Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
HYDRIDES
HYDRIDES
PATRICK C. DAM
EDITOR
Copyright © 2018 by Nova Science Publishers, Inc.
All rights reserved. No part of this book may be reproduced, stored in a retrieval system or transmitted
in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical photocopying,
recording or otherwise without the written permission of the Publisher.
We have partnered with Copyright Clearance Center to make it easy for you to obtain permissions to
reuse content from this publication. Simply navigate to this publication’s page on Nova’s website and
locate the “Get Permission” button below the title description. This button is linked directly to the title’s
permission page on copyright.com. Alternatively, you can visit copyright.com and search by title, ISBN,
or ISSN.
For further questions about using the service on copyright.com, please contact:
Copyright Clearance Center
Phone: +1-(978) 750-8400 Fax: +1-(978) 750-4470 E-mail: info@copyright.com.
Independent verification should be sought for any data, advice or recommendations contained in this
book. In addition, no responsibility is assumed by the publisher for any injury and/or damage to persons
or property arising from any methods, products, instructions, ideas or otherwise contained in this
publication.
This publication is designed to provide accurate and authoritative information with regard to the subject
matter covered herein. It is sold with the clear understanding that the Publisher is not engaged in
rendering legal or any other professional services. If legal or any other expert assistance is required, the
services of a competent person should be sought. FROM A DECLARATION OF PARTICIPANTS
JOINTLY ADOPTED BY A COMMITTEE OF THE AMERICAN BAR ASSOCIATION AND A
COMMITTEE OF PUBLISHERS.
Additional color graphics may be available in the e-book version of this book.
Preface vii
Chapter 1 Crystal Structures of Alkali and Alkaline Earth
Tetra Borohydrides 1
P. Vajeeston
Chapter 2 Bonding Nature and Its Implications with the
Energy Storage and Ionic Transport in Hydrides 81
P. Vajeeston and H. Fjellvåg
Chapter 3 The Application of Metal Hydride Based on
Zr-V Alloy in Hydrogen Plasma 149
Ihor Sereda, Alexandr Tseluyko
and Nikolay Azarenkov
Chapter 4 Preparation and Application of Metal Hydrides 193
N. E. Galushkin, N. N. Yazvinskaya
and D. N. Galushkin
Index 231
PREFACE
the plasma source based on Penning trap with metal hydride cathode appears
to generate current-compensated ion beams with the ability to control the
energy of the extracted ions. Lastly, the book discusses metal hydrides
obtaining in sintered electrodes of nickel-cadmium batteries with
electrochemical methods by the way of electrolyte decomposition onto
hydrogen and oxygen. It was shown that as a result of electrolyte
decomposition, oxygen releases from batteries, while hydrogen partly
releases and is partly (in virtue of its high diffusion permeability)
accumulated in sintered matrices of electrodes of nickel-cadmium batteries
in the metal hydrides form.
Chapter 1 - Hydrogen gas is particularly relevant to energy applications
as it is a naturally abundant source of clean energy, with a high density per
mass unit. Moreover, it is the lightest fuel and, unlike electricity, it can be
easily stored. Hydrogen gas is the most promising fuel for a large number of
applications, and it is already in use for space programs around the world.
Metal hydrides are a fascinating class of compounds because the small mass
and size of hydrogen and its medium electronegativity causes a large
flexibility in terms of metal-ligand interactions, resulting in a vast variety of
possible compositions, chemical bonding, crystal structures and physical
properties. Metal borohydrides rank among the most hydrogen-rich groups
of compounds (e.g., LiBH4 has an extreme gravimetric hydrogen content of
18.5 wt %) and attract attention as potential chemical hydrogen stores.
However, the borohydrides explored so far reveal serious drawbacks, e.g.,
either too high or too low temperature of hydrogen release, difficult
reversibility and contamination of evolved hydrogen with B2H6, which
obstruct their application but also triggers the research of new systems.
Consequently, various novel borohydride-based materials have been
prepared and investigated. Numerous unsolved problems remain on our way
towards a sustainable, carbon free energy system based on renewable energy
and on hydrogen as a future energy carrier. Novel materials form the back-
bone in most emerging energy related technologies. In this chapter, the
authors presented the structural details of alkali, alkali earth based selected
tetra-boro hydrides.
Preface ix
and show how these affect the properties of these materials, focussing on
application in the energy storage and in the transportation sector.
Chapter 3 - The features of gas discharge and plasma sources based on
Penning trap with metal hydride cathodes are presented. In such devices,
metal hydrides fulfill the functions of both a cathode and the solid-state
generator of working gas. Their advantages are high purity of gas injected
(99.99 – 99.999%), along with the safety and compactness in storage.
Hydrogen is injected (desorbed) locally under the influence of ion
bombardment of metal hydride surface, which fact provides return coupling
between the intensity of gas desorption and the parameters of gas discharge.
The rate of sputtering for those materials by plasma ions significantly
reduces as well as heat loads. Above effect is achieved due to the creation
of protective gas target as a result of both the thermal decomposition of metal
hydride and ion stimulated desorption. The feature of metal hydride cathode
under the conditions of gas discharge is a decrease in the ionization potential
of desorbed hydrogen by 0.3-0.5 eV due to the molecules desorption in the
vibrationally/rotationally excited state. This permits a substantially increase
in ionization efficiency and the formation of negative ions by the mechanism
of dissociative attachment in plasma volume. However, hydrogen desorbed
from metal hydride significantly changes the properties of the discharge.
This is expressed, for example, in the fact that the plasma source based on
Penning trap with metal hydride cathode appears to generate current-
compensated ion beams with the ability to control the energy of the extracted
ions. There is also the opportunity of longitudinal extraction of negative
hydrogen ions against the traditional method of extraction across the
magnetic field.
Chapter 4 - This chapter discusses metal hydrides found in sintered
electrodes of nickel-cadmium batteries with electrochemical method by the
way of electrolyte decomposition onto hydrogen and oxygen. It was shown
that as a result of electrolyte decomposition, oxygen is released from
batteries, while hydrogen is partly released and partly – in virtue of its high
diffusion permeability – is accumulated in sintered matrices of electrodes of
nickel-cadmium batteries in metal hydrides form. It was established that in
a case of a long-term operation of batteries (more than five years), in their
Preface xi
Chapter 1
P. Vajeeston*
Center for Materials Science and Nanotechnology,
Department of Chemistry, University of Oslo, Oslo, Norway
ABSTRACT
drawbacks, e.g., either too high or too low temperature of hydrogen release,
difficult reversibility and contamination of evolved hydrogen with B2H6,
which obstruct their application but also triggers the research of new
systems. Consequently, various novel borohydride-based materials have
been prepared and investigated. Numerous unsolved problems remain on
our way towards a sustainable, carbon free energy system based on
renewable energy and on hydrogen as a future energy carrier. Novel
materials form the back-bone in most emerging energy related
technologies. In this chapter, we presented the structural details of alkali,
alkali earth based selected tetra-borohydrides.
INTRODUCTION
14]. The concept aims to lower the overall reaction enthalpy by introducing
extra (endothermic) chemical reaction(s), i.e., by utilizing the well-known
Hess law from chemistry, also denoted as Medima principles.
Metal borohydrides are also a multifunctional class of materials that may
also be used as fast ion conductors for new types of batteries or for gas
adsorption [7, 15]. They may also have optical, electronic and magnetic
properties and can be used as reducing agents in organic synthetic chemistry
[16-18]. For this reason, mono-, bi- and tri-metallic borohydrides have been
the object of intense research efforts aimed at developing new synthesis
strategies, at performing structural characterization, and at investigating
physical and chemical properties [9, 19]. In this chapter, we critically review
the structures of selected metal borohydrides (only alkali, alkaline earth, and
their mixed tetrahydroborate) and their derivatives, which may form the
basis of future technological applications. The focus of this chapter is to
discover structures and provide perspectives for further research that may
act as inspiration for the development of novel materials towards rational
materials design.
DIBORANE
two crystalline phases of B2H6 in the temperature region from 4.2 to 100 K,
naming them α and β. The α phase was formed by deposition from gaseous
B2H6 at 4.2 K. It transformed slowly to the β phase above 60 K, which in
turn was obtained by deposition at 77 K and annealing to 90 K. Interestingly,
neither of the α and β phases nor the additional phase found by passing B2H6
through a microwave discharge showed a diffraction pattern that
corresponded to the phase reported earlier by Mark and Pohland [37]. The
refinement of the diffraction data for the three phases they found, was,
however, not carried out by Bolz, Mauer, and Peiser. In 1965, using single-
crystal X-ray diffraction, Smith and Lipscomb [39] successfully
characterized the β phase in atomic detail (see Figure 1). The structures of
all other crystalline phases of B2H6 remain unknown. The β phase is
monoclinic (space group P21/n, Z = 2, a = 4.40, b = 5.72, c = 6.50 Å, and γ
= 105.1°). Recently, Yao and Hoffmann [36] identified a new polymorph of
B2H6 from the theoretical simulation with lattice parameters for the P21/n
structure are, a = 4.46, b = 8.68, c = 4.55 Å, and β = 120.5° (monoclinic
cell), in good agreement with the measured lattice parameters a = 4.54, b =
4.54, c = 8.69 Å, and γ = 120.0° (hexagonal cell) [37]. Recent theoretical
simulation shows that various molecular diborane structures, including β
diborane and the P21/c structure, are all within 5 meV/molecule of each
other in the calculations at P = 1 atm. It is therefore not surprise that there
are several polymorphs of diborane [36].
BOROHYDRIDES (TETRAHYDROBORATE)
LiBH4
1. -LiBH4:
At ambient conditions LiBH4 (-LiBH4) crystallizes [43, 44] with an
orthorhombic (KGaH4-type) structure in which each [BH4]⎺ anion is
surrounded by four lithium Li+ and each Li+ by four [BH4]⎺, both in
tetrahedral configurations (Figure 2a). The calculated positional parameters
for -LiBH4 gives a nearly ideal tetrahedral [BH4]⎺ complex (with H–B–H
bond angles ranging between 108° and 112°) whereas the experimental
results suggest that the complex is highly distorted (bond angles: 85–120
[43]). The theoretically derived B–H distances in -LiBH4 turned out to be
almost equal (ca. 1.22 Å), whereas experimental studies gave different
results (1.28–1.44 Å [45], 1.01–1.28 Å [43]). Assuming that the thus
exposed mutual discrepancy is real it seems likely that the structural
arrangements of LiBH4 as a whole or in parts are sensitive to temperature.
Cell parameters: a = 7.1786 Å; b = 44.4369 Å; c = 6.8032 Å; Atomic
positions: Li(4c): 0.1568, ¼, 0.1015; B(4c): 0.3040; ¼, 0.4305; H1(4c):
0.90, ¼,0.9560; H2(4c): 0.4040, ¼, 0.28; H3(8d): 0.1720, 0.0540, 0.4280.
2. β-LiBH4:
At ca. 381 K, -LiBH4 undergoes a structural transition to β-LiBH4, but
conflicting results have been reported concerning the symmetry of the phase
formed. First β-LiBH4 was believed [46, 47] to belong to the tetragonal
class, but the recent experimental study of Soulié et al. shows that the
symmetry is hexagonal [43]. The structure of the hexagonal high-
10 P. Vajeeston
3. hp1-LiBH4:
The structure of phase hp1-, observed between 1.2 and 10 GPa, has been
determined ab initio from high-quality powder diffraction data. hp1- has a
new structure type in the ABX4 family [48]. Moreover, no isomorphous
derivatives of this Ama2 structure exist. The Li and [BH4]⎺ sublattices
interpenetrate so that Li atoms occupy tetrahedral voids in the [BH4]⎺
network (see Figure 2c). Most interestingly, the [BH4]⎺ anions have a nearly
square-planar coordination comprising four Li atoms, in striking contrast to
the -phase. A square-planar coordination of [BH4]⎺ anions has no
analogues in the crystal chemistry of metal borohydrides. The shortest H···H
distance between two neighboring [BH4]⎺ anions is 1.92 Å, as obtained from
the crystallographic data at 2.4 GPa. These short contacts serve as links
within chains formed by borohydride anions. This geometry is unique, and
such a short H···H distance is unprecedented in metal borohydrides.
Crystal Structures of Alkali and Alkaline Earth … 11
4. hp2-LiBH4:
The high-pressure phase hp2 was identified by comparing experimental
data with theoretical powder patterns calculated for known ABX4 structure
types [49]. The best description of the experimental diffraction pattern has
been achieved with the cubic Fm3̅m structure having a disordered
arrangement of [BH4]⎺ anions (-NaBH4-type) [49], with a = 5.109 Å at 18.1
GPa [48]. This phase exhibits the shortest H···H contacts between [BH4]⎺
anions (1.67 Å in the optimized model; see Figure 2d). However,
considering the high pressure (greater than 10 GPa) of the transition into
hp2, it is less likely that this phase can be stabilized at ambient conditions.
Figure 2. Crystal structures of (a) -LiBH4, (b) β-LiBH4, (c) hp1-LiBH4, and (d) hp2-
LiBH4 (partial H site occupancy is marked as dark brown colour).
NaBH4
1. -NaBH4:
The crystal structure of sodium borohydride was first proposed by
Soldate in 1947 [57] who predicted that it is based on a face centered cubic
lattice and suggested that the structure consisting of tetrahedral [BH4]⎺ ions
and Na+ ions. Abrahams et al. [49] repeated the experiment below the
transition point (−83°C) and demonstrated that sodium borohydride
becomes tetragonal at −95°C. This transition point was according to
Stockmayer and Stephenson [52], the result of an order–disorder transition
involving different orientations of [BH4]⎺. Later on, in 1985, Davis and
Kennard [58] determined the structure of NaBD4 using neutron diffraction,
which was found to have the NaCl type structure, in space group F-43m with
D atoms tetrahedrally oriented about B and along all cube diagonals. This
gives a random distribution of [BD4]⎺ tetrahedra in two different
configurations. Although in a recent study the structure of NaBD4 was found
to belong to space group P421c [59] but a later experimental [55] and
theoretical [44] study showed that the high pressure of 6.3 GPa can alter the
crystal structure of -NaBH4 (cubic; Fm3̅m; see Figure 3a) to β-NaBH4
(tetragonal; P-421c), which further undergoes with a transition to an
orthorhombic phase (Pnma) at 8.9 GPa, which is stable up to 30 GPa. The
structural arrangement of -NaBH4 is similar to the hp2-LiBH4 phase.
Cell parameters: a = 6.1308 Å; b = 6.1308 Å; c = 6.1308 Å; Atomic
positions: Na(4a): 0,0,0; B(4b): ½, ½, ½; H1(32f): -0.1030, -0.3970, 0.1030
(50% occupancy).
Crystal Structures of Alkali and Alkaline Earth … 13
2. β-NaBH4:
On cooling below 190 K [54] or upon a compression to 6 GPa at room-
temperature (RT) [55] a phase with closely related ordered tetragonal
structure appears (β-NaBH4; see Figure 3b). It was first reported in the space
group P-421c [54] but later revised in a higher symmetry P42/nmc [60].
Comparing to the cubic phase, the [BH4]⎺ groups in the tetragonal phase are
ordered in two different orientations. As a result, the number of the shortest
H···H contacts in the tetragonal phase is reduced by one third comparing to
the disordered cubic phase, and the network of the repulsive H···H contacts
changes from isotropic three-dimensional to two-dimensional oriented in the
ab plane. Consequently, the c/a ratio changes discontinuously from 1 to
0.964 upon the cubic-to-tetragonal transition at 186 K [60] due to the
contraction of the c-axis.
Cell parameters: a = 4.3706 Å; b = 4.3706 Å; c = 5.9509 Å; Atomic
positions: Na(2a): ¾, ½, ¾; B(2b): ¾, ½, ½; H(8g): ¾, 0.0290, 0.1444.
3. -NaBH4:
Above 9 GPa a new phase was detected [55] and its structure has been
solved from synchrotron powder diffraction data measured at 11.2 GPa in
BaSO4 structure type [61]. Both the cubic and the tetragonal phases are
strongly textured in diamond anvil cells [54, 62]. For the successful solution
of the tetragonal structure, it was essential to model the texture, including
14 P. Vajeeston
one parameter in the global optimization. The Pnma phase has a BaSO4-type
structure see Figure 3c. This structure type rarely occurs in the family of
light metal hydrides. It is important to note that orthorhombic LiBH4 is
sometimes erroneously attributed to the KGaH4 and BaSO4 structure type.
These two structure types have the same symmetry and Wyckoff sequence
and similar unit cell volumes. Nevertheless, they differ significantly with
respect to the coordination numbers (CN) of the cations and the anions sites
CN=6 with octahedral coordination for the BaSO4-type structure and CN=4
with tetrahedral coordination in LiBH4. In the -NaBH4 phase, the Na atom
is surrounded by six borohydride anions at Na-B distances of 2.7633–2.8493
Å. They are slightly shorter than Na-B distances in the P-421c phase at 10
K 2.93–3.06 Å and in the -NaBH4 phase at ambient conditions 3.07 Å. 16
H2 and H3 atoms form nine short contacts Na-H at 2.09–2.35 Å and H1
atom forms three longer contacts at 2.48–2.52 Å. Thus, with respect to
hydrogen atoms, CN for Na is 12, and [BH4]⎺ units are connected to Na+
cations via tetrahedra edges. The same CN is observed in the β- and -
phases and the Na-H distances, 2.44–2.59 and 2.58 Å, respectively. The
[BH4]⎺ units are very close to the tetrahedral geometry, with B-H distances
centered on 1.17 Å. The shortest H···H distance between [BH4]⎺ anions is
2.31 Å.
Cell parameters: a = 7.2970 Å; b = 4.1166 Å; c = 5.5692 Å; Atomic
positions: Na(4c): 0.1540, ¼, 0.2107; B(4c): 0.5924, ¼, 0.7841; H1(4c):
0.4676, ¼, 0.9161; H2(4c): 0.7298,¼, 0.8926; H3(8d): 0.4135, 0.9825,
0.3363.
KBH4
1. -KBH4:
This KBH4 compound shows a similar behaviour to NaBH4. Its cubic
phase has been studied at room temperature by X-ray diffraction on single
crystals and by neutron powder diffraction [63] in both cases it was
described in the space group Fm-3m.
Crystal Structures of Alkali and Alkaline Earth … 15
2. β-KBH4:
The cubic phase transforms to tetragonal (P421c; space group 114) at
65–70 K with a structure equivalent to the tetragonal NaBH4. The K atoms
are located at 2a, B atoms at 2b, and H atoms at 8e positions [64]. The cation
size in the cubic alkali borohydrides, MBH4 increases as M goes from Na to
Cs, thus the unit cell expands and the shortest H···H distances between the
neighbouring [BH4]⎺ anions increase. As a result, the weaker H···H
repulsion leads to lower temperatures of the cubic-to-tetragonal transition
[39].
Cell parameters: a = 4.4754 Å; b = 4.4754 Å; c = 6.3630 Å; Atomic
positions: K(2a): 0,0,0; B(2b): 0,0, ½; H(8e): 0, 0.7906, 0.4150.
3. -KBH4:
Above 6.8 GPa the tetragonal phase transform into Pnma symmetry
(space group 62) with K and B atoms at 4c and H atoms at 4c and 8d [64].
Cell parameters: a = 6.9760 Å; b = 4.9330 Å; c = 5.1110 Å; Atomic
positions: K(4a): 0.2056, ¼, 0.2456; B(4c): 0.4559, ¼, 0.4321; H1(4c):
0.3920, ¼, 0.7070; H2(4c): 0.1490,¼, 0.8200; H3(8d): 0.2070, 0.0459,
0.5400.
RbBH4
1. RT()-RbBH4:
The crystal structure at ambient-pressure phase is cubic, with space
group Fm-3m and 4 formula units in the unit cell [49]. It has a NaCl-type
arrangement of Rb cations and [BH4]⎺ anions, where the latter are
orientationally disordered over the two positions around the inversion center
(see Figure 4a). The disordered centrosymmetric structure for the cubic
RbBH4 was assumed by the analogy to the well-established structure of
NaBH4. The [BH4]⎺ group is disordered over two positions. Its partial
ordering at low temperature (LT) was detected at 44 K, [66] without a
reconstruction or a change of the atomic coordination numbers.
Cell parameters: a = 7.0290 Å; b = 7.0290 Å; c = 7.0290 Å; Atomic
positions: Rb(4a): 0,0,0; B(4b): ½, ½, ½; H(32f): 0.410, 0.410, 0.401 (50%
occupancy).
2. HP1-RbBH4:
A transition from the cubic to P4/nmm high-pressure phase starts at 3
GPa and completes at 3.4 GPa. In this high-pressure structure Rb and B
atoms are well ordered. Hydrogen atoms cannot be directly detected from
synchrotron X-ray diffraction data due to the presence of the much heavier
Rb atoms. They can however be located as a part of the rigid tetrahedral
[BH4]⎺ group. Due to the change of the [BH4]⎺ coordination mode, the
corresponding interatomic distances are segregated into four short Rb···B of
3.05−3.2 Å (12 long Rb···H of 2.7−3.1 Å) distances for the η3-BH4
coordination and four long Rb···B of 3.76−3.95 Å (four short Rb···H of
2.6−2.7 Å) distances for the η1-BH4 coordination (η1 and η3 stand
respectively for the coordination via a vertex and a face). The P4/nmm
structure is derived from the cubic by shifting the Rb atom and the [BH4]⎺
group along the c axis of the resulting tetragonal cell. The P4/nmm structure
is intermediate between the NaCl and CsCl types, and the transitions occur
by the mechanism of atomic displacements.
Crystal Structures of Alkali and Alkaline Earth … 17
3. HP2-RbBH4:
Above 10.4 GPa the P4/nmm phase starts to transform into a new phase
(C222), and this transition completes at ∼12.2 GPa. The number of [BH4]⎺
neighbors increases to eight in the P4/nmm phase. Rb and B atoms were
reliably located in the special positions. The atomic arrangement represents
an orthorhombically distorted CsCl-type structure (see Figure 4c). The ideal
CsCl-type structure is not met in RbBH4 owing to the better packing of the
anisotropic anions and spherical cations in the distorted structures. The
arrangement of the [BH4]⎺ group inside the Rb8 cube, with η1-BH4 pointing
the B−H bonds toward the midpoint between two Rb atoms, is likely the
densest one, as it is practically the same in the higher pressure C222 and I-
42m phases (see Figure 4b and c).
Figure 4. Crystal structures of (a) -RbBH4 (hydrogen vacancy are marked as dark
brown colour), (b) hp1-RbBH4, (c) hp2-RbBH4, (d) hp3-RbBH4.
4. HP3-RbBH4:
Above 17 GPa the I-42m phase is observed. This transition is not fully
completed even at the highest reported pressure (up to 25 GPa). The
octahedral coordination in the ambient-pressure phase gradually transforms
to a more or less distorted cubic arrangement at higher pressures. The
18 P. Vajeeston
average Rb···B distances correlate with the change of the CN: upon the first
transition, CN goes from 6 to 4 + 4 and the average Rb···B distance increases
by ∼7% and then smoothly decreases at higher pressures. A combination of
different [BH4]⎺ coordination modes together with the increase in CN leads
to an increase in density by 8.8% at the cubic-to-tetragonal phase transition,
in spite of the fact that the average Rb···B distance increases. The specific
Rb···B and Rb···H distances show a large spread, related to the occurrence
of the three different Rb···BH4 coordination modes: via a face, an edge, and
a vertex of the tetrahedral anion [65].
Cell parameters: a = 5.0057 Å; b = 5.0057 Å; c = 7.8161 Å; Atomic
positions: Rb(4e): 0,0,0.2329; B(4c):0, ½, 0; H(16j): 0.125, 0.375, 0.080.
CsBH4
1. α-CsBH4
Similar to NaBH4, KBH4 and RbBH4 at room temperature CsBH4
crystallize with a cubic high-temperature (HT) structure having Fm3̄m
symmetry in which the [BH4]⎺ complexes are disordered [49].
Cell parameters: a = 7.4190 Å; b = 7.4190 Å; c = 7.4190 Å; Atomic
positions: Cs(4a): 0,0,0; B(4b): ½, ½, ½; H(32f): 0.5948, 0.5948, 0.5948
[63].
In general, the room temperature structures of MBH4 (M=Na, K, Rb, Cs)
have cubic symmetry and are isotypic with their sodium congener NaBH4.
The alkali cations and [BH4]⎺ anions form a NaCl-type arrangement. The
bond distances those within the disordered [BH4]⎺ complexes increase as the
alkaline radius increases: B-H=1.178 Å (Na), 1.196 Å (K), 1.206 Å (Rb) and
1.217 Å (Cs), effect which is expected in view of matrix effects.
Simultaneously, the M-H distances increases. Calculating a ‘pseudo
hydrogen ionic radius’ from the formula r(H−)=d(M-H)−r(M+) by using the
ionic radius of the alkaline in 12 coordination as given by Shannon [67]
(Na+: 1.39 Å, K+: 1.64 Å, Rb+: 1.72 Å, Cs+: 1.88 Å) yields r(H−)=1.19 Å
(NaBD4), 1.20 Å (KBD4), 1.26 Å (RbBD4) and 1.28 Å (CsBD4) [63].
Crystal Structures of Alkali and Alkaline Earth … 19
These values compare well with the tabulated 1.22 Å for hydride ions in
coordination four [67] as in the present compounds (H surrounded by three
M and one B) while illustrating the strong dependence of M–H interactions
on cation electronegativity [67].
Be(BH4)2
Figure 5. Crystal structures of (a) -Be(BH4)2, (b) -Mg(BH4)2, (c) β-Mg(BH4)2, (d) -
Mg(BH4)2, (e) 1-Mg(BH4)2, and (f) -Mg(BH4)2.
Mg(BH4)2
1. -Mg(BH4)2:
The first polymorph, -Mg(BH4)2, has been reported in P61 space group
symmetry by two independent groups [5, 88] using powder diffraction data.
