Hydrogen-generating solid-state hydride/hydroxide reactions

Article abstract of DOI:10.1016/j.jallcom.2004.08.042

Solid-state reactions of ionic hydrides with alkaline hydroxides are shown to produce hydrogen gas and metal oxides. These reactions are analogous to the well-known hydrolysis reactions of ionic hydrides. Both classes of reactions are generally exothermic

Full text of DOI:10.1016/j.jallcom.2004.08.042

Journal of Alloys and Compounds 390 (2005) 55–61
Hydrogen-generating solid-state hydride/hydroxide reactions
a,∗
a
b
b
John J. Vajo , Sky L. Skeith , Florian Mertens , Scott W. Jorgensen
a
HRL Laboratories, LLC, 3011 Malibu Canyon Road, Malibu, CA 90265, USA
b
General Motors Researchand Development Center, 30500 Mound Road, Warren, MI 48090, USA
Received 21 June 2004; received in revised form 16 August 2004; accepted 26 August 2004
Abstract
Solid-state reactions of ionic hydrides with alkaline hydroxides are shown to produce hydrogen gas and metal oxides. These reactions are
analogous to the well-known hydrolysis reactions of ionic hydrides. Both classes of reactions are generally exothermic and are thermody-
◦
namically favored; ꢀG < 0 near room temperature. However, solid-state hydride/hydroxide reactant mixtures are kinetically stable at room
temperature and can be prepared by mechanical milling without appreciable reaction. Thus, optimally stoichiometric mixtures are possible and
◦
nearly theoretical amounts of hydrogen can be generated. Reaction occurs upon heating with H
reaction occurring by 200–300 C. The reaction rate can be enhanced with additives such as TiCl
2
evolution beginning at ∼50 C and complete
◦
3
. Specifically, we discuss the reactions LiH
+
LiOH, 2LiH + NaOH, LiBH
4
+ 4LiOH, and 3LiBH
4
+ 4LiOH·H
2
O. The 3LiBH
4
+ 4LiOH·H
2
O reaction generates approximately 10 wt.%
◦
hydrogen with more than 5 wt.% produced at temperatures below 100 C.
©
2004 Elsevier B.V. All rights reserved.
Keywords: Hydrogen storage materials; High-energy ball milling; X-ray diffraction
1
. Introduction
The best studied of these reactions is saline metal hydride hy-
drolysis[2–9]. Forbinaryhydrides, theproductsofhydrolysis
are hydrogen gas and hydroxide salts and the reaction can be
considered as an acid–base reaction between the acidic pro-
tons in water and the basic hydride anions. Hydrolysis can
occur spontaneously upon contact of the hydride and water
or more slowly in a solution of the hydride. For example,
the binary hydride LiH reacts spontaneously with water [3,4]
while the complex hydride NaBH4 dissolves and hydrolyzes
slowly unless catalyzed [6–8]. In both cases, achieving sto-
ichiometric reaction can be difficult. Spontaneous reactions
typically produce alkali hydroxides which are hygroscopic
and bind excess water. The high pH and viscous physical
consistency of the product mixture can retard further reaction
unless there is excess water. For reactions such as NaBH4 +
2H2O, the NaBH4 must be dissolved again requiring excess
water.
Hydrogen-generatingreactionsarecurrentlyofinterestfor
a future hydrogen-based economy. Broadly, hydrogen gener-
ation can be endothermic or exothermic [1]. Endothermic re-
actions are thermodynamically directly reversible with a suf-
ficient hydrogen overpressure. These reactions are reversible
because the reverse H2 absorption reaction is exothermic and
the heat released compensates forthe entropy loss, hence, free
energy gain that accompanies incorporation of H2 into a con-
densed phase. Exothermic hydrogen-generating reactions are
much more difficult to reverse. Slightly exothermic reactions
can be reversed by application of extreme hydrogen pressures
which reduces the H2 gas entropy. However, most exother-
mic reactions, in which hydrogen absorption is endothermic,
are not directly reversible and require coupling to at least one
additional exothermic reaction to compensate for the entropy
loss. Although not directly reversible, exothermic reactions
have been investigated for hydrogen-generating applications.
