Non-solvated aluminum hydride. Crystallization from diethyl ether-benzene solutions

Article abstract of DOI:10.1007/s11172-007-0199-5

Crystallization of non-solvated aluminum hydride from a diethyl ether-benzene mixed solvent was studied. The desolvation of AlH ₃?(Et₂O)_x etherate in solution and the crystallization of α-AlH₃ during polythermal heating of the solution occur only in the presence of >10 wt.% LiAlH₄. The process is multistage, and the crystallization begins with the formation of the AlH₃?0.25Et₂O solvate, which recrystallizes in the solid phase into γ-AlH₃ and then α-AlH₃. Four crystalline modifications of aluminum hydride were characterized by X-ray diffraction and electron microscopy.

Full text of DOI:10.1007/s11172-007-0199-5

Russian Chemical Bulletin, International Edition, Vol. 56, No. 7, pp. 1305—1312, July, 2007
1305
Nonꢀsolvated aluminum hydride. Crystallization
from diethyl ether—benzene solutions
a
a
a
a
b
B. M. Bulychev,a V. N. Verbetskii, A. I. Sizov, T. M. Zvukova, V. K. Genchel´, and V. N. Fokin
ꢀ
aDepartment of Chemistry, M. V. Lomonosov Moscow State University,
1 Leninskie Gory, 119992 Moscow, Russian Federation.
Fax: +7 (495) 932 8846. Eꢀmail: B.Bulychev@highp.chem.msu.ru
bInstitute of Problems of Chemical Physics, Russian Academy of Sciences,
1 prosp. Akad. Semenova, 142432 Chernogolovka, Moscow Region, Russian Federation.
Fax: +7 (496) 522 1743
Crystallization of nonꢀsolvated aluminum hydride from a diethyl ether—benzene mixed
solvent was studied. The desolvation of AlH3•(Et2O)x etherate in solution and the crystallizaꢀ
tion of αꢀAlH3 during polythermal heating of the solution occur only in the presence of ≥10 wt.%
LiAlH4. The process is multistage, and the crystallization begins with the formation of the
AlH3•0.25Et2O solvate, which recrystallizes in the solid phase into γꢀAlH3 and then αꢀAlH3.
Four crystalline modifications of aluminum hydride were characterized by Xꢀray diffraction
and electron microscopy.
Key words: aluminum hydride, alanes, etherates, synthesis, phase transitions, crystalline
modifications.
The development of fuel cells using hydrogen as a
combustible component requires search for economical
and convenient methods for obtaining and storing this
gas. It is most convenient to store hydrogen bound into
hydrides. Among the latter, nonꢀsolvated aluminum
hydride alane is the best in terms of hydrogen weight
content and volumetric density, decomposition temperaꢀ
ture, and safety.
Alane in the form of adducts with various donor molꢀ
ecules with the composition AlH3•nL (L designates terꢀ
tiary amines, ethers, phosphines, etc.) has been known
since the 1940s.1 However, although all these substances
are aluminum hydride derivatives, they are inappropriate
as a hydrogen source for several reasons, first of all, beꢀ
cause of the low percentage of "active" hydrogen and high
fire risk.
The first perfect procedure for the synthesis of pure
(nonꢀsolvated, i.e., containing no donor molecules) alane
was described in Ref. 2. The method proposed for the
synthesis is the crystallization of aluminum hydride from
an ether—benzene mixed solvent in the presence of
10—20 wt.% LiAlH4 and (sometimes) the same amount
of LiBH4. Rhombohedral (hexagonal) αꢀAlH3 is formed
with solvent distillation, decreasing the weight fraction of
ether and increasing the temperature of the solution. In
addition to αꢀAlH3, several other modifications of nonꢀ
solvated alane (α´, γ, ε, δ, and ξ) were obtained.2 These
substances are formed upon both crystallization from soꢀ
lution and heating in vacuo of a fine dry mixture of alumiꢀ
num hydride etherate with lithium tetrahydridoaluminate.
Unlike adducts, the substances obtained react unreadily
under normal conditions with atmospheric oxygen, acids,
and water, but vigorously interact with alkalis, as alumiꢀ
num metal does.