However, DFT studies suggested a possibly higher P6122 symmetry
(see Figure 5b) [89], which was confirmed by single-crystal diffraction at
100 K [77]. Analysis of the published P61 models shows that the location of
the H-atoms from powder data posed the main problem for the identification
of the correct symmetry. Analysis of the initial P61 models shows that the
determination of H-atom locations from powder data posed the main problem
22 P. Vajeeston
2. β-Mg(BH4)2:
The -phase with space group P61, is stable below 453 K. Upon heating
above that temperature it transforms to an orthorhombic phase, with space
group Fddd (β-Mg(BH4)2; Figure 5c), stable to 613 K at which point it
decomposes with hydrogen release [88]. The latter phase is metastable upon
cooling, showing an anomalous temperature dependence of the cell
parameter [77]. The β-Mg(BH4)2 phase has a complicated structure with five
distinct types of MgB4 tetrahedra. This phase is consisting of corner-sharing
tetrahedra with Mg atoms at the center and [BH4]⎺ units at the vertices. Mg3
are located close to the c axis, while the other four Mg atoms are relatively
close to the ab planes. No two Mg atoms of the same type do not share a
[BH4]⎺ bridge. Mg1 tetrahedra have one face nearly parallel to the bc plane,
whereas Mg2 tetrahedra have one edge nearly parallel to a and one nearly
parallel to c. Each type of Mg tetrahedron is bonded to one tetrahedron of
the same type, and to three tetrahedra of the other type. Mg—Mg distances
of linked tetrahedra range from 4.61 to 4.72 Å, and the Mg—B—Mg angles
range from 150 to 171°.
Cell parameters: a = 37.0720 Å; b = 18.6476 Å; c = 10.9123 Å; Atomic
positions: Mg1(32h): 0.07749, 0.2501, 0.5381; Mg2(32h): 0.05290, 0.9298,
0.7464; B1(32h): 0.4460, 0.1533, 0.410; B2(32h): 0.1078, 0.0085, 0.2636;
B3(16e): 0.3124,0, 0; B4(32h): 0.4441, 0.6472, 0.560; B5(16g): 0, 0, 0.773;
H1(32h): 0.4284, 0.1778, 0.3524; H2(32h): 0.4593, 0.1908, 0.4600;
H3(32h): 0.4639, 0.1239, 0.3612; H4(32h): 0.4321, 0.1204, 0.4679;
H5(32h): 0.1220, 0.9662, 0.2304; H6(32h): 0.0950, 0.9938, 0.3422;
H7(32h): 0.0891, 0.0239, 0.2006; H8(32h): 0.1247, 0.0499, 0.2812;
H9(32h): 0.3281, -0.0444, 0.0027; H10(32h): 0.2965, 0.0016, 0.0759;
H11(32h): 0.4199, 0.6724,0.5519; H12(32h): 0.4409, 0.5942, 0.5406;
H13(32h): 0.4532, 0.6526, 0.6473; H14(32h): 0.4620, 0.6692, 0.5009;
H15(32h): 0.0218, 0.00992, 0.8267; H16(32h): 0.0049, -0.0433, 0.7192.
3. β'-Mg(BH4)2
The porous γ-Mg(BH4)2 undergoes thermally-induced phase transitions
to ε [79, 80, 85] and, subsequently, to β′-Mg(BH4)2 (allegedly a disordered
Crystal Structures of Alkali and Alkaline Earth … 25
phase of β [79, 85]). The structure for the β′-phase has not been published,
and the phase was identified by comparison with the literature [79, 80, 85].
4. -Mg(BH4)2:
The crystal structure of γ-Mg(BH4)2 has a much lower density (0.5431
g cm−3, as calculated from the crystal structure) compared with the α- and
β-phases, resulting in a porous structure with voids surrounded by [BH4]⎺
groups. Recently, Filinchuk et al. investigated the propensity for nitrogen
and hydrogen gas adsorption into the pores of the γ-phase [7]. The
monoclinic solvate structure of Mg(BH4)2⋅1/2S(CH3)2 is a 3D framework
containing two Mg atoms: one atom is tetrahedrally coordinated to four
[BH4]⎺ groups, and the other to four [BH4]⎺ groups and one S(CH3)2 ligand
to form a trigonal bipyramid. Removal of the S(CH3)2 ligand does not break
the integrity of the framework but leads to a highly symmetric cubic
structure of γ-Mg(BH4)2 (space group Id-3a, no. 230: see Figure 5d), where
a single Mg site has a tetrahedral environment of the [BH4]⎺ groups. Its
structure has a 3D net of interpenetrated channels, thus making γ-Mg(BH4)2
the first reported hydride with a large permanent porosity. The empty
volume in the structure amounts to 33 %. The narrowest part of the channel
is defined by a distance of 5.8 Å between hydrogen atoms, while a point at
(1/8, 1/8, 1/8) is 3.56 Å away from the nearest H atom, 4.12 Å from the B,
and 4.82 Å from the Mg atoms. The framework topology of γ-Mg(BH4)2 is
isomorphous to a hypothetical zeolite-type polymorph of SiO2 [93] and to a
porous zinc imidazolate framework ZIF-72, [94] while no other materials
show any similarities.
Cell parameters: a = 15.7575 Å; b = 15.7575 Å; c = 15.7575 Å; Atomic
positions: Mg(24d): 0.25, 0.125, 0.5; B(48g): 0.3098,0.0598, 0.375;
H1(96h): 0.2857,0.0207,0.4373; H2(96h): 0.2974,0.1355,0.3856.
5. 1-Mg(BH4)2:
α- and γ-Mg(BH4)2 have 6.4% and 33% of empty space, respectively, at
elevated pressures. A remarkable volume collapse of approximately 20 %
upon the transition from the α- to a new high-pressure polymorph of
Mg(BH4)2was observed at 1.1–1.6 GPa when using diamond anvil cells [7].
26 P. Vajeeston
This new polymorph, denoted δ1-Mg(BH4)2 (see Figure 5e), has a tetragonal
structure consisting of two interpenetrated Mg(BH4)2 frameworks. Each
framework resembles the cristobalite structure (a polymorph of SiO2), while
their doubly interpenetrated arrangement has a Cu2O topology, which is
typical for metal organic frameworks (MOF). This structural organization is
very robust, as the δ-phase is stable up to 15 GPa and upon a decompression
to 1 bar, and even upon heating to approximately 373 K at ambient pressure,
where it transforms back to α-Mg(BH4)2. δ1-Mg(BH4)2 possesses no empty
voids and has the second highest volumetric hydrogen density (147 g H2/L
at ambient conditions) among all the known hydrides; this value is slightly
below Mg2FeH6 with the hydrogen density of 150 g H2/L. The latter
compound has a much lower gravimetric hydrogen density of 5.5 %,
compared to 14.9 wt % in Mg(BH4)2. The second highest volumetric
hydrogen density in borohydrides, 127 g H2/L, is recorded for the toxic
Be(BH4)2, which has an extreme gravimetric hydrogen density of 20.7 wt %.
In this structure Mg atoms have a tetrahedral environment of four [BH4]⎺
groups, and the [BH4]⎺ groups are coordinated by two Mg atoms through the
opposite edges. The Mg⋅⋅⋅BH4 interaction is highly directional and results in
linear Mg⋅⋅⋅BH4⋅⋅⋅Mg fragments, which can be considered as fundamental
building units in all the structures. Here the [BH4]⎺ group acts as a directional
ligand, similar to the organic ligands (“linkers”) in MOFs.
Cell parameters: a = 5.4361 Å; b = 5.4361 Å; c = 6.1468 Å; Atomic
positions: Mg(2a): 0, 0, 0; B(4c): 0.2584,0.2584,0.2432; H1(8d):
0.2050,0.4641,0.3048; H2(4c): 0.1122,0.1122, 0.3185; H3(4c):
0.2525,0.2525, 0.0460.
6. 2-Mg(BH4)2:
The XRD patterns of the high-pressure phase were indexed in the
hexagonal P63 space group at 5.5 GPa, with lattice parameters a = 8.35(5)
Å, c = 4.68(4) Å, and V = 283.47(8) Å3 [85]. The complete structural
description was however not solved.
Crystal Structures of Alkali and Alkaline Earth … 27
7. -Mg(BH4)2:
The porous γ-Mg(BH4)2 undergoes thermally-induced phase transitions
to ε [79, 80, 85] and, subsequently, to β′-Mg(BH4)2 (allegedly a disordered
phase of β [78, 79]). The ε-to-α phase transition upon cooling of ε-Mg(BH4)2
has been also reported [79, 81]. It is however not always observed. The
structure of ε-phase is unpublished, and it has been identified by comparison
with the literature [79, 80, 85].
8. -Mg(BH4)2:
The structure of ζ-Mg(BH4)2 was refined from SR-PXD data measured
at 224°C [71]. Realizing that the diffraction pattern of ζ-Mg(BH4)2 is similar
to that of α-Mn(BH4)2 (space group P3112) [6]. In this structure, both
independent Mg atoms are surrounded by four [BH4]⎺ tetrahedra (see Figure
5f) in deformed tetrahedral coordinations (Mg-B distances within 2.42-2.52
Å, B-Mg-B angles within 64.3-153.19°). A similar tetrahedral coordination
of a cation by four [BH4]⎺ anions is known for Li+ in three of the four known
phases of LiBH4 [43, 48] and in LiK(BH4)2 [5]. On the other side, each
[BH4]⎺ is approximately linearly coordinated by two Mg2+ (Mg-B-Mg angles
within 159° to 170.0°), similarly to manganese borohydride [6]. The
coordination of manganese to hydrogen atoms cannot be determined
unambiguously because of the limited accuracy of the [BH4]⎺ tetrahedra
orientation, which is strongly affected by Mg-H and Mg-B antibump
restraints used in the Rietveld refinement. The configuration resulting from
such restraint refinement can be understood as the Mg-H2BH2-Mn bonding
scheme giving 8-fold coordination of Mg by hydrogen atoms, with Mg-H
distances varying within 1.86-2.304 Å. A single B-H distance of 1.08-1.22
Å was refined for rigid [BH4]⎺ tetrahedra. It is consistent with those in related
borohydrides such as Mn(BH4)2 (1.13 Å [6]).
Cell parameters: a = 10.424 Å; b = 10.424 Å; c = 10.729 Å; Atomic
positions: Mg1(6c): 0.2313,0.9181, 0.1253; Mg2(3a): 0.562, 0.12544,
0.6667; B1(6c): 0.040, 0.6990, 0.006; B2(3b): 0.471, 0.9415, 0.16667;
B3(3a): 0.204, 0.1021, 0; B4(3a): 0.133, 0.8671, 0.3333; B5(3b): 0.706,
0.2944, 0.8333; H1(6c): -0.082, 0.658, 0.001; H2(6c): 0.071, 0.645, 0.929;
H3(6c): 0.069, 0.669, 0.098; H4(6c): 0.104, 0.823, 0.994; H5(6c): 0.428,
28 P. Vajeeston
0.014, 0.21; H6(6c): 0.39, 0.869, 0.091; H7(6c): 0.132, 0.115, -0.072;
H8(6c): 0.277, 0.064, -0.046; H9(6c): 0.020, 0.827, 0.375; H10(6c): 0.125,
0.787, 0.259; H11(6c): 0.753,0.415, 0.859; H12(6c): 0.72, 0.23, 0.915.
In general, all the experimentally observed polymorphs and the lowest-
energy theoretical Mg(BH4)2 structures contain very similar local
configurations: Mg atoms have a tetrahedral environment of four [BH4]⎺
groups, and these [BH4]⎺ groups are coordinated to two Mg atoms through
the opposite edges. The Mg⋅⋅⋅BH4 interaction is highly directional and
results in linear Mg⋅⋅⋅BH4⋅⋅⋅Mg fragments, which can be considered as the
fundamental building units in all the structures. Here the [BH4]⎺ group acts
as a directional ligand, similar to the organic ligands (“linkers”) in MOFs.
On the other hand, Mg atoms form a limited set of MgH8 polyhedra.
Interestingly, out of all the possible eight-vertex polyhedral [95], only the
less uniform Johnson solids are found in the experimental structures. The
stability of MgH8 coordination polyhedra can presumably be linked to the
relative stability of the polymorphs. Both experiments and theoretical
predictions suggest a vast polymorphism of Mg(BH4)2. Moreover, the
experimentally observed phases are stable over relatively wide temperature
and pressure ranges, in particular under ambient conditions, thus indicating
that the reconstruction of strongly bound Mg(BH4)2 coordination
frameworks is kinetically hindered. This behavior is likely due to the high
stability of the linear Mg⋅⋅⋅BH4⋅⋅⋅Mg units, which link the MgH8 nodes into
various framework structures, similar to the partly covalently bonded MOFs.
The relatively small charge transfer from Mg to [BH4]⎺ makes the bonding
partly covalent and is essentially the reason for the MOF-like behavior of
Mg(BH4)2, namely the rich polymorphism and metastability, the large
pressure-induced volume collapses, and the amorphization under pressure.
A possible bonding scheme involves a formation of molecular orbitals
between Mg, H, and B atoms, similar to those in diborane, B2H6.
Ca(BH4)2
space group Fddd and later revised in F2dd by Filinchuk et al. [97]. Besides
the overall improved fit to the powder diffraction data, the non-
centrosymmetric structure shows a group-subgroup relation with the
symmetry of the high-temperature phase. The α′-phase forms via a second
order phase transition at ~495 K and was described in the space group I-42d
[77]. Above 450 K, the closely related F2dd and I-42d phases kinetically
transform into a completely different β-phase. The determination of its space
group symmetry is ambiguous, and its structure was described in two similar
models, corresponding to the lowest binding energies as evaluated by DFT,
namely in the space groups P42/m [98] and P-4 [77] It is likely that the
[BH4]⎺ group in β-Ca(BH4)2 is intrinsically disordered, similar to the high-
temperature phase of LiBH4 [48, 99]. According to the systematic absences
of the diffraction peaks the true symmetry of the disordered structure might
be P42nm [97]. The HT phase is 4–6% denser than the F2dd and I-42d
phases, it is metastable upon cooling, and at room temperature it slowly
transforms back into the orthorhombic phase [53]. Another phase, labelled
-Ca(BH4)2, obtained by wet chemical synthesis, presents an orthorhombic
structure [100]. The crystal structure of the orthorhombic γ-phase of
Ca(BD4)2 is identified by combined synchrotron X-ray powder diffraction,
neutron powder diffraction, and ab initio calculations. It is metastable at all
temperatures and irreversibly transforms into the β-phase at ~590 K.
1. α-Ca(BD4)2:
The α-Ca(BH4)2 structure crystallizes in the noncentrosymmetric space
group F2dd, containing an ordered [BH4]⎺ anion (Figure 6a). The crystal
structures of α-Ca(BH4)2 contain calcium cations nearly octahedrally
coordinated to six borohydride anions. CN for the alkaline-earth atoms
correlate with their radii, and change from 3 in Be(BH4)2 to 4 in Mg(BH4)2,
and finally to 6 in the three Ca(BH4)2 structures. The six Ca-B distances in
α-Ca(BH4)2 within range of 2.816–2.967 Å at 91 K. The local environments
of the Ca atoms differ between the α-, α’-, β-phases by different [BD4]⎺
orientations and deformation of the [BD4]⎺ octahedral coordination (see
Figure 6). The α-phase exhibits more distorted octahedra with respect to the
more regular and higher-symmetry β-phase. This can be seen from the
30 P. Vajeeston
differences between the shortest and the longest B−B interatomic distances
between the [BD4]⎺ groups surrounding the Ca atoms. These differences are
of 11% (3.91−4.366 Å) for the β-phase and 14% (3.807−4.442 Å) for the
α-phase. It is worth noting that, in spite of the similarities in coordination
and calculated formation energy, differences in densities can be as large as
∼4% between the density of the α-phase and the densities of the β-phase and
γ-phase (the density of the β-phase and γ-phase are almost identical. The
coexistence of the α-, β-, and γ-phases as well as the kinetics of the phase
transitions between these phases are very much dependent on the synthesis
method of the samples [76, 101].
Cell parameters: a = 8.7759 Å; b = 13.0234 Å; c = 7.4132 Å; Atomic
positions: Ca(8a): 0, 0, 0; B(16b): 0.0056, 0.2243, 0.0081; H1(16b): 0.0190,
0.1648, 0.1272; H2(16b): 0.0345,0.1817,-0.1256; H3(16b): -0.1180,
0.2553, 0.012; H4(16b): 0.0910, 0.2922, 0.028.
Figure 6. Crystal structures of (a) -Ca(BH4)2, (b) ′-Ca(BH4)2, (c) β-Ca(BH4)2, and
(d) -Ca(BH4)2.
Crystal Structures of Alkali and Alkaline Earth … 31
2. α′-Ca(BD4)2:
Ca atoms form a T-shaped arrangement around the B atoms in the α-
and α′-phases (see Figure 6a and b) and a nearly triangular arrangement in
the β-phase (Ca···B···Ca angles 95.5–133.7°). The coordination mode for
the [BH4]⎺ anion with respect to the Ca cations differs correspondingly. In
the α′-phase each Ca atom coordinates two [BH4]⎺ groups via the edge
(Ca…H2B, with Ca–H distances 2.52 Å), and other four via the corner
(Ca…HB, with Ca–H distances 2.02 Å). Coordination via the edge is typical
for alkaline and alkaline-earth borohydrides, such as Be(BH4)2 [69] and
Mg(BH4)2 [7, 88]. The six Ca-B distances in α′-Ca(BH4)2 within range of
2.941–2.976 Å for the α′-phase at 495 K.
Cell parameters: a = 5.8446 Å; b = 5.8446 Å; c = 13.2279 Å; Atomic
positions: Ca(4a): 0, 0, 0; B(8c):0, 0, 0.2250; H1(16e):-0.1608, -
0.035,0.1762; H2(16e): 0.032,-0.1613,0.2752.
3. β-Ca(BD4)2:
The crystal structure of β-Ca(BD4)2 is quite ambiguous. Majzoub and
Rönnebro [102] recently determined the β-polymorph to be best described
in space group P-4 (No. 81) or P42/m (No. 84) by a combined theoretical
and experimental approach, using the Prototype Electrostatic Ground State
method to find the lowest energy structures. A similar conclusion is drawn
from combined experimental and ab initio calculations showning that the
best description of the structure is with the space group P42/m based on
appropriate size/geometry of the [BD4]⎺ tetrahedra, the lowest calculated
formation energy, and real positive vibrational energy, indicating a stable
structure [98]. On the other hand, the Rietveld refinements indicated that the
structure could preferably be described in P-4 [97]. This structure model of
the β-phase was verified by a single-phase sample, obtained by annealing
the α + β mixture at 573 K. At this temperature, the conversion from the α-
to the β-phase is complete. The P-4 model showed slightly better fit than the
P42/m one, but more importantly, more reasonable interatomic distances.
Thus, finally the P-4 structure was confirmed by the Rietveld method with
FullProf [97]. Hence we assume the real β-phase is with P-4 space group
and the second one is assumed as β2-Ca(BH4)2.
32 P. Vajeeston
In the β-phase (P-4; Figure 6c); similarly to the α′-phase, two Ca atoms
coordinate the BH4 groups via edges, and the third one via a tetrahedral
vertex. The experimentally determined Ca–H distances vary from 2.08 to
2.56 Å. The six Ca–B distances in β-Ca(BH4)2 within range of 2.923–2.939
Å for the β-phase at 305 K. The β-phase exhibits more regular octahedra
with respect to the distorted low-symmetry α- and γ-phases. This can be seen
in differences between the shortest and longest B−B interatomic distances
between the BD4 groups surrounding the Ca atoms. These differences are of
11% (3.91−4.366 Å) for the β-phase and 14% (3.807−4.442 Å) to 17%
(3.82−4.618 Å) for the α- and γ-phases.
Cell parameters: a = 6.9189 Å; b = 6.9189 Å; c = 4.3471Å; Atomic
positions: Ca(2g): 0.5, 0, 0.240, 3.37; B(4h): 0.790, 0.693, 0.259; H1(4h):
0.626, 0.727, 0.217; H2(4h): 0.833, 0.765, 0.491; H3(4h): 0.808, 0.528,
0.262; H4(4h): 0.873, 0.769, 0.062.
4. β2-Ca(BH4)2:
For the structural description please refer above. Cell parameters: a =
6.924 Å; b = 6.924 Å; c = 4.3492 Å; Atomic positions: Ca(2d): 0,1/2, ½;
B(4j): 0.2102, 0.6952, 0; H1(4j): 0.464, 0.68, 0; H2(4j): 0.163,0.548,0;
H3(8k): 0.17000, 0.75500, 0.8090.
5. γ-Ca(BD4)2:
The γ-phase exhibit more distorted octahedra with respect to the more
regular and higher-symmetry β-phase. This can be seen in differences
between the shortest and longest B−B interatomic distances between the
[BD4]⎺groups surrounding the Ca atoms. These differences are of 11%
(3.91−4.366 Å) for the β-phase and 17% (3.82−4.618 Å) for the γ-phase.
The difference between the shortest and the longest interatomic distance
within a tetrahedron is less than 14.5% for the B−D distances (for the two
independent tetrahedra: 1.06−1.24 Å and 1.17−1.33 Å), and less than
10.8% for D−D distances (for the two independent tetrahedra: 1.86−1.95 Å
and 1.90−2.13 Å). Like for the α- and the β-phases of Ca(BD4)2, in the
γ-phase, the Ca atoms are surrounded by six [BD4]⎺groups.
Crystal Structures of Alkali and Alkaline Earth … 33
Sr(BH4)2
1. o-Sr(BH4)2:
The structure of Sr(BH4)2 was solved simultaneously and independently
from SR-PXD data measured for the sample prepared by solvent-mediated
synthesis [103]. Strontium borohydride, crystallizes in an orthorhombic unit
cell with space group symmetry Pbcn (see Table 2). The structural drawing
of the strontium borohydride Sr(BH4)2 is shown in Figure 7a. The basic
building unit for the structure of Sr(BH4)2 is the borohydride anion [BH4] ⎺,
which is a regular tetrahedron. This structure is similar to the α-PbO2-type
[104]. The structure is built from half-occupied brucite-like layers of slightly
distorted [Sr(BH4)6] octahedra stacked along the a-axis direction [103]. The
octahedra share two edges and all vertices with eight other octahedra in total.
The [BH4]⎺ units are coordinated by three Sr in a nonplanar triangle. There
are three different strontium–boron distances, each appearing twice within
an octahedron either adjacent or opposite to each other; these are 2.96
(adjacent), 3.058 (opposite), and 3.12 Å (adjacent). These strontium–boron
distances are consistent with the distance of 2.960 Å reported by Bremer et
al. in Sr(BH4)2·2THF [105]. There are 3 B–Sr–B angles within 169.2–173.9°
and 12 within 83.7–97.3°. The Sr–B–Sr angles within the nonplanar [BSr3]
triangle range from 96.3° to 129.9°, and the distance of the B atom from the
Sr plane is 0.36 Å. Four [BH4]⎺ units coordinating via the tetrahedral edge
and two via the tetrahedral face. Thus, Sr is coordinate to 14 hydrogen atoms.
Cell parameters: a = 6.978 Å; b = 8.937 Å; c = 7.559 Å; Atomic
positions: Sr(4ca):0.5, 0.3495, 0.25; B(8d):0.7751, 0.1313, 0.0796; H1(8d):
34 P. Vajeeston
2. t-Sr(BH4)2:
Upon heating the - polymorph of Sr(BH4)2 transformed into the
tetragonal t-Sr(BH4)2 polymorph increasing its coordination number from 6
to 8. It then transformed at 445°C into the P41212 polymorph [106]. The
latter is a non-centrosymmetric superstructure (doubled along the c-axis) of
Crystal Structures of Alkali and Alkaline Earth … 35
the ht-ZrO2 type (P42/nmc), a deformation of the cubic CaF2 type which is
stabilized by even higher temperatures in ZrO2. The Sr atoms in this
polymorph are coordinated to eight [BH4]⎺ (Figure 7b) forming a cube, and
each borohydride is surrounded by four Sr atoms in a tetrahedral
coordination. Sr–B distances are between 3.0911 and 3.1424 Å.
Cell parameters: a = 5.4208 Å; b = 5.4208 Å; c = 11.709Å; Atomic
positions: Sr(4a): 0.06484, 0.06484, 0; B(8b): 0.9401, 0.4249, 0.3648;
H1(8b): 0.0998, 0.4164, 0.4199; H2(8b): 0.959, 0.5845, 0.3101; H3(8b):
0.9311, 0.255, 0.3159; H4(8b): 0.7705, 0.4436, 0.4133.
3. c-Sr(BH4)2:
At temperatures above 450°C, t-Sr(BH4)2 transformed into the cubic
CaF2 type polymorph (c-Sr(BH4)2) [106]. Wtihin this structure the [BH4]⎺
units are embedded into the face centered Sr lattice (see Figure 7c) and the
cation coordination number increases from 6 to 8. The calculated Sr–B, B–
H, and Sr–H distances are 3.078, 1.13, and 2.904 Å, respectively.
Cell parameters: a = 7.10870 Å; b = 7.10870 Å; c = 7.10870Å; Atomic
positions: Sr(4a): 0, 0, 0; B(8c): 0.25, 0.25, 0.25; H(32f): 0.1583, 0.3419,
0.1583.
Ba(BH4)2
1. o1-Ba(BH4)2:
Ba(BH4)2 crystalize in orthorhombic Pnnm (o1-Ba(BH4)2) and it
transforms at 395°C into the o2-Ba(BH4)2 polymorph, isostructural to the
orthorhombic modification in Eu- and Sr-borohydrides [106]. At 445°C it
further transforms into the tetragonal t-Ba(BH4)2 polymorph. At 460°C the
cubic polymorph also appears. On cooling down to RT the tetragonal and
cubic polymorphs with 8-fold coordination of Ba2+ do not transform to the
orthorhombic polymorphs with octahedral coordination, due to the slow
kinetics as the cooling was rather fast (20°C/min).
Cell parameters: a = 6.98476 Å; b = 7.2284 Å; c = 4.8494 Å; Atomic
positions: Ba(2a): 0, 0, 0; B(4g): 0.1308, 0.2460, 3/6; H1(4g): 9582, 0.2160,
3/6; H2(4g): 0.2204, 0.0998, 1/2; H3(8h): 0.1689, 0.3344, 0.2913.
