, %
Oretl| tes| I1 !! Il l 1. 1.I
!teete tes| I1 1I1 1 l. 1.
Kell tes| II I1 l l.l 1.!
1=eLc . Mcci=ic=L =nc=cn+ice c= +ic ee,
ecLLeiLizco =+ e<cc =cn & icLne ==+cn
c=e+ic, cLccico i ic+ w=+cn =o =cco =+
eec =cn & icLne i=ncm nc=. &)
Fig. 6: DSC diagrams for two samples T and S of
the AA356 alloy described n the text.
These Authors attribute the above results to the reduction of the number
and sizes of the pores, promoted by the applied pressure.
The values reported in Table 1 suggest that there is plenty of room for
further improvement of the elongation to fracture of the alloy under
study.
From the findings of paragraphs 3 and 4, the LHIP-ing time seems to be
sufficiently long for the healing of the dendrites at the microscopic scale.
Evidence for the beneficial effect of the healing of the dendrites on the
mechanical properties of interest in the present study is reported in G.I.
Eskin [1998].
Remarkable advantages are expected from the adoption of higher values
of the solubilization temperature.
In particular, in order to assess the maximum value of the solubilization
temperature, which could be safely adopted without damaging the alloy
provided that it was possible to exert an ideally accurate temperature
control not only in the laboratory but also during production, two sam-
ples of the AA356 alloy described below have been subjected to tests of
Differential Scanning Calorimetry (DSC), aiming to measure the mini-
mum melting temperature of the eutectic:
Sample T. As cast, cooled as fast as possible till 400C, kept at this tem-
. c~cvL&ccM&~1e
We are grateful to Professor Sergio Gallo for his advice and stimulating
discussions. We thank Professor M. F. Ahsby for having made available to us
his Technical Report HIP 6.0, Cambridge University, U.K. (1990), and
Professor Giuseppe Riontino of the Universit di Torino for his DSC tests.
perature for about 1 minute while being removed
from the die, quenched in water and stocked in
freezer before testing;
Sample S. As cast, cooled till 400C, removed from
the die, air cooled, solubilized at 540 for 8 hours,
quenched in water and stocked in freezer be-
fore testing.
Notwithstanding the quite different thermal
treatments experienced by these two samples, a
unique common feature is evident from the DSC
diagrams of Fig. 6: the minimum melting tempera-
ture of the eutectic, whose value, for both, doesnt
drop below 563 C. This temperature is much
higher than the solubilization temperature of 520
C explored so far.
Further tests will be performed in order to
ascertain the effects of higher solubilization
temperatures according to the aforementioned
practical constraints and DSC results.
Thanks are due to Mr. Antonio Mantegazza
(Politecnico di Milano) for his assistance in the
mechanical tests, and to Mr. Olimpio Di Tollo of
the Centro Ricerche Fiat Orbassano, for his work
at the opti cal and the scanni ng el ectron
microscopes.
n&r&n&~c&e
1. M. Rosso, C. Mus, and G. Chiarmetta, Liquid Hot Isostatic Pressing
process to improve properties of thixoformed parts, Metallurgical
Science and Technology, 18, N 2, pp. 16-18 (2000)
2. P.A. Karnezis, G. Durrant and B. Cantor, Effects of Processing on the
Microstructure and Tensile Properties of A356SiC
p
MMCs, Materials
Science Forum Vols. 217-222 (1996) pp.341-346
3. J. Weertman and J.R. Weertman (1970), Mechanical Properties Strongly
Temperature Dependent, in Physical Metallurgy, Second revised edition,
ed. R. W. Cahn, North Holland, pp. 985-1010
4. A. S. Argon (1996), Mechanical Properties of Single-Phase Crystalline
Media: Deformation in the Presence of Diffusion, in Physical Metallurgy,
Third edition, Eds. R. W. Cahn and P Haasen, Elsevier, pp. 1958-2007
5. J.L. Bocquet, G. Brebec and Y. Limoge, Diffusion in Metals and Alloys in
Physical Metallurgy, 4-th revised edition, R.W. Cahn and P Haasen Eds.,
Elsevier Science, (1996) Ch. 7
6. M. F. Ashby (1990), HIP 6.0: Background
Reading, Sintering and Isostatic Pressing
Diagrams, Technical Report, Department of
Engineering, University of Cambridge
7. H. E. Exner and E. Arzt (1996), Sintering
Processes, in Physical Metallurgy, 4-th revised
edition, R.W. Cahn and P. Haasen Eds., Elsevier
Science, (1996).
8. T. Vicsek (1986), Formation of Solidification
Patterns in Aggregation Models, in Fractals
in Physics, Eds. L Pietronero and E. Tosatti,
North Holland, Amsterdam, pp. 247-250
9. G.I. Eskin, Ultrasonic Treatment of Light Alloy
Melts, Gordon and Breach Science Publishers,
1990
Mc+=LLLncic=L
eciccc =o 1ccicLcc-
JcLn=L =LeLisico
e- 1csio
+wicc = -c=n
Micncs+nLc+Ln=L cvvcc+ cv +ic L...=.
=ncccss c +ic ee =LLmiiLm =LLc-
H. CabIbbe and . vaneIIrta
lKFH - 0ept. ef HethanItr, 0nIverrIty ef Antena, Antena
Micnce+nLc+Ln=L cvvcc+ cv +ic L...=.
=ncccee c +ic ee =LLmiiLm =LLc
H. CabIbbe and . vaneIIrta
lKFH - 0ept. ef HethanItr, 0nIverrIty ef Antena, Antena
Abstract
An extended investigation of the microstructure of A356
both as cast and processed by Liquid Hot Isostatic Pressing
(LHIP) was performed using TEM, SEM and X-ray diffraction
techniques. Specimens of the alloy in both conditions were
solutionised for 8h at 520C and aged for 6h at 160C.
TEM investigations revealed no significant influence of
the L.H.I.P. process on particle precipitation kinetics and
morphology. Pores were investigated and characterised
by means of light and electron microscopy and statistical
techniques. L.H.I.P. treatment reduced the pores mean
size by about 1/3, the volume fraction by one order of
magnitude and the shrinkage voids in number and volume.
Riassunto
La lega A356 da getto, dopo trattamento di L.H.I.P. stata studiata per mezzo di
tecniche di microscopia elettronica; TEM, SEM e di diffrazione a raggi X. Il materiale
stato solubilizzato a 520C per 8 ore e quindi invecchiato a 160C per 6 ore. Le indagini
al TEM non hanno mostrato alcuna influenza significativa del processo di L.H.I.P. sulla
cinetica e sulla morfologia di precipitazione. La percentuale di micro pori prodotti da
gas intrappolati durante il riempimento dello stampo stata calcolata per mezzo di
osservazioni al microscopio ottico ed elettronico in scansione (SEM). I risultati sono
il frutto di elaborazioni statistiche dei dati ottenuti. La dimensione media dei pori,
dopo L.H.I.P., raggiunge valori inferiori ad 1/3 rispetto al materiale non trattato, e la
frazione in volume ridotta di un ordine di grandezza, inoltre i vuoti da ritiro subiscono
una drastica diminuzione in numero e volume.
~1nccUc1c~
In an effort to reduce the costs and the fuel con-
sumption of transportation means, new manu-
facturing technologies and light-weight metals,
such as aluminium and magnesium alloys, are be-
ing used with increasing frequency. In automo-
tive applications, weight reduction and high-qual-
ity components, coupled with lower costs, are a
must. New production processes aim at improv-
ing mechanical properties, especially tensile
strength and fatigue behaviour. Moreover, cast-
ings suffer from voids due to gas entrapped (H
2
,
O
2
, N
2
) by the liquid alloy during turbulent filling
of the die and to subsequent shrinkage during
solidification. The different coefficients of ther-
mal expansion of primary particles and matrix
also induce voids at interfaces. Both shrinkage
voids and gas inclusion pores impair strength and
fatigue resistance, as they can easily coalesce to
premature fracture. Hot Isostatic Pressing (H.I.P.)
has successfully overcome the susceptibility to
shrinkage porosity of highly performing cast ma-
terials [1]. Unfortunately, its complexity makes H.I.P. a technology suitable
for a narrow range of applications. Recently, this process has been modi-
fied by using, instead of a gas, a liquid at elevated temperatures to reach
high pressures in few seconds (L.H.I.P.). On this basis, Metal Casting Tech-
nology has successfully engineered the process and extended it to mass
production by increasing the volume of the room available for the press-
ing process [2]. During the L.H.I.P. process, the combination of pressure
and temperature is able to close the voids due to shrinkage and the pores
due to gases entrapment (such as H
2
and O
2
) through the dissolution of
gaseous inclusions within the lattice [2,3,4]. The novel development of
H.I.P. process has constituted the starting point for Teksids experimenta-
tion on castings of aluminium alloys using a mixture of salts particularly
suitable for its low melting point, low density and viscosity, operating at
temperatures ranging between 410 and 535C. The castings are preheated
at the temperature of the liquid and then introduced in the salt bath
chamber, whose pressure is maintained for about 200 seconds over 1000
bars. The final step consists of quenching the castings to make them suit-
able for the subsequent hardening heat treatments [2-7].
The present paper aims to investigate the effect of the L.H.I.P. process on
the microstructural properties of an A356 Aluminium alloy. To this end,
we compared the microstructure of specimens before and after L.H.I.P.
treatment aged at the same conditions.
&=&nM&~1L =ncc&cUn&e
The A356 aluminium alloy, with the chemical composition reported in
Table I, was sand cast at 715C in the form of cylindrical specimens. One
set of specimens (hereafter A) was solutionised at 520C for 8h, quenched
and aged at 160C for 6h; another set (hereafter B) was solutionised,
subjected to L.H.I.P. and then aged.
Light (LM) and scanning electron microscopy (SEM) investigations were
conducted on specimens polished and etched
with a 0.5% solution of HF in methanol. For TEM,
slices were mechanically ground to about 120 mm,
then etched with a 25% HNO
3
solution in metha-
nol to about 80 mm. Discs were punched and
finally thinned by double-jetting with a solution
of a 1/3 HNO
3
in methanol at -30C and 18/20V.
5i Fe Cu Mn Mg Zn Ii Al
. 1- l. 1 1. 1 1. I 1. 1 1. I1- 1. 1 1. 1 1. I |el .