Up to recent time, of the six modifications described
in Ref. 2 only the structure of one of them, namely, the
rhombohedral (hexagonal) αꢀphase, has been known.3
Perhaps, this is caused by both difficulties in the isolation
of phases other than αꢀAlH3 in the pure state and the
absence of practical interest from the scientific society.
Despite all attempts, all prepared samples contain2 some
amounts of other alane modifications as admixtures. The
structural data for γꢀ and δꢀAlH3 have recently been pubꢀ
lished.3 As has been assumed,6 all of them are inorganic
polymers, whose simplest structural units, viz., {AlH6}
octahedra, are crossꢀlinked in a way other than in the case
of the αꢀphase.
Since all aluminum hydride phases are isolated, as a
rule, from solutions, in the present work we studied
their formation during crystallization from a diethyl
ether—benzene mixed solvent.
Experimental
All procedures involved in the purification, preparation, and
storage of the starting substances, the preparation of aluminum
hydride solutions, crystallization, and treatment of the crystalliꢀ
zation products (washing, drying, packaging) were carried out
Published in Russian in Izvestiya Akademii Nauk. Seriya Khimicheskaya, No. 7, pp. 1259—1265, July, 2007.
1066ꢀ5285/07/5607ꢀ1305 © 2007 Springer Science+Business Media, Inc.

1306
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 7, July, 2007
Bulychev et al.
under dry argon in a dry box or in vacuo using Schlenk techꢀ
niques.
using these solutions for the synthesis of the nonꢀsolvated subꢀ
stance, 10—15 wt.% LiAlH4 were introduced into the solutions,
because in the absence of this admixture only the AlH3•0.31Et2O
solvate was obtained.
Benzene, toluene, and octane were purified using standard
procedures and were dried over LiAlH4 prior to use. After drying
and distillation from alkali, diethyl ether was treated with soꢀ
dium benzophenone ketyl, distilled under argon, refluxed over
LiAlH4, and distilled again. The final procedure of ether prepaꢀ
ration was the dissolution of aluminum chloride in the ether and
solvent redistillation.
Since the desolvation mechanism and the stability of nonꢀ
solvated hydride depend very strongly on the quality of the startꢀ
ing substances and, primarily, on admixtures of compounds of
transition metals (such as titanium, iron, and chromium) even
in concentrations <10–3 wt.%, all crystalline starting substances
and apparatus were thoroughly purified and cleaned, respecꢀ
tively.
The powder Xꢀray diffraction analysis of the samples obꢀ
tained, which were placed in sealed in vacuumꢀsealed glass capꢀ
illaries together with a standard (silicon powder), was carried
out on a Guinier G670 HUBER diffractometer. Diffraction
patterns were indexed using the WinXpow (Stoe) program packꢀ
age followed by parameter refinement by the leastꢀsquares
method. Thermal analysis was carried out on a Q1500D system
at a heating rate of 10 °C min–1. Electron microscopic examinaꢀ
tions were carried out using a JSM 100C instrument.
The thermobaric studies have been described earlier.9
Results and Discussion
Commercial LiAlH4 was recrystallized from an ether—toluꢀ
ene (1 : 1, concentration ~30 g L–1) mixture. For this purpose,
alanate was dissolved at 35—40 °C, the undissolved solid was
filtered off, and the solvent was removed from the transparent
solution by heating in an oil bath. The process ceased when the
temperature in the flask approached 95—98 °C. The suspension
was cooled down, and the precipitate was separated from the
mother liquor on a special frit funnel, washed with pentane, and
dried in vacuo with slight heating. Thus obtained LiAlH4 crysꢀ
tallizes as prisms up to 1.5 mm in length, raises no dust, has
lower fire risk, and is soluble in diethyl ether up to 98—99%.
Commercial AlCl3 was subjected to double sublimation from
the melt over aluminum metal shavings. The mixture was placed
in a Pyrex tube (≤5 vol.%), which was vacuumized using a fore
pump, sealed off, placed into a tubular furnace to 4/5 of the tube
length, and heated to 200—220 °C. The sublimate was conꢀ
densed on a cooler part of the tube. It was soluble in diethyl
ether.