36 P. Vajeeston
2. o2-Ba(BH4)2:
In the o2-Ba(BH4)2 polymorph the octahedra share all six vertices and
two edges, thus building zigzag chains along the c-axis. Each chain is again
linked by vertices to four other chains. The borohydride tetrahedron is
surrounded by three Ba atoms and forms a nearly perfect trigonal planar
coordination. The Ba–B distances in this polymorph are between 3.127 and
3.289 Å. This structure type is the room temperature polymorph of Eu(BH4)2
and Sr(BH4)2 [106].
Cell parameters: a = 7.4043 Å; b = 8.0935 Å; c = 8.8201 Å; Atomic
positions: Ba(2a): 0.5,0.25, 0.1622; B(4g): 0.7617, 0.0768, 0.3906; H1:
0.7949, 0.1644, 0.4864; H2: 0.8525, -0.0352, 0.3976; H3: 0.6158, 0.0367,
0.4000; H4: 0.7835, 0.1414, 0.2784.
3. t-Ba(BH4)2:
Upon heating the - polymorph of o2-Ba(BH4)2 transformed into the
tetragonal t-Ba(BH4)2 polymorph, thus increasing its coordination number
from 6 to 8. It further ansformed at 718K into the P41212 polymorph [106].
This structure is a non-centrosymmetric superstructure (doubled along the
c-axis) of ht-ZrO2 type (P42/nmc), a deformation of the cubic CaF2 type
which is stabilized at even higher temperatures in ZrO2. The Ba atoms are
coordinated in this polymorph by eight [BH4]⎺ (Figure 7b) forming a cube,
and each borohydride is surrounded by four Ba atoms in a tetrahedral
coordination. Ba–B distances are between 3.052 and 3.499 Å [106].
Cell parameters: a = 5.5805 Å; b = 5.5805 Å; c = 12.5727 Å; Atomic
positions: Ba(4a): 0.0282, 0.0279, 0; B(8b): 0.9401, 0.4249, 0.3648; H1(8b):
0.0998, 0.4164, 0.4199; H2(8b): 0.9590, 0.5845, 0.3101; H3(8b): 0.9311,
0.2550, 0.3159; H4(8b): 0.7705, 0.4436, 0.4133.
4. c-Ba(BH4)2:
At higher temperatures above 773K, t-Ba(BH4)2 transformed into the
cubic CaF2 type polymorph (c-Ba(BH4)2). In this structure [BH4]⎺ units are
embedded into the face centered Sr lattice (see Figure 7c) and the cation
coordination number increases from 6 to 8 with the temperature. The
calculated B-H, and Ba-H distances are 1.13, and 1.941 Å, respectively.
Crystal Structures of Alkali and Alkaline Earth … 37
Most of the following alkali metal mixed tetra borides are characterized
with the help of XPD experiments for structure determination, and the final
structural refinement proceeds via first principles solid-state calculations,
that allowed us to identify minute structural details, such as the hydrogen
coordination mode in all compounds.
LiK(BH4)2
Nickels et al. report [107] the first synthesis and crystal structure
determination of a mixed alkali metal borohydride, LiK(BH4)2 from 1:1 ratio
mixture of LiBH4 and KBH4. Importantly in this new material, the observed
decomposition temperature lies between that of the constituent phases.
LiK(BH4)2 crystalize in space group Pnma with lattice parameters a = 7.9134
Å, b = 4.4907 Å, and c = 13.8440 Å (see Figure 8a). The b-axis lattice
parameter is very similar to that of orthorhombic LiBH4 (4.4368 Å [43]),
suggesting a degree of structural similarity between these phases. The
[BH4]⎺ units in LiK(BH4)2 form an approximately tetrahedral coordination
around the lithium ion, which is similar to that found in orthorhombic LiBH4.
The Li-B bonds are more elongated in LiK(BH4)2 than in LiBH4 but with a
narrower range of angles. The larger Li⋅⋅⋅B separations observed in the new
phase may originate from the presence of potassium cations in the structure,
which are considerably larger than their lithium counterparts (Li+ ionic
radius 0.59 Å, K+ ionic radius 1.38 Å in tetrahedral coordination). The
arrangement of the [BH4]⎺ units in LiK(BH4)2 and KBH4 differs
considerably. KBH4 has an octahedral arrangement of [BH4]⎺ units, whereas
those in LiK(BH4)2 might be best described as monocapped trigonal prisms.
The K⋅⋅⋅B distances in KBH4 are 3.364 Å, whereas in LiK(BH4)2 the (seven)
distances are 3.404 Å (twice), 3.409 Å (twice), 3.431 Å (twice), and 3.475
38 P. Vajeeston
Figure 8. Crystal structures of (a) LiK(BH4)2; (b) m-LiRb(BH4)2; (c) o-LiRb(BH4)2 and
(d) Li2Rb(BH4)3 (isostructural with Li2Cs(BH4)3).
Crystal Structures of Alkali and Alkaline Earth … 39
LiRb(BH4)2
1. m-LiRb(BH4)2:
The formation of monoclinic m-LiRb(BH4)2 is observed in the ball-
milled 1:1 ratio mixture of LiBH4 and RbBH4, together with unreacted
RbBH4 [108]. In the m-LiRb(BH4)2 structure, (4,4)-connected chains
[Li(BH4)2] are made from two 4-rings which share one Li in a square
coplanar coordination, the other Li atoms are in a tetrahedral [BH4]⎺
environment (see Figure 8b). Such chains do not occur in several other
oxides. In particular, the rare square planar coordination is observed in
intercalated lithium ion electrode materials [109]. The same environment is
found in the perovskite-type ionic conductor La0.5Li0.5TiO3 [110].
Substituting for the small Li+ here creates A-site vacancies which account
for the high mobility, while Li itself occupies a square planar position face-
centred on the cubic unit cell. This phase might be a potential candidate for
the Li-ion battery as a solid electrode.
Cell parameters: a = 23.5991 Å; b = 4.6178 Å; c = 17.893 Å; β=97.776o;
Atomic positions: Rb1(2c):0, 0, 0.5; Rb2(4i):0.2862, 0, 0.3584; Rb3(4i):
0.0684,0, 0.2099; Rb4(4i): 0.35229,0.5, 0.0686; Li1(2b): 0, 0.5, 0; Li2(4i):
0.435, 0, 0.264; Li3(4i): 0.203, 0.5, 0.144; Li4(4i): 0.163, 0.5, 0.429; B1(4i):
0.2495, 0, 0.1477; B2(4i): 0.1673, 0, 0.4844; B3(4i): 0.4706, 0, 0.1492;
B4(4i): 0.0314, 0.5, 0.3878; B5(4i): 0.4014,0.5, 0.3164; B6(4i): 0.1910, 0.5,
0.2985; B7(4i): 0.1107, 0.5, 0.083; H1(4i): 0.214,0, 0.185; H2(4i): 0.231, 0,
0.084; H3(8j): 0.2769,-0.199,0.161; H4(4i): 0.140, 0, 0.425; H5(4i): 0.214,
0, 0.477; H6(8j): 0.157, -0.199, 0.518; H7(4i): 0.5186, 0, 0.166; H8(8j):
0.452, -0.199, 0.174; H9(4i): 0.001, 0.5, 0.434; H10(4j): 0.006, 0.5, 0.327;
H11(8j): 0.0594, 0.301, 0.395; H12(4i): 0.365, 0.5, 0.266; H13(4i): 0.444,
0.5, 0.292; H14(8j): 0.399,0.301, 0.354; H15(4i): 0.145, 0.5, 0.269;
H16(4i): 0.221, 0.5, 0.252; H17(8j): 0.199,0.301, 0.3366; H18(4i):
0.067,0.5, 0.104; H19(4i): 0.105, 0.5, 0.017; H20(8j): 0.135, 0.301, 0.106;
H21(4i): 0.460, 0, 0.083.
40 P. Vajeeston
2. o-LiRb(BH4)2:
The monoclinic polymorph m-LiRb(BH4)2 transforms to o-LiRb(BH4)2
at 403 K, which then melts at 414 K. Recrystallization takes place upon
cooling at 409 K, and the high-temperature phase transforms back to m-
LiRb(BH4)2 at 375 K, and it is stable at room temperature [108]. The
hysteresis of 28 K of the m–o transformation is compliant with a first-order
character of the transformation in LiRb(BH4)2. In this structure puckered
[Li(BH4)2] ⎺ sheets built from vertex-sharing tetrahedra of both Rb+ counter-
cations (Figure 8c).
3. h-Li2Rb(BH4)3
The hexagonal phase Li2Rb(BH4)3 is observed in the ball-milled mixture
of LiBH4 and RbBH4 with 2:1 ratio [108]. The hexagonal phase Li2Rb(BH4)3
is stable up to 392 K where it melts and this is the stable polymorph of
Li2Rb(BH4)3 at room temperature. In this phase each Rb+ is surrounded by 8
H atoms (4 with 2.87Å and 4 with 2.2913 Å; Figure 8d) and Li is surrounded
by 6 H atoms with the distance vary from 1.864 to 2.256 Å.
Cell parameters: a = 7.474; b = 7.474Å; c = 11.585 Å; Atomic positions:
Rb(3c):0.5, 0.5, 1/3; Li(6i):0.85588, 1.7118, 0; B1(3b):0, 0, 0.5; B2(6f):0.5,
0, 0.2943; H1(12k):-0.03678, 0.10050, 0.55574; H2(12k):0.6366, 0.1022,
0.2386; H3(12k):0.5391,-0.0988, 0.3501.
4. o-LiCs(BH4)2
The nature of the system Li–Cs–BH4 is much more complex and
presumably subject to various metastable phase equilibria [108]. Schouwink
et al. studied the thermal phase stabilities of Li–Cs-BH4 phases and are
determined from in-situ powder diffraction are schematized for three
different samples LiBH4 and CsBH4 with 1:1, 2:1 and 3:1 ratios. LiBH4 and
Crystal Structures of Alkali and Alkaline Earth … 41
CsBH4 with 4:1 mixture, produced a single phase sample of the most Li-rich
phase Li3Cs(BH4)4 [108]. This phase is already a major phase in the mixture
3:1. Hence, this is the only of the stable phase in the Li-Cs-BH4 system.
LiCs(BH4)2 crystalize in orthorhombic Cmc21 space group [o-LiCs(BH4)2].
Similar to the o-LiRb(BH4)2 phase this structure also have puckered
[Li(BH4)2]– sheets built from vertex-sharing tetrahedra of both Cs+ counter-
cations (see Figure 9a).
Figure 9. Crystal structures of (a) LiCs(BH4)2; (b) Li2Cs(BH4)3; (c) Li3Cs(BH4)4; (d)
LiCs2(BH4)3; and (e) Li2Cs3(BH4)5 and (f) Li3Cs2(BH4)5.
5. h-Li2Cs(BH4)3
Li2Cs(BH4)3 crystalize in hexagonal [h-Li2Cs(BH4)3; P6222] and
monoclinic [m-Li2Cs(BH4)3; Cc] form (see Figure 8d and Figure 9b) [108].
The 3-connected triangular [Li2(BH4)3] – sheet in m-Li2Cs(BH4)3 has the
topology of a graphene-like Li sheet expanded by borohydride linkers.
Topologically this is called a honeycomb layer. In m-Li2Cs(BH4)3 the
triangular sheets are not linked by layers of larger coordination polyhedra.
Here the connectivity between sheets is ensured by Cs+ cations occupying
sites within the plane defined by the [Li2(BH4)3]– sheet. The symmetry of the
honeycomb layer [Li2(BH4)3] – is lowered over an orthorhombic c-centered
cell down to the monoclinic Cc, monoclinic angle β = 90o, by the rotation
and translation of [BH4]⎺ groups. Strong pseudosymmetry of the sheet in m-
Li2Cs(BH4)3 as a vertical mirror plane passing through Li atoms and leading
to the orthorhombic supergroup Cmc21. In m-Li2Cs(BH4)3 all BH4 groups
are linear linkers, bridging two Li atoms. This situation does not allow for a
general Li–H coordination scheme where all H–H distances are maximized.
Therefore, the monoclinic distortion has its origin in different rotations of
the three H–H edges from the vertical direction, and two symmetrically
independent Li(BH4)3 triangles. The [Li2(BH4)3]– net may be rationalized as
a frustrated net with respect to the ideal conrotation of the three vertical H–
H edges. Such a situation may promote specific dynamics of the [BH4]⎺
groups, such as 180o jump motions around the twofold operation of the 4
axes.
Cell parameters: a = 7.6634; b = 7.6634; c = 12.1279; Atomic positions:
Cs(3c): 0.5, 0.5, 1/3; Li(6i): 0.8614, 0.723, 0; B1(3b): 0, 0, 0.5; B2(6I): 0.5,
0, 0.3076; H1(12k): -0.0363, 0.099, 0.5542; H2(12k): 0.602, 0.134,0.2534;
H3(12k): 0.596, -0.041, 0.3618.
6. m-Li2Cs(BH4)3:
Cell parameters: a = 7.574; b = 7.948; c = 12.880; β = 89.97o; Atomic
positions: Cs(4a): 0.5, 0.3418, 0.83210; Li1(4a): 0.062, 0.657, 0.135;
Li2(4a): 0.07, 0.246, -0.013; B1(4a): 0.811, 0.600, 0.042; B2(4a): 0.360,
0.694, 0.114; B3(4a): 0.561, 0.794, 0.811; H1(4a): 0.88, 0.58, -0.038;
H2(4a): 0.88, 0.52, 0.108; H3(4a): 0.66, 0.56, 0.036; H4(4a): 0.82, 0.745,
Crystal Structures of Alkali and Alkaline Earth … 43
0.062; H5(4a): 0.27, 0.75, 0.178; H6(4a): 0.508, 0.71, 0.141; H6(4a): 0.34,
0.77, 0.035; H7(4a): 0.33, 0.550, 0.10; H8(4a): 0.52, 0.70, 0.747; H9(4a):
0.44, 0.82, 0.866; H10(4a): 0.61, 0.92, 0.771; H11(4a): 0.68, 0.74, 0.861.
7. m-LiCs2(BH4)3
LiCs2(BH4)3 is prepared mechano-chemically by ball-milling and
crystalize in monoclinic C2/c symmetry (m-LiCs2(BH4)3; see Figure 9d)
[108]. In this structure isolated chains [Li(BH4)3]2- of tetrahedra sharing two
vertices are found in m-LiCs2(BH4)3. The vertex-sharing tetrahedral chain
[Li(BH4)3]2- in m-LiCs2(BH4)3 is an analogue of the 2D-periodic (zweier)
chain underlying the architecture of inosilicates. This chains [Li(BH4)3]2- of
tetrahedra sharing two vertices. The average H···H distances is 2.1315 Å.
The Cs-H bond lengths vary from 2.654 to 3.48Å; Cs–B varies from 3.59 to
3.80Å.
Cell parameters: a = 7.6021; b = 15.160; c = 17.243; β = 92.693o;
Atomic positions: Cs1(8f): 0.2444, 0.3917, 0.0689; Cs2(8f): 0.7563, 0.2004,
0.1495; Li1(8f): 0.2422, 0.4902, 0.3285; B1(4e): 0.0, 0.0062, 0.25; B2(8f):
0.7453, 0.4078, 0.0448; B3(8f): 0.3060, 0.1763, 0.1324; B4(4e): 0.0, 0.4250,
0.25; H1(8f): 0.1299, 0.0530, 0.2563; H2(8f): 0.0065, -0.0409, 0.1928;
H3(8f): 0.8418, 0.3462, 0.0296; H4(8f): 0.8366, 0.4727, 0.0612; H5(8f):
0.6582, 0.3879, 0.1002; H6(8f): 0.6448, 0.4247, -0.0107; H7(8f): 0.3883,
0.2007, 0.0768; H8(8f): 0.3188, 0.2333, 0.1828; H9(8f): 0.3742, 0.1081,
0.1579; H10(8f): 0.1516, 0.1644, 0.1118; H11(8f): 0.1135, 0.4720, 0.2238;
H12(8f): -0.0622, 0.3785, 0.1978.
8. o-Li3Cs2(BH4)5
Li3Cs2(BH4)5 is prepared mechano-chemically by ball-milling and
crystalizes in orthorhombic (Fddd) symmetry [o-Li3Cs2(BH4)5; see Figure
9f] [108]. O-Li3Cs2(BH4)5 contains (4,3)-connected tetrahedral–triangular Li
sheets. The phase O-Li3Cs2(BH4)5, built from the sheets of [Li3(BH4)5]2-
containing tetrahedra and triangles in the ratio 1:1, and sharing all vertices
[108]. No other phase with this composition and topology was observed for
Rb containing borohydrides. The (4,3)-connected tetrahedral–triangular
sheet in o-Li3Cs2(BH4)5 may be compared to the tetrahedral–triangular
44 P. Vajeeston
9. m-Li2Cs3(BH4)5
Li2Cs3(BH4)5 is prepared mechano-chemically by ball-milling and
crystalize in monoclinic (P21/c) symmetry [m-Li2Cs3(BH4)5; see Figure 9e]
[108]. Increasing the Li content leads to the phase m-Li2Cs3(BH4)5
containing isolated chains [Li2(BH4)5]3- of tetrahedra sharing one edge and
one vertex. The edge- and vertex-sharing tetrahedral chain [Li2(BH4)5]3- in
m- Li2Cs3(BH4)5 does not find a match amongst inorganic compounds. In
this phase the average H···H distances is 1.89 Å. The Cs–H bond lengths
vary from 2.363 to 3.152 Å and Cs–B distance varies from 3.211 to 3.699
Å.
Cell parameters: a = 14.444 Å, b = 7.4218 Å, c = 14.7907 Å, β=104.470;
Atomic positions: Cs1(4e):0.2691, 0.747, 0.6341; Cs2(4e):0.0016, 0.784,
0.8502; Cs3(4e):0.5020, 0.248, 0.0988; Li1(4e):0.79, 0.75, 1.02;
Li2(4e):0.68, 0.83, 0.20; B1(4e):0.759, 0.68, 0.350; B2(4e):0.504, 0.85,
0.156; B3(4e):0.73, 0.55, 0.13; B4(4e):0.75, 0.03, 0.095; B5(4e):0.041,
Crystal Structures of Alkali and Alkaline Earth … 45
10. m-Li3Cs(BH4)3
Li3Cs(BH4)3 is prepared mechano-chemically by ball-milling of LiBH4
and CsBH4 with 4:1 ratio, which produced a single phase sample of the most
Li-rich phase of Li3Cs(BH4)4 [crystalize in monoclinic (P21/c) symmetry m-
Li3Cs(BH4)3; see Figure 9c] in the Li-Cs-BH4 mixture [108]. This phase is
already a major phase in the 3LiBH4:1CsBH4 mixture. This implies that the
maximum Li content in the phase diagram of LiBH4–CsBH4 is reached with
this phase. The framework [Li3(BH4)4]⎺ observed in the most Li-rich
compound m-Li3Cs(BH4)4 can be derived from the o-LiBH4 framework by
replacing a quarter of the Li atoms by Cs, thus breaking the infinite stripes
of edge-sharing tetrahedra into fragments of three tetrahedra. In this phase
the average H···H distances is 1.89 Å. The Cs–H bond lengths vary from
2.910 to 3.647 Å and Cs–B distance varies from 3.71 to 3.929 Å.
Cell parameters: a = 8.1992 Å, b = 12.3044 Å, c = 11.9009 Å,
β=123.767; Atomic positions: Cs(4e):0.1766, 0.17545, 0.1542;
Li1(4e):0.297, 0.927, 0.485; Li2(4e):0.718, 0.333, 0.276; Li3(4e):0.691,
1.010, 0.185; B1(4e):0.992,0.403, 0.261; B2(4e):0.654, 0.0902, 0.955;
B3(4e):0.568, 0.6284, 0.956; B4(4e):0.226, 0.8667, 0.677; H1(4e):0.976,
0.475, 0.196; H2(4e):0.136, 0.361, 0.299; H3(4e):0.867, 0.344, 0.200;
H4(4e):0.991, 0.433, 0.350; H5(4e):0.713, 0.023, 0.922; H6(4e):0.599,
0.159, 0.880; H6(4e):0.774, 0.120, 0.058; H7(4e):0.529, 0.057, 0.959;
H8(4e):0.51, 0.67, 0.01; H9(4e):0.54, 0.68, 0.87; H10(4e):0.49, 0.55, 0.92;
H11(4e):0.73, 0.61, 0.03; H12(4e):0.241, 0.935, 0.621; H13(4e): 0.373,
0.850, 0.774; H14(4e):0.17, 0.791, 0.612; H15(4e):0.116, 0.890, 0.702.
46 P. Vajeeston
K2Mg(BH4)4
Figure 10. Crystal structures of (a) K2Mg(BH4)4; (b) LT-K3Mg(BH4)5 (c) and HT-
K3Mg(BH4)5. The partial occupancy was marrkes as duel colour.
K3Mg(BH4)5
1. LT-K3Mg(BH4)5:
Similarly to K2Mg(BH4)4, K3Mg(BH4)5 is also prepared mechano-
chemically by ball-milling [113]. The structure of K3Mg(BH4)5 is tetragonal
(P42bc) and has an intriguing nature that can be described as an intergrowth of
KBH4 and K2Mg(BH4)4. Structure of K3Mg(BH4)5 is closely related to
Cs3CoCl5, which crystallizes in space group I4/mcm. The loss of body
centering for K3Mg(BH4)5 is probably caused by the non-spherical nature of
[BH4]⎺ occupying the chlorine positions of the prototype Cs3CoCl5.
Interestingly, the [BH4]⎺ group occurs in two different environments,
forming similar complex anions [Mg(BH4)4]2– as observed in the structure
of K2Mg(BH4)4, on one hand, while the second boron position only
coordinates to potassium. This [BH4]⎺ is octahedrally coordinated, strongly
resembling the local environment of [BH4]⎺ in KBH4. The B–Mg–B angle
with its bisector along c is 96.64° and the average H···H distances is 2.1315
Å. Average bond lengths Mg–B of the complex anion is 2.4369 Å.
Cell parameters: a = 8.9693 Å; b = 8.9693 Å; c = 15.9501Å; Atomic
positions: K1(8h): 0.32714, 0.67704, 0; K2(8f): 0, 0.5, 0.72739 (50%
occupancy); Mg(4b): 0, 0, 0.75; B1(16i): 0.1181, 0.835, 0.1484; B2(8f): 0,
0.5, 0.482 (50% occupancy); H1(16i): 0.225, 0.76, 0.16; H2(16i): 0.021,
Crystal Structures of Alkali and Alkaline Earth … 49
2. HT-K3Mg(BH4)5:
At higher temperature (at 367K) tetragonal K3Mg(BH4)5 (P42bc)
transform into another tetragonal high symetric P42/mbc phase. The B–Mg–
B angle with its bisector along c is 95.02° and the average H···H distances
is 2.1 Å. Average bond lengths Mg–B of the complex anion is 2.4079 Å.
Cell parameters: a = 8.9693 Å, b = 8.9693 Å, c = 15.9501 Å; Atomic
positions: K1(8h):0.3271, 0.6770,0; K2(8f):0, 0.5, 0.7274; Mg(4b):0, 0,
0.75; B1(16i):0.1181, 0.8351,0.1484; B2(8f):0, 0.5, 0.4821; H1(16i):0.2259,
0.761, 0.161; H2(16i):0.021, 0.796,,0.1928; H3(16i):0.081, 0.8235, 0.081;
H4(16i):0.1463, 0.961, 0.161; H5(16i):-0.1042, 0.4739, 0.5251; H6(16i):-
0.0279, 0.6042, 0.4401.
KCa(BH4)3
1. LT-KCa(BH4)3:
At room temperature, KCa(BH4)3 stabilizes in an orthorhombic structure
(LT-KCa(BH4)3; Pba2; BH4 orientationally ordered) and crystallizes with
the octahedral rotation pattern a0a0c+, which is characteristic of
centrosymmetric tetragonal P4/mbm perovskites [117]. In the symbol a0a0c+,
superscripts refer to the sense of rotation and letters to the axis of rotation.
The space group symmetry P4/mbm, however, is lowered because of the
ordering of hydrogen atoms to the polar space group Pba2.
50 P. Vajeeston
2. HT-KCa(BH4)3
Before the temperature-induced phase transition at 340 K, KCa(BH4)3
approaches tetragonal symmetry, following the intuitive structural
evolution of perovskites. The transition, however, is orthorhombic–
orthorhombic, hence without symmetry increase. The unit cell axis c is
doubled in the HT-polymorph (HT-KCa(BH4)3; space group Pbn21),
introducing a further octahedral tilt (a−a−c+), characteristic of the
orthorhombic Pnma perovskites. Space group symmetry is lowered to
Pbn21, again due to hydrogen ordering [115].
The lattice instabilities and structural dynamics properties of LT-/HT-
KCa(BH4)3 described in the literature [115, 118]. But the structural details
(atomic positions) are not reported in the literature including the structural
data bases [119, 120].
3. α-KSr(BH4)3
The KSr(BH4)3 was synthesized via an addition reaction between Sr(BH4)2
and KBH4 by mechanochemical treatment. At RT KSr(BH4)3 crystalizes into
orthorhombic P21cn [121]. The structures of α-KSr(BH4)3 are built from
distorted octahedra of [Sr(BH4)6], which contain tetrahedral borohydride anions,
[BH4]−. Each [Sr(BH4)6] octahedron shares all corners with other [Sr(BH4)6]
octahedra, thereby forming a three-dimensional framework, while BH4 units are
linearly coordinated to two Sr atoms by edge sharing, as imposed by the
perovskite structure type, see Figure 11a. In KSr(BH4)3, the Sr–B distance varies
between 2.710–3.537 Å; the bond angles Sr–B–Sr are vary from 119.001° to
158.792° and B–Sr–B vary from 70.619° to 176.924°.