1=eLc . ccm=cei+ic cv +ic ee =LLmiiLm
=LLc
n&eUL1e ~c cecUeec~
X-ray Energy Dispersive Spectrometry (EDS) investigations on SEM and
Electron Diffraction Pattern (EDP) indexing on TEM were used to iden-
tify the hardening phases in specimens A and B. The microstructure of A
and B aged specimens is shown in Figure 1a, b. The Si particles, which had
an aspect ratio ranging between 1.2 and 2.2, appeared dark and typically
Fig. 2: Micrographs of second phases in A, (a): Fe
2
Si
2
Al
9
, (b): FeMg
3
Si
6
Al
8
,(c): all the present dark-grey rounded particles are Si
(1) The equivalent diameter is a one-dimensional measure of the diameter
of spherical shaped particles having the same volume of non-spherical ones
Fig. 1: Light microscopic micrographs of the same specimen:
(a) solutionised at 520C for 8h and then aged at 160C for 6h,
(b) after L.H.I.P. under the same ageing conditions as (a).
Fig. 1a Fig. 1b
spherical or slightly elongated. The Fe
2
Si
2
Al
9
par-
ticles appeared as light-grey needle-like and the
FeMg
3
Si
6
Al
8
phase appeared as chinese script-like
and darker grey particles (see also Figure 2). A
statistical evaluation of the volume fraction and
dimensional variation of all the phases present in
A and B specimens was carried out on TEM mi-
crographs. The results - summarised in Table II -
show no significant differences in particle volume
fraction. The difference observed regarded the
mean equivalent diameter
(1)
of the Si hardening
particles. The mean equivalent diameter differ-
ence between A and B has been found to be: (d
A
- d
B
) / d
A
= |(6.6 - 8.2) / 6.6| = 24%, which shows
Fig. 2a Fig. 2c
100 m 100 m
|e
2
!i
2
A|
9 |eNg
3
!i
6
A|
&
6oseoes potes
20 m
Fig. 2b
A35 Volume Equivolenl A35 Volume Equivolenl
(Cosl) Froclion* Diomeler* (L.H.l.P.) Froclion* Diomeler*
(%) ( m) (%) ( m)
!i .&0 6.6 !i .55 &.2
|eNg
3
!i
6
A|
&
.I5 32 |eNg
3
!i
6
A|
&
.65 30
|e
2
!i
2
A|
9
.95 32 |e
2
!i
2
A|
9
2.0 32
1=eLc &. vcLLmc rn=c+ic =o &cLiv=Lc+
ci=mc+cn cv =LL +ic oc+cc+co eccco =i=ece
i e=ccimce =o e. LL +ic nceLL+e i=vc
ecc =nccceeco e +ic ocoic=+co ecv+w=nc:
L=+imc+ ecc
*: associated error within one standard deviation
Fig. 3: X-ray diffraction analysis on the A356 aluminium alloy, and EDP of
the different second phases detected, (a): Si, (b): Fe
2
Si
2
Al
9
, (c): FeMg
3
Si
6
Al
8
.
The peaks of the cast and the L.H.I.P. materials practically overlap.
Volume Equivolenl
Froclion* Diomeler*
(%) ( m)
A356 - [A) 2.& I.0
A356 |.H.I.P. - [8) 0.6 2.I
1=eLc . vcLLmc =n=c+ic =o ccLiv=Lc+
mc= oi=mc+cn c= +ic c=e+ =o L...=.
m=+cni=Le vcioe
*: associated error within one standard deviation
the B specimens particles to be greater than in
A.
The different phases, identified in specimens A
and B, are shown in Figures 2 and 3. TEM investi-
gations were also focused on the hardening
microstructural precipitation. The hardness of the
A and B samples was practically identical (92 2
HB and 91 2 HB, respectively), reflecting the
lack of effectiveness of L.H.I.P. on hardening ki-
netics. The particles dimension of either A and B
was bigger than GP and the relative difference
did not affect significantly the hardness.
The LM micrographs of A (figure 1a) showed
clearly a high concentration of pores and shrink-
age voids, which were reduced by the L.H.I.P. proc-
ess (figure 1b). The high pressures involved in the
process produced a denser material by reducing
micro and macroporosity. A detailed microstruc-
tural investigation allowed to identify the prefer-
ential location and the mechanism of microcracks
formation. Figure 4 shows an intergranular re-
gion where the primary phases are located and
this microstructural feature is common on A as
well in B. In particular, the iron-rich particles
(Fe
2
Si
2
Al
9
and FeMg
3
Si
6
Al
8
) showed micro-voids
and shrinkage cracks (the little dark striped zone)
along the boundary with the matrix. Such a micro-
structural features at interfaces, under high stress
concentration, can coalesce into cracks.
The presence of discontinuities at boundaries of
the primary iron rich (>0.4wt.%) particles in the
cast A356 alloy has been observed by Taylor and
co-authors who analysed the role of the Si con-
tent [8]. They pointed out that the porosity at
interfaces is minimised when solidification pro-
ceeded through ternary diagram Al-Si-Fe
2
Si
2
Al
9
.
In case of Si deficiency, the solidification pro-
ceeded thorough the binary Al-Fe
2
Si
2
Al
9
and the
porosity volume fraction increased. An explana-
tion of the primary particles on discontinuities at
interfaces was the restricted feeding theory,
proposed by Mascr [9]. The theory basically sug-
gests that Fe
2
Si
2
Al
9
and FeMg
3
Si
6
Al
8
intermetallic
platelets formed within the interdendritic region
during solidification and caused physical restric-
tions to the movement of the feeding liquid. It
follows that voids are formed since interdendritic
regions cannot be adequately fed.
The statistical evaluation of the porosity, on speci-
mens A and B, has been focused on mean dimen-
sion (i.e. the equivalent diameter) and volume
fraction. The results are reported in Table III and
in Figure 5. The mean micro-pores dimension due
to H
2
and O
2
[4] was up to 1/3 lower after L.H.I.P.
and the porosity level (i.e. the micro-pores vol-
I
n
t
e
n
s
i
t
y
,

c
p
s
0
100
200
300
400
500
600
700
800
900
d-Scale
(a) (b) (c)
Si Fe
2
Si
2
Al
9
FeMg
3
Si
6
Al
8
Fig. 4: Intergranular region where the two iron-rich second phases Fe
2
Si
2
Al
9
and
FeMg
3
Si
6
Al
8
are located. A dark striped zone (a microcrack) appears along the
boundary between the particle and the matrix
Fig. 5: Microvoid distribution. The cast and L.H.I.P. material are compared. The
equivalent diameter is a one-dimensional measure of the magnitude of an object,
and measures the diameter that the object would have if spherical (the equivalent
particle volume is obtained by multiplying the square of the equivalent diameter by
p/4). Left-side set of data: L.H.I.P. material; right-side data: cast specimens
cc~cLUec~e
The effect of LH.I.P. on the microstructure of
aged samples of A356 was investigated by means
of light and electron microscopy, and X-ray dif-
fraction techniques. The results can be summa-
rised as follow:
a) L.H.I.P. did not significantly influence the pre-
cipitation kinetics of Si particles that control-
led the hardening of the alloy; the Fe
2
Si
2
Al
9
and FeMg
3
Si
6
Al
8
primary particles were stable
at the temperatures of exposure.
b) L.H.I.P. dramatically reduced porosity from en-
trapped H
2
and O
2
in volume fraction and di-
mension. Morover, L.H.I.P. was capable of elimi-
nating micro and macro voids.
c~cwL&co&M&~1e
The research was supported by funds from the Ministry of University and
Scientific Research. The authors are indebted to S. Gallo and C. Mus of
Teksid for supplying the material and for useful discussions.
n&r&n&~c&e
1. S. Gallo, C. Mus, S.P. Barone, G. Scholl and G. Mortari liquid Hot
Isostatic Pressing - A breaking Through Technology, SAE Technical
Papers Series, 2001-01M-140.
2. S.Gallo, C.Mus, S.P.Barone, G.Scholl and G.Mortari Liquid Hot Isostatic
Pressing A Breaking Through Technology, SAE Technical Papers Se-
ries, 2001-01M-140.
3. E.L.Roory, Improving Casting Properties and Integrity with Hot
Isostatic Pressing, Modern Casting, 1983, 10, 18.
4. M. Rosso, C. Mus and G. Chiarmetta, Liquid Hot Isostatic Pressing
Process to Improve Properties of Thixoformed Parts, 6
th
Intl Con-
ference on Semi-Solid Processing of Alloys and Composites, G.L.
Chiarmetta and M. Rosso editors, Edimet, Brescia, 2000, 209.
5. G.R. Wakefield and R.M.Sharp, Effect of Casting Technique on Fa-
tigue Properties of Hot Isostatically Pressed A.l10Mg castings, Ma-
ume fraction) was down from 2.8% to 0.6%
throughout the specimen volume, entailing a re-
duction in porosity by about 80%. The N
2
pores
were not affected by L.H.I.P. [4]. Thus, the major-
ity of macro pores (i.e. those having an equiva-
lent diameter of the order of 0.7-0.6 mm) were
reduced to micro pores of the order of 0.2-0.3
mm, and the former micro pores eventually dis-
appeared.
The beneficial effect of the reduction in volume
fraction of the mean pores size on tensile and
fatigue behaviour produced by the L.H.I.P. proc-
ess has been described in [4]. In particular, the
fatigue properties of B specimens improved two-
threefold compared with the A ones.
terials Science & Technology, 1996, 6, 12.
6. S.J.Marshal and J.C.Hebeisen, Hot Isostatic
Pressing of A356 Alloys and 380/383 Alu-
minium Alloys: An Evaluation of Porosity,
Fatigue Properties and Processing Costs,
SAE Technical Papers Series, 2000-01-0062.
7. C.H.Caceres, M.B.Djurdjevic, T.J.Stockwell and
J.H.Sokolowski, The effect of Cu Content
on the Level of Microporosity in Al-Si-Cu-Mg
Casting Alloys, Scripta Mater., 1999, 40, 5, 631.
8. J.A.Taylor, G.B.Schaffer and D.H.StJohn, The
Role of Iron in the Formation of Porosity
in Al-Si-Cu-Based Casting Alloys Part I, II &
III, Metallurgical & Materials Transactions A,
1999, 30A, 1643.
9. M.C.Mascr, Fonderie, 1955, 8, 4, 225.
|eNg
3
!i
6
A|
& |e
2
!i
2
A|
9
Mc+=LLLncic=L
eciccc =o 1ccicLcc-
JcLn=L =LeLisico
e- 1csio
+wicc = -c=n
nc=scs +c ocvcLc= LicLio ic+
iscs+=+ic =ncssic
*I. 0aIIe,*C. Hur - ** 0. HertarI
*1ekrId I.p.A. - 8erarette (1O} ltaIy - ** ldra, 8rertIa, ltaIy
~!nccDC!c~
The well established castings process (gravity die-
casting , low-pressure die-casting and sand cast-
ing) can produce good quality casting having non
uniform microstructure.
These processes can lead to different directional
solidification path in different locations. The foun-
dry expertise helps in defining where solidifica-
tion can start (very fine microstructure and good
properties),having these starting points close to
the critical area of the component and moving
any segregations and micro shrinkage cavities to
areas of the component less subjected to stress.