In this work we extend the acid–base reaction in hydrol-
ysis to solid-state metal hydride/hydroxide reactions. These
reactions also generate hydrogen exothermically. However
the reactants can be mixed stoichiometrically without appre-
ciable reaction. Hydrogen production occurs upon heating
∗
Corresponding author. Tel.: +1 310 317 5745; fax: +1 310 317 5483.
E-mail address: vajo@hrl.com (J.J. Vajo).
0
925-8388/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2004.08.042

5
6
J.J. Vajo et al. / Journal of Alloys and Compounds 390 (2005) 55–61
and can be accelerated by additives such as TiCl3, which
may function as catalysts. We illustrate this class of reactions
using the examples of LiH + LiOH, LiH + NaOH, 2LiH +
NaOH, LiBH4 + 4LiOH, and 3LiBH4 + 4LiOH·H2O.
2
. Experimental details
Lithium hydride (95% purity from Aldrich and 97% pu-
rity from Fluka) and LiOH (98% purity), LiBH4 (95% pu-
rity), LiOH·H2O (99.995% purity), and TiCl3 (99.999% pu-
rity), all from Aldrich, were used without further purifica-
tion. Anhydrous NaOH was prepared by decomposing the
recrystallized alcoholate of reagent grade NaOH. All sample
handling was performed in an Ar filled glove box. Reactant
mixtures were prepared using mechanical milling. Approxi-
Fig. 1. Hydrogen generation in the LiH/LiOH system. Curves (a) and (b)
show hydrogen generation from mixtures of LiH + LiOH and LiH + LiOH
+ 0.1TiCl3, respectively, after milling for 1 h. Curve (c) shows the sample
◦
◦
temperatureprofile. Sampleswereheatedat2 C/minto100, 200, and250 C
◦
◦
with 2 h dwells at 100 and 200 C. Reaction begins at 50–60 C for both
systems. For the LiH + LiOH mixture, 0.39 wt.% hydrogen is produced at
mately 1.2 g mixtures were milled with a Fritsch Pulversette
3
6
planetary mill at 400 rpm in an 80 cm hardened steel vessel
◦ ◦
00 C and 5.73 wt.% at 250 C. For the LiH + LiOH + 0.1TiCl3 mixture,
.82 wt.% is produced at 100 C but only 3.94 wt.% at 250 C. Hydrogen
1
2
with thirty 7 mm diameter chrome–steel balls. Milling times
varied from 10 min to 3 h.
Temperature ramp hydrogen-generation measurements
were performed in a custom Sieverts apparatus, described in
detailelsewhere[10]. Briefly, theall-metalstainless-steelsys-
temwaspumpedusinganoil-freepumpingstationwithabase
◦
◦
generation during the milling likely accounts for the reduced amount of
hydrogen in the LiH + LiOH + 0.1TiCl3 mixture.
3. Results
−
8
−6
pressure at the pump inlet of <1 × 10 Torr (1.3 × 10 Pa).
Fig. 1 shows H2 evolution from mechanically milled mix-
tures of LiH + LiOH with and without TiCl3 added as a
−
6
The base pressure at the sample was <1 × 10 Torr (1.3 ×
−4
◦
1
0
Pa). Hydrogen pressures, gas temperatures, and sample
catalyst during heating in steps to 250 C. Hydrogen evolu-
®
◦
temperatures were recorded by computer, using LabView
software. Experiments were conducted using 0.8–1.0 g of the
milled mixtures loaded into an approximately 20 cm sample
tion begins at 50–60 C and without TiCl3 reaches 0.39 wt.%
of hydrogen at 100 C. Upon further heating additional H2
is produced. The maximum amount produced at 250 C is
◦
3
◦
volume. Samples were evacuated for at least 1 h prior to re-
action. Hydrogen generation occurred into an evacuated and
sealed volume of approximately 220 cm . Only the volume
5.73 wt.%. Assuming complete dehydrogenation of the LiH
+LiOHmixture, themaximumamountofH2 producedwould
be 6.27 wt.%. The hydrogen collected thus represents a yield
of 91%. Addition of TiCl3 at a loading of 10 mol% increases
3
containing the sample was heated. After reaction, the com-
position of desorbed gas was monitored using a residual gas
analyzer connected to the desorption volume through a leak
valve. In all experiments the evolved gas was predominately
H2. Small amounts, <1%, of water were also detected. Hy-
drogen gas pressures, typically below approximately 3 bar,
were converted into moles using the ideal gas equation by as-
◦
the quantity of H2 produced at 100 C, with respect to the
weight of LiH + LiOH only, to 2.82 wt.%, but decreases the
maximumamountproducedto3.94 wt.%. Thedecreaseinthe
final amount of hydrogen produced appears to result from hy-
drogen production during milling. As shown in Table 1, the
◦
amount of H2 generated at 100 and 250 C varies with TiCl3
3
suming that a portion of the gas, approximately 20 cm , was
loading and milling time. Fig. 2 shows X-ray diffraction pat-
terns for a mechanically milled mixture of LiH + LiOH before
at the sample temperature and that the remainder was at the
ambient gas temperature. The amount of hydrogen generated
was expressed as weight percent (wt.%) with respect to the
initial sample weight. For experiments with TiCl3 as an ad-
ditive, the weight of the TiCl3 was not included in the wt.%
calculations. The uncertainties in the amount of hydrogen
generated are ± 0.05 wt.%.