Cell parameters: a = 11.5078 Å, b = 8.2953 Å, c = 7.8967 Å; Atomic
positions: K(4a): 0.23979, 0.54281,0.96526; Sr(4a): 0.46099, 0.50716,
0.48652; B1(4a): 0.19014, 0.99238, 0.09103; B2(4a): 0.55245, 0.88349,
0.32359; B3(4a): 0.48618, 0.21503, 0.69107; H1(4a): 0.12999, 0.97787,
0.20842; H2(4a): 0.16173, 0.10055, 0.01475; H3(4a): 0.18850, 0.88432,
0.01575; H4(4a): 0.28312, 1.01438, 0.14054; H5(4a): 0.55954, 0.81557,
0.44048; H6(4a): 0.62484, 0.85705, 0.24562; H7(4a): 0.54852, 0.01656,
0.36290; H8(4a): 0.47102, 0.84315, 0.25868; H9(4a): 0.51842, 0.30082,
Crystal Structures of Alkali and Alkaline Earth … 51
4. Rb3Mg(BH4)5
The diffraction peaks originating from Rb3Mg(BH4)5 have been indexed
in tetragonal unit cells with an extinction class of I4cm. K3Mg(BH5)5
crystallizes in a similar unit cell, however, in this case a loss of body
centering is indicated by a set of weak Bragg signals observed below 367 K
[122]. This structure has been described as an intergrowth of RbBH4 and
Rb2[Mg(BH4)4]; it contains the [Mg(BH4)4]2⎺ anions (where Mg2+ is
coordinated tetrahedrally by four bidentate [BH4]⎺ groups) and the
disordered [BH4]⎺ anions (coordinated octahedrally by Rb+ cations, like in
some RbBH4 phases). The geometry of the [Mg(BH4)4]2⎺ anions for Rb-
containing compounds closely resembles that found in the K analogue; the
shortest Mg–B distance is 2.434 Å. The equatorial M1–B2 distance is 3.4402
Å and the apical Rb2–B2 distance is significantly larger than the equatorial
one, 3.9983 Å.
Cell parameters: a =9.2996 Å, b = 9.2996; c = 15.993 Å; Atomic
positions: Rb1(8h): 0.67310, 1.17310, 0; Rb2(4a): 0, 0, ¼; Mg(4b): 0, ½,
¼; B1(16l): 0.1400, 0.6400(3), 0.1499; B2(4c): 0, 0, 0; H1(32m): 0.02416,
0.66693, 0.16947; H2(16l): 0.1509; 0.6509; 0.0785; H3(16l): 0.21825;
0.71825; 0.1821; H4(32m): -0.10097(4), 0.00000, 0.0415 (50% occupancy).
5. RbCa(BH4)3
The ionic radius of Rb+ lies in between those of K+ and Cs+, and the
transformation-sequence of RbCa(BH4)3 unifies atomic displacements and
order–disorder mechanisms. Identically to CsCa(BH4)3, the RT-polymorph
of RbCa(BH4)3 is described as cubic Fm-3c (mistakenly described as P-43m
in [115]) [118]. The cubic-orthorhombic transformation at 370 K, results in
a 16-fold supercell metric of HT1-RbCa(BH4)3, space group A2122, which
has no precedence in the literature and it is stable only in a small temperature
range of approximately 30 K (370–400 K). The commensurate modulation
in the HT1-polymorph is attributed to H···H interactions, and is
accompanied by a negative volume anomaly of –1.2%. The subsequent
52 P. Vajeeston
6. RbSr(BH4)3
RbSr(BH4)3 wasa synthesized via an addition reaction between Sr(BH4)2
and RbBH4 by mechanochemical treatment [121]. At RT RbSr(BH4)3
crystalizes into orthorhombic Pbn21. The structures of RbSr(BH4)3 are built
from distorted octahedra of [Sr(BH4)6], which contain tetrahedral
borohydride anions, [BH4] ⎺. Each [Sr(BH4)6] octahedron shares all corners
with other [Sr(BH4)6] octahedra, thereby forming a three-dimensional
framework, while [BH4]⎺ units are linearly coordinated to two Sr atoms by
edge sharing, as imposed by the perovskite structure type, see Figure 11b.
In RbSr(BH4)3, the Sr–B distance varies between 2.878–3.165 Å; the bond
angles Sr–B–Sr are vary from 147.141° to 158.107° and B–Sr–B vary from
74.5° to 178.606°.
Cell parameters: a = 8.3341; b = 8.0835; c = 11.6600; Atomic
positions: Rb(4a): 0.5136, 0.0126, 0.7428; Sr(4a): 0, 0,
0;B1(4a):0.3246,0.7129,0.0665; B2(4a): 0.71156, 0.1951, 0.01258; B3(4a):
0.99067, 0.1057, 0.23879; H1(4a): 0.4506, 0.7579, 0.0995; H2(4a): 0.2606,
0.6499, 0.1385; H3(4a): 0.2536, 0.8139, 0.0325; H4(4a): 0.3536, 0.6179,
0.9955; H5(4a): 0.7512, 0.1787, -0.0839; H6(4a): 0.7594, 0.0962,4 0.0662;
H7(4a): 0.7565, 0.3132, 0.0501; H8(4a): 0.5771, 0.1929, 0.0217; H9(4a):
0.9946, -0.0323, 0.2587; H10(4a): 0.9210, 0.1727, 0.30901; H11(4a):
0.1183, 0.1564, 0.2347; H12(4a): 0.9287, 0.1259, 0.15271.
Crystal Structures of Alkali and Alkaline Earth … 53
7. CsMg(BH4)3
The CsMg(BH4)3 structure is identical to that found for a nano-
crystalline monoclinic high-pressure polymorph of ReO3 [123] and an
intermediate low-temperature modification of BaTiO3 [124]. The supercell
metric (√6/√2/√2) have been reported for CsMg(BH4)3 and no subcell to the
superstructure is found in the investigated temperature range between 100 K
and its decomposition [118]. In the CsMg(BH4)3 phase the Mg-atom is
shifted off-centre towards a square pyramidal coordination. The lattice
instabilities and structural dynamics properties of CsMg(BH4)3 are described
in the literature [115, 118]. The structural details (atomic positions)
are however not found in the literature including the structural data bases
[119, 120].
8. Cs3Mg(BH4)5
The diffraction peaks originating from Cs3Mg(BH4)5, have been indexed
to a tetragonal unit cells with an extinction class of I4cm [122]. K3Mg(BH5)5
also crystallizes in a similar unit cell, however, in this case a loss of body
centering is indicated by a set of weak Bragg signals observed below 367 K
[113]. The geometry of the [Mg(BH4)4]2⎺ anions for Cs-containing
compounds closely resembles that found in the K and Rb analogue; the
shortest Mg–B distance is 2.410 Å. The equatorial Cs1–B2 distance is
3.6003 Å and the apical Cs2–B2 distance is significantly larger than the
equatorial one, 4.0631 Å.
Cell parameters: a = 9.7115; b = 9.7115(4); c = 16.2540; Atomic
positions: Cs1(8h): 0.6711, 0.1711,0; Cs2(4a): 0, 0, 0.25; Mg(4b):0, 0.5,
0.25; B1(16l): 0.1370,0.6370, 0.1581; B2(4c): 0,0,0; H1(32m): 0.02031,
0.6570, 0.1583; H2(16l): 0.1852, 0.6852, 0.1012; H3(16l): 0.1852, 0.6852,
0.2162; H4(32m): -0.0967, 0.0408, 0.5.
54 P. Vajeeston
CsCa(BH4)3
1. LT-CsCa(BH4)3:
LT-CsCa(BH4)3 crystallizes in space group Fm-3c in the as-milled
samples [118]. This corresponds to the symmetry of the high-temperature
modification of Y(BH4)3 [125]. The high-temperature polymorph is
described as fully disordered Pm-3m, explaining the absence of
superstructure reflections in the low-temperature phase [115]. The ordered
polymorph of CsCa(BH4)3 is an 8-fold supercell of the fully disordered one.
The space-groups Fm-3c and Pm-3m are thus related by doubling of the
cubic cell parameter. The relationship between the F-centred and the
primitive cubic unit cell can account for an identical x-ray diffraction
pattern, in spite of different systematic extinctions, as was the case for
Y(BH4)3. The origin of the unit-cell axis doubling is therefore exclusively
related to the positional ordering of [BH4]⎺ groups, which occupy a
disordered position 4/mm.m (3d) in the Pm-3m HT-phase, and order on
position -4m.2 (24c) in the Fm-3c LT-phase [115, 118].
The lattice instabilities and structural dynamics properties of
CsCa(BH4)3 are described in the literature [115, 118]. The structural details
(atomic positions) are however not found in the literature including the
structural data bases [119, 120].
2. CsSr(BH4)3
CsSr(BH4)3 was synthesized via an addition reaction between Sr(BH4)2
and CsBH4 by mechano-chemical treatment [121]. At RT, CsSr(BH4)3
crystalize into orthorhombic P22121 [121]. The structures of α-CsSr(BH4)3
are built from distorted octahedra of [Sr(BH4)6], which contain tetrahedral
[BH4]⎺ anions, each [Sr(BH4)6] octahedron shares all corners with other
[Sr(BH4)6] octahedra, thereby forming a three-dimensional framework,
while [BH4]⎺ units are linearly coordinated to two Sr atoms by edge sharing,
as imposed by the perovskite structure type, see Figure 11d. In RbSr(BH4)3,
the Sr–B distance varies between 2.84–3.24 Å; the bond angles Sr–B–Sr are
vary from 149.7° to 180° and B–Sr–B vary from 82.4° to 180°. The bond
angles found in CsSr(BH4)3 are more similar to the ones in Sr(BH4)2, while
Crystal Structures of Alkali and Alkaline Earth … 55
3. LiKMg(BH4)4
LiKMg(BH4)4 is built up of single-layered sheets organized in an ABAB
stacking sequence along the [010] direction of the standard setting Aba2
[126]. Ignoring K+ counter-cations we can rationalize a single sheet as
[LiMg(BH4)4]−. Li and Mg atoms are arranged in a pseudohexagonal
symmetry (Figure 12a) and are both tetrahedrally coordinated by [BH4]⎺.
This leads to two different boron sites, both on the general position 8b in
Aba2. One position links Mg/Li(BH4)4 tetrahedra via common corners and
56 P. Vajeeston
the other via edges, forming dimeric units. Since double-bonds are counted
as a single link in topology, these sheets correspond to honeycomb layers.
The edge-sharing BH4 groups are located outside the plane defined by the
sheet of pseudohexagonally arranged metal atoms. The corner-linking
[BH4]⎺ groups come to lie within the plane. If we consider a dimeric unit of
[LiMg(BH4)6]3−, then the distances between metal atom centers and those
Bb-atoms that form bridging [BH4]⎺ groups within the unit are longer than
those to the terminal groups Bt, which link to the next dimer and in this way
form the rings from which the sheet is built. For the experimental structure
the refined distances Li–Bb and Li–Bt are 2.56 and 2.54 Å, and the respective
values for Mg–Bb and Mg–Bt are 2.52 and 2.37 Å. The K position is located
in the center of the rings in a distorted cubic coordination by [BH4]⎺. The
single-layered sheets are interconnected by K–BH4 interactions with the
[BH4]⎺ groups lying out of plane.
Cell parameters: a = 8.16633; b = 9.1679; c = 12.5395; Atomic
positions: Li1(4a):0, 0, 0.0538(85.3% occupancy); Li2(4a):0, 0, 0.33645
(14.7% occupancy); K(4a):0, 0, 0.69254; Mg1(4a): 0, 0, 0.0538 (14.7%
occupancy); Mg2:0, 0, 0.0538(85.3% occupancy); B1(8b): 0.7384, 0.5090,
0.4444; B2(8b): 0.60323, 0.8235,0.6973; H1(8b): 0.6741, 0.5952, 0.3954;
H2(8b):0.8444, 0.4629, 0.3972; H3(8b):0.6495, 0.4186, 0.4643;
H4(8b):0.7858, 0.5594, 0.5206; H5(8b):0.5334, 0.7823, 0.6256;
H6(8b):0.6486, 0.9379, 0.6812; H7(8b):0.7111, 0.7496, 0.7133; H8(8b):0
0.5198, 0.8245, 0.7692.
Figure 12. Crystal structures of (a) LiKMg(BH4)4; (b) Li2K3Mg2(BH4)9 (c) and
LiRbMg(BH4)4.
Crystal Structures of Alkali and Alkaline Earth … 57
4. Li2K3Mg2(BH4)9
Li2K3Mg2(BH4)9 is made up of double-layered sheets which are stacked
along the monoclinic axis of the space group P2/c (Figure 12b). The new
compound is formed by a reaction between KLiMg(BH4)4 and the KBH4
from molten/decomposed LiK(BH4)2. The Li : Mg ratio is still 1 : 1, which
means that the cation-sequence within a sheet should still correspond to Li–
Mg–Li–Mg [126]. The structure contains one Li and one Mg atom, both
occupying the general position 4g. The sheets contain 4-fold rings that are
interrupted by terminal ligands. The 4-fold rings are interconnected between
both layers of the same double-layered sheet via a Li–B–Mg sequence.
However, within the layer, rings are connected via Li–B–Li within the layer,
while the Mg-atoms have terminal BH4 groups. The potassium atoms occupy
three different positions, 2a, 2c and 2f. 2a and 2f are distorted cubes, sharing
edges in chains along c, while 2c is octahedrally coordinated and forms one-
dimensional octahedral chains connected by vertices in the c direction. Each
octahedron is connected to 2 adjacent octahedra in trans-position and shares
common edge with four cubes in the equatorial plane. There are five
crystallographically distinct [BH4]⎺ groups in the structure, all of which are
bridging ligands, with the exception of B2.
Cell parameters: a = 11.3307; b = 10.20598; c = 11.74449; β = 121.210o;
Atomic positions: Li(4g): 0.2136, 0.7561, 0.7986; K1(2f): 1/2, 0.0005, 3/4;
K2(2c): 0, 1/2, 0; K3(2a): 0, 0, 0; Mg(4g): 0.30230, 0.71402, 0.4039;
B1(4g): 0.2359, 0.7963, 0.5496; B2(4g): 0.2097, 0.2229, 0.6604; B5(2e): 0,
0.1428, 1/4; B3(4g): 0.7725, 0.4899, 0.1175; B4(4g): 0.5342, 0.8060,
0.5194; H1(4g): 0.3332, 0.7392, 0.5782; H2(4g): 0.2242, 0.8042, 0.6391;
H3(4g): 0.1432, 0.7443, 0.4662; H4(4g): 0.2443, 0.8978, 0.5153; H5(4g):
0.1573, 0.190, 0.7142; H6(4g): 0.2094, 0.1401, 0.5963; H7(4g): 0.3202,
0.2515, 0.7352; H8(4g): 0.1545, 0.3103, 0.5963; H9(4g): 0.8518,
0.5651,0.1832; H10(4g): 0.8231, 0.3901, 0.1382; H11(4g): 0.7292, 0.5172,
0.0096; H12(4g): 0.6861, 0.4882, 0.1391; H13(4g): 0.5122, 0.7013, 0.4832;
H14(4g): 0.5812, 0.8602, 0.4682; H15(4g): 0.6092, 0.8071, 0.6305;
H16(4g): 0.4351, 0.8551, 0.4962; H17(4g): -0.0262, 0.2067, 0.1615;
H18(4g): 0.09166, 0.0789, 0.2752.
58 P. Vajeeston
5. LiRbMg(BH4)4
LiRbMg(BH4)4 is built up of single-layered sheets organized in an
ABAB stacking sequence along the [010] direction of the standard setting
Aba2. Ignoring Rb+ counter-cations we can rationalize a single sheet as
[LiMg(BH4)4]−. Li and Mg atoms are arranged in a pseudohexagonal
symmetry and are both tetrahedrally coordinated by [BH4]⎺. This leads to
two different boron sites, both on the general position 8b in Aba2. One
position links Mg/Li(BH4)4 tetrahedra via common corners and the other via
edges (see Figure 12c), forming dimeric units. Since double-bonds are
counted as a single link in topology, these sheets correspond to honeycomb
layers. The edge-sharing [BH4]⎺ groups are located outside the plane defined
by the sheet of pseudohexagonally arranged metal atoms. The corner-linking
[BH4]⎺ groups come to lie within the plane. If we consider a dimeric unit of
[LiMg(BH4)6]3−, then the distances between metal atom centers and those
Bb-atoms that form bridging [BH4]⎺ groups within the unit are longer than
those to the terminal groups Bt, which link to the next dimer and in this way
form the rings from which the sheet is built. For the experimental structure
the refined distances Li–Bb and Li–Bt are 2.56 and 2.54 Å, and the respective
values for Mg–Bb and Mg–Bt are 2.44 and 2.42 Å. The Rb position is located
in the center of the rings in a distorted cubic coordination by [BH4]⎺ The
single-layered sheets are interconnected by Rb–BH4 interactions with the
[BH4]⎺ groups lying out of plane.
Cell parameters: a=8.2342; b= 9.4852; c= 12.5767; Atomic positions:
Rb(4a): 0, 0, 0.70087; Li1(4a): 0, 0.0637, 0.606; Li2(4a): 0, 0.3411, 0.394;
Mg1(4a): 0, 0.0637, 0.394; Mg2(4a): 0, 0.3411, 0.606; B1(8b):0.59950,
0.8333, 0.6960; B2(8b): 0.7726, 0.5366, 0.4453; H1(8b): 0.854, 0.6221,
0.4781; H2(8b): 0.752, 0.5552, 0.3576; H3(8b): 0.6522, 0.538, 0.4884;
H4(8b): 0.832, 0.4305, 0.4571; H5(8b): 0.529, 0.80, 0.7700; H6(8b): 0.524,
0.81, 0.6233; H7(8b): 0.72, 0.7720, 0.6932; H8(4a): 0.63, 0.9505, 0.697.
6. RbMgxCa1-x(BH4)4
The compound RbMgxCa1-x(BH4)3 was identified as a by product of the
RbBH4+Mg(BH4)2+Ca(BH4)2 mixture the cell dimension and the space
group number is reported in Ref. [115]. The structural details (atomic
Crystal Structures of Alkali and Alkaline Earth … 59
positions) are however not found in the literature including the structural
data bases [119, 120].
Remark
REFERENCES
[108] Schouwink P, Smrcok L, Cerny R. Role of the Li+ node in the Li-
BH4 substructure of double-cation tetrahydroborates. Acta
Crystallogr B. 2014;70(5):871-8.
[109] Perez-Flores JC, Baehtz C, Hoelzel M, Kuhn A, Garcia-Alvarado F.
Full structural and electrochemical characterization of Li2Ti6O13 as
anode for Li-ion batteries. Physical chemistry chemical physics:
PCCP. 2012;14(8):2892-9.
[110] Alonso JA, Sanz J, Santamaría J, León C, Várez A, Fernández-Díaz
MT. On the Location of Li+ Cations in the Fast Li-Cation Conductor
La0.5Li0.5TiO3 Perovskite. Angew Chem Int Ed 2000;39(3):619-21.
[111] Ravnsbæk DB, Sørensen LH, Filinchuk Y, Besenbacher F, Jensen
TR. Screening of Metal Borohydrides by Mechanochemistry and
Diffraction. Angew Chem Int Ed 2012;51(15):3582-6.
[112] Černý R, Schouwink P, Sadikin Y, Stare K, Smrčok Lu, Richter B, et
al. Trimetallic Borohydride Li3MZn5(BH4)15 (M = Mg, Mn)
Containing Two Weakly Interconnected Frameworks. Inorg Chem
2013;52(17):9941-7.
[113] Schouwink P, D’Anna V, Ley MB, Lawson Daku LM, Richter B,
Jensen TR, et al. Bimetallic Borohydrides in the System M(BH4)2–
KBH4 (M = Mg, Mn): On the Structural Diversity. J Phys Chem C.
2012;116(20):10829-40.
[114] Aeberhard PC, Refson K, Edwards PP, David WIF. High-pressure
crystal structure prediction of calcium borohydride using density
functional theory. Phys Rev B. 2011;83(17):174102.
[115] Schouwink P, Ley MB, Tissot A, Hagemann H, Jensen TR, Smrčok
Ľ, et al. Structure and properties of complex hydride perovskite
materials. Nat Commun. 2014;5:5706.
[116] Černý R, Ravnsbæk DB, Schouwink P, Filinchuk Y, Penin N,
Teyssier J, et al. Potassium Zinc Borohydrides Containing Triangular
[Zn(BH4)3]⎺ and Tetrahedral [Zn(BH4)xCl4–x]2– Anions. J Phys
Chem C. 2012;116(1):1563-71.
[117] Glazer A. The classification of tilted octahedra in perovskites. Acta
Crystallogr B. 1972;28(11):3384-92.
Crystal Structures of Alkali and Alkaline Earth … 75
[137] Broach RW, Chuang IS, Marks TJ, Williams JM. Metrical
characterization of tridentate tetrahydroborate ligation to a transition-
metal ion. Structure and bonding in Hf(BH4)4 by single-crystal
neutron diffraction. Inorg Chem 1983;22:1081-4.
[138] Hoekstra HR, Katz JJ. The Preparation and Properties of the Group
IV-B Metal Borohydrides. J Am Chem Soc. 1949;71(7):2488-92.
[139] Bernstein ER, Hamilton WC, Keiderling TA, La Placa SJ, Lippard
SJ, Mayerle JJ. 14-Coordinate uranium(IV). Structure of uranium
borohydride by single-crystal neutron diffraction. Inorg Chem
1972;11(12):3009-16.
[140] Charpin P, Nierlich M, Vigner D, Lance M, Baudry D. Structure of
the second crystalline form of uranium(IV) tetrahydroborate. Acta
Crystallogr C. 1987;43(8):1465-7.
[141] Banks RH, Edelstein NM, Spencer B, Templeton DH, Zalkin A.
Volatility and molecular structure of neptunium(IV) borohydride. J
Am Chem Soc. 1980;102(2):620-3.
[142] Hagemann H, Longhini M, Kaminski JW, Wesolowski TA, Černý R,
Penin N, et al. LiSc(BH4)4: A Novel Salt of Li+ and Discrete
Sc(BH4)4− Complex Anions. J Phys Chem A. 2008;112(33):7551-5.
[143] Jaroń T, Wegner W, Fijałkowski KJ, Leszczyński PJ, Grochala W.
Facile Formation of Thermodynamically Unstable Novel
Borohydride Materials by a Wet Chemistry Route. Chem Eur J.
2015;21(15):5689-92.
[144] Černý R, Chul Kim K, Penin N, D’Anna V, Hagemann H, Sholl DS.
AZn2(BH4)5 (A = Li, Na) and NaZn(BH4)3: Structural Studies. J Phys
Chem C 2010;114(44):19127-33.
[145] Dovgaliuk I. PhD thesis: Synthesis, structure and properties of Al-
based borohydrides for hydrogen storage. http://hdl.handle.net/
2078.1/169194: Prom.: Filinchuk, Yaroslav; Devillers, Michel; 2015.
[146] Lindemann I, Domènech Ferrer R, Dunsch L, Filinchuk Y, Černý R,
Hagemann H, et al. Al3Li4(BH4)13: A Complex Double-Cation
Borohydride with a New Structure. Chem Eur J. 2010;16(29):8707-
12.
78 P. Vajeeston
Chapter 2
ABSTRACT
*
Corresponding Author address. Email: ponniahv@kjemi.uio.no Web: folk.uio.no/ponniahv.
82 P. Vajeeston and H. Fjellvåg
INTRODUCTION
Though the periodic table has only 118 or so elements, there are
obviously more substances in nature than 118 pure elements. Compounds
are formed when two or more atoms chemically bond together, the resulting
compound is unique both chemically and physically from its parent atoms.
A stable compound occurs when the total energy of the combination has
lower energy than the constituent atoms. The properties of a compound are
determined by the chemical bonds that hold the constituents together. A
Bonding Nature and Its Implications with the Energy Storage … 83
be noted that a system with strong ionic bond usually possesses high
symmetric structures.
Another major type of bonding in hydride is the covalent category where
electrons are equally shared between atoms. As opposite to ionic bonding in
which a complete transfer of electrons occurs, covalent bonding occurs when
two (or more) atoms share their electrons. An example of this type of
bonding is found in diamond (Figure 1b, h). Hydrogen is covalently bonded
to the neighbours in most of the organic molecules. The covalent bonding
has a quantum mechanical origin and occurs because the atoms in a given
compound have a similar “desire” for electrons. In order to have covalent
interaction between electrons they should be energetically and spatially
degenerate and also their spins should be “opposite”. Unlike ionic
compounds, covalent molecules exist as true molecules. Because electrons
are shared in covalent bond, it is the strongest bond in solids. Further, due to
the sharing of electrons, there is no net charge build-up in atomic sites and
hence covalent molecules are not strongly attracted to one another. As a
result, covalent molecules move about freely and tend to exist as liquids or
gases at room temperature. As the covalent bond is having directional
character, the covalently bonded solids often possess low symmetric
structures.
The third type of bonding is called metallic, where the valence electrons
are delocalized, i.e., homogeneously distributed in the space in between the
atoms. The properties of metals suggest that their atoms possess strong
bonds, yet the ease of conduction of heat and electricity suggest that
electrons can move freely in all directions in a metal. The general
observations give rise to a picture of “positive ions in a sea of electrons” to
describe metallic bonding. Examples of this type of bonding is found most
typically in alkali metals where each atom contributes only one electron to
the common pool (see Figure 1c, i, o for Na as an example). The materials
with metallic bond usually have metallic behaviour i.e., they possess good
thermal and electrical conductivity as well as Fermi surface in contrast to
ionic or covalent solids where semiconducting or insulating behaviour
prevails.
Bonding Nature and Its Implications with the Energy Storage … 85
classification does not always clarify the typical features of the compound
in question. For example, lithium hydride is classified as a saline hydride,
whereas it actually exhibits some covalent characteristics. The rare-earth
hydrides are usually classified under the metallic hydrides, whereas they
exhibit some characteristics similar to those of volatile and saline hydrides
(e.g., noticeable charge transfer interaction with high heat of formation). In
solids, hydrogen can occupy the interstitial or substitutional sites depending
upon the chemical environment which changes the bonding behaviour
substantially.