Process selection can be done when component
specification are defined.
nc=scs +c ccvcLc= LicLic ic+
iscs+=+ic =ncssic
*I. 0aIIe,*C. Hur - ** 0. HertarI
*1ekrId I.p.A. - 8erarette (1O} ltaIy - ** ldra, 8rertIa, ltaIy
Abstract
Following the oil crisis in the 1970s and consumer
awareness of the great harm caused to the environment
by motor vehicle emissions, automakers began to overhaul
their design criteria with a view to reducing fuel
consumption. One of the many solutions adopted has
been the extensive use of Aluminum and Magnesium
instead of steel and cast iron. This has resulted in great
weight savings and hence a reduction in fuel consumption.
It has also given Aluminum and Magnesium foundries access
to new and attractive markets. In the first substantial
applications, namely the fabrication of Aluminum cylinder
heads and engine blocks, quality requirements were
primarily concentrated on static strength and water and
oil tightness, in addition, of course, to dimensional
precision. New and important uses are being found for
Aluminum alloys: steering knuckles and suspension links,
levers and cross members. For these families, the main
quality requirements relate to fatigue and impact
resistance. New alloys and new manufacturing processes
are necessary to ensure that suspension components are
endowed with these qualities.
The HIP Hot Isostatic Process has been introduced
several years ago has been proven one interesting and
effective way of Aluminum treatment to increase fatigue
strength. An evolution of the concept, known as Liquid
HIP- LHIP, perfected in conjunction with Metal Casting
Technology Inc., Idra Presse and Teksid is now coming on
stream. It ensures quality levels and trimmed costs difficult
to obtain with conventional processes.
Riassunto
In conseguenza alla crisi petrolifera degli anni 70 le industrie automobilistiche
cominciarono a porre particolare attenzione alla progettazione di veicoli pi leggeri
con lobiettivo di ridurne i consumi. Tra le soluzioni adottate vi fu una progressiva
sostituzione di materiali dalla ghisa e acciaio allalluminio e al magnesio. Si aprirono
interessanti prospettive di crescita per le fonderie in grado di trasformare le leghe
leggere. Per le prime applicazioni (teste cilindro e basamenti in lega di alluminio) le
specifiche di prodotto richiedevano al getto, oltre alla precisione dimensionale,
unicamente la resistenza alla prova di tenuta idraulica e resistenza statica. Linteresse
ad applicare le leghe di alluminio ad altri componenti, ad esempio particolari di sicurezza
quali montanti, bracci e traverse sospensione, richiede un inseverimento delle specifiche,
in particolare resistenza a fatica ed impatto. Per soddisfare le nuove esigenze risulta
necessario sviluppare nuove leghe e processi di trasformazione.
Il processo HIP Pressatura isostatica a caldo, stato introdotto molti anni fa e si
dimostrato efficace nellincrementare la resistenza a fatica delle leghe leggere.
Unevoluzione del processo , LHIP Pressatura isostatica a caldo in liquido, sviluppata
da Metal Casting Technology, Idra Presse e Teksid ha fornito interessanti risultati che
sono qui descritti.
If we take into consideration as an example the engine cylinder heads and
we analize our findings from tests carried out on more than 70 types of
engines over the last 30 years [1], we can reach the conclusion that the
sand-cast process is not suitable for cylinder heads on diesel engines and
certain types of highly stressed petrol engines since in these application,
under heavy loading conditions, there are high risk of failure due to ther-
mal fatigue, even when using primary alloys. This is attributable to the
microstructure (dendritic arm space) and to the presence of micro shrink-
age cavities between the dendrites, which are unavoidable due to the
slow speed of solidification typical of this casting processes [2],[3]. Our
findings also show that cylinder heads cast with the Lost Foam process
present similar problems.
New and important uses are being found for Aluminum alloys: steering
knuckles and suspension links, levers and cross members. For these fami-
lies, the main quality requirements relate to fatigue and impact resistance.
New alloys and new manufacturing processes are necessary to ensure
Fig. 1: Solid shape (left) and hollow shape (right) components with cores
that suspension components are endowed with these qualities [4].
These components are currently made from ductile iron, steel forging or
stamping of sheets of welded steel. The new solutions made from Aluminum
alloys are subdivided into two categories:
solid shapes (e.g. steering knuckles, arms, cross members)
hollow shapes (arms with cores ) Fig. 1
Several technologies are available for manufacturing the components of
the first category like permanent mould die casting, low-pressure die casting,
semisolid processing, squeeze-casting, Vacural, forging.
Due to the presence of the cores, only the permanent mould die castings
or the low-pressure die castings are viable solutions for manufacturing
the components of the second category.
Fig. 2 : LHIP schematic concept
c! ece!!C =n&ee~o =ncC&ee
As it is well known from the extensive experience in production, the
treatment of Hot Isostatic Pressing (HIP) [5], [6], can eliminate the micro
shrinkage cavities between the dendrites and the porosity attributable to
hydrogen present in the castings, defects that considerably limit the re-
sistance to fatigue of the components.
This process has only been applied on aeronautic components due to the
high costs involved attributable to the long cycle time and the burden-
!& L= (LDc c!
ece!!C =n&ee~o)
=ncC&ee
The process principle is based on the idea of ap-
plying the isostatic pressure over the casting
through a liquid instead than through a gas in
order to overcome the HIP cost process issues.
It can be easily understood that the cycle time
can be dramatically reduced (from hours to min-
utes) and the risk of explosion of the high pres-
sure working vessel can be reduced to zero (the
liquid pressure will immediately drop in case of
leakage or failure).
Considerable weight savings are obtainable by
opting for the solutions made from Aluminum
alloys. This ranges from 30 to 40% depending on
the type of component and the degree of opti-
mization of the drawing. Given the considerable
difference in terms of the cost of the raw mate-
rials, the solution employing the Aluminum alloy
is more expensive than the one employing the
ferrous material. Automakers are highly inter-
ested in the solutions using Aluminum since they
lead to valuable weight reductions of the sus-
pended components with increased energy sav-
ings and added driving comfort. However,
automakers are much less inclined to accept con-
siderable increases in costs. For those reasons,
the foundry industry needs to pay considerable
attention on the choice of the production proc-
ess, in order to obtain a component that per-
forms optimally minimising costs. In order to
achieve high resistance to fatigue and crash and
guarantee the quality requirements requested for
safety components such as those of the suspen-
sions, new alloys and new treatments and con-
trol systems are necessary and have been devel-
oped.
some safety measures required for the risk of
explosion when working with compressed gases
at high pressures (1000 atm) and high operating
temperatures.
A new process, which we called Liquid HIP - LHIP,
was devised from the works of Dixon Chandley
[7] and perfected and industrialized thanks to a
close collaboration between Metal Casting Tech-
nology Inc, Idra Presse, and Teksid [8].
Fig. 3: LHIP process conditions and related effect
The selected liquid had to fulfill the following re-
quirements:
low cost
recyclable and easily washable
non corrosive for the aluminum alloy and for
the vessel material
low temperature melting point (250- 300 C)
high temperature boiling point (above 600C)
A long period time has been dedicated to the
testing of different solutions and a family of
eutectics salts has been defined and verified. The
vessels material has been object of a deep inves-
tigation with the goal of guarantee thightness
under operating conditions (Fig. 2). The selection
study went through a best compromise check-
loop among high temperature mechanical prop-
erties (thermal fatigue and creep at 500-600C)
and corrosion performances. A cost/perform-
ances analysis ended-up with the selection of an
high strenght steel grade, able to reach the one
million cycle life, instead of a infinite life nickel
based super alloys.
As the HIP process, LHIP is capable of eliminating
some of the typical casting defects (fig. 3) like
micro and macro shrinkage porosity and hydro-
gen inclusion. Defects connected with the sur-
face (i.e. cold shots, surface cracks), as well as
nitrogen inclusions and oxides, cannot be elimi-
nated: these kind of defects can be slightly modi-
fied in shape.
Process normal running parameters to obtain
such results on Aluminum castings are:
1000-1200 ATM pressure
400-540C salt temperature
20-35 seconds pressure applied
3-4 min. cycle time (charging, treating, quench-
ing).
The LHIP effect on the microstructure of the
treated castings improves the material mechani-
cal properties and increases density (fig. 4).
In particular on sand casting, tensile and yield
strength are slightly increased when compared
with the untreated material, while fatigue strength
can be dramatically increased up to two times.
Elongation can be positively affected by LHIP, even
if the casting process is the main driver in the
achievable level of such an important material
property. As well known oxide and DAS have a
strong influence on the ductility: LHIP is not af-
fecting those two features. Table 1 shows results
achieved on A356 sand-cast specimen, with no
chills, when treated LHIP, and in figure 5 the frac-
ture surface analysis shows how micro-shrinkage
porosities are eliminated after LHIP.
5ond cosl 5ond cosl LHlP
I I
Tonsilo slronglh 230-250 MPo 250-300 MPo
Yiold slronglh T0-2T0 MPo 2T0-250 MPo
Elongolion T-2 4-
Foliguo Slronglh 80-T00 MPo T20-T80 MPo
!eL& .
=nc=cn+ics ccm==nisc
Fig. 5: Comparison of fracture surface before (left) and after LHIP treatment
Fig. 4: Comparison of suspension arms density before ad after LHIP treatment
An industrial pilot plant (internal vessel diameter
470 mm and depth 1420 mm) is installed at the
TEKSID Technical Centre (fig. 6-a). Running in
automatic conditions the achieved cycle time is
in the range of 3 5 minutes.
A full scale industrial plant has been defined (in-
ternal vessel diameter 900 mm and depth 2000
mm) and will be placed in between the solution
and the ageing furnace. The LHIP stage will be
fully integrated in the heat treatment process
flow (fig. 6- b).
The castings will be introduced in the salt bath,
always at high temperature and water quenched
before ageing. This approach can be considered
the most suitable to cut treatment costs since
cycle time, energy and manpower can be dra-
matically reduced.
The target LHIP costs including the full T6 heat
treatment are in a very competitive range given
by a shape factor of the castings to be treated:
compact castings can more efficiently fill the avail-
able vessel volume for a lower cost per kg of
material treated. Those target range costs make
the process attractive for the automotive indus-
try opening a bright future for this technology.
As an example suspension components could be
produced as sand casting on fully automated cast-
ing machines (i.e. Disamatic) and then treated LHIP
with the achievable goal of high fatigue strength
similar and higher than other more expensive
casting techniques (i.e. low pressure casting). Test-
ing programs are in progress: preliminary results
are very encouraging both on specimen and on
components (fig.7).