Table 1
Effect of milling time and TiCl3 addition on hydrogen generation in LiH +
LiOH
Milling time (h)
TiCl3 loading
H2 produced at
100 C (wt.%)
H2 produced at
250 C (wt.%)
◦
a
◦
a
(mol%)
0
1
1
0
1
3
1
.15
.5
0
0
2
5
5
5
0.39
0.56
0.56
0.89
2.17
0.83
2.82
5.73
5.34
5.38
4.84
4.58
1.31
3.94
X-ray data were obtained with a Philips PW3040/60
X’Pert Pro diffractometer using Cu K␣ radiation with a com-
bination of an X-ray mirror and a 2-crystal Ge(220) 4-bounce
monochrometer. Samples were mounted into 1 mm glass cap-
illary tubes in an Ar glove box and sealed prior to data acqui-
10
sition. A broad feature seen the X-ray diffraction patterns at
◦
a
1
5–40 two theta originates from the glass wall of the capil-
Weight percent H2 for reactions including TiCl3 calculated with respect
to the LiH + LiOH weight only.
lary tubes.

J.J. Vajo et al. / Journal of Alloys and Compounds 390 (2005) 55–61
57
Fig. 2. X-ray diffraction patterns of LiH + LiOH mixtures after milling for
Fig. 4. X-ray diffraction patterns from a mixture of 2LiH + NaOH after
milling for 1 h and after reaction at 250 C. Curve (a) shows that milling
◦
◦
1
h and after reaction at 250 C. Milling produces a physical mixture of
◦
LiH + LiOH, curve (a). Heating to 250 C yields predominately Li2O with
some residual LiOH, curve (b). Symbols show location of major peaks for
LiH, LiOH, and Li2O from ICDD cards 01-078-0839, 00-032-0564, and
produces NaH and small amounts of LiOH and Li O in addition to the
2
reactants LiH and NaOH. X-ray diffraction peaks for LiH are not identified
in curve (a) because LiH scatters X-rays relatively weakly, as shown in Fig. 1,
and because of overlap with NaOH. Curve (b) shows the reaction products.
The main products are Li2O and NaH, only a small residue of NaOH remains.
Symbols show the location of major peaks for LiOH, NaOH, Li2O and NaH
from ICDD cards 00-032-0564, 00-035-1009, 01-077-2144, and 01-076-
0
1-077-2144, respectively.
◦
and after heating to 250 C. Milling produces a physical mix-
ture of LiH and LiOH. No new crystalline phases appear to
0171, respectively.
◦
have formed. The crystalline products after heating to 250 C
are predominately Li2O and some residual LiOH.
◦
Li2O. Heating to 250 C produces Li2O and NaH. Heating
Although expected to be exothermic, see calculations be-
low, and therefore, not directly reversible, a rehydrogenation
experiment was performed. A dehydrogenated mixture of
◦
a 1:1 mixture of LiH + NaOH to 250 C yields Li2O, NaH,
and unreacted NaOH (data not shown). The presence of un-
reacted NaOH from a 1:1 mixture indicates that the optimal
stoichiometry for the LiH/NaOH system is 2:1. As discussed
below, Na2O does not form at the temperatures used in these
experiments, <250 C. A similar amount of H2, 1.90 wt.%
and products, Li2O, NaH, and NaOH, are produced from a
NaH + LiOH mixture (data not shown).