Ionic Hydrides
Covalent Hydrides
Covalent hydrides are found in the solid, liquid or gaseous phases. The
bonding between hydrogen and the other constituent(s) in this class of
hydride is characterized by sharing of the valence electrons on a fairly equal
basis. In general, molecules with covalent bonding are not strongly attracted
to each other and the absence of strong intermolecular forces results in a
degree of volatility and a low melting point. Covalent hydrides are generally
somewhat thermally unstable and this instability increases with increasing
atomic weight of the non-hydrogen element(s). In covalent hydrides,
hydrogen can formally be regarded to appear in the oxidation states +1 or –
1 of which +1 is the most abundant occurrence (as in the single bond between
O and H in H2O). Electron deficient molecules usually have hydrogen in the
formal valence state –1 and these are also known to form hydrogen bridge
bonds similar to those found in diborane (B2H6). Typical covalent hydrides
are carbo-hydrides, borohydrides, germanium hydrides, etc. The kinetics of
hydrogenation and dehydrogenation reactions among members in this group
are usually slow and generally follow a few specific mechanisms if at all
any. Metastable forms of covalent hydrides are to be considered common
observations rather than exceptions. Covalent hydrides generally exhibit low
symmetric structures. In general, most of the complex hydrides (with higher
H wt%) are having dominant covalent characters.
Metallic Hydrides
of metallic hydrides may also be distinctly different from those of the parent
matrices. Metal hydrides are characterized by a rather high hydrogen bulk
diffusion rate and hence they find application in rechargeable batteries.
Many of these materials are exceedingly complicated, showing
compositional non-stoichiometry and sometimes complex phase relations.
Several metal hydrides exhibit a low-hydrogen-concentration phase where
the hydrogen ions occupy a small portion of the available interstitial sites.
The lattice strain in such phases is low enabling the metal to retain its
macroscopic shape. However, larger percentage of hydrogen filling into the
interstitial sites brings lattice stresses which results in the formation of
micro-meter sized hydride particles. Since the metals which form the metal
hydrides are heavy and take low valence states (due to the low
electronegativity of hydrogen), the result is a low weight percentage of
stored hydrogen. The energy storage devices based on metal hydrides are
consequently not suitable for mobile hydrogen storage applications.
CHARACTERIZATION OF BONDING IN
HYDRIDES EXPLORED BY THEORETICAL TOOLS
Figure 1. Calculated valence-electron-density maps (a), (b), (c), (d), (e), and (f); charge
transfer plots (g), (h), (i), (j), (k), and (l); and ELF plots (m), (n), (o), (p), (q), and (r) for
NaCl, C (diamond structure), Na, LiH, B2H6, and ScH2, respectively. Atoms are labelled
on the illustration scales are given on the right hand side.
Charge Density
covalent (charges are also present in between the B-B and polarised
electronic distribution around the H site) and metallic bonding (uniform
charge distribution at the interstitial region), respectively. However, the
charge-density distribution in ScH2 cannot be simply viewed as that for a
pure metallic phase as Figure 1f conveys significant ionic and covalent
bonding contributions.
(a)
(b)
Figure 2. The valence electron charge-density plot between the Th and H atoms for
Th2AlH4 in the ac plane (through H) with 40 contours drawn between 0 and 0.25
electrons/a.u.3.
Bonding Nature and Its Implications with the Energy Storage … 91
Charge density analysis has also been used to understand the anisotropic
volume expansion, the high decomposition temperature of certain hydrides,
and to explain the unusual short H-H separations and novel electrical, optical
and magnetic properties. In particular the charge density analysis is useful
to understand the formation of nanophases of solids and their
surface/interface characteristics. We have used charge density analysis along
with structure analysis to explain the anisotropic volume expansion in
Th2Al. The experimental [8, 9] and theoretical studies showed high
anisotropic changes in lattice expansion on hydrogenation of Th2Al.
According to the crystal structure of Th2Al, the inter-atomic distance
between the interstitial where one can accommodate H in the ab plane is
2.4 Å. Hence, a large flexible space is available for accommodation of H
atoms in the ab plane without a need to expand the lattice. In contrast, the
inter-atomic distance between the interstitial regions in the ac plane is only
1.65Å. So, a large expansion of the lattice along c is necessary to
accommodate H atoms within the ac plane. As a result, even with a short
H-H separation of 1.95Å, a lattice expansion of 12.41% along c-axis is
required when Th2AlH4 is formed from Th2Al. The experimental
observation of 0.105% lattice expansion along a and 12.15% along c is
found to be in excellent agreement with these considerations. The bonding
between Th and H is predominantly covalent as evidenced by the finite
charge between these atoms. The H-s electrons are tightly bound to the
Th-d states and the arrangement of Th and H atoms forms an H-Th-H dumb-
bell pattern (see Figure 2).
Charge Transfer
The charge density displays the chemical bonds in a given system, but
remains rather featureless with charge maxima around the atoms similar to
the charge density of overlapping free atoms. The charge-transfer contour is
the self-consistent electron density in a particular plane, comp, minus the
electron density of the corresponding overlapping free atoms, o.f.a as
defined below,
92 P. Vajeeston and H. Fjellvåg
1
t 2
ELF 1
p(r )
(2)
t p ,h
where:
n 2
i (4)
i
2
1 n
i (5)
2 i
are computed from the orbitals, i . In both equations, the index i runs over
all occupied orbitals.
According to Eqn.1, the ELF takes the value one either for a single-
electron wave function or for a two-electron-singlet wave function. In a
many-electron system, ELF is close to one in regions where electrons are
paired such as covalent bonds, or for unpaired lone electrons of dangling
bonds, while the ELF is small in low-density regions. In a homogeneous
electron gas system ELF equals 0.5 at any electron density and ELF values
of this order in homogeneous systems indicates regions where the bonding
has a metallic character. The ELF distributions of typical examples for the
bonds in the three test cases are given in Figure 1n-p. For the NaCl case ELF
94 P. Vajeeston and H. Fjellvåg
is around 0.9 at the Cl site and only a non-significant ELF value is found at
the Na site, thus clearly reflecting ionic bonding. Further, the spherically
symmetric distribution of ELF at the atomic sites along with negligible ELF
in between Na and Cl are further confirmation for ionic bonding. In the
diamond case, a ELF value of around one is found where the shared electrons
are present (in regions between the C atoms). Moreover, the ELF distribution
around and between the C atoms are anisotropic and this is the typical feature
of covalently bonded solids. In the third case where the metallic bonding is
dominant, a uniform distribution of ELF is present around the interstitial
region with the value of nearly 0.5 where the delocalized free electrons are
found. From the above examples, one can clearly visualize different types
of bonding in simple solids based on results from DFT calculations. We
extended the corresponding analyses to hydrides with different bonding
interactions and found almost similar ELFs for LiH as NaCl and for B2H6
like C. In the metallic ScH2 hydrides, we obtain lower value of ELF
compared to that in metallic Na. It may be noted that our test calculations
for the transition metals such as Ni, Co, and Cu also show such a low ELF
values and this is the general characteristic for materials with d electrons in
the valence state. We have used ELF to interpret the short H-H separation in
the RTInH1.333 (R = La, Ce, Pr, Nd; T = Ni, Pd, Pt) series [12-14]. ELF based
analysis is also conducted for the complex hydrides in order to identify the
reason for their high decomposition temperature (for example see ref. [15]).
localization on other sites may either increase or decrease. In the former case
a new site may be available for hydrogen filling, say, when exposed to high
hydrogen pressures. This type of analysis has been used to identify the H
position in metal matrix in several systems. [17-21]
The ZrNiAl-type structure (see Figure 3a; maintaining the general
formula RTM) has eight different interstitial sites centered around 2d, 2e, 3f,
3g, 4h, 6i, 6k, and 12l. Of these, the tetrahedral holes around 4h (surrounded
by 3R and 1T), the trigonal-bipyramidal holes around 2d (surrounded by 3R,
1T, and 1M), and the distorted octahedral holes around 3g (surrounded by
3R, 1T, and 2M) are the most interesting in relation to hole size [22] and
symmetry. We have searched the site preference of hydrogen in all these
possible interstitial sites for around a hundred ZrNiAl-type compounds.
Even though we have examined a large number of ZrNiAl-type phases, it is
convenient to concentrate the presentation of the findings on two typical
examples, where hydrogen prefers to occupy different sites. From the total
energy calculation one can identify the likely hydrogen position. In a
particular framework which contains two or more possible sites for location
of hydrogen, we must perform total energy calculations for all the
alternatives. The configuration with the lowest total energy will determine
the actual H position in the framework under consideration. The electron
localization function is another useful tool to spot hydrogen positions in a
particular matrix. [10] Recently we have investigated hydrogen position in
ZrNiAl-type structures, where H occupies different interstitial positions
depending upon the chemical environment (see Figure 3a-f). For example,
H occupies the 2d position in ZrNiAl (Figure 3g and 3h) with the
composition ZrNiAlH0.6667 whereas H occupies the 4h position in LaNiIn
(Figure 3i and 3j) with the composition LaNiInH1.3333. We have correctly
reproduced the experimental findings for these phases and also predicted
hydrogen position in several other compounds of the ZrNiAl series. This
type of theoretical exploration is sometimes very useful for experimentalists
working on crystal structure determination (or to settle contradictions
between different experiments). [23]
Bonding Nature and Its Implications with the Energy Storage … 97
Figure 3. The ZrNiAl-type crystal structure (RTM). Legends for the different kinds
of atoms are given in the illustration. The 2d (trigonal-bipyramidal interstice), 4h
(tetrahedral interstice), and 3g (distorted octahedral interstice) sites are indicated by open
circles and pointed by arrows. (b) The empty trigonal-bipyramidal interstice. (c) The
empty tetrahedral interstice. (d) The trigonal-bipyramidal interstice filled with hydrogen.
(e) Two face-sharing tetrahedral interstices filled with hydrogen. (f) The empty
(distorted) octahedral interstice. Calculated ELF in the (001) plane for (g) ZrNiAl, (h)
ZrNiAlH0.667, (i) LaNiIn, (j) LaNiInH1.333. Different crystallographic sites (2d, 4h, and
3g) are marked by arrows. The iso-surface values correspond to 0.6 ELF value. Legends
for the different kinds of atoms are given in the illustration.
According to the band theory of solids [24, 25], when atoms are brought
together to form a crystal, the discrete electronic energy states of the isolated
atoms merge into energy bands which represent the allowed energies for
electrons in the crystal. These bands may be separated by forbidden regions
or gaps. Based on distribution of electrons in the allowed energy bands the
conductivity of solids changes and hence they can be classified as metal,
98 P. Vajeeston and H. Fjellvåg
No. Compound Space group Band gap (Eg) Band gap type
1 -CaB2H2 Cmc21 (36) 0.32 indirect
2 SrSiAlH P3m1 (156) 0.65 (0.63)[26] indirect
3 Mg2CoH5 P4/nmm (129) 0.71 indirect
4 Sr2FeH6 FM-3M (225) 1.24 direct
5 Ca2FeH6 FM-3M (225) 1.41 direct
6 Mg2NiH4 C12/C1 (12) 1.65 indirect
7 NaBeH3 Pm-3m (221) 1.79 indirect
8 Mg2FeH6 FM-3M (225) 1.97 direct
9 -CaB2H2 P-3m1 (164) 1.98 indirect
10 Sr2RuH6 FM-3M (225) 2.19 direct
11 CsH FM-3M (225) 2.3 direct
12 -AlH3 R-3c (167) 2.53 (3.5)[28] indirect
13 SrLiH3 Pm-3m (221) 2.74 indirect
14 Na3AlH6 P 1 21/n1 (14) 3.0 direct
15 Mg2RuH6 FM-3M (225) 3.02 direct
16 KMgH3 Pm-3m (221) 3.13 indirect
17 BaAlH5 Pna21 (33) 3.38 direct
18 NaMgH3 Pnma (62) 3.45 direct
19 LiMgH3 R-3c (167) 3.98 indirect
20 NaAlH4 I41/aZ (88) 4.98 (6.41)[28] indirect
21 LiAlH4 P121/c1 (14) 4.76 indirect
22 MgH2 P42/mnm (136) 4.83 (5.6)[29] indirect
23 BeH2 Ibam (72) 5.30 direct
24 KH FM-3M (225) 5.51 direct
25 Mg(AlH4)2 P-3m1 (164) 5.60 indirect
26 KAlH4 Pnma (62) 5.97 indirect
27 KGaH4 Pnma (62) 6.05 indirect
28 NaBH4 Pnma (62) 6.21 direct
29 Mg(BH4)2 Pmc21 (26) 6.54 indirect
30 -LiBH4 Pnma (62) 7.0 indirect
100 P. Vajeeston and H. Fjellvåg
Figure 4. The calculated total density of states for NaCl, C, Na, LiH, B2H6, and ScH2.
The valence bands are shaded; the Fermi levels are set at zero energy and marked by
dotted vertical lines.
The calculated total DOS for the studied test cases are shown in
Figure 5. The magnitude of the Eg for these systems shows that both ionic
(NaCl, LiH) and the covalent compounds (C, B2H6) are of the non-metallic
type (the valence band and conduction bands are separated by 3.2 eV for
B2H6 and 4.5 eV for LiH). In Na and ScH2 systems a finite number of
electrons are at the EF in accordance with their metallic character. A good
example for the DOS of metallic hydride is found in LaNiInH1.333 and such
Bonding Nature and Its Implications with the Energy Storage … 101
Figure 5. The calculated total density of states for LaNiInH1.333 (metallic; bottom
panel), SrSiAlH (semiconductor; middle panel), and LiAlH4 (insulator; top panel).
Fermi levels are set at zero energy and marked by dotted vertical lines.
hydrides. The partial DOS (PDOS) for Li shows (see Figure 6) that the s and
p states of -Li3AlH6 are energetically degenerate in the whole energy range,
while Al-s and -p states are well separated. Apart from the ionic contribution
to the bonding, the energetic degeneration of the Al-p and H-s states together
with the spatially favourable constellation of Al and H facilitates covalent
bonding between Al and H. Due to the electron transfer from the Li site to
the AlH6 unit the magnitude of the Li PDOS is much smaller than the Al
PDOS and the interaction between Li and H is ionic.
Figure 6. Calculated total and partial DOS for -Li3AlH6. The Fermi level is set at zero
energy and marked by the vertical dotted line and s states are shaded.
Bonding Nature and Its Implications with the Energy Storage … 103
Figure 7. Calculated COHP curves for (a) NaCl, (b) LiH, (c) diamond, (d) B2H6, (e)
Na, and (f) ScH2. Fermi levels are set at zero energy and marked by dotted lines.
strength is observed for all X-H (X = B, Al, Ga) interactions in this series]
compared to the other interactions. Moving from A = Li to Cs the X-H
interaction (covalent bond strength) is reduced (both in the Al and Ga series).
According to the calculated ICOHP for these phases, the strength of the Al-
H bond is stronger than that of the Ga-H bond. This indicates that the
stability (the decomposition temperature) in the Al series is higher than that
in the Ga series. It implies that the partial substitution of Li by other alkaline
elements in materials like Li1-xNaxAlH4, Li1-xKxAlH4, or Li1-xKxAl1- yGayH4
may give rise to phases with lower decomposition temperatures (for more
details see Ref. [15]).
QUANTIFICATION OF CHARGES
unique definition of how to extract how many electrons are associated with
an atom in a molecule or a structural sub-unit of a solid, it has nevertheless
proved useful in many cases to perform population analyses. Due to its
simplicity the Mulliken population scheme has become the most familiar
approach to count electrons associated with a given atom. However, this
method also rather qualitative than quantitative; the results obtained from
this approach are sensitive to the atomic basis. Mulliken effective charges
(MEC; defined as the difference between the Mulliken charge and the atomic
charge) are listed in Table.2 for representative materials which may provide
benchmarks for systems with well recognized chemical bonding. For a pure
ionic NaCl the calculated MEC clearly reflect that Na donates one electron
to the Cl atom and the bond overlap population (BOP) between Na and Cl is
almost zero. LiH is a reasonably pure ionic compound and the calculated
Mulliken charges reflect approximately the expected ionic picture with Li+
and H-. The overlap population between Li+ and H- is also close to zero, as
expected for an ionic compound. Similarly, in MgH2, BeH2, and AlH3 the
bonding interaction is mainly ionic, but the degree of ionicity is reduced
from MgH2 to BeH2 and further to AlH3, viz. these compounds exhibit some
covalent character as evidenced by the non-zero overlap population.
Similarly, for the covalently bonded diamond, the calculated BOP value is
much higher. For the CH4 molecule, the overlap population takes a value of
0.384 consistent with the well known covalent character of this molecule.
The Mulliken effective charges for Li, Al, and H in LiAlH4 and Li3AlH6
indicate that the interaction between the Li and AlH4/AlH6 is ionic
(one electron transferred from Li to AlH4/AlH6).
In the Bader charge (BC) analysis, each atom of a compound is
surrounded by a surface (called Bader regions) that run through minima of
the charge density and total charge of an atomic site is determined by
integration within the Bader region. The calculated BC for the selected
compounds is given in Table 2. For LiH phase, similar to the Mulliken
analysis, BC also reflects an almost pure ionic picture i.e., around one
electron transferred from Li to H. This finding is consistent with the MEC
and charge density analyses. However, in most of the cases the magnitudes
of the charges obtained from MEC and BC analysis are considerably
108 P. Vajeeston and H. Fjellvåg
different from each other. This is partly due to the artefact of Bader’s “atom
in molecule” approach that involves making the boundaries to integrate
charges in each atomic basin and, also the basis set dependent behaviour of
MEC. However, consistent with the charge, charge transfer, and ELF
analyses, the BC always qualitatively shows the bonding nature in materials.
We thus find that the bonding nature of the hydrides (both metallic and
complex) does not exhibit a simple ionic or covalent character. In fact, the
bonding present in these compounds have mixed bonding character. In the
complex hydrides like LiAlH4 and Li3AlH6 the interaction between Li and
AlH4/AlH6 is ionic whereas the interaction between Al and H comprises of
both ionic and covalent character. A similar type of bonding situation
prevails in all complex hydrides, but the magnitude of the ionic/covalent
mixture will exhibit considerable individual variation.
Born-Effective-Charge Analysis
q Born
(6)
i
i
where the first term on the right-hand side is the core charge of the ion and
the second term is the corresponding component of the product of the core-
shell second-derivative matrix with the inverse of the shell-shell second-
derivative matrix, scaled by the vector of shell charges, qshell. Physically, the
second term corresponds to the response of all the shells present to the
atomic displacement of atom i, or, in other words, the electronic
contribution. Hence, for a rigid ion model, the Born effective charge tensor
is equal to a diagonal matrix with all the diagonal elements equal to the core
charge. Increasingly, the Born effective charges are being determined from
ab initio calculations, thus creating a new avenue for determining shell
model parameters beyond the fitting of dielectric constants.
For the LiH and MgH2 cases (see Table 3) the diagonal components of
the effective charges in a Cartesian frame satisfy the relation Zxx = Zyy = Zzz,
whereas the off-diagonal components turn out to be negligible. This is as
expected for ionic compounds due to the spherical character of ionic bonds.
For LiAlH4 the diagonal component of the effective charges at the Li site is
also almost equal (Zxx ≈ Zyy ≈ Zzz) and the off-diagonal components are
negligibly small (see Table 3). At the Al and H sites the diagonal
components are also almost equal, but the off-diagonal components take
small but definite values. This probably originates from some exchange
owing to sharing of electrons between H and Al within the AlH4 units. The
BEC analyses accordingly reconfirms that the Li sites of the LiAlH4 series
give up nearly one electron, whereas the H atoms correspondingly gain 0.52-
0.89 electron each. This conclusion therefore largely concurs with the mixed
iono-covalent picture between the Al and H. A similar type of bonding
character was present in most of the complex hydrides (ex. alanates, boro
hydrides, amides, etc.). A methodological comment seems appropriate: Our
experience from studies of chemical bonding by both Mulliken-population
and Born-effective-charge analyses shows that these two approaches yield
almost the same conclusion. Our personal view is that Born-effective-charge
calculations normally demand larger computations. Hence, we recommend
110 P. Vajeeston and H. Fjellvåg
Position xx yy zz xy xz yx yz zx zy
LiH
ZLi 4a 1.06 1.06 1.06 0 0 0 0 0 0
ZH 4b -1.06 -1.06 -1.06 0 0 0 0 0 0
NaH
ZNa 4a 1.0 1.0 1.0 0 0 0 0 0 0
ZH 4b -1.0 -1.0 -1.0 0 0 0 0 0 0
MgH2
ZMg 2a 1.96 1.96 2.01 0.02 0.0 0.02 0.0 0.0 0.0
ZH 4f -0.98 -0.98 -1.01 -0.02 0.0 0.02 0.0 0.0 0.0
-AlH3
ZAl 6b 2.72 2.72 2.96 -0.6 0 0.62 0 0 0
ZH 18e -0.77 -1.04 -0.99 -0.23 -0.23 -0.23 -0.39 -0.23 -0.39
-LiAlH4
ZLi 4e 1.09 1.14 1.1 0.16 -0.02 -0.01 -0.06 -0.01 0.04
ZAl 4e 1.54 1.69 1.75 -0.13 -0.02 0.13 -0.03 -0.04 0.03
ZH 4e -0.52 -0.66 -0.89 0.05 0.10 0.33 -0.16 0.09 -0.15
ZH 4e -0.66 -0.83 -0.73 -0.25 0.07 -0.24 -0.01 0.06 -0.01
ZH 4e -0.61 -0.67 -0.76 0.04 -0.14 0.0 0.13 -0.11 0.16
ZH 4e -0.84 -0.68 -0.49 0.14 0.03 0.19 0.08 0.02 -0.01
-NaAlH4
ZNa 4a 1.23 1.23 1.12 0.02 0.0 -0.02 0.0 0.0 0.0
ZAl 4b 1.57 1.57 1.87 0.0 0.0 0.0 0.0 0.0 0.0
ZH 16f -0.75 -0.65 -0.75 -0.09 0.13 -0.09 0.09 0.11 0.11
Bonding Nature and Its Implications with the Energy Storage … 111
Position xx yy zz xy xz yx yz zx zy
Li3AlH6
ZLi 18f 1.05 1.12 1.02 0.10 -0.08 0.07 0.11 -0.03 -0.01
ZAl 3a 2.11 2.11 2.41 -0.05 0.0 0.05 0.0 0.0 0.0
ZAl 3b 2.38 2.38 1.76 -0.08 0.0 0.08 0.0 0.0 0.0
ZH 18f 1.07 -0.75 -0.88 -0.08 -0.21 -0.12 -0.03 -0.19 -0.08
ZH 18f -0.74 -1.08 -0.88 -0.07 0.08 -0.11 0.12 0.03 0.21
Na3AlH6
ZNa 2b 1.06 1.04 1.07 -0.03 0.01 0.0 0.07 -0.03 0.04
ZNa 4e 1.07 1.1 1.09 -0.06 -0.07 0.04 -0.04 0.07 -0.01
ZAl 2a 2.15 2.20 2.16 0.01 0.03 0.08 -0.06 -0.03 0.02
ZH 4e -0.77 -0.71 -1.20 0.0 -0.17 -0.03 -0.07 -0.17 -0.09
ZH 4e -1.07 -087 -0.74 -0.2 0.07 -0.22 0.03 0.07 0.09
ZH 4e -0.83 -1.11 -0.76 -0.22 0.11 -0.23 0-15 0.10 0.13
SrAlSiH
ZSr 1a 2.64 2.64 2.20 0.0 0.0 0.0 0.0 0.0 0.0
ZAl 1c 1.69 1.69 1.44 0.0 0.0 0.0 0.0 0.0 0.0
ZSi 1b -3.22 -3.22 -2.01 0.0 0.0 0.0 0.0 0.0 0.0
ZH 1c -1.1 -1.1 -1.63 0.0 0.0 0.0 0.0 0.0 0.0
behaviour and it has the mixed bonding nature. On the otherhand, one can
also imagine that, within one Li3AlH6 unit, the three electrons from Li will
fill three of the six half-filled H-s orbitals and the remaining three H-s
orbitals are filled by three electrons from the Al atom (i.e., pure ionic
situation). In most of the complex hydrides (like NaAlH4, Li3AlH6, etc.) the
latter type (pure ionic) of bonding was believed. [30, 31] The arguments for
this interpretation are based on DOS analysis and long-range Coulomb
interactions according to calculations using the linearized-augmented plane-
wave method. As more H-s states are present in the valence band than in the
conduction band, it is concluded that the interaction between Al and H must
be ionic.
The calculated DOS, from different codes are perfectly matching each
other. Hence, there is no ambiguity between the different computational
methods, whereas the interpretation/understanding of the results differs. It is
commonly recognized that it is difficult to characterize the nature of the
chemical bonding (in particular for hydrides) from DOS and integrated
charges inside each atomic spheres alone. Each and every theoretical tool
has some additional flexibility/facility to evaluate bonding behavior. We
have therefore now going to see how one can use different approaches to
gather as much information as possible regarding the chemical bonding in
Li3AlH6.
The calculated partial DOS of Al and H in Li3AlH6 are shown in Figure
6, illustrating the following three main features: (1) The VB and CB are
separated by a band gap of 3.8 eV, confirming that this compound is an
insulator. (2) The VB is split into two separate regions by a 1.3 eV energy
gap. The lowest-lying band (at −7.5 to −5.8 eV) mainly originates from
Al-s states with appreciable contributions from H-s states. The second
region, from −4.5 to −2.2 eV, comprises energetically degenerate H-s and
Al-p states, which we focused on as a favorable situation for formation of
covalent bonding. The spherically symmetric nature of s orbitals, together
with the energetic degeneration of H-s and Al-s states, suggests that there
should be a high probability for formation of covalent-type bonding between
these atoms. (3) The very small contribution from H-s in the unoccupied
states above the Fermi level (EF) is explained with the help of COHP
Bonding Nature and Its Implications with the Energy Storage … 113
Figure 8. Calculated COHP for H-H (note: multiplied by 6) and Al-H in Li3AlH6.
If the chemical bonding between Al and H was purely ionic one would
expect that Al-p and H-s states should be energetically well separated. [32]
Furthermore, in a purely ionic situation one should not expect the finite
electron-density distribution between Al and H seen in Figure 9. Moreover,
the charge-density distribution should be spherically symmetric around the
Al and H sites if the bonding is purely ionic. Hence, we came to our initial
conclusion that the bonding between Al and H in Li3AlH6 is largely covalent.