Fig. 6-a: LHIP pilot plant
Fig. 7: Comparison of fatigue test results on A356 components and cut-out specimen
5pecimen dolo Low 5ond 5ond
pressure cosling cosling
wilh chills no chills
LHlP
Fe|lte llml|
!D [Pe] l1. ! 1!. 1 l. !
5pecimen Low 5ond 5ond
componenls pressure cosling cosling
wilh chills no chills
LHlP
/teree ll|e
[#e| ttles} 11. 111 111. 111 11. 111
Fig. 6-b: LHIP integration in the H.T. cycle
Cc~CLDec~e
The needs to reduce fuel consumption and emis-
sions will drive the growing request of light com-
ponents in the automotive industry. An increased
demand will come for Aluminum safety parts. In
order to get the business in this emerging mar-
ket, foundries have to be ready in delivering high
quality and high performances castings at competitive prices. Research
and development is the key factor to be the winner in this future rush
since most of the well established casting techniques cannot satisfy these
demands. Breaking-through technologies will be the answer to these fu-
ture needs. Based on the results achieved so far the LHIP technology can
be considered one interesting alternative to produce high performances
aluminium components at competitive costs both for safety components
both for heavy duty engine components.
n&r&n&~C&e
1. G.Gorrea, R.Medana. Prove di shock termico
su teste cilindri in leghe di Alluminio.
Alluminio Magazine, 10 (1987) 30-33
2. T.S. Piwonka and M.C.Flemings, Dendrite
Arm Spacing and Solidification, Trans. AIME,
236 (1966 1157-1165
3. M.C.Flemings, T:Z:Kattamis, B.P:Bardes, Den-
drite Arm Spacing in Aluminum Alloys, Sil-
ver Anniversary Paper, Aluminum, Div. 2,
Trans. A.F.S, 99 (1991) 501-506
4. M.C.Flemings and H.F.Taylor, Super-Strong Light-Alloy Castings,
Machine Design (June 12, 1958) 22-25
5. G.E.Wasiliewski and N.R. Lindblad, Elimination of casting defects us-
ing HIP, Proceedings of the 2
nd
International Conference, MCIC
Sept.1972
6. Hot Isostatic Pressing of A356- SAE Paper 2000-01-0062
7. Dickson Chandley , Optimizing processing for quality castings, Pro-
ceedings, Merton Flemings Symposium, Cambridge, June 2000.
8. S.Gallo, C.Mus,S.Barone, G.Scholl, G.Mortari, Liquid Hot Isostatic
Pressing A breaking through tecnology SAE Paper 2001-01M-140
Mc+=LLLncic=L
eciccc =o 1ccicLcc-
JcLn=L =LeLisico
e- 1csio
+wicc = -c=n
1ic cvvcc+ cv LicLio ic+ iscs+=+ic =ncssic
c v=+icLc =nc=cn+ics cv L e=sco c=s+ics
*. kemane, *H. kerre, - **C. Hur
*eIItetnIte dI 1erIne, ltaIy - **1ekrId, 1erIne, ltaIy
~!nccDC!c~
LHIP (Liquid Hot Isostatic Pressing) is a process that subjects a compo-
nent to both elevated temperature and liquid isostatic pressure in a suit-
able vessel. Under these conditions of heat and pressure, internal pores
or defects within a solid body collapse and weld up.
This process represents the evolution of HIP (Hot Isostatic Pressing), in
which a gas, argon, is used as the pressuring medium. The HIP process was
invented in the mid-1950s at Battelle Columbus Laboratories. The proc-
ess was specifically developed as a gas pressure diffusion bonding tech-
nique for cladding nuclear fuel elements. In 1965 the use of hot isostatic
pressing for improving the fatigue life of cast aluminium diesel engine was
investigated. Further investigations into the effect of HIP on titanium and
superalloy castings were conducted in the 1970s and they demonstrated
that significant improvement in mechanical properties could be obtained
[1]. Initially most of the casting being hot isostatically pressed were aero-
space components, such as turbine blades and structural hardware for
!ic c==cc+ c= LicLic ic+ iscs+=+ic =ncssic
c ==+icLc =nc=cn+ics c= L e=scc c=s+ics
*. kemane, *H. kerre, - **C. Hur
*eIItetnIte dI 1erIne, ltaIy - **1ekrId, 1erIne, ltaIy
Abstract
Because of weight and cost savings automotive manufactures are making more extensive
use of cast aluminium alloys to replace forged and cast iron or steel components. The
presence of pores in certain locations may reduce the fatigue life and it is cause of a
large scatter in the mechanical properties of cast parts. This is of special concern for
safety critical components.
In this work A356 T6 heat-treated specimens were drawn from a suspension arm, that
was manufactured with two different low pressure foundry processes, permanent
mould and Disamatic process, and they were subjected or not to liquid hot isostatic
pressing in a salt bath at 120 MPa and 540 C for 30 s. Then high-cycle axial fatigue tests
were performed on smooth plane samples at room temperature with a tension
compression ratio of 1, in order to obtain fatigue S-N curves in the different conditions.
After LHIP it was found that there is a large increase of fatigue resistance and a
remarkable decrease of the scatter of fatigue data, which is due to the presence of
internal discontinuities. Optical microscopy and scanning electron microscopy (SEM)
were used to document the elimination of porosity after LHIP and to show the different
nucleation sites of fatigue cracks on the fracture surfaces in no-LHIP and LHIP conditions.
Riassunto
Numerosi componenti di meccanica dellauto vengono
oggi realizzati in leghe di alluminio in alternativa allacciaio
e alla ghisa per ragioni di alleggerimento. Come noto la
vita a fatica dei getti in lega leggera fortemente influenzata
dalla presenza di porosit. Questo fenomeno risulta
particolarmente critico per i particolari di sicurezza. Con
il presente lavoro sono stati analizzati campioni di lega
A356-T6 ricavati da componenti della sospensione
prodotti con due tecnologie di colata in bassa pressione
: conchiglia e Disamatic. Sui campioni sono stati valutati gli
effetti del processo LHIP in bagno di sali a 120 MPa e
530C per 30 secondi. In particolare sono stati osservati
gli effetti del processo sulla vita a fatica ad alto numero di
cicli e temperatura ambiente. Il processo LHIP aumenta
notevolmente la resistenza a fatica dei getti e riduce la
dispersione dei dati relativi.
Lanalisi micrografica ed al microscopio a scansione ha
permesso di documentare la riduzione di porosit nei
getti trattati LHIP e leffetto sulla fase di innesco della
frattura.
aircraft gas turbine engines. These expensive com-
ponents could accept the high-added cost of the
HIP process. Also the casting industry originally
perceived HIP as a means to reduce its scrap rates.
Prior to the development of HIP no other non-
destructive technique was available for the eco-
nomic repair of castings containing internal po-
rosity. Recently, however the cost of HIP has been
reduced (increased number and size of vessels
dedicated to HIP of castings and improvements
in equipment) to the point at which it is afford-
able for less critical applications. [2]. Anyway the
HIP is a discontinuous process and it is hard to
expect its integration into a productive cycle, e.g.
of an automotive casting, also because of the
elevate cycle times (4-6 hours).
LHIP presents two significant advantages because
Fig. 1: Defects, which can be closed after LHIP:
a) H
2
pore, b) Solidification shrinkage
it is a continuous process and it enables a drastic
reduction of the process time, thus to make pos-
sible a high grade of automation. In this process
the pressurizing medium is a salt bath, with a much
higher density than argon at process tempera-
ture, and so it is able to reach high pressures in a
shorter time.
This process is currently under development with
the aim to upgrade the quality of castings. In fact
castings may have internal defects such as gas
porosity (due to trapped gases, such as hydrogen
or air), interdendritic microporosity (due to
shrinkage) and microcracks that form during so-
lidification and that have an adverse effect on
performance. Such pores tend to concentrate
stress and therefore act as yield and crack initia-
tion points. The more sharply angled a pore, the
more the stress will be concentrated. The LHIP
process heals these defects first through creep
and plastic deformation and then by diffusion
bonding of the surfaces of the collapsed area,
creating a casting with a fully dense omogeneous
microstructure. During processing the treatment
temperatures are maintained with the plastic
range of the material concerned, or more pre-
cisely, high enough for the diffusion bonding to
occur but at the same time low enough to avoid
undesidered microstructural modification such as
grain growth.
The higher integrity and homogeneity of the cast-
ing after LHIP improves its mechanical charac-
teristics with the subsequent increase of fatigue
resistance but also of tensile and yield strengths,
ductility and, creep life [3,4]. It also reduces the
scatter of these properties. As the volume nor-
mally occupied by defects is limited, LHIP does
not significantly modify the dimensions or the
shape of the casting. In addition if large voids ex-
ist in the casting, surface depressions may develop
as a result of material displacement during void
closure. If small, uniformly distributed voids are
present in a casting, a dimpled surface may be
evident after hot isostatic pressing. Such a condi-
tion is typical in many sand-cast aluminium parts.
LHIP treatment does not heal surface related
defects. For this type of healing the part requires
a coating to act as barrier against the pressu-
rized liquid [3]. Furthermore, diffusion bonding
does not occur when metal/metal contact is ob-
structed, if for example the surfaces of the de-
fect are oxidized or if there is a gas inside the
pore that does not diffuse in the material, e.g. air.
In the figures 1 and 2 the possible situations are
shown.
At the same time the use of aluminium castings in the automotive indus-
try has seen ever increasing growth due to the economic savings of a
reduced number of production steps compared to other manufacturing
methods such as forging. These production advantages have lead to strong
interest in aluminium castings for applications where stringent mechanical
requirements must be met.
The fatigue properties of Al-Si castings alloys have been related to micro-
Fig. 2: Defects, which can not be removed after LHIP:
a) Air entrapped, b) Surface porosity
structural features such as secondary dendrite arm spacing (SDAS) [5]
and porosity morphology and location [6], to variations in heat treatment
[7] and to loading parameters such as the load sequence or the load ratio
[8].
It is generally accepted that the fatigue process consists mainly of four
important stages: microcrack initiation, microcrack coalescence and growth
(small-crack propagation), macrocrack growth and final fracture. From
the point view of fatigue life, the crack propagation dominates the low-
cycle fatigue life and crack initiation controls the high-cycle fatigue life.
Fatigue cracks in aluminium alloys have been found to nucleate from slip
bands, inclusion breakage and debonding, and grain-boundary separation,
voids and notches [9]. Anyway in hypoeutectic Al-Si casting alloys, fatigue
cracks usually nucleate from interdendritic shrinkage pores at or close
to the specimen surface [5,7] and so cast pores constitute the main influ-
ence on fatigue properties. S-N fatigue curves are remarkably insensitive
to the heat-treatment condition (either solution treatment or ageing
treatment) in contrast to the behaviour of wrought alloys in which the
fatigue resistance is usually closely related to the yield stress or the ten-
sile strength [7]. The effect of hot isostatic pressing on the mechanical
properties of aluminium alloys, especially fatigue behaviour, is not well
understood.