◦
LiH + LiOH was cycled from room temperature to 200 C and
back to room temperature over a period of ∼12 h in 100 bar
H2. No absorption of hydrogen was observed.
◦
Hydrogen evolution from mechanically milled mixtures
of LiH + NaOH and 2LiH + NaOH is shown in Fig. 3. Heat-
◦
ing to 250 C yields 1.92 and 3.20 wt.% H2 for LiH + NaOH
Fig. 5 shows H2 evolution from mechanically milled mix-
tures of LiBH4 + 4LiOH and 3LiBH4 + 4LiOH·H2O. In these
and 2LiH + NaOH mixtures, respectively. These yields are
only ∼1/2 and ∼2/3 of the total hydrogen contents for these
systems. X-ray patterns for a mixture of 2LiH + NaOH af-
−
reactions, the simple hydride anion, H , is replaced with the
−
complex anion BH4 . For LiBH4 + 4LiOH, H2 production
does not begin until ∼250 C and a maximum of 6.64 wt.%
◦
ter milling and after reaction at 250 C are shown in Fig. 4.
◦
After milling, in addition to the reactants LiH and NaOH,
the mixture contains NaH and small amounts of LiOH and
Fig. 5. Hydrogen generation in the LiBH4/LiOH and LiBH4/LiOH·H2O
systems. Curves (a) and (b) show hydrogen generation from mixtures of
LiBH4 + 4LiOH and 3LiBH4 + 4LiOH·H2O, respectively, after milling for
1 h. Curve (c) shows the sample temperature. For the LiBH4/LiOH system,
Fig. 3. Hydrogen generation in the LiH/NaOH system. Curves (a) and (b)
show hydrogen generation from mixtures of LiH + NaOH and 2LiH + NaOH,
respectively, after milling for 1 h. Curve (c) shows the temperature profile.
Both reactions begin to generate H2 at 50–60 C. The LiH + NaOH mixture
produces 1.92 wt.% H2 and the 2LiH + NaOH mixture produces 3.20 wt.%
H2.
◦
H2 generation begins at 250 C and a maximum of 6.64 wt.% H2 is pro-
◦
◦
duced at 350 C. For the LiBH4/LiOH·H2O system, H2 generation begins
◦
◦
at 40–50 C, >5.5 wt.% H2 is produced at temperatures <100 C, and the
maximum amount of H2 produced is 10.05 wt.%.

5
8
J.J. Vajo et al. / Journal of Alloys and Compounds 390 (2005) 55–61
Fig. 6. X-ray diffraction patterns of LiBH4 and LiBH4 + LiOH mixtures.
Curve (a) shows the X-ray pattern for the LiBH4 starting material. Most of
the peaks match the reference pattern from ICDD card 00-027-0287. The
Fig. 8. X-ray patterns for LiBH /LiOH and LiBH /LiOH·H O systems after
4
4
2
◦
reaction at 350 C. Curve (a) shows that the products of the LiBH + 4LiOH
4
reaction are Li O, Li BO , and Li B O . Curve (b) shows that the products
2
3
3
4
2
5
◦
◦
◦
peaks at 24.7 , 26.0 , and 35.6 , also appear to correspond to LiBH4. The
of the 3LiBH + 4LiOH·H O reaction are Li BO and Li B O . Symbols
4
2
3
3
4
2
5
◦
peak at 32 may result from a B2O3 impurity. Curve (b) shows the pattern
2 3 3 4 2 5
show the major peaks for Li O, Li BO , and Li B O from ICDD cards
for LiBH4, after milling for 1 h. The intensities of the peaks change, but
the patterns are similar. Curve (c) shows the pattern after milling a mixture
of LiBH4 + LiOH for 1 h. The overall pattern changes. Several broadened
peaks correspond to LiOH, ICDD card 00-032-0564. The remaining peaks
do not correspond to LiBH4.
01-007-2144, 00-018-0718, and 00-018-0720, respectively.
a milled mixture of 3LiBH4 + 4 LiOH·H2O. Several peaks
from anhydrous LiOH are evident which are similar to those
◦
observed from a milled mixture of LiBH + 4LiOH. Again,
the remaining peaks do not correspond to LiBH4 and more-
over, they are different from the unassigned peaks in Fig. 6c.