The crystallization of aluminum hydride from a diꢀ
ethyl ether—benzene solution starts (depending on the
concentration of the solution, the amount of LiAlH4 taken
over the stoichiometric amount for the Schlesinger reacꢀ
tion, and the amount of ether in the mixed solvent) in the
temperature interval 74—76 °C to form semitransparent
spherulites 0.8—1.5 mm in diameter, which inflamed imꢀ
mediately on contact with water. According to the elꢀ
emental analysis data, the substance composition apꢀ
proaches to the formula AlH3•0.25Et2O (let us name it
"preꢀhydride"). Since at this temperature LiAlH4 begins to
crystallize simultaneously with the solvate, 2 to 6% alanate
(according to Xꢀray powder diffraction and elemental
analysis data) are observed in the etherate as an impurity.
The formation of a solvate (without indication of its comꢀ
position) mixed with LiAlH4 has been observed previꢀ
ously.2 At the same time, an etherate of analogous comꢀ
position, to which the formula Al(AlH4)3•Et2O was asꢀ
cribed, was described in more detail.10—12 However, its
physicochemical properties differ radically from those of
the solvate isolated by us.
A solution of alane with an AlH3 concentration of 3—5 g L–1
was prepared via the Schlesinger reaction7
3 LiAlH4 + AlCl3•xEt2O
4 AlH3•x Et2O+ 3 LiCl
by mixing a solution of AlCl3 (50 mL) in a diethyl ether—benzene
(1 : 1.5) mixed solvent (in some experiments, nꢀoctane was used
as the second component) containing 1.5—2.5 g of the subꢀ
stance, with a solution of LiAlH4 (550 mL) in the same solvent
containing lithium tetrahydridoaluminate (1.5—2.5 g, 10—15%
stoichiometric excess). The solution with the precipitate was
stirred for 20—25 min, and the precipitate was then let to settle
down and decanted. The flask with the transparent solution was
placed in an oil bath heated to 110 °C and was kept in the bath
with continuously distilling off the solvent until the onset of the
crystallization (74—76 °C). From the moment of crystallizaꢀ
tion, the suspension (40—50 mL) was taken at an interval of
15—20 s from the flask into a frit funnel. The solid phase was
immediately separated from the mother liquor, dried in a low
vacuum, and transferred into tubes. Thus prepared samples, five
to eight in number, were taken from one reactor in one entry.
In particular entries, aluminum hydride etherate was synꢀ
thesized using NaAlH4. In this case, the synthesis was carried
out in special reactors equipped with highꢀspeed (cavitationꢀ
type) stirrers, whose design has been described.8 However, when
The diffraction data obtained for "preꢀhydride" are
given in Table 1 in comparison with the characteristics of
the solvate of composition AlH3•0.31Et2O most freꢀ
quently mentioned in literature, which was figuratively
named13 "dead" (most likely, due to its "unwillingness" to
turn into the nonꢀsolvated hydride upon thermal treatꢀ
ment; however, it reacts vigorously with water causing
ignition as all compounds of this type do), and "alumiꢀ
num alumohydride monoetherate." 11 The last two subꢀ
stances were synthesized as the products of the Schlesinger
reaction carried out with the strictly stoichiometric ratio
of the reactants with subsequent concentrating by evapoꢀ
ration of the ethereal solution and drying of the residue
in vacuo.