In this work A356 T6 heat-treated specimens were drawn from a suspen-
sion arm, that was manufactured with two different low pressure foundry
processes, permanent mould and Disamatic, and subjected or not to liq-
uid hot isostatic pressing in a salt bath at 120 MPa and 540 C for 30 s.
Then high-cycle axial fatigue tests were per-
formed on smooth plane samples at room tem-
perature with a tension compression ratio of 1,
in order to obtain fatigue S-N curves in the dif-
ferent conditions.
After LHIP i t was found that there i s a
large increase of fatigue resistance and a remark-
able decrease of the scatter of fatigue data, which
is due to the presence of internal discontinuities.
Optical microscopy and scanning electron
microscopy (SEM) were used to document the
elimination of porosity after LHIP and to show
the different nucleation sites of fatigue cracks on
the fracture surfaces in no-LHIP and LHIP condi-
tions.
&=&nM&~!L =ncC&cDn&
To understand the effect of LHIP on microstructure and fatigue behav-
iour of an Al based automotive part, A356 suspension arms were pro-
duced by through two different low pressure casting techniques, perma-
nent mould and Disamatic process. In the figure 3 the automotive com-
ponent is shown.
The main features required to a suspension arms are light weight and
stiffness, fatigue resistance and elongation, yield and tensile strength. Such
mechanical properties could be obtained by means of premium-quality Al
alloys and tightly controlled process.
Aluminium-silicon alloys combine the casting advantages of high corrosion
resistance, good weldability, and low specific gravity; aluminium-silicon al-
loys hardened by Mg
2
Si have excellent casting characteristics, weldability,
pressure tightness and corrosion resistance. These alloys are heat treat-
Fig. 3: Two different views of the suspension arm (not in scale)
able to provide various combinations of tensile
and physical properties that are attractive for
many applications, including aircraft and automo-
tive parts. Among the aluminium-silicon magne-
sium-silicide group, A356 (Al-6.5-7.5Si0.2Fe0.20-
0.45Mg0.10Zn0.10Mn0.20Cu0.20Ti, [10]) is capa-
ble of much higher ductility than its lower purity
counterpart, 356. A356 is among the premium-
quality sand and permanent mould casting alloys
specified for military and aircraft applications and
for safety automotive components.
As the casting processes concern the rapid so-
lidification rates associated with low-pressure
casting result in castings with finer grain size,
smaller secondary dendrite arm spacings, and
enhanced mechanical properties, such as strength
and ductility. The main advantage related to
Disamatic process is the high productivity rate.
The cooling property of the sand mould materi-
als doesnt enable to obtain very small dendrite
arm spacings compared to other processes in
chilled moulds.
The suspension arms were T6 heat-treated and
were partially subjected to LHIP process, accord-
ing to the following cycle:
Solubilization: 8h at 540 C
LHIP
Quenching: water at 60 C
Artificial ageing: 6h at 160 C
The LHIP conditions are shown in the following
figures.
Then the four categories of low pressure cast-
ings (permanent moulding T6, permanent mould-
ing T6 and LHIPped, Disamatic T6, Disamatic T6
and LHIPped) were cut into pieces for the
smooth plane fatigue specimens; the specimen
geometry used in axial fatigue testing was from
ASTM E466-72 [11]. Then high-cycle axial fatigue
tests were performed on smooth plane samples
at room temperature with a tension compres-
sion ratio of 1, in order to obtain fatigue S-N
curves in the different conditions.
Optical microscopy and scanning electron
microscopy (SEM) were used to document the
elimination of porosity after LHIP and to show
the different nucleation sites of fatigue cracks on
the fracture surfaces in no-LHIP and LHIP condi-
tions.
Fig. 4: Scheme of an element of material, which is subjected to
LHIP in a melted salt bath
Fig. 5: Pressure in function of time during LHIP process
n&eDL!e ~c
ceCDeec~
The first result of LHIP process is the increase of
the density of the treated parts, by means of the
elimination of internal porosity. In fact radio-
graphic inspections have shown that every auto-
motive component subjected to LHIP is charac-
terized by a 0-1 ASTM grade, defining the level of
porosity. Instead not treated parts have a 1-2
ASTM grade.
Such resul t has been confi rmed by a
metallographic analysis of the sections extracted
from the suspension arms in the different condi-
tions, as it results clear in the following micro-
graphs.
The not LHIP treated parts are characterized by
two different kind of porosity: a more irregular
one, that is interdendritic porosity, which is a
consequence of the solidification from the liquid
state and a rounded one, that is gas porosity. The
samples from parts obtained by Disamatic proc-
ess show a significantly bigger porosity in respect
of the samples produced with permanent mould
process, which also shows a finer microstructure.
This is the typical microstructure of A356 T6 and
micrographs show dendrites as the primary struc-
ture, wi th a eutecti c mi xture fi l l i ng the
interdendritic spaces. The silicon particles are dis-
tributed fairly uniformly in the microstructures
observed before LHIP and they have a fine, spheri-
cal shape because of the modifying addition to
the melt.
Fig. 6: The microstructure of A356 in the different casting conditions:
permanent mould + T6 (a) and (b), permanent mould + T6 + LHIP (c) and (d),
Disamatic + T6 (e) and (f), Disamatic +T6 + LHIP (g) and (h).
(a) (b)
(c) (d)
(e) (f)
(g) (h)
After the LHIP treatment there is a nearly complete closure of the inter-
nal pores, while the microstructure has not significantly changed, even if
the micrographs show a little increase of the concentration of silicon
particles in some zones. Anyway during LHIP the temperatures required
for densification are low and the times at elevated temperature short. So
the possible occurrence of secondary effect such as grain growth changes
in precipitate distributions and changes in segregation pattern can be pre-
vented. LHIPping pressures can also crack those brittle particles associ-
ated with porosity in ductile matrix, i.e. intermetallic compounds in Al
alloys. Such effects underline the need for careful control of pressure
and temperature ramp rates.
However the major effect of LHIP is the removal of porosity. The driving
force for pore closure is the reduction in the surface area associated
with pores. The driving force for the removal of small pores is greater
than that for the removal of larger pores [3]. If the pores contain gas, the
internal pressure of the gas will tend to oppose the driving force for
shrinkage. In LHIP the external applied pressure adds to and completely
swamps this driving force and almost inexorably causes any gases in a pore
to dissolve into the matrix. The solubility of such gases increases with
increasing pressure within the pore, which increases during the early stages
of LHIP. Under pressure the gas diffuses to the surface, rather than to
another pore as in sintering. The pore than collapse. It is only when a pore
reduces to a diameter of perhaps 40 nm that the driving force due to
surface energy becomes comparable with that due to the externally ap-
plied pressure. Yield stress decrease for most metals with increasing tem-
perature. LHIP conditions are generally chosen
so that the salt bath pressure is greater than the
reduced yield point of the material at that tem-
perature. Plastic flow can then occur on a micro-
scopic scale; under LHIP conditions creep proc-
esses such as Nabarro-Herring creep (diffusion
through grain interiors), Coble creep (diffusion
around grain boundaries) and dislocation creep
operate at relatively high rates. Finally the joining
of the collapsed surface occurs through a diffu-
sion bonding mechanism [3].
Data showing the effect of LHIP on static me-
chanical properties is given in table 1 for the dif-
ferent casting conditions. The averaged values of
yield strength and ultimate tensile strength are
better for permanent mould samples than
Disamatic samples, because of a finer microstruc-
ture and a reduced porosity , as it is visible in the
previous micrographs. LHIP does not improve the
yield stress and ultimate tensile strength signifi-
cantly, but the decrease of the internal porosity
has an important effect on the tensile ductility
with a nearly 50 % improvement.
In the following figures the results of axial fa-
tigue tests, as Wohler curves, are shown for the
different casting conditions. These curves were
obtained first by estimating the fatigue limit (50
% level of probability of fracture) with Staircase
method (15 tests with a 10 MPa step and 10
7
cycles as reference number of cycles). Then other
fatigue proofs were performed at different lev-
els of tension, so it was possible to estimate the
characteristic straight line of the finite endurance
zone by means of a linear interpolation.
In the figure 7 the effect of LHIP is evident on
the fatigue curves for permanent mould and
Disamatic specimens. LHIP improves the S-N fa-
Cosling 0,2% eld Ullimole Ullimole
condilion 5lress (MPo) slrenghl (MPo) 5lroin (%)
Pornononl Mould T 250 305 7,7
Pornononl Mould T LHlP 23 3T0 TT,T
Disonolic T 2T2 25 3,
Disonolic T LHlP 2T8 270 ,
!eL& .
ncsLL+s c= +csiLc +cs+s (=vcn=ccc v=LLcs) i
+ic L= =c c+ L= ccci+ics
Fig. 7: S-N axial fatigue curves for A356 T6 in low pressure permanent mould (a) and Disamatic (b)condition
subjected or not subjected to LHIP treatment
tigue life of cast A356 T6. The three LHIP curves
(relative to the different probability of survival)
are shifted up and the 50% fatigue limit for the
specimens subjected to LHIP is 35% higher. Also
the 90% (probability of survival) curve of the
specimens subjected to LHIP is over the 10%
(probability of survival) curve of not LHIP treated
samples.
The standard deviation is much higher for not
LHIPped material, so there is an important de-
crease of fatigue scatter after LHIP treatment.
This scatter is explained by the repetition of rela-
tively small loads, which enables weak properties
of the materials to become dominant. Such prop-
erties include the local fluctuations of the com-
position; the different size and orientations of
grain; the distribution and sizes of slag inclusions
and of segregations; cavities, blisters, blow holes
and microstructural defects in general. So the
reduction in scatter is an experimental evidence
of the removal of porosity by LHIP. Finally the
slope of the straight line in finite endurance zone
is higher for LHIPped specimens; this means that
LHIP increases the reliability of the component
because a little variation of the applied stress gives
rise to a lower variation in the number of cycles
to failure in respect of not LHIP treated mate-
rial.
Fatigue crack initiation is very sensitive to the
void content of a material. LHIP helps to close
the voids present in the cast materials and, in this
way, enhances their fatigue properties. This as-
pect has been confirmed by SEM analysis of frac-
ture surfaces after fatigue tests.
Above all the fractographic observations have
shown the possible nucleation sites of fatigue
cracks for the specimens in the different casting
conditions. Fatigue failure for not LHIPped test
pi eces general l y tends to ori gi nate from
interdendritic shrinkage below, or at, the surface.