The appearance of new peaks at lower angles may indicate
4
H2 is generated at 350 C. The H2 yield is 98% of the total
hydrogen content of the reactants. In contrast, beginning with
hydrated LiOH greatly accelerates the reaction rate. Hydro-
gen production begins at 40–50 C and >5.5 wt.% is gener-
ated at temperatures <100 C. A maximum of 10.05 wt.% H2
was obtained after heating to 350 C, which is also 98% of
◦
◦
intercalation of LiBH into the layered structure of LiOH.
4
◦
Fig. 8 shows X-ray patterns for milled mixtures of LiBH +
4
◦
4
LiOH and 3LiBH4 + 4LiOH·H2O after heating to 350 C.
the calculated hydrogen content.
Both reactions produce lithium borates with compositions of
X-ray patterns of the LiBH4 starting material, the LiBH4
starting material after mechanical milling, and a milled mix-
ture of LiBH4 + 4LiOH are shown in Fig. 6. Milling of LiBH4
broadens all of the peaks and significantly reduces the in-
Li3BO3 and Li4B2O . Reaction of LiBH4 + 4LiOH also pro-
5
duces Li2O. No LiBO2 appears to form from either reaction.
◦
tensity of the prominent peak at 45 two theta. Otherwise,
4. Discussion
the patterns are similar. After milling LiBH4 with LiOH, the
diffraction pattern changes. As indicated in Fig. 6c, several
of the peaks pertain to LiOH, although these peaks are ex-
tremely broadened. The remaining peaks do not correspond
to LiBH4 and thus far they have not been satisfactorily as-
signed. As shown in Fig. 7, a similar change occurs with
The reactions studied in this work generate hydrogen
exothermically. For LiH + LiOH, the overall reaction and
◦
the calculated standard enthalpy of reaction (ꢀH (298 K))
[11] are given by
LiH + LiOH → Li2O + H2
◦
ꢀ
H (298 K) = −23.3 kJ/mol-H2
(1)
Thus, the milled mixtures of LiH + LiOH are thermo-
dynamically unstable at room temperature. However, kineti-
cally the reaction rate at room temperature is slow and practi-
◦
cally, the mixture must be heated to 50–60 C for observable
hydrogen production to occur. Once reaction begins, build-
up of the Li2O product appears to form a barrier between
remaining LiH and LiOH, thereby inhibiting further reac-
tion. This inhibition is seen in Fig. 1 where the production of
◦
hydrogen slows upon dwelling at 100 and 200 C, although
considerable LiH and LiOH remain.
Fig. 7. X-ray pattern for a 3LiBH4 + 4LiOH·H2O mixture, after milling for
Addition of 10 mol% TiCl3 during milling increases the
1
h. Symbols show peaks from anhydrous LiOH. The remaining peaks do
◦
amount of H2 produced at 100 C from 0.39 to 2.82 wt.%. The
not correspond to LiBH4 and also differ from a milled mixture of LiBH4 +
LiOH as shown in Fig. 6c.
4
mechanism for this enhanced reactivity is not known. After

J.J. Vajo et al. / Journal of Alloys and Compounds 390 (2005) 55–61
59
milling with TiCl3, LiCl was detected by X-ray diffraction
data not shown). However, all of the peaks in the diffraction
240 kJ/mol, ∼120 kJ of heat per mole of H2 oxidized must
be dissipated. Depending on the size of the system, removal
of that much low-grade heat can be challenging. Exothermic
reactions can significantly increase this heat load, thereby ex-
acerbating the heat removal problem. Reaction (1) generates
less than 1/3 of the heat of comparable hydrolysis reactions.
This includes the NaBH4 + 2H2O reaction which generates
75 kJ/mol-H2 and currently represents the most developed
exothermic hydrogen-generating reaction for fuel-cell appli-
cations [7].
In addition, any large-scale application of exothermic re-
actions for hydrogen generation will require conversion of
the oxide products back into hydrides [14]. The low effi-
ciency and therefore large energy requirements for this re-
processing is currently a major obstacle to large-scale use of
hydrolysis reactions for hydrogen production. At least in a
thermodynamic sense, the reduced amount of heat generated
in hydride/hydroxide reactions implies a lower required en-
ergy input for reprocessing. Moreover, for Reaction (1), only
half of the Li2O needs to be converted into LiH in an energy
intensive process. The remainder can be converted into LiOH
by an exothermic reaction with water.