If there is strong covalent bonding, one should expect that the H character
114 P. Vajeeston and H. Fjellvåg
should be distributed between the three s-p manifolds. As the H-s character
is very small in the CB (Figure 6), one can conclude that the covalent
contribution is negligible. However, in the first place one cannot judge the
character of chemical bonding in complex materials based on DOS analyses
alone (see, e.g., [12]). Owing to iono-covalent interactions, all H-s orbitals
will be filled and therefore both bonding and antibonding states of the s-p
hybrid are within the VB, as is evident from our COHP analysis (see, e.g.,
Figure 8). If the bonding interaction between all constituents in Li3AlH6 was
purely ionic, one would expect narrow-band features and certainly not the
broad DOS features found for Li3AlH6 (which indicates overlap interaction
between the constituents). In a purely ionic situation, the distinct DOS
manifold around −7 eV with two electrons per formula unit would have to
be contributed by only one of the constituents. However, considerable
amounts of electrons (according to the integrated DOS 36% from Al, 49%
from H, and the remaining 15% from Li) from both Al-s and H-s indicate a
distinct degree of covalent character. If Al was in the 3+ state, negligible
amounts of electrons would be left at the Al site, resulting in a very small
contribution from it to the VB. In fact, the integrated DOS yielded 0.78
electrons at the Al site and shows that the bonding interaction is not purely
ionic (partial conclusion from DOS alone).
In order to gain further understanding of the nature of the bonding, we
have calculated the charge density, charge transfer, energy projected charge
density, and ELF for Li3AlH6. The results are shown in Figure 9, where we
show only the relation between Al and H since there is no ambiguity with
regard to the ionic interaction between Li+ and AlH6−. The charge-density
distribution and charge-transfer plots indicate that a finite number of
electrons are present between Al and H, which means that there is a clear-
cut degree of covalent interaction between Al and H within the AlH6 subunit.
The charge-transfer plot [Figure 9b] clearly indicates that charges are
depleted from Al and Li. The depletion is not spherically symmetric at the
Al site (in a pure ionic case it should be spherically symmetric), which
implies that the bonding can be regarded as ionic with an appreciable
covalent-type interaction (directional bonding) between Al and H
(additional insight from charge-density and charge-transfer plots). The
Bonding Nature and Its Implications with the Energy Storage … 115
Figure 9. Calculated plots of (a) valence-electron charge density, (b) charge transfer,
(c) energy (range -7.5 to -5.8eV) projected charge density, and (d) ELF for Li3AlH6.
The illustrations refer to the (001) plane.
The calculated ELF for Li3AlH6 in Figure 9d shows that the H electrons
are polarized toward the Al site. A similar character of the ELF is found for
the molecules C2H6 and C2H4, where the interaction between C and H is
commonly recognized as covalent. [10] Hence, one finds that there is a
directional (covalent) bonding character between Al and H in Li3AlH6
(subconclusion from ELF). The maximum value of ELF at the H sites along
with the nonspherical distribution of ELF in Figure 9d also confirming the
iono-covalent bond between Al and H in Li3AlH6.
Mulliken charges are reported in Table 2 for a few H-based
representative materials that may provide benchmarks for systems with well-
recognized chemical bonding characters. LiH is a nearly pure ionic
116 P. Vajeeston and H. Fjellvåg
as well as enhance sorption kinetics, in alane one should enhance its kinetics
to use it for practical applications. In general, lightweight metal hydrides are
considered as storage media, but the kinetic constraints limit their
application. A promising approach to address this issue is to reduce the
particle size of the metal hydride to the nanometer range, resulting in
enhanced kinetics without the need of a catalyst [39, 40]. For example, in
MgH2, a numerous studies have been focused on improving the problematic
sorption kinetics, including mechanical ball milling [41-43] and chemical
alloying [44, 45]. The smallest particle sizes (20 nm) obtainable by these
methods still primarily display bulk desorption characteristics [46]. Though
AlH3 is one of the promising candidate for hydrogen storage application, it
is not a stable compound and its property changes with time when it is stored
ambient condition and especially releases the hydrogen around 100°C
[37, 38]. In order to increase the stability of this system one must reduce the
particle size below its critical value. In our study we suggested that particle
size of α-AlH3 clusters below 1 nm might have the required
physical/chemical properties for the practical applications. [47]
Figure 10. Calculated total energy as a function of particle size (a) for the AlH3 nano-
clusters (in filled circle) with nano particles of Al plus H2 molecule [i.e.,
EAl(nano)+(3/2)EH2 (mol.)] (in open circle) and (b) for NaBH4 and KBH4.
120 P. Vajeeston and H. Fjellvåg
The critical particle size of the nano clusters and nano whiskers of
LiBH4, NaBH4 and KBH4 are identified from the calculated the total energy
as a function of the cluster size and that are shown in Figure 10b (only for
NaBH4 and KBH4). If the cluster size decreases the total energy becomes
more positive (i.e., the formation energy decreases with decreasing cluster
size). In particular, there is a steep increase in the total energy when the size
of the cluster is below 1.35 and 1.8 nm for NaBH4 and KBH4, respectively.
Similarly, the calculated total energy as a function of nano-whisker diameter
shows that when we reduce the diameter below certain critical value (2.6 nm
for NaBH4; the corresponding value for KBH4 is 2.9 nm) the total energy
increase drastically. If one reduces the cluster size and nano-whisker
diameter, the relative energy of the clusters/nano-whiskers becomes less
negative, which indicates the destabilization of small particles. This also
suggests that the thermodynamical properties and in particular the hydrogen
sorption temperature is expected to reduce in nanophases compared with that
in bulk materials. It should be noted that a similar trend has also been
observed in LiBH4 where the critical size of the cluster was 1.75 nm. On the
other hand, in the previous section we have seen that in AlH3 the particle
size reduction enhances the stability. This finding clearly indicates that
particle size plays an important role on decomposition temperature of the
complex hydrides. The reduction of the particle size in some cases enhances
the stability and in some other cases deceases it. In order to identify the
stability upon particle size reduction we have carried out detailed study on
39 complex hydrides and the calculated deformation energies are calculated
using the following equation
where, E(bulk) is the total energy of the bulk phase (solid state), E(nano)
total energy of the ultra-small nano-particles, and n represent the number of
formula units of the specific chemical composition. If the deformation
energy is positive then the chosen compound becomes stable upon particle
Bonding Nature and Its Implications with the Energy Storage … 121
size reduction. On the other hand, if the deformation energy is negative then
the chosen compound becomes unstable upon particle size reduction. The
calculated deformation energies for the studied 39 compounds are displayed
in Figure 11. From Figure 11 it is evident that most of the complex hydrides
are more stable and only few compounds are unstable upon particle size
reduction. This finding clearly demonstrates that nano particles of the
complex hydrides may not the suitable candidates for the reversible
hydrogen storage application.
Figure 11. Theoretically calculated stability diagram (energy difference between bulk
and the nano-particle; positive and negative values corresponding to the particle size
reduction can stabilize and or destabilize the system, respectively) for ultra-small
nanoparticles made of different high hydrogen content metal/complex hydrides.
noticeable covalent bond within the BH4 units. [50] The substantial
difference in the electronegativity between Na/K and H suggests the
presence of strong ionic character (i.e., the Na/K valence electrons
transferred to the H sites) and the small difference in the electronegativity
between B and H suggests the presence of strong covalent character. In order
to gain further understanding about the bonding situation in nano objects of
β-Na/KBH4, we turn our attention to the charge density analysis. According
to the charge-density distribution at the Na and K sites in the center of the
nano-phases (Figure 12 (a–f)), it is evident that the highest charge density
resides in the immediate vicinity of the nuclei. The spherical charge
distribution at the Na and K sites (Figure 12 (b, c, e and f)) clearly reflects
the ionic character. On the other hand, due to the covalent interaction a finite
charge density was present in between B and H. In order to get more insight
into the chemical bonding we have visualized the charge density with the
iso-surface value of 0.6 e/volume, which clearly shows the sharing of
electrons in between B and H (see Figure 12 (b and e)). From the Figure 12,
it is evident that the charge density distribution in clusters and whiskers are
almost similar to that in the bulk phase. A closer examination at the edges
of the clusters and whiskers B and H sites shows however a significant
difference in the charge density distribution. In particular, the charges are
more directional dependent at the surface H sites towards the B sites and the
charge distribution is more asymmetric. It can be hence concluded that
considerable more covalent interaction is present between B and H at the
BH4 surface than that in the inner part of the BH4 units.
In order to have a better understanding about the bonding interaction
between the constituents the bond overlap population (BOP) values are
calculated on the basis of Mulliken population analysis. The BOP values can
provide useful information about the bonding property between atoms. A
high BOP value indicates a strong covalent bond, while a low BOP value
indicates an ionic interaction. For the practical use of NaBH4 and KBH4 as
hydrogen storage materials one must weaken their B–H bonds. The
calculated BOP values for the B–H bonds in different modifications of
NaBH4 and KBH4 are listed in Table 4. Table 4 shows that the BOP values
for the B–H bonds are considerably weaker in the nano-phases. It is
Bonding Nature and Its Implications with the Energy Storage … 123
interesting to note that the BOP value for B–H bond in the bulk phase is
around 1.13, which is almost equivalent to that of covalent C–C bond in
diamond (the BOP value is 1.08). This might be one of the reasons why these
phases are able to release H only at elevated temperature (above the melting
point). In all the considered phases the calculated Na/K–H BOP values are
almost zero, owing to the high degree of the ionic character.
Figure 12. Charge density distribution in β-KBH4 derived nano-clusters, and nano-
whiskers. Charge distribution in volumetric data view with isosurface value of 0.6
e/volume. [(a) for dot and (d) for whiskers], only the isosurfaces with value of 0.6 e
[(b) for dot and (e) for whiskers], and two dimensional view [(c) for dot and (f) for
whiskers]
Figure 13. Schematic representation of the calculated hydrogen site energies for the
optimized NaBH4 and KBH4 nano-clusters and whiskers compared with the
corresponding bulk phase. H1, H2, H3 and H4 refers to H in between Na/K; H between
Na/K and B; H connected with B; and H in the center of the cluster, respectively. The
red (left side) and black (right side data) color represent the NaBH4 and KBH4 phases,
respectively.
𝑒𝑄𝑉𝑧𝑧
𝐶𝑄 = ℎ
(10)
(𝑉𝑥𝑥 −𝑉𝑦𝑦 )
𝜂𝑄 = 𝑉𝑧𝑧
(11)
where σ refers to the chemical shielding tensor in the principal axis frame.
This is an absolute value of the isotropic chemical shielding, not relative to
any standard. Vzz is the largest component of the diagonalized EFG tensor,
Q is the nuclear quadrupole moment, and h is Plank’s constant.
The calculated values of σiso, CQ, and ηQ for the different polymorphs are
displayed in Table 5. The calculated σiso values for Mg are scattered between
549 to 556 ppm. Similarly, for the hydrogen atoms in different polymorphs
these values are scattered between 25.4 to 26.6 ppm and one cannot compare
these values directly to the experimental values. Because the calculated
values are absolute chemical shielding whereas the experimental value are
concerned with shifts relative to a known standard. However, one can
compare the comparative shifts between the different peaks. [49] Similarly
the CQ and ηQ values are also scattered over a wide range and these values
are directly comparable with experimental values. It should be noted that the
NMR parameters are strongly related to the atomic environment (near
neighbors) and quadrupolar coupling constants are directly related to the
bond strength and anisotropy in the charge density distribution around the
probe nuclei. In this connection, we have calculated NMR related
parameters only for the stable α-derived nanowhiskers (for size 1.6 nm) and
β-derived nanoparticles (of 2.4 nm size cluster). As calculation of the NMR
parameters requires higher accuracy for the wave functions, we have
performed very accurate computation to obtain reliable values.
The calculated values of σiso, CQ, and ηQ for the nano-objects are given in
Table 5. As we have shown in the previous section, the bond length values
are very scattered in the nanophases and also the NMR-related parameters
are highly dependent on the local environment of the system. Hence, the
calculated σiso, CQ, and ηQ values for both nano-objects are highly scattered
(see Table 5). For example, the calculated σiso value for Mg in the
nanowhisker varies between 548 to 593 ppm and the corresponding value in
Bonding Nature and Its Implications with the Energy Storage … 127
nano particles varies between 542 to 581 ppm. Similarly, the calculated σiso
value for H in the nanowhisker varies between 22 to 26.9 ppm and the
corresponding value in nano particles between 21.4 to 27.2 ppm. To
understand the changes on NMR parameters as a function of particle size,
we have calculated the NMR parameters for the selected nanowhiskers (α-
derived; size: 0.6, 1, and 1.6 nm) and nanoparticles (β-derived; size: 0.6, 1.6,
and 2.4 nm). Due to the close similarity in the chemical shielding of Mg and
H in nanowhiskers and nanoparticles, we have displayed only the
nanowhiskers chemical shielding corresponding to Mg and H in Figure 14.
It is clear that depending on the particle size the calculated chemical
shielding varies considerably and the values are well scattered. If the system
becomes considerably larger, the calculated chemical shielding is well
broadened owing to the bulk-like core structure formation. On the other
hand, in a smaller whisker/particle the chemical shielding becomes very
narrow. The present study suggests that from the shift in the chemical
shielding spectrum one can qualitatively classify the size of nanoparticles.
Figure 15. Crystal structures of (a) LT-LiBH4, (b) HT-LiBH4, and (c) HT-
LiBH4+LiI+P2S5.
Electrochemical Li Storage
approximately 200 mAh g−1 at 298 K. The study provided a novel design
principle for the development of the electrolytes for use in rechargeable
batteries. All these finding clearly indicates that [BH4]−-[BH4]− interaction
place an important role on Li ion conduction.
The aim of our study is to determine the energy barriers for Li ions
hopping between different sites and how these barriers are influenced by
incorporation of LiI, Li2S and/or P2S5 molecular units and [I-] ions into the
LiBH4 structure. In this work, only the addition of the P2S5 phase to the HT-
LiBH4+LiI system (LPS) at low local concentration is considered, in order
to have better view ability. The calculated activation barriers for the LT and
HT phases of LiBH4 are 0.69 eV and 0.53 eV respectively, in good
agreement with previous theoretical work. [105, 106] Upon introduction of
LiI in the HT-LiBH4 matrix, these activation barriers are lower (0.42 eV to
0.54 eV) and the minimum values are found in proximity of the chemical
few paired valence electrons left at the Na and B sites. A certain polarized
character is found in the ELF distribution at the H sites in all the complex
hydrides we have investigated earlier. When we introduced LI, Li2S, or P2S5
in the HT-LiBH4 phase the charge distributions are considerably unchanged.
On the other hand, for the LiI added, LLS and LPS systems the calculated
MEC and BOP values are varying considerably. Especially the BOP value
for the B-H interaction is notably different within the BH4 units and this
value is minimum in the pure HT-LiBH4 and maximum in the LPS system.
This finding clearly demonstrated that the BH4 ···BH4 interaction plays a
crucial role on the Li transport. It should be highlighted that for the hydrogen
storage application one might weaken the B-H bonds within the [BH4]− unit.
On the otherhand, for obtaining a good Li-conductors the B-H interaction
should be increase in order to reduce the BH4 ···BH4 interaction.
Figure 16. Li ion migration in different possible path ways in (a) HT-LiBH4 doped with
LiI and P2S5 as obtained from climbing Nudged Elastic Band method and (b) energy
barrier heights as function of the local concentration. The connected lines are given for
visual help of the sequence of images. The symbols represent calculated data points.
136 P. Vajeeston and H. Fjellvåg
Figure 17. Calculated valence electron charge density distribution (a and d; color code:
red-high and blue-low electron density.) charge transfer (b and e; isosurface value +0.2
e) and ELF (c and f; isosurface values of 0.9) for low temperature LiBH4 and high
temperature LiBH4.
In general, in the present chapter we have seen that the bonding in the
hydrides like LiAlH4 and Li3AlH6 does not exhibit a simple ionic or covalent
character. In fact, the bonding interaction in these compounds is rather
complicated. The interaction between Li and AlH4/AlH6 is ionic and the one
between Al and H manifests both ionic and covalent characters. We believe
that similar types of bonding situations prevail in all similar complex
hydrides, but the magnitude of the ionic/covalent mixture will exhibit a
considerable individual variation. This study clearly demonstrates that
several analysis tools based on ab initio calculations must be consulted in
order to make more assured conclusions regarding the bonding nature of
multicomponent systems.
Bonding Nature and Its Implications with the Energy Storage … 137
REFERENCES
[90] P. Manoravi and K. Shahi, “Fast ion transport and phase diagram of
the KBr-NaI system,” J. Phys. Chem. Solids, vol. 52, no. 3, pp. 527-
535, 1991.
[91] J. F. M. Oudenhoven, L. Baggetto and P. H. L. Notten, “All-Solid-
State Lithium-Ion Microbatteries: A Review of Various Three-
Dimensional Concepts,” Adv. Energy Mater., vol. 1, no. 1, pp. 10-33,
2011.
[92] M. Armand and J. M. Tarascon, “Building better batteries,” Nature,
vol. 451, pp. 652, 2008.
[93] K. Takada, “Progress and prospective of solid-state lithium
batteries,” Acta Materialia, vol. 61, no. 3, pp. 759-770, 2013.
[94] Y. Kato, K. Kawamoto, R. Kanno and M. Hirayama, “Discharge
Performance of All-Solid-State Battery Using a Lithium Superionic
Conductor Li10GeP2S12,” Electrochemistry, vol. 80, no. 10, pp. 749-
751, 2012.
[95] J. W. Fergus, “Ceramic and polymeric solid electrolytes for lithium-
ion batteries,” J. Power Sources, vol. 195, no. 15, pp. 4554-4569,
2010.
[96] M. Matsuo and S.-i. Orimo, “Lithium Fast-Ionic Conduction in
Complex Hydrides: Review and Prospects,” Adv. Energy Mater., vol.
1, no. 2, pp. 161-172, 2011.
[97] K. Miwa, N. Ohba, S.-i. Towata, Y. Nakamori and S.-i. Orimo,
“First-principles study on lithium borohydride LiBH4,” Phys. Rev. B,
vol. 69, no. 24, pp. 245120, 2004.
[98] E. Orgaz, A. Membrillo, R. Castañeda and A. Aburto, “Electronic
structure of ternary hydrides based on light elements,” J. Alloys
Compd., vol. 404-406, no. Supplement C, pp. 176-180, 2005.
[99] H. Oguchi, M. Matsuo, J. S. Hummelshøj, T. Vegge, J. K. Nørskov,
T. Sato, Y. Miura, H. Takamura, H. Maekawa and S. Orimo,
“Experimental and computational studies on structural transitions in
the LiBH4–LiI pseudobinary system,” Appl. Phys. Lett., vol. 94, no.
14, pp. 141912, 2009.
148 P. Vajeeston and H. Fjellvåg
Chapter 3
ABSTRACT
*
Corresponding author: Email: igorsereda@karazin.ua.
150 Ihor Sereda, Alexandr Tseluyko and Nikolay Azarenkov
ion stimulated desorption. The feature of metal hydride cathode under the
conditions of gas discharge is a decrease in the ionization potential of
desorbed hydrogen by 0.3-0.5 eV due to the molecules desorption in the
vibrationally/rotationally excited state. This permits a substantially
increase in ionization efficiency and the formation of negative ions by the
mechanism of dissociative attachment in plasma volume. However,
hydrogen desorbed from metal hydride significantly changes the properties
of the discharge. This is expressed, for example, in the fact that the plasma
source based on Penning trap with metal hydride cathode appears to
generate current-compensated ion beams with the ability to control the
energy of the extracted ions. There is also the opportunity of longitudinal
extraction of negative hydrogen ions against the traditional method of
extraction across the magnetic field.
1. INTRODUCTION
holder (5) with water-cool to keep the temperature unchangeable. The third
one was a copper cathode with balloon hydrogen supply for the simulation
of hydrogen desorption (Figure 2). Balloon hydrogen was injected locally
through a number of thin apertures in the surface of the cathode. The ratio
between the local flow of hydrogen (through the cathode) and the additional
one (in the vacuum chamber) were picked the same as in the discharge with
the MH-cathode. The fourth one was solid copper cathode, and it was the
same as cathode-reflector (3). All the types of the cathodes had the same
spatial dimensions: 2.0 cm in diameter and 0.5 cm thick. The copper
cathode-reflector (3) had an aperture at the center 0.5 cm in diameter. Behind
the aperture a collector (4) was set.
When studying the energy spectra of charged particles, collector (4)
were replaced by multigrid energy analyzer. The high-frequency component
of the registered current was carried out by a spectrum analyzer. The plasma
parameters on the axis of the discharge cell were measured by Langmuir
probes (7) located at half distance between the anode (7) and the cathodes
(1 and 3). The analyzer of negative ions (8) had a collector under the positive
potential of +400 V and a grid under the positive potential of +300 V.
The residual pressure in the vacuum chamber did not exceed the value
of 5×10-6 Torr. The pressure higher, than the residual one was achieved by
initial balloon hydrogen supply into the vacuum chamber.
The metal-hydride cathodes were produced from hydride-forming alloy
Zr50V50 by method including its melting, activation and filling with
hydrogen, hydride crushing and mixing it with a binder, and pressing. The
filling of the alloys with hydrogen was performed after its activation. It was
activated in dynamic vacuum at the pressure of 10-2 Torr under the
temperature of 600 K during an hour. Afterwards, the volume with the
sample cooled down to room temperature and was filled with hydrogen
under the pressure of 2 – 4 standard atmospheres. The hydride powder was
loaded in an argon medium, mixed up with copper binder (30 – 60% by
weight) and then it was loaded into the press-form. Pressing was carried out
at room temperature at the pressure of 3 – 5 tf/cm2 during 5 – 10 min.
The saturation coefficient of the alloy with hydrogen was about 180
3
cm /g under normal conditions. The quantity of accumulated in our MH-
cathode hydrogen was about 900 cm3 under normal conditions. Heating the
cathode from 400 to 700 K realizes hydrogen desorption under the pressure
of 10-6 – 1 Torr. The purity of hydrogen desorbed from metal hydride was
better, than 99.999%.
154 Ihor Sereda, Alexandr Tseluyko and Nikolay Azarenkov
desorption curve (Figure 3). As one can see, almost 50% of hydrogen stored
in the metal hydride fits on the high-temperature phase, characterized by the
decomposition temperature of 700 K. Therefore, it was originally expected
that under the conditions of vacuum thermal insulation, the further
enhancement of discharge current will lead to the increase in cathode
temperature with following discharge transition into an arc mode.
Figure 3. The differential curve of hydrogen thermal desorption from Zr50V50 sample.
800 copper
cathode
700
Тc , K
600
500
MH-cathode
400
300
0 10 20 30 40 50 60 70 80
t, min
where Us is the space potential of ion formation, Ii is the ion current on the
MH-cathode, S is the surface area of the MH-cathode, kg is the coefficient of
working gas thermal conductivity, Tc is the temperature of the MH-cathode,
T0 is the temperature of discharge chamber walls, σ is the Stefan-Boltzmann
constant, Q is a term that takes into account the total thermal effect of metal
hydride formation/decomposition, as well as convective heat transfer in gas
phase, λc is the thermal conductivity of cathode holder, ke is the emissivity
coefficient of the MH-cathode surface.
The left side of balance equation (1) is the power of heat source, which
is released at the cathode volume when it is bombarded by ions. The first
and the second terms on the right side of the equation describe the loss of
thermal energy due to the thermal conductivity of hydrogen and the elements
of cathode-holder respectively. The third term is the thermal radiation power
of the cathode. The last term takes into account the total thermal effect of
hydride phases formation/decomposition, as well as the convective heat
transfer, which accompanies hydrogen sorption/desorption processes.
The Application of Metal Hydride Based on Zr-V Alloy … 157
following high-speed cooling after the plasma pulse action. Phase analysis
has revealed hydrogen loss on the level of 5% per pulse from total amount
stored in the material.
current registered and the third mode is not observed. All of the modes were
observed in the pressure range of 10-5 – 10-4 Torr.
At small Ud the rate of hydrogen desorption from metal hydride is small,
therefore the existence of the first two modes does not depend on the cathode
material, which was observed earlier by a number of authors [19-20]. An
increase in the introduced power leads to the intensification of hydrogen
desorption from the MH-cathode. The interaction of the additional flux of
neutral gas with charged particles substantially changes the discharge
characteristics and it transits to the third mode. The first two modes have
been thoroughly studied, so the main attention was paid to the third mode of
the discharge.
The investigation of the high-frequency oscillation component of
discharge voltage and the axial collector currents revealed a number of
characteristic features. The modulation of collector current is at the same
frequency as the oscillations of discharge voltage. In the experiments with
two copper cathodes, when only the ionic axial current is observed, the
dependence of the oscillations frequency on discharge voltage is close to
linear (Figure 6c). But in case of the MH-cathode, the transition from mode
to mode is accompanied by a sharp change in the oscillations frequency by
the value of 6 MHz (Figure 6c). The frequency decreases during the
transition from the first mode to the second one and increases with the
transition from the second to the third mode. The axial electrons yield from
Penning discharge corresponds to the increased oscillation frequency.
Since a linear dependence of oscillation frequency on discharge voltage
is clearly expressed [16], it could be made an assumption of diocotron
instability development. Taking into account well known ratio for the
frequency of diocotron oscillations f ~ Er/H [21], the dependence of
oscillation frequency (f) on discharge voltage could be explained in the
following way. At transition to third mode the frequency (f) increases due to
both effects. The first one is an increase in radial electric field (Er) due to
discharge voltage rise (linear growth). And second one is owing to
redistribution of axial electric field (Ez) towards Er (frequency jumps,
followed by an increase in discharge current).
The Application of Metal Hydride Based on Zr-V Alloy … 163
Figure 6. The dependence of collector current (a), oscillation frequency (b), discharge
current (c) and floating potential (d) on discharge voltage at P = 3×10-5 Torr, H = 0.1
T.
164 Ihor Sereda, Alexandr Tseluyko and Nikolay Azarenkov
Figure 6d shows the reason for electrons to yield from the side of copper
cathode-reflector only. One can see, when the discharge goes to the third
mode, a significant decrease in floating potential is observed on the
discharge axis by both cathodes. However, near the copper cathode-reflector
the potential drops to values of 50 V, while near the MH-cathode it drops
only to 450 V. In other words, axial electric field (Ez) falls with the radial
one (Er) rapid growth.