By contrast, when LHIPped test pieces eventu-
ally fail, they do so at large intermetallics, inclu-
sions, surface-connected porosity or internal
pores which could not be closed by LHIP, since
these are the largest defects present within the
structure. Thus to improve the properties of
LHIPped materials further, attention should be
focused on the refinement of cast microstruc-
ture and the removal of large inclusions.
Some scanni ng el ectron mi croscopy
fractographies of failed test pieces are shown in
the following figures. In figure typical fatigue-
crack nucleation sites are related for A356 T6
failed specimens, that were not subjected to LHIP.
The most significant observation is that in nearly
all samples the preferential nucleation site of fa-
Fig. 8: Fatigue-crack nucleation sites for: (a) permanent mould specimen,
tested at 80 MPa and failed at 3992807 cycles, (b) permanent mould specimen,
tested at 130 MPa and failed at 161764 cycles, (c) Disamatic specimen, tested at
80 MPa and failed at 4013477 cycles, (d) Disamatic specimen, tested at 80 MPa
and failed at 1151496 cycles
(a) (b)
(c) (d)
tigue cracks is shrinkage interdendritic porosity close or at the surface.
This is true for permanent mould failed specimens (figure 8a and 8b) and
also for Disamatic failed specimens (figure 8c and 8d). This kind of discon-
tinuity seems to be more deleterious than other defects, which are present
in the material, e.g. gas porosities or oxides. Reducing the size of shrink-
age defects will increase the fatigue life, but only up to the stage at which
initiation from persistent slip bands on the surface becomes operative
[7]. Other investigators [9,12] have observed that casting discontinuities
of size similar to the silicon eutectic particles in Al-Si alloys will not have
a detrimental effect on fatigue strength.
In the figure 9 initiation sites are shown for LHIPped samples, obtained by
permanent mould process (figure 9a and 9b) and Disamatic process (fig-
ure 9c and 9d). The fatigue crack initiated mainly at the specimen surface
since LHIP suppressed the voids in the material. Anyway also fatigue crack
nucleation from oxides (figure 9c) and from pores, which were not closed
by LHIP (figure 9a) are possible.
In the figure 10 the fatigue propagation and abruptly rupture zones are
shown for two failed samples. This kind of surface morphology has been
found for almost every sample, which was or was not subjected to LHIP
Fig. 9: Fatigue-crack nucleation sites for: (a) permanent mould LHIPped specimen,
tested at 120 MPa and failed at 1249953 cycles, (b) permanent mould LHIPped
specimen, tested at 140 MPa and failed at 780406 cycles, (c) Disamatic LHIPped
specimen, tested at 90 MPa and failed at 4710397 cycles, (d) Disamatic LHIPped
specimen, tested at 110 MPa and failed at 1205000 cycles
(a) (b)
(c) (d)
Fig. 10: Typical propagation (a) and final fracture (b) zones of failed specimens
(a) (b)
treatment. These zones are typical of the fracture surface of Al-Si alloys.
The fatigue fracture surface near the initiation site has a flat aspect, with
Cc~CLDec~e
The effect of liquid hot isostatic pressing (LHIP) on the fatigue properties
of Al based automotive parts (A356 suspension arms) was investigated.
The following conclusions may be drawn from this work:
1. LHIP process increases the density of the treated parts, by means of
the elimination of the internal porosity. The parts subjected to LHIP
have a 0-1 ASTM grade against a 1-2 ASTM grade of parts, which were
not LHIPped.
2. The removal of porosity has been confirmed by a metallographic analysis
for both low pressure permanent mould and Disamatic samples. This
occurs through plastic flow, creep and diffusion bonding mechanism.
3. The decrease of the internal discontinuities improves (50%) the ten-
sile ductility.
4. The S-N axial fatigue curves of LHIPped samples are shifted up, with a
30-35 % improvement of 50% fatigue limit, which was determined by
Staircase method.
5. The standard deviation is much higher for not
LHIPped material, so there is an important
decrease of fatigue scatter after LHIP treat-
ment. The better reliability of LHIP treated
component arises also from the higher slope
of the straight line in finite endurance zone.
6. Scanning electron microscopy has shown that
in nearly all samples, which were not subjected
to LHIP, the preferential nucleation site of fa-
tigue cracks is shrinkage interdendritic poros-
ity close or at the surface. For LHIPped sam-
ples the fatigue crack initiated mainly at the
specimen surface since LHIP suppressed the
voids in the material; also fatigue crack nu-
cleation from oxides and from pores, which
were not closed by LHIP, are possible.
C~cwL&co&M&~!e
The authors thank Mr Giunti at CRF of Turin for the statistical elaboration of fatigue data.
n&r&n&~C&e
1. J.M. ERIDON, Hot Isostatic Pressing of Casting, Foundry Equipment
and Processing Vol. 9 ASM Handbook, (1994), p. 538-544.
2. H.D. HANES and J.M. McFADDEN, HIPing of Castings: An Update,
Metal Progress, American Society for Metals (April 1983), p. 1-4.
3. H.V. ATKINSON and B.A. RICKINSON, Hot Isostatic Processing, Adam
Hilger, Bristol, Philadelphia and New York (1991), p. 10-17, p. 44-45, p.
50, p. 74-82.
4. C.S.C. LEI, W.E. FRAZIER and E.W. LEE, The Effect of Hot Isostatic
Pressing on Cast Aluminum, Journal of Material (November 1997), p.
38-39.
5. B. ZHANG, D.R. POIRIER and W. CHEN, Micostructural Effects on
High-Cycle Fatigue-Crack Initiation in A356.2 Casting Alloy, Metallur-
gical and Materials Transactions A, Vol. 30A (October 1999), p. 2659-
2666.
6. M.E. SENIW, M.E. FINE, E.Y. CHEN, M. MESHII and J. GRAY, High Cy-
cle Fatigue of Structural Materials (in Honour of Professor Paul C.
Paris), The Minerals, Metals and Materials Society, Indianapolis (1997),
p. 371-379.
7. M. J. COUPER, A. E. NEESON and J. R.
GRIFFITHS, Fatigue Fract. Engng. Struct., Vol.
13, No. 3 (1990), p. 213-227.
8. S.E. STANZL-TSCHEGG, H.R. MAYER, A.
BESTE and S. KROLL, Int. J. Fatigue, Vol. 17,
No. 2 (1995), p. 149-155.
9. T.L. REINHART, Fatigue and Fracture Prop-
erties of Aluminum Alloy Castings, Fatigue
and Fracture, Vol. 19 ASM Handbook (1996),
p. 814-818.
10. A. KEARNEY and E.L. ROOY, Aluminum Foun-
dry Products, Vol. 2 ASM Handbook, p. 146
11. Standard Practice for Conducting Constant
Amplitude Axial Fatigue Tests of Metallic
Materials, ASTM E466-72, Annual Book of
ASTM Standards, ASTM, Philadelphia (1996),
p.471-474.
12. F.T. LEE, J.F. MAJOR and F.H. SAMUEL, Fatigue
Fract. Engng. Struct., Vol. 18, No. 3 (1995), p.
385-396.
clear signs of the propagation of the crack under
the action of an alternate load (see fatigue
striations in figure 10a). Instead ductile dimples
(figure 10b) cover the final fracture zone. Even if
the final fracture occurred abruptly, the process
is dominated by ductile rupture of the soft alu-
minium matrix involving the pull out of silicon
particles. No evidence of particle cracking was
found. This is in agreement with the work of Lee
[12], which pointed that the propensity of parti-
cle cracking increases as the K increase, that is
low-cycle fatigue, whereas particle decohesion and
fatigue striation formation increase as the K lev-
els decrease, that is high-cycle fatigue.
Mc+=LLLncic=L
eciccc =o 1ccicLcc-
JcLn=L =LeLisico
e- 1csio
+wicc = -c=n
=nc=&n1&e M=ncv&M&~1 cr
1crcnM&c =n1e ev LUc c1
ece11c =n&ee~c =ncc&ee
*H. kerre, *. kemane - **I. 8arene
*HaterIaI ItIente 0epartment - eIItetnIte dI 1erIne - ltaIy
**1ekrId I.p.A. - 8erarette (1erIne} - ltaIy
~!nccDC!c~
The demand of light weight and near net shape
components from the automotive industry to
satisfy the emission legislation is stronger and
stronger. In this contest the aluminum alloys cast-
ings met success and the shaping processes were
continuously improved. Sand, as well as gravity
and low pressure permanent molds are widely
used to cast aluminum automotive components,
especially structural parts. When a mass produc-
tion of light alloys products is required conven-
tional high pressure die casting is unquestionably
the most often selected process, being it rela-
=nc=&n!&e M=ncv&M&~! cr
!crcnM&c =n!e ev LDc c!
ece!!C =n&ee~o =ncC&ee
*H. kerre, *. kemane - **I. 8arene
*HaterIaI ItIente 0epartment - eIItetnIte dI 1erIne - ltaIy
**1ekrId I.p.A. - 8erarette (1erIne} - ltaIy
Abstract
Introduced almost a century ago as a concept the hot
isostatic pressing process has been widely and successfully
used to treat aluminum casting for niche markets.
The reason why this post-processing idea is not applied
to mass production is costs. In fact the media used to
apply pressure on components to be treated in a vessel
at high temperature,where the material softens and
compacts uniformely, is a gas. Long treatment times added
to highly risky plants are adding costs to the processing
route.
Recently the partnership between Teksid and Idra is on
the way to industrialize a new hot isostatic process, called
Liquid Hot Isostatic Pressing LHIP where the media
used to apply pressure on the components to be treated
is a liquid salt.
With this technology it is reasonable to think that mass
production post-treatment of automotive aluminum
castings will be achieved, with the goal of having stronger
parts with higher fatigue properties and reduced scrap
level in the foundry process production chain.
This paper describes the effect of LHIP on aluminum
thixoformed castings, outlining the beneficial process
effects on the mechanical properties.
Riassunto
Il concetto di pressatura isostatica a caldo fu introdotto circa un secolo fa ed applicato
con successo per trattare getti in lega di alluminio per mercati di nicchia.
Tra le ragioni principali che non permettono allindustria di utilizzare la tecnologia per
alti volumi produttivi vi principalmente il costo. Infatti il processo prevede di caricare
i getti da trattare in autoclave dove un gas viene portato a pressione e temperatura
elevata, tanto elevata da rendere il materiale facilmente deformabile e in grado di
compattarsi in modo uniforme. La produttivit e capienza dei forni di trattamento
limitata e la loro gestione molto delicata visti i possibili rischi di esplosione.