(
pattern were broad and nearly indiscernible. No diffraction
peaks from TiCl3 were observed. The observed large reduc-
tion in crystallinity suggests that TiCl3 may facilitate a finer
dispersal and mixing of the LiH and LiOH, leading to an
enhanced reaction rate. Although, the extent of reaction at
◦
◦
1
00 C is enhanced, H2 evolution still begins at 50–60 C
indicating that the TiCl3 does not significantly lower the ac-
tivation energy. A reduction in the activation energy may be
expected if the TiCl3 functioned as a catalyst. The data in
Table1showthat>2 mol%TiCl3 isnecessarytoseeenhanced
◦
reactivity and that enhanced reaction at 100 C reduces the
overall amount of hydrogen collected during the tempera-
ture ramp experiments. This trend suggests that as a mixture
with enhanced reactivity is produced, hydrogen evolution oc-
curs during milling. After milling, we can qualitatively detect
build-up of pressure in the milling vessel but cannot quantify
the amount of any hydrogen produced. At a TiCl3 loading of
5
mol%, milling for 1 h appears optimal. Longer time reduces
the overall amount of H2 produced significantly including the
◦
amount produced at 100 C. Although, the hydrogen produc-
tion temperature did not change significantly, the formation
of LiCl and H2 during milling indicates that at least some of
the TiCl3 is being reduced to Ti metal. In this form, the Ti
may function catalytically by providing diffusion pathways
for evolving hydrogen and by facilitating desorption of hy-
drogen molecules. The catalytic effect of Ti on metal hydrides
and complex hydrides, such as MgH2 and NaAlH4, is well
known [12,13].
The quantities of hydrogen obtained and the crystalline
products identified by X-ray diffraction indicate that a mix-
ture of LiH + NaOH reacts according to
2
LiH + NaOH → Li2O + NaH + H2
◦
ꢀH (298 K) = −48.1 kJ/mol-H2 (4)
Reaction (4) generates 3.58 wt.% H2, which is 2/3 of
The heat released during Reaction (1) may be compared
with the heat released for hydrolysis of LiH. Two possible
hydrolysis reactions are
the total hydrogen content and consistent with the measured
amount of 3.2 wt.%. Complete dehydrogenation of a LiH +
NaOH mixture would generate 4.17 wt.% H2 and result in
the overall reaction
1
2
1
LiH + H2O(l) → Li2O + H2
2
1
1
◦
LiH + NaOH → Li2O + Na2O + H2
ꢀ
H (298 K) = −65.9 kJ/mol-H2 (2)
2
2
◦
ꢀ
H (298 K) = 8.1 kJ/mol-H2
(5)
and
LiH + H2O(l) → LiOH + H2
Reaction (5) is endothermic although the positive enthalpy
is small and ꢀG is negative at 25 C. However, reaction of a
:1 LiH + NaOH mixture, as shown in Fig. 3, generates only
.92 wt.% H2 which is consistent with Reaction (4). Reaction
5) is not observed because the reaction
◦
ꢀ
H (298 K) = −108.6 kJ/mol-H2 (3)
◦
◦
1
1
(
As discussed above, Reaction (2) is not observed because
water is usually present in excess. Both of these reactions
generate considerably more heat than Reaction (1). Heat gen-
eration becomes important if these reactions are used as a
hydrogen source, for example, for a H2/O2 polymer-based
proton exchange membrane (PEM) fuel cell. Exothermic re-
actions are useful hydrogen sources in fuel-cell applications
because no additional energy is needed to liberate the hydro-
gen. However, the heat released during H2 generation must
be continuously removed to keep the source isothermal. This
heat adds to the heat that must be removed from the fuel
cell itself. For example, the current generation of PEM fuel
cells are ∼50% energy efficient and operate at a temperature
NaH + NaOH → Na2O + H2
◦
ꢀ
H (298 K) = 64.3 kJ/mol-H2
(6)
◦
has a large endothermic enthalpy and ꢀG is negative only
◦
at temperatures above 400 C. Thus, the oxide, Na2O, is not
stable in the presence of hydrogen. Similarly, reaction of NaH
LiOH proceeds according to
+
NaH + 2LiOH → Li2O + NaOH + H2
◦
◦
of 80 C. Since the enthalpy for the H2 + 1/2O2 reaction is
ꢀH (298 K) = 1.6 kJ/mol-H2
(7)

6
0
J.J. Vajo et al. / Journal of Alloys and Compounds 390 (2005) 55–61
◦
◦
with ꢀG , again, negative at 25 C. Overall, Reactions (4)
Reactions of LiBH4 + 4LiOH and 3LiBH4 + 4LiOH·H2O
can generate several product mixtures with the required over-
all composition. For example,
and (7) represent the lowest energy pathways for reaction
temperatures below 250 C.