Thus, electrons with energies greater than 50 eV overcome a potential
barrier by the copper cathode-reflector and yield through its central aperture.
Only high-energy electrons can yield from the side of the MH-cathode,
which quantity is very small.
When working with both copper cathodes there is no third mode and
there is no potential drop by the cathodes.
The reason for changing the properties of Penning discharge is the flux
of desorbed hydrogen, which is injected into discharge region locally from
the side of the MH-cathode. To determine its impact on the discharge
properties, the experimental simulation of the MH-cathode was carried out.
It was performed by organizing a balloon hydrogen supply through the one
of the copper cathodes (see Figure 2).
Here hydrogen was injected in two ways: the local one is through the
cathode with H2 supply and the additional one into the vacuum chamber. In
the experiments with both solid copper cathodes hydrogen was injected only
into the vacuum chamber.
The investigations were carried out at the pressure of 10-5 – 10-4 Torr.
The relationship between local and additional hydrogen fluxes, as well as
the magnitude of external magnetic field were chosen to be the same as in
the previous section.
The Application of Metal Hydride Based on Zr-V Alloy … 165
Figure 7. The dependence of collector current (a), discharge current (b) and floating
potential (c) on discharge voltage at P = 3×10-5 Torr, H = 0.1 T.
166 Ihor Sereda, Alexandr Tseluyko and Nikolay Azarenkov
Figure 8. The dependences of pressure (a), discharge current (b) and collector current
(c) on discharge voltage at P = 5×10-6 Torr, H = 0.1 T.
170 Ihor Sereda, Alexandr Tseluyko and Nikolay Azarenkov
One can see from the Figure 8c that the discharge behaves in the same
way as in the previous experiments with the MH-cathode and with additional
external hydrogen supply into the vacuum chamber. Moreover, an increase
in pressure does not sufficiently shift the voltage of the discharge transition
to the third mode. All these issues give an opportunity to work only on
hydrogen desorbed from the MH-cathode without external supply using the
data carried out in previous experiments [18]. The water-cooling of the MH-
cathode is only the tool for pressure stabilization here.
Thus, the experiments confirm that forced cooling of the MH-cathode
allows to stabilize the pressure in discharge cell and eliminates the hydrogen
waste due to the uncontrolled thermal decomposition of hydride phases. The
low temperature of the MH-cathode (below the decomposition temperature
of hydride phases) ensures the desorption of hydrogen only due to ion-
stimulated processes. This allows to effectively control the rate of hydrogen
desorption by discharge current. It also significantly reduces the
consumption of hydrogen and increases the time of continuous operation of
the cathode. Any qualitative changes are not observed in the discharge
operation in comparison with the uncooled MH-cathode.
Figure 9. The electron energy distribution functions for Ud = 3.5 kV in case of constant
pressure P = 5×10-6 Torr (a) and in case of constant magnetic field H = 0.1 T (b).
Figure 10. Ion energy distribution function at H = 0.1 T for different pressures, cathode
types and discharge voltages.
Figure 11. Plasma density on the axis of Penning discharge with different cathodes
depending on discharge voltage, P = 5×10-6 Torr, H = 0.1 T.
Figure 12. Plasma density on the axis of Penning discharge with different cathodes
depending on magnetic field intensity, P = 5×10-6 Torr.
Figure 13. Plasma density on the axis of Penning discharge with different cathodes
depending on discharge voltage, P = 3×10-5 Torr, H = 0.1 Т, 1 and 2 – by MH-cathode
and by opposite side correspondingly, 3 – by MH water-cooled cathode, 4 – by cathode
with balloon H2 supply, 5 – by both solid copper cathodes.
Thus, the MH-cathode is not only the source of local hydrogen supply,
but also increases the efficiency of plasma device as a whole [23]. When the
discharge operates on the hydrogen desorbed from the MH-cathode, the
plasma density n sharply increases. Moreover, the largest values of n ≈
1.2×1010 см-3 are achieved by the MH-cathode without water-cool,
regardless of initial gas environment. On the opposite side, by the cathode-
reflector, the plasma density is somewhat lower, so that a density gradient
forms in the MH-cathode direction. The additional cooling of the MH-
cathode reduces hydrogen desorption rate, plasma density takes on lower
values and the density gradient along the axis practically disappears. When
hydrogen supplied only from a balloon, regardless of the method of its
injection, the mode with axial electron yield is not observed and the plasma
density is much lower, than in case of the MH-cathode application.
The Application of Metal Hydride Based on Zr-V Alloy … 177
sharply along with magnetic field increasing from 0.02 T to 0.04 T. Such a
behavior was the same for all investigated hydrogen pressures. The sharp
decrease in the current of extracted particles at H < 0.04 T is due to the
retraction of accompanying electrons along the external magnetic field.
To determine the contribution of accompanying electrons to the total
registered current, argon was used as working gas, which does not form
negative ions. To keep the conditions of experiment unchangeable, the MH-
cathode had been not saturated with hydrogen in this case.
In order to analyze the obtained dependences, the current I– was plotted
as a function of hydrogen pressure in the chamber for the different types of
hydrogen injection (Figure 15). As one can see from the figure, the internal
hydrogen injection from the MH-cathode under the impact of high-energy
ions leads to a significant increase in the current I– in comparison with
external hydrogen supply.
One can see from the picture, that the complete cutoff of accompanying
electrons is observed at H > 0.04 T.
Figure 14. The dependence of negative current on external magnetic field at Ud = 700
V, IMH = 4.3mA, Р = 4×10-2 Torr.
180 Ihor Sereda, Alexandr Tseluyko and Nikolay Azarenkov
Figure 15. The dependence of negative current on the pressure in vacuum chamber at
Н = 0.06 T, Ud = 700 V, IMH = 4.3 mA.
In the range of magnetic field H > 0.04 T and hydrogen pressure from
1.3·10-2 Torr to 4·10-3 Torr the current of negative particles does not
practically change at magnetic field increasing. Therefore, the measured
current at H > 0.04 T can be associated with negative hydrogen ions, which
are formed in this type of discharge.
This is due to the desorption of hydrogen molecules from the MH-
cathode in vibrationally/rotationally excited state. As a consequence, the rate
of H– ions formation increases and their contribution to the total extracted
current of negative particles rises. An increase in the ion current to the MH-
cathode leads to an increase in the yield of negative particles (Figure 16).
Moreover, the higher power transferred to the MH-cathode, the more I–
current registered.
The experiments were carried out only at water-cooled MH-cathode,
because it was the only way to stabilize hydrogen pressure in the cell. If it’s
not, the temperature of metal hydride will rapidly grow up to 400 – 500 K
and the pressure will significantly jump more, than in order of magnitude.
The Application of Metal Hydride Based on Zr-V Alloy … 181
Figure 16. The dependence of negative current on the pressure in vacuum chamber at
different ion current on MH-cathode at Н = 0.06 T, Ud = 700 V.
The magnetic filter was set on the axis of the discharge behind the
aperture in the copper cathode-reflector so, that all reverse magnetic field of
the coil (4) was concentrated outside the discharge cell. For convenience the
distance between the cathode (3), the grid (1) and the electron collectors (2)
were the same and were 0.4 cm. The ion collector (5) were at the distance of
1.8 cm from the copper cathode-reflector (3).
The cathodes and collectors were under ground potential. The grid (1)
was supplied with +3 kV for positive particles removing. The whole
electrodes system was placed in external uniform longitudinal magnetic field
Hzo0 with intensity that could be changed in the range of Hzo0 = 0 – 0.1 T.
The idea is to create reverse magnetic field in the gap between the
cathode (3) and the collector (5) to deflect electrons on the collector (2), but
not impact on H– ions being registered by collector (5).
We calculated its construction basing on the analysis of the trajectories
of negatively charged particles by numerical solution of a motion equation
in axially symmetric electric and magnetic fields [28]. Taking into account
the configuration of the electrodes, which ensures the registration of only
184 Ihor Sereda, Alexandr Tseluyko and Nikolay Azarenkov
the paraxial group of particles, the equation of motion can be obtained in the
paraxial approximation in similar way as it was done in [29]. The result is a
paraxial trajectory equation in meridian plane r = r(z):
One can see that reversal magnetic field has a little effect on the
trajectory of H– ions, whereas, electron trajectories significantly curves, and
at the value of total magnetic field Hz0 = 0 at z = 1.8 cm (collector edge) the
electrons are diverted.
It should be noted that the end of electron trajectories essentially
depends on the initial value of the constant r0, where r0 is the radius of
electron entry in the gap. For example, at zero magnetic field on the collector
edge and r0 > 0.08 cm all the electrons are diverted. Accordingly, electrons
that moves strictly in the coaxial tube 0.16 cm in diameter slightly diverted
by magnetic field and registered by the ions collector. Estimates show that
this current is around an order of magnitude smaller, than the total current
of negative particles registered by the collector. In other words, if the
collector is taken exclusively electron beam, the effect of such a reverse
magnetic field will reduce electron current in about 10 times.
In the same time the trajectories of ions had been little effect on magnetic
field.
An experimental check of the filter operation was carried out using an
electron gun, which simulated the electron flow characteristic for the
discharge cell [22]. The electron gun was set on the axis of the system
instead of the MH-cathode. It created a cylindrical electron beam 1.2 cm in
diameter with a current 10 mA and energy 100 eV. The experimental results
are shown in Figure 19.
One can see that at Hcoil/Hzo0 = 1 the electron beam is diverted almost
completely on the electrons collector. Only a small group of paraxial
particles passes through with a current by order of magnitude smaller than
the total current. Thus, the obtained data are in good agreement with the
calculation. Large errors are caused by using a non-stabilized emission
power source of the electron gun.
So, as a result of numerous calculations, it was built a model that allows
to choose the best external parameters for the efficient separation of H– ions
from the axial flow of charged particles. Good coincidence between the
experimental and the calculated data shown the possibility to apply the
model for the interpretation of following experiments.
186 Ihor Sereda, Alexandr Tseluyko and Nikolay Azarenkov
This figure is also shown the typical curve in the case of discharge work
(dot line) at Ud = 5 kV and Hz0 = 0.08 T. The registered current at Hcoil/Hzo0
= 1 turns out to be higher than the electron current. Taking into account all
the foregoing, in our opinion, it will be justified to consider the collector
current at Hcoil/Hzo0 = 1 as the current of negative hydrogen ions.
Figure 18. The trajectories of electrons and ions as well as the normalized profile of
magnetic field in the gap cathode (z = 0 cm)-collector (z = 1.8 cm).
At Hcoil/Hzo0 > 1 the effect of the axial electrons decreases, but the
magnetic field of the filter Hcoil is already distorts the external magnetic field
Hzo0 in the discharge cell. Therefore, we have not carried out data analysis
for Hcoil/Hzo0 > 1 in the framework of this paper.
The dependence of the collector current Iion col on the discharge voltage
Ud when the filter is switched on and the filter is switched off is shown in
Figure 20. The filter has the optimal distribution of Hz0 when Hcoil/Hzo0 = 1
and the retarding grid has a potential +3 kV.
One can see when the filter is switched on, the ion current H– is
registered only in those modes when the total extracted current also takes on
negative values. The behavior of the total extracted current is described
previously.
The Application of Metal Hydride Based on Zr-V Alloy … 187
Figure 19. Current on H– ion collector 5 (Iion col) and on electron collector 2 (Iel. col)
depending on Hcoil/Hzo0, solid and dash lines – experiments with electron gun, dot line –
experiment with discharge.
Figure 20. Collector current when the filter is switched on Hcoil/Hzo0 = 1 (solid lines)
and switched off (dash lines) depending on discharge voltage.
CONCLUSION
protective gas target under the surface of metal hydride as a result of both
the thermal decomposition of the material and ion stimulated desorption.
Hydrogen sorption by the MH-cathode in plasma occurs only with a
sharp decrease in the ion current on the surface of metal hydride. And much
larger amount of hydrogen is absorbed under the ion impact from plasma in
comparison with the case of ordinary thermal relaxation of metal hydride.
Local hydrogen injection under the impact of ion bombardment of metal
hydride provides return coupling between the intensity of gas desorption and
the parameters of gas discharge. In Penning discharge with the MH-cathode
the effect of desorbed hydrogen is, foremost, an intense yield of electrons
together with ions in longitudinal direction, whereas without metal hydride
there is no longitudinal electron flux. The decisive factor for the emission of
axial electron flux is a significant decrease in near-cathodic potentials during
the development of high-frequency instability under hydrogen desorption in
vibrationally/rotationally excited state from metal-hydride. These electrons
have a weak energy dependence, both on the discharge voltage and on the
magnetic field. The ion energy, on the contrary, decreases when the current
of axial electrons grows, which is due to the widening of intense ionization
region to the discharge axis. Sufficient increase in plasma density is
observed in this case.
The water cool of the MH-cathode allows to stabilize the pressure in the
discharge cell and eliminates the hydrogen waste due to uncontrolled
thermal decomposition of hydride phases. Keeping temperature of the MH-
cathode below the hydride phases decomposition one ensures desorption of
hydrogen only due to the ion-stimulated processes. So, the discharge current
could effectively control the hydrogen desorption rate. It also significantly
reduces the consumption of hydrogen and increases the time of continuous
operation. The qualitative changes in the discharge operation are not
observed in comparison with uncooled MH-cathode.
Hydrogen desorption from the MH-cathode, caused by the ion
bombardment of the surface, leads to a significant increase in the yield of
negative ions in comparison with external balloon hydrogen supply.
Negative ions production takes place in the near-cathode region caused by
the dissociative attachment of thermal electrons to hydrogen, desorbed from
190 Ihor Sereda, Alexandr Tseluyko and Nikolay Azarenkov
REFERENCES
[11] Klochko, YeV; Lototsky, MV; Popov, VV; Shmal’ko, YuF; Borysko,
VN. J. Alloys Compounds, 1997, 261, 259-262.
[12] Borisko, VN; Bobkov, VV; Chebotarev, VV; Garkusha, IYe;
Glazunov, GP; Klochko, YeV; Lototsky, MV; Poltavtsev, NS;
Shmal’ko, YuF; Starovoitov, RI; Tereshin, VI. Hydrogen and Helium
Recycling at Plasma Facing Materials; Hassanein A.; Ed.; Nato
Science Series II: mathematics, physics and chemistry. Kluwer
Academic Publishers: Norwell, MA, 2002, Vol. 54, pp 205-212.
[13] Borisko, VN; Garkusha, IE; Chebotarev, VV; Lototsky, MV; Langner,
J; Sadowski, MJ; Tereshin, VI; Shmal’ko, YuF. J. Nucl. Mat., 2003,
313-316, 465-468.
[14] Barosi, T; Giorgi, A. Vacuum, 1973, 23 (1), 15-19.
[15] Afanas’eva, IA; Borisko, VN; Klochko, YeV; Tseluyko, AF. Probl.
Atomic Sci. and Tech. Series: plasma physics, (13) 2007, 1, 191-193.
[16] Klochko, YeV; Ryabchikov, DL; Sereda, IN; Tseluyko, AF. Probl.
Atomic Sci. and Tech. Series: plasma electronics and new acceleration
methods, (7) 2010, 4, 226-229.
[17] Borgun, IV; Ryabchikov, DL; Sereda, IN; Tseluyko, AF. Probl.
Atomic Sci. and Tech. Series: plasma physics, (83), 2013, 1, 228-230.
[18] Agarkov, AV; Ryabchikov, DL; Sereda, IN; Tseluyko, AF. Probl.
Atomic Sci. and Tech. Series: plasma electronics and new acceleration
methods, (86), 2013, 4, 301-303.
[19] Krendel, YuE; Ionov, AS. Tech. Phys., 1964, 34, 1199.
[20] Borisko, VN; Petrushenya, AA. Tech. Phys., 2003, 48, 80-84.
[21] Khauer, W. J. Appl. Phys., 1966, 37(2), 602-611.
[22] Borgun, IeV; Ryabchikov, DL; Sereda, IN; Tseluyko, AF. J. Physics:
Conference Series, 2014, 514, 012051.
[23] Sereda, IN; Tseluyko, AF; Ryabchikov, DL; Babenko, IV; Hrechko,
YaO; Hetman, VA. Probl. Atomic Sci. and Tech. Series: plasma
electronics and new acceleration methods, (98), 2015, 4, 342-344.
[24] Shmal’ko, YuF; Lototsky, MV; Solovey, VV; Yartys’, VA; Strokach,
AP. Z. Physik. Chem., 1994, 183, 479-483.
[25] Ehlers, KW; Gavin, BF; Hubbard, EL. Nucl. Instrum. and Meth. 1963,
22 (1), 87-92.
192 Ihor Sereda, Alexandr Tseluyko and Nikolay Azarenkov
[26] Bacal, M; Hamilton, GW. Phys. Rev. Lett., 1979, 42, 1538-1540.
[27] Borisko, VN; Borisko, SV; Lapshin, VI; Maslov, VI; Klochko, YeV;
Ryabchikov, DL; Sereda, IN. 29th EPS Conf. Plasma phys. Сontr.
Fusion, Montreux, Switzerland, 2002, 26B, 2.036.
[28] Sereda, IN; Tseluyko, AF; Ryabchikov, DL; Babenko, IV; Hrechko,
YaO; Hetman, VA. Probl. Atomic Sci. and Tech. Series: plasma phys.,
(22), 2016, 6, 241-243.
[29] Brown Ian, G. The physics and technology of ion sources; John Wiley
& Sons: New York, 1989, pp. 85-87.
In: Hydrides: Types, Bonds and Applications ISBN: 978-1-53613-581-7
Editor: Patrick C. Dam © 2018 Nova Science Publishers, Inc.
Chapter 4
N. E. Galushkin*, N. N. Yazvinskaya
and D. N. Galushkin
Department of Technics and Techology,
Don State Technical University, Laboratory of Electrochemical
and Hydrogen Energy, Shakhty, Russia
ABSTRACT
*
Corresponding Author address. Email: galushkinne@mail.ru.
194 N. E. Galushkin, N. N. Yazvinskaya and D. N. Galushkin
1. INTRODUCTION
The studying of metal hydrides started more than 150 years ago, when
it was found that palladium was capable of accumulating lots of hydrogen.
At the present time, binary hydrides have been obtained virtually for every
metal [1, 2]. However for a long time, metal hydrides were out of a wide use
in practice. For the first time, the metal hydrides obtained an essential
commercial application as the negative electrodes in the nickel-metal
hydride batteries [3, 4]. Now, the metal hydrides also have other
applications.
Firstly, they are used in diverse chemical processes. For example,
sodium borohydride, lithium aluminum hydride and lithium
triethylborohydride are usually used as reducing agents in chemical
synthesis; sodium hydride and potassium hydride are used as strong bases in
organic synthesis; hydride complexes represent catalysts and catalytic
Preparation and Application of Metal Hydrides 195
n
MeHn
P
P ,T
1
,T
1
Me
2 2
H2 (1)
2
where Me represents metal or alloy, while P1, T1 and P2, T2 are the pressure
and temperature of the hydrogenation or dehydrogenation processes
respectively.
Preparation and Application of Metal Hydrides 197
At the present time, this method is the main one for hydrogen storage
systems. However it features a number of shortages, too. Firstly, it is inertial.
For a hydrogen release start, the metal hydrides must be heated up to a
certain temperature value, after which with a certain delay (depending on
hydride type), the hydrogen release starts. A similar procedure is needed for
the termination of the hydrogen release. However, the processes of heating-
up/cooldown as well as the processes of hydrogen release initiation/
termination from hydrides are very inertial. Secondly, the launching of the
hydrogen release process from hydrides and the termination of this process
requires supplying an additional energy to the system, which decreases the
power-generating efficacy of the hydrogen storage systems. Thirdly, the rate
of the hydrogen release from metal hydrides in frame of the thermo-chemical
method is not high. To increase it and thereby improve the kinetic
parameters of the system, a temperature increase is needed, which would
worsen the thermodynamic parameters of the system. That is why until now
metal hydrides have not been found, which – in the frame of the thermo-
chemical method – would meet the US DOE requirements on both kinetic
and thermodynamic parameters.
The third method is an electrochemical one consisting of
electrochemical hydrogen accumulation in metal hydrides. During an
electrolyte decomposition, for example, in alkaline medium, on a cathode,
the hydrogen is released.
surface layer of the cathode [23]. That is why the gravimetric capacity of the
cathode as a hydrogen accumulating unit turns out to be rather small. This
method of hydrogen accumulation in cathodes has been well-known for a
long time. In particular, this method of hydrogen accumulation is used in the
nickel-metal hydride batteries for the hydrogen accumulation in negative
electrodes [1, 4].
In this chapter, we shall study the electrochemical method of hydrogen
accumulation in metals and compare it with the thermo-chemical method of
hydrogen accumulation.
2.1. Introduction
Figure 1. The experimental set-up to study the process of gas release from the heated
electrodes. (1) Manometer, (2) intake chamber accumulating gas, (3) thermocouple, (4)
retort, (5) tap, (6) muffle furnace, (7) metal thermal chamber, (8) rubber plug with a
tube for gas extraction, (9) standard coil and (10) heater.
Preparation and Application of Metal Hydrides 201
The intake chamber for collecting the gas 2 was a hermetically sealed
glass reservoir filled with water. Its capacity was 5 L. At the bottom of this
reservoir, a number of tubes were inserted through a rubber plug: a long tube
for supplying gas from the thermal chamber to the top of the reservoir, a tube
for removing excess water, and a leg of a manometer tube 1 for controlling
the pressure of the gas accumulated above the water (Figure 1).
The heating up of the electrode in the thermal chamber resulted in the
gas being released from it. This gas was partially cooled down by its passage
through the standard coil 9 prior to its entering into the gas intake chamber.
As the gas entered the chamber, the pressure above the water increased, and
this was registered by the manometer. In order to balance the internal
pressure in the intake chamber and atmospheric pressure, a part of the water
was removed from the intake chamber into the retort 4. The water level in
the retort gave an indication of the amount of gas that had entered to the
intake chamber during the period studied [25-27].
The electrodes were subjected to decomposition inside of the thermal
chamber at the temperature 800°C, as an essential hydrogen release from
electrodes starts only at the temperature 740°C. The electrodes’
decomposition stopped when a daily hydrogen release from the electrodes
became less than 100 ml. On average, the oxide-nickel electrodes were
decomposed during 13 days, while the cadmium ones lasted 7 days. The
electrodes thermal decomposition was conducted in the mode of 11 hours
per day.
To conduct the experiment, eight batteries KSX-25 were chosen with
different operating lives in the same conditions. From each battery there
were taken three cadmium and three oxide-nickel electrodes. Then, the
electrodes were subjected to the thermal decomposition in the installation
shown in Figure 1. In Table 1, the average values are shown of the amount
of the released hydrogen calculated for three electrodes taken from each
battery.
To find the gas composition, the gas analyzer VOG-2M was used. The
absolute error in the percentage concentrations was 0.3–0.5. The analysis of
the gas released from both cadmium and oxide–nickel electrodes showed
that it consisted only of hydrogen.
202 N. E. Galushkin, N. N. Yazvinskaya and D. N. Galushkin
From the Table 1, it is seen that the hydrogen amount in the electrodes
grows in the course of the batteries’ operation. However after five years of
operation, the hydrogen amount in electrodes stops growing, i.e., peak
(maximum) capacity of the electrodes is reached as the hydrogen is stored.
In the battery KSX-25, there are 14 cadmium and 15 oxide–nickel
electrodes. That is why in one battery KSX-25 after five years of operation,
approximately 820 liters hydrogen are accumulated.
As for this hydrogen amount, it is quite realistic that a battery KSX-25
was able to accumulate it during five years of operation. Indeed, according
to batteries’ Operation Manual, a battery KSX-25 is charged in two stages:
at first by the current 10A during 3 hours and then by the current 5A during
2 hours. Hence, batteries are overcharged 1.6 times in comparison to their
nominal capacity. The overcharge is necessary for the batteries to fully
charge. So during one charging session of the battery KSX-25, there can be
up to 6 L hydrogen and up to 3 L oxygen released. Thus, for the 820 liters
of hydrogen found in the electrodes, theoretically, any battery KSX-25 was
able to accumulate them in the course of 137 cycles of charge-discharge.
During five years of its operation, any battery KSX-25 passes through a ten
times bigger number of cycles of charge-discharge. That is why,
theoretically, it is able to accumulate the found hydrogen amount.
Now, let us evaluate the gravimetric capacity of the oxide-nickel
electrode as a hydrogen accumulating unit. After five years of operation
in an oxide-nickel electrode, about 36 liters hydrogen is accumulated
Preparation and Application of Metal Hydrides 203
electrodes were washed out in the distilled water, dried up and weighed
again. The results of the mass loss of the oxide-nickel electrodes after their
etching in the sulfuric acid are shown in the Table 2.
The experiments showed that during the etching of an oxide-nickel
electrode in sulfuric acid, the hydrogen does not release at all. Hence, in the
active substance of the oxide-nickel electrode, the hydrogen is accumulated
neither in its free state nor in its weakly bound state, which could be
decomposed by the sulfuric acid causing a hydrogen release.
The mass loss of the oxide-nickel electrodes as a result of their etching
in the sulfuric acid makes 33-36% (Table 2). According to the data of the
batteries KSX-25 manufacturer, a positive electrode contains 30-34% nickel
hydroxide and 1.5-2% cobalt hydroxide. The results obtained for the mass
loss of the electrodes fall exactly into this range. Therefore, during batteries’
operation, with the electrodes’ active substance, the hydrogen did not form
any stable chemical compounds unable to be decomposed in the sulfuric
acid.
minutes for removal of the nickel hydroxide. Then the electrodes were
washed out in the distilled water and dried up. Then the electrodes were
subjected to the thermal decomposition inside of the installation Figure 1 at
the temperature 800°C. In the Table 3, the average hydrogen amount inside
of the electrodes is given. It was calculated for three electrodes from each
battery. The obtained experimental results for the amount of the hydrogen
released from the sintered nickel matrix coincide (in limits of an
experimental error) with the similar results obtained for the oxide-nickel
electrode (Table 1). Thus, during batteries’ operation, the hydrogen is
accumulated in the sintered matrices of electrodes.