As can be seen from the data in Table 1, these etherates
have no common features from the viewpoint of Xꢀray
diffraction. The latter two crystallize10,13 in cubic lattices
(no indexing is presented), while "preꢀhydride" is a crysꢀ

Crystallization of aluminium hydride
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 7, July, 2007
1307
Table 1. Xꢀray diffraction data for aluminum hydride etherates
13
10
AlH3•0.25Et2O ("preꢀhydride")
AlH3•0.31Et2O
Al(AlH4)3•Et2O
1/dexp •104
hkl
1/dcalc •104
d/Å
Irel (%)
d/Å
Irel (%)
2
2
d/Å
Irel (%)
12.01
4.62
4.42
4.23
4.02
3.86
3.80
3.56
100
70
15
30
20
20
40
30
69
469
512
559
619
671
691
788
001
200
112
020
003
120
202
212
013
301
221
104
030
131
320
005
105
68.5
469
516
558
619
665
682
821
756
12.0
6.10
5.05
—
—
100
60
40
—
—
40
—
—
—
—
—
40
—
—
—
—
—
11.777
6.804
6.320
4.601
—
—
—
—
—
3.263
—
2.882
—
100
5
10
60
—
—
—
—
—
50
—
40
—
—
15
7
3.85
—
—
—
—
—
3.17
3.10
2.88
2.80
2.65
2.61
2.42
2.35
5
10
40
30
5
5
10
15
995
1038
1203
1266
1429
1468
1713
1809
987
1034
1198
1266
1436
1476
1712
1814
3.03
—
—
—
—
—
—
2.561
2.307
1.941
6
talline product and its diffraction pattern is indexed in the
tetrahedral lattice with the parameters a = 9.90, b = 8.54,
and c = 12.05 ( 0.01) Å. However, since the IR spectra of
all solid etherates, including "preꢀhydride," in the region
of stretching and bending vibrations of the Al—H bond
are similar, it can be assumed that their structures are also
similar. It is most likely that the main elements of
their structure, as in the nonꢀsolvated hydrides, are the
{AlH6} octahedra, in which some hydride atoms are reꢀ
placed with an ether molecule and the differences are
caused by distortions in the regularity of octahedra crossꢀ
linking into the inorganic polymer.
Figure 1 illustrates the thermal decomposition of the
"preꢀhydride." It is seen that the etherate decomposes in
three endothermic stages. At the first stage, in the temꢀ
perature interval 75—95 °C the substance is desolvated in
part to the molecular composition AlH3•(0.05—0.08)Et2O
(elemental analysis and GLC data). The diffraction patꢀ
130
TG (H2)
DTA
160
125
135
Nevertheless, this is sufficient for the thermal behavꢀ
ior of the etherates and "preꢀhydride" to be substantially
different. The thermoanalytical curves of the decomposiꢀ
tion of the etherates synthesized by different methods but
studied under the same conditions are presented in
Figs 1—3.
148
154
70
90
110
130
150
T/°C
Fig. 2. Thermoanalytical traces of the decomposition of the
AlH3•0.31Et2O and Al(AlH4)3•Et2O etherates.
TG (H2)
155
DTA
172
TG (H2)
162
150
109
92
83
104
90
DTA
148
96
162
100
70
90
110
130
150
170
T/°C
70
90
110
130
150
170
T/°C
Fig. 1. Thermoanalytical traces of the decomposition of the
Fig. 3. Thermoanalytical traces of the decomposition of an
AlH3•0.25Et2O etherate ("preꢀhydride").
AlH3•0.31Et2O + LiAlH4 (15 wt.%) mixture.

1308
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 7, July, 2007
Bulychev et al.
tern of the residue shows γꢀAlH3 and some amount of the
amorphous phase (halo at small angles). In the temperaꢀ
ture interval 100—105 °C, the ether is released completely
and most of the γꢀAlH3 is transformed into αꢀAlH3, which
decomposes at 160—165 °C. According to the Xꢀray difꢀ
fraction data, the residue contains only the aluminum
metal phase.
thus obtained γꢀAlH3 differ appreciably from the properꢀ
ties of the product isolated by crystallization from soluꢀ
tion. Fineꢀparticle γꢀAlH3 resulting from the solidꢀphase
transformation is very unstable, easily oxidizable, and
solvatable by diethyl ether; reacts vigorously with dilute
acids; and gradually transforms on storage into the
αꢀmodification even at ambient temperature.
The decomposition of the solvates synthesized by earꢀ
lier described procedures11—14 takes place in a similar way
(see Fig. 2): exotherm at 130—135 °C coinciding with the
onset of hydrogen evolution, which is almost immediately
transformed into the endotherm at 150—155 °C, and comꢀ
plete hydrogen evolution after the end of the endotherm.