◦
Considering the kinetics of Reactions (1), (4), and (7) to-
gether permits some insight into a possible mechanism for all
three reactions. These reactions all begin to evolve hydrogen
LiBH4 + 4LiOH → LiBO2 + 2Li2O + 4H2
◦
ꢀH (298 K) = −21.6 kJ/mol-H2
◦
at ∼60 C, as shown in Figs. 1 and 3. This similarity suggests
(
13)
a common rate-determining step. Reaction of thin corrosion
layers of LiOH on LiH, essentially Reaction (1), has been
suggested to proceed by LiOH dehydration with water as a
transient intermediate [15–17]. For the LiH + LiOH system,
the individual steps are
LiBH4 + 4LiOH → Li3BO3 + Li2O + 4H2
(14)
(15)
LiBH4 + 4LiOH → Li4B2O + 4H2
5
1
1
LiOH → Li2O + H2O(g)
These product mixtures may be combined into a single
expression given by
2
2
◦
ꢀ
H (298 K) = 64.7 kJ/mol-LiOH
(8)
z
LiBH4 + 4LiOH → xLiBO2 + yLi3BO3 + Li4B2O5
and
2
ꢀ
ꢁ
1
2
1
2
3z
LiH + H2O(g) → Li2O + H2
+ 2x + y + 2 Li O + 4H
2
2
(16)
◦
ꢀ
H (298 K) = −87.9 kJ/mol-H2 (9)
where x + y + z = 1. From the X-ray diffraction pattern shown
in Fig. 8, it appears that x is small and that y ∼ z ∼ 0.5.
Similarly, reaction of LiBH4 with LiOH·H2O can follow
Thus, Reaction (8) would likely determine the overall re-
action rate because the activation energy must be at least as
large as the positive enthalpy. For the 2LiH + NaOH system,
the first step could be dehydration of NaOH given by
3LiBH4 + 4LiOH · H2O → 3LiBO2 + 2Li2O + 12H2
◦
ꢀ
H (298 K) = −43.5 kJ/mol-H2
(17)
1
2
1
2
NaOH → Na2O + H2O(g)
or
◦
ꢀ
H (298 K) = 96.1 kJ/mol-NaOH
(10)
3
LiBH4 + 4LiOH · H2O → Li3BO3 + Li4B2O + 12H2
5
(
18)
However, the enthalpy for this reaction is much larger than
for LiOH and therefore, a higher H2 evolution temperature
might be expected. An alternative is an initial exchange re-
action given by
Fig. 8 indicates that no Li2O or LiBO2 is formed and thus
reaction (18) describes the product mixture. For Reaction
13), the heat released per mole of H2 generated is approxi-
(
mately equal to the heat released for Reaction (1). Enthalpy
data for the higher borates are not available. Nevertheless,
the enthalpies for Reactions (14) and (15) should be similar
although slightly more negative. For reaction with hydrated
LiOH, Reaction (18), the heat released should also be slightly
more than 43.5 kJ/mol-H2, which is the enthalpy change for
Reaction (17). This heat release is larger, by approximately
a factor of 2, compared to the heat for anhydrous LiOH. The
increased heat is due to the water although the heat is reduced
by the binding energy of the water in the hydrated LiOH. Re-
action (18) is, therefore, a compromise between hydrolysis
and a hydride/hydroxide reaction.
LiH + NaOH → LiOH + NaH
◦
ꢀ
H (298 K) = −24.8 kJ/mol
(11)
Reaction (11) would be followed by Reaction (8), which
again would be rate determining, and finally by Reaction (9).
Reaction (11) is consistent with the X-ray pattern from a mix-
ture of 2LiH + NaOH milled for 1 h, Fig. 4a. Milling clearly
produces NaH although no H2 appears to be generated. For
Reaction (7), the initial step would be dehydration of two
moles of LiOH, Reaction (8), followed by
NaH + H2O(g) → NaOH + H2
◦
ꢀ
H (298 K) = −127.7 kJ/mol-H2
5. Summary
(12)
In this case formation of Na2O is not favorable at the
temperatures used in our experiments because of the thermo-
dynamics of Reaction (6). Thus, dehydration of LiOH, de-
scribed by Reaction (8), may be a common rate-determining
step in Reactions (1), (4), and (7).