Recentemente Teksid e Idra Presse hanno avviato uno sviluppo congiunto mirato ad
industrializzare un processo innovativo di pressatura isostatica a caldo, chiamato Liquid
Hot Isostatic Pressing, LHIP, in cui lelemento adottato per applicare la pressione sui
getti in alluminio da trattare un liquido, una miscela di sali, anzich un gas. E ragionevole
pensare che con questa tecnologia sia possibile affrontare produzioni di elevati volumi
quali quelle dei componenti di sicurezza dellauto, cui sono richieste elevate
caratteristiche meccaniche.
Larticolo illustra i risultati ottenuti su componenti formati con tecnologia thixo e
successivamente sottoposti a LHIP, evidenziando, in particolare, leffetto positivo sulla
resistenza a fatica.
tively inexpensive and suitable for a broad variety of components, how-
ever it is not capable of the same high product integrity and reliability
achievable through concurrent processes. Among the traditiona1 shaping
technologies forging generally provides the highest properties available in
formed aluminum products and is therefore used for many aerospace
and automotive structural parts. The demand of light, reliable and safe
components together with the continuous increase of competitiveness
worked also as an intense driving force to the development of new tech-
nologies: squeeze casting, thixoforming, thixomolding and rheocasting proc-
esses came in age [1] and for different reasons thixoforming or Semi-
solid forming is probably the most promising process. It is an effective
near net shape forming process in which the metal is formed in the semi-
solid state [2], Prof. Merton Flemings was the inventor of the process in
the late 60s, early 70s.
!crcnM~o =ncC&ee
The thixoforming process is capable of the di-
mensions, details and thin wall of conventional
high pressure die castings while accomplishing the
high integrity generally ascribable only to the high-
est quality gravity and 1ow pressure permanent mold castings. With re-
spect to squeeze casting process, a new competing technology able to
assure high integrity parts having relatively thick sections, it is able to
assure the same level of integrity but also in thinner cast sections. Squeeze
casting works at higher temperatures, the molten metal slowly fills the
die cavity and solidifies under high pressure ( about 100 bars) which avoids
Fig. 1: Microstructure characteristics of a feedstock A356 aluminum alloy
for thixoforming (left) and of the same alloy after thixoforming
Fig. 2: Process flow steps
in Thixoforming
many defects usually noted for die casting, while the Semi-solid forming
works at lower injection temperatures with high solid &action, 0.6 to 0.8
of the material feedstock and the cycle time requirements are strongly
reduced. The main need to realize a semi-solid process is to obtain a solid
plus liquid slurry with consistent theological properties, followed by a
simple transfer of the slurry to fill the die cavity. The ability consists in the
suitable control of the rheological properties of the slurry, immense cost
benefits are available by such a processing route.
The process requires a suitable raw material or feedstock, usually high
purity alloy with a microstructure composed of clustered degenerate-
dendritic particles in a matrix of second (eutectic) phase. To meet these
features stirring methods are used during the cooling of the primary alloy
to broken the dendritic structure and to obtain a globular phase, as
shown in figure 1 for the A356 alloy. These features causes thixotropic
properties, meaning that if a slug of the material is heated into the two
phase range, upon the application of shear, the slug became fluid.
The process flow steps are depicted in figure 2. The feedstock billet is
heated between solidus and liquidus, injected into the die cavity and then
the part is stripped. During the heating process the billet must retain its
Fig. 3: Process flow steps in
Rheocasting, F means feedstock billet
for Thixoforming
FEEDSTOCK BILLET
HEATING BETWEEN
SOLIDUS AND
LIQUIDUS
INJECTION INTO
THE DIE
NEAR NET
SHAPE PARTS
COOLING AND
STIRRING THE ALLOY
INJECTION OF SEMI-
SOLID ALLOY INTO
THE
NEAR NET
SHAPE PARTS
F
shape. Since the semi-solid is composed of solid
particles with the remainder liquid, the particle
size and shape and the fraction liquid are the con-
trolling parameters [3]. The smaller and more
rounded the particles and the larger the liquid
fraction the lower the viscosity, the same param-
eters are also important in the control of the
fluidity together with the benefits of a high la-
tent heat of fusion of the liquid. A definite advan-
tage of the process is that both the alloy solidifi-
cation range and the shrinkage can be controlled
by limiting the percent liquid in the semi-solid.
The advantages of Semi-solid forming [4] are re-
lated to:
low porosity (< 0.1%) with product of ex-
ceptional soundness,
good combination of strength and ductility,
product complexity, close dimensional toler-
ances, near net shape, thin walls and excellent
surface finish,
low processing temperature and therefore
short cycles times and low stress on tooling,
ability to utilize unusua1 alloys that normally
prove difficult in liquid casting processes
(wrought compositions, for instance)
an ability to undergo a T-5 heat treatment
without loosing ductility, thus often achieving
required mechanical properties without the
dangers of blistering, distortion or quench
stress associated with the full T-6 heat treat-
ment,
the ability to form hypereutectic Al-Si alloys
and MMCs while retaining the desirable size
and distribution of primary or second phases
developed during billet manufacturing under
ideal continuous casting conditions (those
phases dont re-melt or dissolve during Semi-
solid forming),
the product consistency that results from us-
ing the pre-cast billet that was manufactured
under the same ideal continuous processing
conditions that are employed to make forg-
ing or rolling stock.
A drawback is actually contrasting the lot of ad-
vantages, the cost of the pre-cast billet. In fact, it
costs about as much as 3540 % over the cost of
similar primary alloy purchased as foundry ingot
for melting. A solution to this problem can came
from the rheocasting process, in which the in-
volved steps concern the melting of the alloy
coupled with stirring effect during the solidifica-
tion. From this point of view the process may be
the same used to produce the feedstock billet
for the thixoforming process. However at that
Fig. 4: Two different views of the suspension arm (not in scale)
point it is possible try to introduce the semi-solid
alloy into the die cavity obtaining directly near
net shape parts, as indicated by the flow diagram
in figure 3.
Recently a new rheocasting process NRC [5]
has been developed, in this process the slightly
superheated melt alloy is taken from a holding
furnace and poured into specially designed steel
crucibles. In the crucible n-grains grow, fo11owing
forced homogeneous nucleation at the cup wall
and in few minutes a stable skeleton of -phase
forms and the alloy reaches the semi-solid state
producing a cylindrical shaped slug. At this point a
high &equency induction heating is used to equalize
the slug temperature before its casting into shape.
This new process can work with conventional alloy
melts, this means that foundries can use a wide
range of alloys and apply their standard melting
and recycling techniques, meeting a really low cost
Semi-solid casting. Taking into account for the
technical evolution, the Semi-solid forming proc-
ess now appears to be very interesting both for
&=&nM&~!L =n!
Sets of A356 alloy thixoformed suspension arm
produced by Stampal Italy for TRW (Figure 4)
were considered for the LHIP process. The parts
have been treated in the TEKSID LHIP pilot plant.
The castings were scrap parts out of normal pro-
duction to have a result on the effect of LHIP on
casting defects and an evaluation of achievable
static mechanical properties. Successively parts
under different heat treatment conditions have
been LHIP treated to study the influence of the
process on the mechanical, as well as on the fa-
tigue properties, each batch was constituted at
least by 15 parts. The different heat treatment
conditions were:
A: T6 treated
B: solution treatment (2 h at 540 C) followed
by LHIP and then aged 3 hours at 160 C
C: solution treatment (4 h at 540 C) followed
by LHIP and then aged 3 hours at 160 C
D: T6 treated, LHIP and then T6 treated again,
always with the same process parameters.
The condition A clearly refers to parts only heat
treated but not subjected to LHIP in order to
have a suitable comparison among the induced
effects. The scrap parts were subjected to X ray
radiography to detect the defect before and af-
ter the LHIP treatment. Tensile test specimens
were obtained by dissections of LHIP treated
castings, two samples each part were taken in
the cost competitiveness and for the high integrity of the produced parts.
It can compete with squeeze casting, as well as with the best permanent
mold castings. It is well known that the properties of aluminum alloy cast
parts, specially when produced by gravity casting, can be improved by ap-
plying HIP [6, 7], mainly for the reduction of defects and particularly of
the porosity, moreover the recent development of the most economic
LHP-Liquid hot isostatic process [8] increase the interest of this tech-
nique. This process is of course helpful also to eliminate some defects in
the thixoformed parts [9]. As the HIP process LHIP is capable of eliminat-
ing some of the typical casting defects like micro and macro shrinkage
porosity and hydrogen inclusion. Defects connected with the surface (i.e.
cold shots, surface cracks), as well as nitrogen inclusions and oxides, can-
not be eliminated. Process normal running parameters to obtain such
results on A356 aluminum castings are:
1000-1200 ATM pressure
500-540C salt temperature
20-35 seconds pressure applied
3 - 4 minutes total cycle time (including heating and cooling)
The mechanical properties attained by the heat treated thixoformed parts
are very satisfying and it is difficult still improve them, however there is
the impression that the LHIP process may be helpful to improve the fa-
tigue properties, in this paper the first results attained on A356 alloy are
presented and discussed.
two different directions, longitudinal and transversal, their average di-
mensions being: total length 87 mm and thickness 3.15 mm, the calibrated
zone was 30 mm length and 8 mm wide. On these samples the tensile
properties, yield strength, ultimate tensile strength and elongation have
been measured. On the samples type D, axial fatigue tests have been per-
formed using a Rumul resonant machine, type TESTROTONIC 100 kN,
working in the Sequence range between 40 and 260 Hz. An alternate
symmetric cycle, R =
min
/
max
, = -1, was adopted. The fatigue limit has
&=&nM&~!L n&eDL!e ~c
ceCDeec~
The main results of the radiography on different casting having cold shots
as defects are summarized in Table 1. Other scraps were caused by gas
5crop As cosl Afler LHlP
porl ZONE kX ZONE kX
A B C A B C
N. T Cold shol ASTM T ASTM T Scro Snoll cold shol --- --- Good
N. 2 --- ASTM 2 ASTM T Scro --- --- --- Good
N. 3 --- ASTM 2 ASTM T Scro --- --- --- Good
!eL& .
!ic c==cc+ c= L= c +iixc=cnmcc c=s+ic cL=Li+
Fig. 5: Radiography of a scrap part, showing the presence of a gas porosity (left
picture) before of LHIP treatment and
the sound part after the execution of the process
porosity, after the LHIP process the pores disap-
peared as can see in figure 5. Also the shrinkage
porosity was eliminated, confirming the capabil-
ity of the process to eliminate most of the de-
fects not connected with the surface without af-
5crop As cosl Afler LHlP
porl Properlies [MPo] Properlies [MPo]
Zone ield k
m
A% ield k
m
A%
N. T 8 27 304 .7 220 30 TT.4
C 23 3TT T0 233 30T .4
N. 2 8 235 288 4.7 23 2 5.7
C 230 2T 8 232 20 5.T
N. 3 8 254 3T0 .7 220 20 .0
C 232 27T 4.7 2T7 285 .7
)8-4)/- " '$ $& $ '" %&
!eL& z.
vicLc s+ncci+, LL+im=+c +csiLc s+ncci+ =c
cLcc=+ic c= +iixc=cnmcc c=s+ics ec=cnc
=c ==+cn +ic L= +nc=+mc+
been obtained using the Stair Case method and the Wohler curve for the
10%, 50% and 90% survival limits has been drawn and compared with
literature data referred to the same alloy. Optical microscopy and SEM
analysis were used to observe the microstruc-
ture features of the alloy and the morphology of
the surface fracture of the samples.
fecting the microstructure of the alloy as was
observed on the polished sections of different
samples by means of optical microscopy.