Now let us assess the gravimetric capacity of the sintered nickel matrix
of the oxide-nickel electrode as a hydrogen accumulating unit. After five
years of operation in an oxide-nickel electrode, about 36 liters of hydrogen
is accumulated (Tables 1, 3). The weight of the oxide-nickel electrode is 24
g and the weight of the nickel hydroxide is 8 g. Therefore, the gravimetric
capacity of the sintered nickel matrix of the oxide-nickel electrode is equal
to 20.1 wt%. This value exceeds by three times, the earlier data obtained by
the traditional thermo-chemical method for any reversible metal hydrides
including magnesium hydride and complex hydrides [1, 40]. As the physical
sizes of the oxide-nickel electrode and ones of its sintered nickel matrix
are the same, their volumetric capacities are also the same, i.e., 400 kg m-3
[25, 26].
studies that lasted more than twenty years [20-30]. Nevertheless, it should
be noted that in the oxide-nickel electrode, a number of factors are present,
which promote the hydrogen accumulation. At the present time, those
factors are under an extensive investigation.
Firstly, the oxide-nickel electrode in the batteries KSX-25 is sintered. It
is produced from finely divided nickel powder. This powder has a badly
destroyed metal crystal structure. As it is known [1], any defects of metal
crystal structure (in particular, dislocations) are traps for hydrogen. Besides,
they decrease the energy of hydrogen atoms in comparison with their
location in normal interstices. All this promotes the hydrogen penetration of
the metal. Hence, any imperfections of a metal crystal structure must cause
a drastic increase of its hydrogen capacity. With this purpose, the hydrides
used in the modern hydrogen storage systems are ground down finely in ball
mills [42, 43].
Secondly, the above experiments (Table 1) showed that the hydrogen
released on a cadmium electrode during battery charging can penetrate into
the pores of both oxide-nickel and cadmium electrodes. Most probably, this
is connected with the dense packing of the electrodes in the batteries KSX-
25. Thus, it is possible that the oxide-nickel electrode accumulates the
hydrogen not electrochemically, but instead because of the high capillary
pressure. According to the papers [44, 45], the concentration CH of the
hydrogen accumulated in metal is connected with a hydrogen external
pressure Р in line with the formula
CH K P . (5)
P = A/R, (6)
The finely divided powder used for the manufacturing of the electrodes’
sintered matrix has micro fissures sized from a few hundreds angstrom up
to the size of the metal crystal structure cell. Meanwhile the pores’ average
size in the sintered electrodes of the batteries KSX-25 is a few tens microns.
So the hydrogen accumulated in the micro fissures because of the electrolyte
decomposition can stay there under the pressure reaches 100 MPa and more
[24].
Thirdly, in the oxide-nickel electrode, the active substance is nickel
oxide. However, it is well-known that the oxides of transition metals act as
catalysts of hydrogen accumulation [46, 47].
And finally, the hydrogen accumulation process in the batteries’
electrodes runs during their whole operating life. There is more than five
years in this research. In the frame of the traditional thermo-chemical
method, the process of hydrides preparation takes time, from a few minutes
up to several hours. While analyzing the papers on hydrides investigation
for the last 20 years, we did not find studies in which metal hydrides were
subjected to high external pressures during a very long period of time.
However, the exposure duration of an external pressure is an important
factor in any processes of the penetration of one substance into another. This
is especially so, when the penetration process is extremely slow as it is in
our case.
Apparently, the high capillary pressure of hydrogen, acting for an
extremely long time on the strongly destroyed crystalline structure of the
electrodes’ metal over the catalyst (an oxide of a transition metal) allowed
the accumulation of such a large amount of hydrogen inside the electrodes.
Of course, this assumption still requires both experimental and theoretical
studies.
It should be noted that the battery KSX-25 is operated in the floating
mode. That is why during most of the time, it stays under charging. Within
this time, its electrolyte is decomposed onto the hydrogen and the oxygen.
The hydrogen is featured with very high diffusion permeability. For
example, at the temperature 20°C, the diffusion coefficient of the hydrogen
in the nickel is about 1010 times higher than that of nitrogen or oxygen [1,
2]. That is why when the electrolyte is decomposed onto hydrogen and
208 N. E. Galushkin, N. N. Yazvinskaya and D. N. Galushkin
3.1. Introduction
The peak value for the hydrogen full flow rate from the system was
found out in the paper [49]. It turned to be equal to 8.333·10-4 wt% s-1.
The kinetic parameter (7) is quite reachable for the modern hydrides
based on magnesium [50] and complex hydrides [51]. Nevertheless, these
hydrides do not meet the US DOE requirements on their thermodynamic
parameters. First of all, this is connected with the fact that the hydrogen
desorption from these hydrides runs at high temperatures.
At the present time, two methods of hydrogen desorption exist.
The first method of hydrogen desorption is the chemical one, i.e.
With use of this method, the hydrogen amount turns out to be twice its
content in the initial hydride. However, the reaction (8) is possible only for
a limited number of hydrides [2]. Besides, it is absolutely irreversible. At the
present time, this method of desorption is considered to be unacceptable for
hydrogen storage systems – first of all, because of economical
considerations.
The second method of hydrogen desorption is the thermo-chemical one
discussed in detail in the section 2. At the present time, this method is the
main one for hydrogen storage systems.
In this section we study thermal runaway as the new method of hydrogen
desorption from hydrides.
Preparation and Application of Metal Hydrides 211
Figure 4. Change in parameters of the KSX-25 battery during thermal runaway: I is the
charging current of the battery, U is the voltage of the battery terminals, H is the
hydrogen desorbed (in wt%), T is the temperature of the battery positive terminal.
As a result of the thermal runaway from the battery, which lasted about
4 minutes, 280 liters hydrogen was released. Thus, the thermal runaway
drastically accelerates the hydrogen escape from the hydrides in comparison
with the thermo-chemical method.
214 N. E. Galushkin, N. N. Yazvinskaya and D. N. Galushkin
Besides, it should be noted that the total hydrogen amount (in wt%)
desorbed during the thermal runaway (Figure 4) was approximately
equal to these electrodes’ gravimetric capacity, experimentally found out in
section 2.
3.3. Discussion
during the thermal runaway, a very large amount of energy is released (about
5012 kJ [30]), which also can be utilized rationally. So on kinetic and
thermodynamic parameters, the thermal runaway exceeds far behind the
minimum requirements set by the US DOE [12] (Table 4).
However, the thermal runaway, as the method of hydrogen desorption
from hydrides, also contains a good deal of essential shortages. Firstly, under
common conditions, the thermal runaway is very unlikely for the cycling of
nickel-cadmium batteries [28-32]. Secondly, this is an episodic and hardly
repeatable phenomenon. However, based on comprehension of the real
mechanism of the thermal runaway, these problems can be solved.
According to the standard point of view [39, 53], the thermal runaway
takes place after a battery’s long charging session at constant voltage (or in
a case of battery work in the floating mode). Along with it, it is thought that
the reason of the thermal runaway is an acceleration of standard
electrochemical reactions to battery charging due to its self-heating [39].
However in the papers [29, 30], a lot of experimental facts are reported,
which contradict the widely accepted mechanism of the thermal runaway
[39].
Firstly, the widely accepted mechanism of the thermal runaway
supposes that during the thermal runaway, a battery heat-up is going on,
which results in a decrease of its internal resistance and subsequently in
charging current growth, etc. Thus, it is supposed that the thermal runaway
is connected with an intensification of electrochemical reactions running in
the course of a battery charging session. Besides, during thermal runaway,
an intensification is also possible of electrochemical reactions of electrolyte
decomposition (2, 3) as the thermal runaway takes place after a long
overcharge of a battery [39]. In this case, as a result of the thermal runaway
from the battery, a steam-gas mixture must release. The steam is formed
because of electrolyte evaporation due to the high temperature of the thermal
runaway. Meanwhile, gas can be formed during the thermal runaway due to
the electrolyte decomposition onto hydrogen and oxygen. Subsequently, in
the released gas, the ratio of hydrogen/oxygen must be 2:1. However, the
direct experimental testing [24-30] shows that the hydrogen makes not less
than 95% in the composition of the released gas mixture in the thermal
216 N. E. Galushkin, N. N. Yazvinskaya and D. N. Galushkin
runaway, the remaining negligible part of the gas is oxygen. Notably, in the
released gas mixture, the hydrogen amount is larger than its content in the
full volume of the battery electrolyte to decompose it onto hydrogen and
oxygen [35]. Besides, the amount of the steam released in the thermal
runaway is approximately equal to the electrolyte amount in the battery. This
fact can be explained only if we suppose that the hydrogen was accumulated
in the battery in the course of its operation, which is exactly what
experimental studies confirm (Table 1 and [25, 26]). It should be noted that
in the course of batteries KSX-25 operation, electrolyte is added into them,
when needed.
Secondly, according to the widely accepted mechanism, the thermal
runaway emerges at the expense of that energy, which batteries obtain from
their charger [39]. However, the direct calorimetric measures [30] showed
that as the result of the thermal runaway, a battery releases energy 140 times
more than it obtains from its charger at the same time. This fact can be
explained based only on the assumption that the thermal runaway is
connected with the running of a very powerful exothermic reaction.
Also, other experimental facts exist that contradict the widely accepted
mechanism of the thermal runaway [29, 30].
In the paper [23], it was shown that the reason of the thermal runaway
in nickel-cadmium batteries is an emergence of the exothermic reaction of a
recombination of atomic hydrogen accumulated in battery electrodes in the
course of operation of the battery, i.e.
The hydrogen is present inside the battery electrodes in the form of metal
hydrides – that is in its atomic form [1]. The reaction (9) is the powerful
exothermic reaction with the heat release of 436 kJ/mol (hydrogen). [30, 54].
Preparation and Application of Metal Hydrides 217
This heat release is more powerful than in the case of the reaction of
hydrogen burning in oxygen (285.8 kJ/mol) [55, 56].
Also in the paper [23], it was experimentally proved that a rate-limiting
step for the reaction (9) is the step of the metal hydrides decomposition.
separator in the form of the regular circles. Indeed, in reality, those circles
are observed in experiments (Figure 3 and papers [23-27]).
Two factors prevent an expansion of the exothermic reaction of the
thermal runaway (9) to the entire electrode. Firstly, there is a gas plug, which
is formed between electrodes in the case of the thermal runaway due to the
electrolyte evaporation (because of the high temperature) and to the
hydrogen release (10, 11). Physically, this gas plug can interrupt the
electrochemical reactions of the thermal runaway (10, 11). Secondly, there
is a depletion of the hydrogen in the electrodes in spots of the thermal
runaway.
It should be noted that the emergence of the regular circles indicating
the separator burning during the thermal runaway is impossible to be
explained [30] in the frame of the widely accepted mechanism of the thermal
runaway [39].
At the common operation conditions of the nickel-cadmium batteries
KSX-25, the current generated by the electrochemical reactions (10, 11) is
rather small; approximately it is equal to 0.22A [23]. However, in the case
of the thermal runaway occurrence, the current generated by the
electrochemical reactions (10, 11) becomes equal to hundreds of amperes;
notably, in this case, the voltage on the batteries terminals falls down to
0.5 - 0.6V (Figure 4) in accordance with the voltage for the reactions
(10, 11) [23].
Thus, knowing the true mechanism of the thermal runaway, it is possible
to create it artificially and reliably. Indeed, for the launching of the
electrochemical reactions of the thermal runaway (10, 11), a powerful local
heating up of the electrodes is needed. In the natural conditions of the
batteries KSX-25 work, this is reached due to the dendrites’ intergrowth
through a separator. However a powerful local heating-up of electrodes can
also be caused artificially, by imposing (on battery’ terminals) a voltage
sufficient for punching-through the distance between the electrodes in the
battery.
In our experiments for the artificial and reliable launching of the process
of the thermal runaway, an impact of charging voltage more than100V was
applied onto the terminals of the battery KSX-25 for 0.5 seconds. Then for
Preparation and Application of Metal Hydrides 219
life that was used in the experiment shown in Figure 4. Along with it, the
initiation of the thermal runaway was conducted in line with the above
described method.
As a result, the peak current of the thermal runaway decreased about
twice the full flow rate of the hydrogen and desorption decreased almost
nine times down to 0.137 wt% s-1 [48].
Thus by regulating the current value of the electrochemical reactions of
the thermal runaway (10, 11), it is possible to set any needed rate of the
hydrogen desorption.
3.4. Conclusion
REFERENCES
[11] Mathiesen, B.V.; Lund, H.; Connolly, D.; Wenzel, H.; Østergaard,
P.A.; Möller, B.; Nielsen, S.; Ridjan, I.; Karnøe, P.; Sperling, K.;
Hvelplund, F.K. Appl. Energy 2015, 145, 139-154.
[12] Hydrogen, Fuel Cells & Infrastructure Technologies Program. Multi-
Year Research,Development and Demonstration Plan. Planned
program activities for 2005–2015; U.S. Department of Energy,
Energy Efficiency and Renewable Energy: Washington, 2009.
[13] Gross, K.J.; Carrington, K.R.; Barcelo, S.; Karkamkar, A.; Purewal,
J.; Ma, S.; Zhou, H-C.; Dantzer, P.; Ott, K.; Burrell, T.; Semeslberger,
T.; Pivak, Y.; Dam, B.; Chandra, D. Recommended Best Practices for
the Characterization of Storage Properties of Hydrogen Storage
Materials; V3.34, Feb. 21; EERE: Washington, 2012.
[14] Eihusen, J.A. In 15th World Hydrogen Energy Conf. "WHEC-15";
Yokohama, Japan, June 27-July 2; 2004; pp 301-307.
[15] Energy Technology Analysis; International Energy Agency: Paris,
2005; pp 70-75.
[16] Ahluwalia, R.K.; Hua, T.Q.; Peng, J.K.; Kumar, R. System Level
Analysis of Hydrogen Storage Options; DOE Hydrogen Program
Review: Washington, DC, 2010.
[17] Shahi, R.R.; Yadav, T.P.; Shaz, M.A.; Srivastva, O.N. Int. J.
Hydrogen Energy 2010, 35, 238-246.
[18] Zhang, F.; Zhao, P.; Niu, M.; Maddy J. Int. J. Hydrogen Energy 2016,
41, 14535-14552.
[19] Lototskyy, M.V.; Yartys, V.A.; Pollet, B.G.; Bowman Jr, R.C. Int. J.
Hydrogen Energy 2014, 39, 5818-5851
[20] Li, L.; Jiang, G.; Tian, H.; Wang, Y. Int. J. Hydrogen Energy 2017,
42, 28464-28472.
[21] Pyle, D.S.; Gray, E.M.; Webb, C J. Int. J. Hydrogen Energy 2016, 41,
19098-19113
[22] Hoogers G. Fuel Cell Technology Handbook; CRC Press: Boca
Raton, FL, 2003.
[23] Galushkin, N.E.; Yazvinskaya, N.N.; Galushkin, D.N. J. Electrochem.
Soc. 2015, 162, A2044-A2050.
Preparation and Application of Metal Hydrides 223
[56] Suban, M.; Tusek, J.; Uran, M. J. Mater. Process. Technol. 2001, 119,
193-198.
BIOGRAPHICAL SKETCHES
Professional Appointments:
He is a professor at Don State Technical University, Russia. He heads a
research laboratory "Electrochemical and Hydrogen Energy".
Professional Appointments:
She is an associate professor at Don State Technical University, Russia.
She is a senior researcher in the research laboratory "Electrochemical and
Hydrogen Energy".
atomic hydrogen accumulated in the electrodes, which runs in line with the
electrochemical mechanism.
Professional Appointments:
He is a professor at Don State Technical University, Russia. He is a
senior researcher in the research laboratory “Electrochemical and Hydrogen
Energy”.
B
C
band gap, 98, 99, 101, 112, 132
Ca2+, 52
batteries, viii, x, 5, 20, 74, 88, 132, 133,
cadmium, viii, x, 193, 199, 201, 202, 205,
137, 144, 145, 147, 193, 194, 198, 199,
206, 208, 211, 214, 215, 216, 217, 218
201, 202, 203, 204, 205, 206, 207, 208,
calcium, 29, 72, 74, 195
carbon, vii, viii, 2, 44, 144, 195, 226
232 Index
catalyst, 118, 196, 207 covalent bond, ix, 3, 82, 84, 87, 90, 92, 93,
charge density, ix, 82, 89, 91, 92, 107, 114, 102, 105, 112, 113, 115, 116, 122
115, 122, 126, 134, 136, 146 critical value, 118, 120, 123
chemical bonding, vii, viii, ix, 1, 82, 83, 85, crystal structure, vii, viii, 1, 2, 3, 6, 7, 11,
107, 109, 111, 112, 113, 115, 122, 137, 12, 13, 16, 17, 19, 20, 25, 29, 30, 31, 34,
138 37, 38, 41, 46, 48, 55, 56, 60, 64, 68, 69,
chemical bonds, ix, 82, 83, 91, 138 70, 72, 74, 75, 76, 79, 87, 89, 91, 95, 96,
chemical method, xi, 194, 196, 197, 198, 97, 125, 130, 138, 145, 146, 206, 207
205, 207, 208, 209, 211, 214, 220, 221 crystalline, ix, 3, 6, 21, 53, 66, 77, 82, 86,
chemical properties, 5, 118 117, 207
chemical reactions, 3, 196 cycles, 133, 199, 202, 212
classification, 74, 85, 86 cycling, 199, 212, 215
clusters, 118, 119, 120, 122, 123, 124, 125
complex hydrides, xi, 2, 3, 46, 87, 92, 94,
D
101, 102, 104, 108, 109, 111, 112, 117,
120, 121, 124, 130, 132, 135, 136, 137,
decomposition, viii, ix, x, 20, 37, 53, 76, 81,
144, 146, 194, 203, 205, 210, 225
91, 94, 104, 117, 119, 120, 129, 155,
composites, 4, 65, 76, 143, 144, 148
156, 157, 170, 189, 193, 197, 199, 201,
composition, 43, 96, 120, 161, 201, 212,
207, 211, 212, 215, 217, 220, 227
216
decomposition temperature, ix, 20, 37, 81,
compounds, vii, viii, 1, 3, 4, 8, 19, 37, 44,
91, 94, 104, 117, 119, 120, 155, 170
47, 51, 53, 64, 84, 85, 88, 92, 96, 100,
deformation, 29, 34, 36, 120
103, 106, 107, 108, 109, 111, 116, 121,
degenerate, 84, 98, 102, 112
134, 136, 137, 160, 204
dendrites, 217, 218
conduction, 78, 84, 98, 100, 101, 112, 132,
density functional theory, 74, 92
137, 145, 146, 148
Department of Energy, xi, 194, 195, 208,
conductivity, 97, 130, 132, 146, 148, 156,
222, 225, 226
157
derivatives, 5, 10, 75, 78, 108
conductor(s), 5, 39, 82, 98, 132, 135
desorption, x, xi, 118, 119, 150, 151, 155,
configuration, 27, 95, 96, 151, 183, 188
157, 160, 161, 162, 167, 168, 170, 172,
constituents, 82, 83, 85, 92, 110, 111, 114,
174, 175, 176, 177, 180, 181, 182, 187,
116, 119, 122
188, 189, 194, 208, 209, 210, 211, 212,
construction, 183, 208
214, 215, 219, 220, 221, 225, 226
consumption, 170, 189
DFT, 21, 29, 92, 94, 98, 137
conversion reaction, 130, 144
dielectric constant, 109
cooling, 12, 13, 21, 24, 27, 29, 35, 40, 158,
diffraction, 6, 10, 11, 12, 13, 14, 21, 27, 29,
170, 173, 176
40, 51, 53, 54, 69, 70, 77, 79, 138
copper, 152, 153, 155, 156, 159, 161, 162,
diffusion, viii, x, 88, 95, 117, 139, 193, 196,
164, 166, 167, 168, 171, 174, 175, 176,
207, 220
182, 183
diffusion permeability, viii, x, 193, 196, 207
correlation, 66, 67, 98
distilled water, 203, 205
covalency, 103, 110
Index 233
distribution, ix, 12, 82, 85, 89, 92, 94, 95, environment, ix, 19, 25, 26, 28, 39, 47, 48,
97, 113, 114, 115, 116, 122, 123, 126, 82, 83, 85, 86, 95, 96, 124, 126, 171,
134, 136, 166, 170, 171, 172, 173, 186 176, 181
distribution function, 98, 170, 171, 172, 173 etching, 203, 204
donors, vii, ix, 82 evaporation, 215, 218
evolution, 15, 50
excitation, 187, 190
E
exothermic reaction, 216, 217, 218, 226,
228
electric field, 110, 162, 164, 182, 184, 188
exposure, 155, 207
electrical conductivity, 84, 87, 98, 133
external magnetic fields, 187
electricity, viii, 1, 84
extraction, x, 150, 178, 182, 200
electrochemical method, viii, x, 193, 198,
199, 208
electrode, xi, 39, 130, 137, 144, 145, 154, F
193, 194, 200, 201, 202, 203, 204, 205,
206, 207, 208, 213, 214, 217, 218, 226, Fermi level, 98, 100, 101, 102, 104, 112
228 flexibility, vii, viii, 1, 8, 112
electrolyte, viii, x, 133, 193, 197, 198, 199, formation, x, 28, 30, 31, 39, 76, 86, 88, 91,
206, 207, 212, 215, 218, 227 101, 112, 120, 144, 150, 151, 156, 157,
electromagnetic, 183 161, 167, 168, 175, 180, 181, 187, 190
electron diffraction, 6, 75 formula, 16, 18, 60, 96, 114, 120, 128, 206
electron microscopy, 157 fragments, 22, 26, 28, 45
electronic structure, ix, 82, 83, 140 free energy, vii, viii, 2
electrons, 83, 84, 87, 88, 90, 91, 92, 93, 95, fuel cell, 68, 196, 209
97, 98, 100, 103, 106, 109, 111, 114,
115, 116, 122, 134, 151, 161, 162, 164,
G
166, 167, 168, 170, 171, 173, 177, 178,
179, 182, 183, 185, 186, 187, 188, 189
gadolinium, 65, 78
emission, 160, 167, 168, 182, 185, 189
geometry, 10, 14, 19, 31, 51, 53, 125
energy, vii, viii, ix, x, 1, 2, 4, 28, 30, 31, 65,
gravimetric capacity, xi, 193, 194, 195, 198,
70, 79, 81, 82, 88, 92, 93, 96, 97, 98,
202, 205, 208, 214
100, 101, 102, 103, 104, 112, 114, 115,
growth, 2, 162, 164, 195, 212, 215, 217
118, 120, 121, 123, 124, 128, 129, 133,
135, 137, 138, 150, 151, 152, 154, 156,
157, 161, 164, 168, 170, 171, 172, 173, H
177, 178, 179, 181, 185, 187, 189, 190,
195, 197, 206, 215, 216, 221, 226, 227 Hamiltonian, 85, 103, 113
energy density, 2, 93, 157 heat release, 151, 217
energy transfer, 154, 181, 190 heat removal, 151
engineering, 46, 225 heat transfer, 156
heavy metals, 46
234 Index
magnitude, 98, 100, 102, 108, 116, 136, negative ions, x, 83, 150, 152, 168, 177,
164, 180, 185 178, 179, 181, 182, 187, 189
materials, vii, viii, ix, x, 2, 3, 5, 25, 39, 46, neutral, 151, 154, 158, 162, 177, 182
66, 74, 75, 79, 81, 83, 84, 85, 88, 89, 94, NH2, 3, 128, 131, 132, 145, 146
98, 105, 107, 108, 111, 114, 115, 117, nickel, viii, x, 193, 194, 196, 198, 199, 201,
118, 122, 124, 125, 128, 129, 130, 132, 202, 203, 204, 205, 206, 207, 208, 211,
137, 139, 140, 141, 146, 149, 150, 156, 213, 214, 215, 216, 217, 218
157, 160, 188, 195, 224, 226 nickel-cadmium, viii, x, 193, 199, 205, 208,
matrix, xi, 18, 94, 96, 103, 105, 109, 134, 211, 214, 215, 216, 217, 218
193, 203, 204, 205, 207 nitrogen, 25, 196, 207
melting, 86, 87, 123, 153, 157 NMR, 69, 125, 126
metal hydride(s), vii, x, xi, 9, 14, 20, 86, 88, nuclei, 122, 126, 134
92, 94, 117, 130, 137, 139, 140, 141,
149, 150, 151, 153, 154, 156, 157, 159,
O
160, 161, 162, 167, 175, 177, 180, 188,
189, 190, 193, 194, 195, 196, 197, 203,
optical properties, vii, ix, 82, 140
205, 207, 208, 212, 215, 217, 219, 220,
optimization, ix, 14, 46, 82
221, 225, 226
optoelectronic properties, ix, 82
metal ion, 77
optoelectronics, vii, ix, 81
metal oxides, 46, 67
oscillation, 162, 163
metals, 8, 84, 86, 88, 98, 140, 196, 198, 220
oxide-nickel, xi, 193, 194, 201, 202, 203,
Mg2+, 27, 51, 52, 128
204, 205, 206, 207, 208, 211, 214
models, 21, 29, 226
oxygen, viii, x, 193, 196, 198, 199, 202,
modifications, 46, 122, 146
207, 212, 216, 217
molecular orbital, 28
molecular structure, 5, 77
molecules, x, 6, 84, 87, 92, 115, 134, 150, P
151, 168, 177, 178, 180
molybdenum, 139 particle creation, 184
phase diagram, 45, 147
phase transformation, 69, 71, 72
N phase transitions, 15, 21, 24, 27, 30, 70, 75
physical properties, vii, viii, 1, 46, 117
Na+, 12, 14, 18, 83, 128, 137
PIG, 150
NaCl, 12, 16, 18, 83, 86, 88, 89, 92, 93, 100,
polymorphism, 28, 75
104, 106, 107
population, ix, 2, 82, 85, 103, 106, 109, 113,
nano phase hydrides, 82
116, 122, 124, 133, 134
nanocomposites, 141
porosity, 8, 20, 25
nanoparticles, 118, 121, 126
potassium, 37, 48, 57, 68, 78, 194
nanophases, 91, 120, 124, 125, 126, 137
purification, 150, 154, 196
nanoscale materials, 117
purity, x, 149, 153, 196
nanostructures, 142, 195, 226
236 Index
zinc, 25
zirconium, 72, 150, 177