After decomposition the residue contains the aluminum
metal phase and a minor amount of alumina (corundum,
after heating of the sample to 550 °C).
The unexpectedly strong dependence of the properties
of the etherates on the ether content attracts attention. At
first glance, the difference of 0.06 М of ether between the
"dead" solvate AlH3•0.31Et2O and the "preꢀhydride"
AlH3•0.25Et2O is insignificant, but this change in the
composition is accompanied by drastic changes in the
properties of these substances. Therefore, it is not surprisꢀ
ing that numerous and very contradictory data concernꢀ
ing both the synthesis of alane and the synthesis and propꢀ
erties of the etherates themselves were published.
The character of decomposition of solvates of this
type remains virtually unchanged, and they are syntheꢀ
sized with 5—8% deficiency or 5—8% excess LiAlH4 comꢀ
pared to the stoichiometry of the Schlesinger reaction.
However, if at least 15% LiAlH4 (or an equivalent amount
of LiH, which can turn into alumohydride under
thermolysis conditions) are added to the dry etherate
AlH3•0.31Et2O and the mixture is ground in an agathe
mortar or a ball mill for 15—20 min, the decomposiꢀ
tion thermoanalytical curves and the composition of
the resulting intermediate and final products change. Unꢀ
der these conditions, during heating at the first endoꢀ
thermic stage at 90—100 °C (see Fig. 3) the etherate
is partially desolvated and a mixture of γꢀAlH3 and
AlH3•(0.11—0.14)Et2O amorphous to Xꢀray with the
component ratio ~1 : 2 is formed. This mixture also deꢀ
composes at 150—160 °C but exothermically. It is most
likely that the relatively large amount of oxygen in it (in
the ether moiety) is sufficient to initiate the oxidation of
the nonꢀsolvated product and to change the sign of the
thermoanalytical feature. This assumption is indirectly
confirmed by the appearance of the thermolysis residue,
which is not a friable powder as in the case of the "preꢀ
hydride" but a blackꢀbrown "spongy" fusion cake.
Thus, our studies confirm completely the earlier
conclusions2 about the desolvating role of LiAlH4 for
the synthesis of alane from its etherate and about the
minimum amount of this substance necessary for deꢀ
solvation.
The next stage of crystallization detected by macroꢀ
methods during polythermal heating of the solution is
associated with the disappearance of the spherulites, which
are recrystallized almost immediately with a 0.5—1.0 °С
increase in the solution temperature. The appearance of
γꢀAlH3 as needles with split ends is observed in the preꢀ
cipitate both visually and by the Xꢀray diffraction method
(Fig. 4, a). Almost simultaneously fine crystals of αꢀAlH3
appear on the surface of the needles (Fig. 4, b, c).
They grow rapidly to a size of 10—35 µm but still reꢀ
main wrapped with particles of the needleꢀlike γꢀphase
(Fig. 4, d). At the end of the process (in 10—15 min when
the solvent temperature increases to 76—77 °C), γꢀAlH3
disappears almost completely (Fig. 4, e), but particular
"needles" of the α´ꢀphase (in the center of the ellipse in
Fig. 4, e) are sometimes observed in the bulk of αꢀAlH3
crystals.
The introduction of lithium borohydride approxiꢀ
mately in the same amounts as LiAlH4 into a solution of
aluminum hydride etherate (the same approach was proꢀ
posed in Ref. 2) does not favor desolvation but makes all
stages of the process more pronounced and increases the
yield of the nonꢀsolvated substance, the degree of crystalꢀ
linity of all intermediate products, and the thermal stabilꢀ
ity of the resulting alane. When solutions, in which the
ratio of three major components LiAlH4, LiBH4, and
AlH3 (2 : 1 : 2—2 : 1 : 3) corresponds to the compositions
of the {2LiAlH4•AlH3•4Et2O} and {LiBH4•2AlH3•
•3Et2O} complexes formed in the LiAlH4—AlH3—Et2O
and LiBH4—AlH3—Et2O systems, are used, the αꢀAlH3
crystals obtained14 differ from the "standard" crystals synꢀ
thesized with 15% additives of LiAlH4 and LiBH4 by both
their shape (cubes of similar sizes with the smooth surface
(Fig. 4, f )) and very high thermal stability and purity. The
crystallization rate of αꢀAlH3 in this method is so high
that, most likely, all intermediate stages cannot be obꢀ
served by the sampling method.