Hydrogen generation has been studied in the solid-state
metal hydride/hydroxide reactions LiH + LiOH, 2LiH +
NaOH, LiBH4 + 4LiOH, and 3LiBH4 + 4LiOH·H2O. These
reactions are exothermic and the reactant mixtures are ther-
◦
modynamically unstable, ꢀG < 0, at room temperature.

J.J. Vajo et al. / Journal of Alloys and Compounds 390 (2005) 55–61
61
However, the reactants are kinetically stable and fine mix-
tures were prepared by mechanical milling. A mechanically
milled mixture of LiH + LiOH generated ∼5.5 wt.% H2 dur-
[3] V.C.Y. Kong, F.R. Foulkes, D.W. Kirk, J.T. Hinatsu, Int. J. Hydrogen
Energy 24 (1999) 665.
[
4] A.W. McClaine, et al., Hydrogen Transmission/Storage with Metal
Hydride-Organic Slurry, Proceedings of the 2000 US DOE Hydrogen
Program review, NREL/CP-570-28890, 2000.
◦
ing heating to 250 C and produced Li2O. The reaction rate
was enhanced by addition of 2–10 mol% TiCl3. The added
TiCl3 may have functioned as a catalyst facilitating diffusion
and desorption of hydrogen or as a dispersant, leading to finer
mixing of the LiH and LiOH phases. A mixture of 2LiH +
NaOH generated 3.2 wt.% H2 and produced Li2O and NaH.
Sodium oxide, Na2O, was not produced in this reaction be-
[
5] Y. Kojima, K.I. Suzuki, Y. Kawai, J. Mater. Sci. 39 (2004)
2227.
[6] Y. Kojima, et al., Int. J. Hydrogen Energy 27 (2002) 1029.
[7] S.C. Amendola, et al., J. Power Sources 85 (2000) 186.
[8] S.C. Amendola, et al., Int. J. Hydrogen Energy 25 (2000) 969.
[9] R. Aiello, J.H. Sharp, M.A. Matthews, Int. J. Hydrogen Energy 24
(1999) 1123.
◦
cause it isunstable inhydrogenat temperatures below 400 C.
Reaction of LiBH4 + 4LiOH generated 6.6 wt.% H2 and
[10] J.J. Vajo, F. Mertens, C.C. Ahn, R.C. Bowman Jr., B. Fultz, J. Phys.
Chem. B 108 (2004) 13977.
11] Estimates of thermodynamic parameters were made using Out-
[
produced Li2O, Li3BO3, and Li4B2O . Reaction of 3LiBH4
5
okumpu HSC Chemistry for Windows, version 4.0 ChemSW Inc.,
◦
+
4LiOH·H2O begins at 40–50 C and generates 10 wt.% H2
1999.
◦
below 270 C. The products are Li3BO3 and Li4B2O . Al-
5
[12] G. Liang, J. Huot, S. Boily, A. Van Neste, R. Schultz, J. Alloys
Compd. 292 (1999) 247.
13] B. Bogdanov ´ı c, G. Sandrock, MRS Bull. 27 (2002) 712.
14] Y. Kojima, T. Haga, Int. J. Hydrogen Energy 28 (2003) 989.
15] L.N. Dinh, M.A. Schildbach, W.J. Siekhaus, W. McLean II, Ad-
vanced Materials for Energy Conversion II, TMS, Pennsylvania,
though exothermic, these reactions generate less heat than
comparable hydrolysis reactions, which facilitates their use
as hydrogen sources for fuel-cell applications.
[
[
[
2
004.
References
[
16] S.M. Myers, J. Appl. Phys. 45 (1974) 4320.
[17] C.E. Holcombe, G.L. Powell, J. Nucl. Mater. 47 (1973) 121.
[
[
1] G. Wojciech, P.P. Edwards, Chem. Rev. 104 (2004) 1283.
2] E. Wiberg, E. Amberger, Hydrides of the Elements of Main Groups
I–IV, Elsevier, New York, 1971.