The effect on mechanical properties evaluated
on samples obtained by dissections of the scraped
parts in the as cast state or aAer the LHIP proc-
ess is shown in Table 2, where the only affected
value on the static properties is elongation, with
a gain of about 15% over the as cast material.
Zones B and C in this case refers to the direc-
tion of dissection of castings, mainly longitudinal
or transversal, they demonstrate no influence of
the direction.
The properties measured on sound parts after
different heat treatment conditions are summa-
rized in Table 3, the reported values are the aver-
age of at least five tested specimens each consid-
ered direction, longitudinal and transversal. Also
in this case no influence of the direction of the
dissection sample was observed.
Table 2: average mechanical properties, yield
strength (YS), ultimate tensile strength (UTS) with
related Standard Deviation (o) and elongation of
parts after different heat treatment conditions.
The influence of LHIP on the strength is not well
defined, even if slightly lower values were meas-
ured for the samples not treated with usual T6
cycle, probably this is due to the low tempera-
ture value and the short time adopted for the
solution treatment. However a great influence
Fig. 6: Gauss distribution of the ultimate tensile strenght
on the elongation properties is evident, with an
increase of this property close to about 40 %
measured on the samples type C, that is solution
treated 4 hours at 520 C, LHIP processed and
aged 3 hours at 160 C. This sequence of treat-
ments is also one of the cheapest because the
expensive T6 treatment is practically included in
the LHIP process. The large number of tests al-
lowed also a statistical approach and in Table 3the
standard deviation of the tensile strength prop-
erties are also shown, it is evident that the LHIP
process reduced the standard deviation of yield
strength and of ultimate tensile strength. Moreo-
ver, plotting the dispersion of these properties a
Gauss curve was obtained (figure 6), showing that
the LHIP met the goal to reduce the dispersion
of mechanical characteristics. Being this a func-
tion of the defect level of the casting, the posi-
tive effect of the treatment is confirmed.
The observation of the fracture surface by SEM
shows for the samples without defects mainly a
dimple morphology, indicating the toughness char-
acteristics caused by the performed processes,
figure 7.
The fatigue tests were performed on the sam-
ples type D, that is T6 treated before and after
the LHIP process, about 30 samples were used
to obtain the Wohler curve drawn in figure 8,
the fatigue limit was about 95 MPa. It is quite
interesting to make the comparison (figure 9)
between the obtained Wohler curve and the
curves of A356 alloy components obtained in dif-
ferent ways, that is:
Permanent mold and T6 treated
Permanent mold + LHIP and T6 treated
Thixoformed and T6 treated
As it was expected the lowest fatigue resistance
is presented by the permanent mold castings T6
treated, while fatigue behavior of the thixoformed
and T6 treated parts are higher. The execution
of the LHIP on these castings causes a great im-
provement of fatigue resistance especially in the
case of the first type of castings, this is a proof of
the effect of the process and of its advantages.
Concerning the thixoformed parts the fatigue
improvement after LHIP is not so high as it was
expected, probably this is due to two different
concomitant effects. Test samples used for the
fatigue tests in the case of the T6+LHIP+T6
thixoformed castings were different with respect
to the other ones and not conforming the ASTM
Standards, this can be influent on the obtained
resuls. The second aspect is mainly theoretical and
concerns the relation between the LHIP treat-
Fig. 7: Fracture morphology of specimens after treatments:
520 C, 4h+LHIP+ageing 3h at 160 C
Condilion 5
5
UI5
UI5
E
[MPo] [MPo] [%]
A. T 224.3 T2.87 25.T T2 .4
8. 520 C,
2 h LHlP ogoing. T.T 8.7 273.3 8.25 T2.5
C. 520 C,
2 h LHlP ogoing. 2T7. .38 287. 7. T3.T.
D. T LHlP T 230.0 8.T 25.5 8.4 T0.4
!eL& e.
vcn=cc mcci=ic=L =nc=cn+ics, icLc
s+ncci+ (ve), LL+im=+c +csiLc s+ncci+ (D!e)
wi+i ncL=+cc e+=c=nc ccvi=+ic ( ) =c
cLcc=+ic c= ==n+s ==+cn ci==cnc+ ic=+
+nc=+mc+ ccci+ics
n&r&n&~C&e
1. D. Apelian, A roadmap for Semi-Solid processing, Proceedings of
the 6
th
Int. Conf. Semi-Solid Processing of Alloys and Composites
Turin, September 27
th
-29
th
, 2000, EDIMET, 2000, p.
2. M. Flemings, R. Riek et al., AFS Int. Cast Metals Journal, Sept. 1976, p.
11-21.
3. Winterbottom, Proceedings of the 6
th
Int. Conf. Semi-Solid Process-
ing of Alloys and Composites Turin, September 27
th
-29
th
, 2000, EDIMET,
2000, p.
4. J.L. Jorstad, Proceedings of the 6
th
Int. Conf. Semi-Solid Processing of
Alloys and Composites Turin, September 27
th
-29
th
, 2000, EDIMET,
2000, p.
5. H. Kaufmann, H. Wabusseg, P.J. Uggowitzer, Proceedings of the 6
th
Int.
Conf. Semi-Solid Processing of Alloys and Composites Turin, Sep-
ment and the defect level of the casting. It seems
that LHIP is more effective on parts with an
higher defect level. In fact, the thixoformed parts
are characterized by lower number of pores with
smaller size, smaller the pore higher the surface
specific energy, consequently the pressure re-
quired to close a small pore will be higher than in
the case of a large pore, this can explain a more
important effect of LHIP on permanent mold
castings rather than on thixoformed parts.
Fig. 8: Wohler curve for the T6+LHIP+T6 trated castings
Fig. 9: Comparison between Wohler curves at 50% survival probality
of different A356 castings
Cc~CLDec~e
Liquid hot isostatic pressing is a new post treat-
ment process for aluminum casting that can im-
prove the soundness of the parts and increase
the material mechanical properties at competi-
tive costs. It can be advantageously applied also
to thixoformed parts, in particular the treatment
is able to:
allow to recuperate scraps affected by the
presence of gas (not nitrogen) and shrinkage
porosity
improve the ductility properties, without
modifying the microstructure characteristics
improve the fatigue resistance
tember 27
th
-29
th
, 2000, EDIMET, 2000, p.
6. G.E.Wasiliewski, N.R. Lindblad, Elimination
of casting defects using HIP Proceedings of
the 2
nd
International Conference, MCIC
Sept.1972
7. Hot Isostatic Pressing of A356- SAE Paper
2000-01-0062
8. S. Gallo, C. Mus Current quality needs for
casting in the automotive Merton Flemings
Symposium Boston, July 2000.
9. M. Rosso, C. Mus, G.L. Chiarmetta, Proceed-
ings of the 6
th
Int. Conf. Semi-Solid Process-
ing of Alloys and Composites Turin, Septem-
ber 27
th
-29
th
, 2000, EDIMET.
~e1nUc1c~e rcn U1cne
To ensure that space can be provided for the publication of several papers in
each issue of the Metallurgical Science and Technology, Authors are kindly
requested, to limit their contributions to a maximum of fifteen typed pages
prepared in accordance with the following instructions:
1. Texts must be submitted on an MS-DOS or Word for Windows compatible
floppy disc in the following format: double line space, and divided into
paragraphs; each page should contain 30-35 lines, each composed of 50-60
characters and spaces.
Authors are also asked to supply a reserve copy on A4 paper in the same
format to cover the possibility that their disc may not be legible.
2. Each paragraph should be indented three spaces.
3. Notes must be numbered and typed on separate sheets.
4. All measurements must be expressed in ISO 1000-81 units.
5. Please set out the following details clearly at the end of the typescript: first
name and surname, university department or firm where employed, full
address and telephone number. This information will help the editorial staff
to get in touch with you if necessary.
6. A short summary should be provided (not more than 150 words). This
should give an account of the scope of the paper and summarise its
conclusions.
7. Tables, graphs, and figures must be set out on separate sheets. Their posi-
tion should be indicated in the text in the form table 1, fig. 3, etc. If data
collected by other persons are used in their preparation, acknowledge-
ment of the source should be made in a note at the foot of the table or
figure concerned. The titles of tables, graphs, figures, etc. must provide a
specific indication of their contents. All photographs should be supplied in
their original form.
8. Bibliographical references should be set out on a separate sheet or sheets
in the order in which they appear in the text. The following rules must be
followed;
a) Authors names: surname of the first author, comma, initials, comma, initials
and surnames of remaining authors; and between the final pair of authors;
finished with full point.
b) Article title: initial caps and lower case (except in German, where nouns
are always capitalized); finished with full point.
c) Journal name: underlined in the typescript; abbreviations may be used; ini-
tial caps; comma; issue number underlined followed by year in parentheses,
comma; page number (without p), final full point.
For book titles the style is similar as follows:
a) Authors names: as for journals.
b) Chapter title: as for article title.
c) Book title: In followed by initials, surname of editor and (Ed.), comma
and then title underlined, comma.
d) Publisher, comma, location, comma, year, comma, volume, part, etc., comma.
e) Page number preceded by p, or pp. and followed by full point.
Examples:
Slatcher, S., and J.F. Knott. The ductile fracture of high strength steels. In D.
Francois (Ed.), Advances in fracture mechanics, Pergamon Press, Oxford,
1981, vol. I, pp. 201-207.
Griffiths, J.R. and D.R.J. Owen. An elastic-plastic analysis for a notched bar in
plane strain bending. J. Mech. Phys. Solids, 19 (1971), 419-431.
The authors are strongly recommended to submit their papers on a MS-
DOS compatible floppy disc prepared using the Word for Windows
programs, besides the written version.
Ihe leoding monufoclurer of melol componenls
Teksid S.p.A. Viu Piunezzu, 123 - 10151 Torino (Iuly} Tel. +3.011.854.111

Metallurgical Science and Technology, Vol. 19/1

Caricato da

Informazioni sul documento

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Metallurgical Science and Technology, Vol. 19/1

Caricato da

Copyright:

Formati disponibili

METALLURGCAL

Potrebbero piacerti anche