At the moderate rate of transformations usually obꢀ
served for hydride crystallization from etherate solutions
with the AlH3 concentration 2—6 g L–1 and 15% excess
LiAlH4, the composition of the distilledꢀoff solvent passes
through two or three maxima depending on the etherate
concentration rather than changing monotonically as in
The ratio of the solvated (amorphous to Xꢀrays) and
nonꢀsolvated (crystalline) components in the mixture can
be changed toward the latter by prolonged isothermal
heating in vacuo and/or by an increase in the mole fracꢀ
tion of LiAlH4. However, in any case, the properties of

Crystallization of aluminium hydride
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 7, July, 2007
1309
a
d
10 µm
10 µm
b
e
40 µm
2 µm
c
f
10 µm
2 µm
Fig. 4. Phase composition changes upon the polythermal crystallization of aluminum hydride under standard conditions (a—e) and
the crystal shape in the αꢀphase upon crystallization from solutions with the ratio LiAlH4 : LiBH4 : AlH3 = 2 : 1 : 2 ( f ).
the pure ether—benzene system. These changes coincide
in time and temperature with the aboveꢀdescribed proꢀ
cesses of etherate desolvation. As can be seen from the
data in Fig. 5, the first, diffused maximum is observed
immediately before the solid phase formation and is not
related to any visible changes in the solution. Perhaps, its
appearance is caused by the preꢀcrystallization of etherate
molecules accompanied by the loss of a portion of ether,
but this phenomenon can be ascribed to other factors.
The beginning of the second peak coincides in time and
temperature with the vigorous boiling up of the solution,
opalescence, and the formation of flakes, from which
spherulites or "preꢀhydride" are formed very rapidly. This
peak is not always manifested (see Fig. 5): upon hydride
crystallization from dilute solutions (<2 g L–1) it disapꢀ
pears and, most likely, merges with the third maximum.
The appearance of the latter coincides in time with spheruꢀ
lite crystallization and with the settling of the denser
γꢀphase on the bottom of the reactor. It is seen that two
most intense maxima shift toward higher temperatures

1310
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 7, July, 2007
Bulychev et al.
[Et2O] (%)
preparation, this time can be enough for the partial or
even complete conversion of the γꢀ and α´ꢀphases into
αꢀAlH3. As can be seen from the data in Fig. 6, in the
reactor only the hexagonal modification is usually obꢀ
served 15—25 min after the crystallization onset in the
reactor. The use of a solution of complexes as the initial
one imparts specific features to the isothermal crystallizaꢀ
tion method. As in the above case, even the first portions
of the resulting solid phase contain only αꢀAlH3 crystals
with cubic habitus and appreciable amounts of crystalline
LiAlH4 and LiBH4.
70
60
50
40
30
20
10
4
3
2
1
70
72
74
76
78
Tcr/°C
It should be mentioned that, as the authors of Ref. 2,
we achieved the formation of absolutely pure singleꢀphase
samples except for αꢀAlH3 in none of the cases of alane
crystallization from solution. Moreover, out of the whole
described2 diversity we distinguished only four phases
mentioned in Ref. 6. The Xꢀray diffraction data and inꢀ
dexing results for these phases are given in Tables 2 and 3,
and the general view of the phases is shown in Fig. 7. In
the framework of the synthesis method used, cubic βꢀAlH3
appeared occasionally in the crystallization products, alꢀ
ways as a mixture with other phases, and reasons for its
formation and stabilization remained unclear. This modiꢀ
fication differs from the other modification isolated from
solution by the absence of any pronounced cutting and
a very small particle size (<1 µm), i.e., the complete abꢀ
sence of visual differences. This is the reason for the abꢀ
sence of its microphotograph. The other phases are easily
identifiable even with a simple optical microscope. The
shape of the α´ꢀphase crystallizing as a carnation bud is
especially characteristic, which is seen in Fig. 4, e.
Nevertheless, among the four aluminum hydride modiꢀ
fications obtained by us, the βꢀphase has one property
that distinguishes it from the γꢀ and α´ꢀphases: it does not
transform into αꢀhydride on simple heating, as it is obꢀ
served for other less dense modifications. However, at
pressures below 5 GPa and temperatures 100—150 °C, all
phases, including the βꢀphase, undergo the phase transiꢀ
tion to αꢀAlH3. In this case, no other more dense alumiꢀ
num hydride modifications are observed under pressures
Fig. 5. Change in the Et2O percentage in the gas phase during
crystallization in the absence (1) and presence of aluminum
hydride in the concentration 2.0 (2), 3.7 (3), and 6.0 g L–1 (4).
with a decrease in the concentration of the solution, which
is quite natural.
When the desolvation is carried out in an isothermic
regime, in which an etherate solution is fed into the reacꢀ
tor with the pure solvent heated to 76—77 °C, on the one
hand, alane crystallization is somewhat easier and, on the
other hand, becomes more complicated. At the initial
state of solid phase formation, no stage of spherulite forꢀ
mation is observed. However, this does not mean that the
spherulites are not formed at all. In this case, most likely,
the rates of their formation and recrystallization are so
high that they cannot be taken quickly from the reactor
and kept. At the same time, even the first portions of
crystals contain two (γ, α´) or three (γ, α´, α) phases, the
quantitative ratio between which depends on the crystalꢀ
lization duration (Fig. 6). It is most likely that the αꢀphase
appears not immediately but some time after the formaꢀ
tion of alane crystals in the form of the γꢀ and α´ꢀphases.
Since sampling from the reactor requires some time for
I/I0
1.0
α
Table 2. Unit cell parameters ( 0.01 Å) and the symmetry type
of the aluminum hydride phases obtained by crystallization from
solutions
0.5
Modification
а
b
c
Z
ρ*/g cm–3
II
Å
I
γ
Rhombohedral
(hexagonal) α
Orthorhombic γ
Orthorhombic α´
Cubic β
4.45
11.80
6
1.47 1.45
α´
5.34 7.33 5.80
6.62 6.48 5.62
9.04
6
6
16
1.31 1.30
1.24 1.22
1.22 —**
0
10
20
30
t/min
—
—
Fig. 6. Change in the proportions of the AlH3 αꢀ, α´ꢀ, and
γꢀphases during crystallization (according to the Xꢀray diffracꢀ
tion data).
* I is calculation, II is experiment.
** A mixture with αꢀAlH3 is formed.

Crystallization of aluminium hydride
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 7, July, 2007
1311
Table 3. Results of indexing the diffraction patterns of the α´ꢀ and βꢀphases of aluminum hydride
α´ꢀAlH3
βꢀAlH3
1/dexp •104
hkl
1/dcalc •104
d/Å
Irel (%)
1/dexp •104
hkl
1/dcalc •104
2
2
2
2
d/Å
Irel (%)
5.63
3.30
3.23
2.845
2.810
100
80
10
40
80
315
917
956
1235
1266
001
200
020
210
002
021
220
202
221
311
222
316
918
952
1230
1266
5.235
3.731
2.617
2.520
2.260
100
80
80
10
60
365
1341
1460
1575
1960
111
311
222
320
400
365
1341
1462
1580
1950
2.314
2.138
40
60
1868
2188
1870
2185
2.195
2.130
10
10
2076
2204
410
330
2072
2194
1.964
1.785
15
10
2594
3139
2608
3136
2.080
1.742
40
40
2311
3295
331
333
511
440
531
443
622
551
711
—
3216
3291
1.650
1.621
1.582
1.558
1.456
15
10
12
10
10
3673
3806
3996
4120
4717
400
040
401
123
240
3674
3806
3990
4116
4724
1.604
1.531
1.418
1.365
1.269
40
40
20
20
20
3887
4266
4973
5367
6210
3901
4266
4998
5363
6217
1.423
1.405
1.293
1.265
10
8
8
4938
5066
5981
6249
402
004
204
124
4940
5064
5982
6246
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
6
a
b
c
5 µm
5 µm
40 µm
Fig. 7. Typical shapes of crystals of αꢀ (a), α´ꢀ (b), and γꢀAlH3 (c).
up to 10 GPa and short heating to 800 °C. Thus, it can be
concluded that all the three phases obtained in our and
earlier2 works are metastable relative to the αꢀAlH3 phase,
which is also metastable (∆G = 11.11 kcal mol–1).15 Since
this conclusion agrees well with the known rule of
Oswald´s steps, we can claim that, in the framework of
the synthesis method considered, even when the solid
phase contains no less dense aluminum hydride modifiꢀ
cations (first of all γꢀphase), this indicates a high rate of
transition to the more stable αꢀphase rather than their
absence.
When octane was used as the second component of
the mixed solvent, mixtures of the γꢀ and α´ꢀphases were
obtained in the best case in a yield <20%. In the presence
of other lowꢀboiling alkanes, only the "dead" solvate
AlH3•0.31Et2O crystallizes from the solution. This fact
indicates a certain role of the πꢀsystem of the aromatic
hydrocarbon in the mechanism of etherate desolvation
and the crystal growth of the inorganic polymer.
This work was financially supported by the INTAS
(Grant 05ꢀ1000005ꢀ7665 "New Alane. Novel Reversible
Hydrogen Storage Materials Based on the Alloys of Al").
References
1. H. Nöth and E. Wiberg, Forschr. Chem. Forsch., 1966,
8/3, 321.
2. F. M. Brower, N. E. Matzek, P. F. Reigler, H. W. Rinn,
C. B. Roberts, D. L. Schmidt, J. A. Snover, and K. Terada,
J. Am. Chem. Soc., 1976, 98, 2450.
3. J. W. Turley and H. W. Rinn, Inorg. Chem., 1969, 8, 18.

1312
Russ.Chem.Bull., Int.Ed., Vol. 56, No. 7, July, 2007
Bulychev et al.
4. V. A. Yartys, R. V. Denys, J. P. Maehlen, C. Frommen,
M. Fichtne, B. M. Bulychev, and H. Emerich, Inorg. Chem.,
2007, 46, 1051.
5. H. W. Brinks, W. Langley, C. M. Jensen, J. Graetz, J. J.
Reilly, and B. C. Hauback, J. Alloys Compd., 2007, 433, 180.
6. B. M. Bulychev, Zh. Neorg. Khim., 2000, 45, 220 [Russ.
J. Inorg. Chem., 2000, 45, 170 (Engl. Transl.)].
7. A. E. Finholt, A. C. Bond, and H. E. Schlesinger, J. Am.
Chem. Soc., 1947, 69, 1195.
11. P. Pierre, E. Jaeques, and B. Bernard, Rev. Chim. Miner.,
1972, 9, 511.
12. P. Claudy, B. Bonnetot, J. Stienne, and G. Turck, J. Therm.
Analysis, 1975, 8, 255.
13. K. N. Semenenko, Kh. A. Taisumov, A. P. Savchenkova,
and V. N. Surov, Zh. Neorg. Khim., 1971, 16, 2076 [J. Inorg.
Chem. USSR, 1971, 16 (Engl. Transl.)].
14. B. M. Bulychev, A. G. Burlakova, and P. A. Storozhenko,
Zh. Neorg. Khim., 1998, 43, 829 [Russ. J. Inorg. Chem., 1998,
43, 752 (Engl. Transl.)].
8. B. M. Bulychev, A. V. Golubeva, and P. A. Storozhenko,
Zh. Neorg. Khim., 1998, 43, 1061 [Russ. J. Inorg. Chem.,
1998, 43, 1765 (Engl. Transl.)].
15. P. Claudy, B. Bonnetot, and J. M. Letoffe, Thermochim.
Acta, 1978, 27, 205.
9. S. K. Konovalov and B. V. Bulychev, Inorg. Chem., 1995,
34, 172.
10. J. Bousquet, J.ꢀJ. Choury, and P. Claudy, Bull. Soc. Chim.
Fr., 1967, 3848.
Received March 14, 2007;
in revised form May 10, 2007