Pressure-composition isotherms and thermodynamic properties of TiF 3-enhanced Na2LiAlH6

Article abstract of DOI:10.1016/j.jallcom.2005.01.012

The mixed alanate Na₂LiAlH₆ was prepared by ball-milling and subsequent heat-treatment under H₂ pressure. After the synthesis, 2 mol% TiF₃ was added by ball-milling. Pressure-composition isotherms were measured for the Ti-enhanced material in the temperature range of 170-250 °C. A van't Hoff plot was constructed using the equilibrium desorption plateau pressures. From this plot, a dissociation enthalpy of 56.4 ± 0.4 kJ/mol H₂ and a corresponding entropy of 137.9 ± 0.7 J/K mol H₂ was found for Na ₂LiAlH₆.

Full text of DOI:10.1016/j.jallcom.2005.01.012

Journal of Alloys and Compounds 397 (2005) 135–139
Pressure–composition isotherms and thermodynamic
properties of TiF -enhanced Na LiAlH
3
2
6
∗
A. Fossdal , H.W. Brinks, J.E. Fonneløp, B.C. Hauback
Department of Physics, Institute for Energy Technology, P.O. Box 40, NO-2027 Kjeller, Norway
Received 25 December 2004; accepted 5 January 2005
Available online 10 February 2005
Abstract
The mixed alanate Na
2
6 2
LiAlH was prepared by ball-milling and subsequent heat-treatment under H pressure. After the synthesis, 2 mol%
TiF
3
was added by ball-milling. Pressure–composition isotherms were measured for the Ti-enhanced material in the temperature range of
◦
1
5
70–250 C. A van’t Hoff plot was constructed using the equilibrium desorption plateau pressures. From this plot, a dissociation enthalpy of
6.4 ± 0.4 kJ/mol H
2
and a corresponding entropy of 137.9 ± 0.7 J/K mol H
2 2 6
was found for Na LiAlH .
©
2005 Elsevier B.V. All rights reserved.
Keywords: Hydrogen storage materials; Alanates; Ball-milling; Pressure–composition isotherms; Thermodynamic properties
1
. Introduction
accomplished at much lower hydrogen pressures (absorption
plateau pressure <12 bar at 170 C) [1]. On the other hand,
◦
The reversibility and hydrogen storage properties of
the relatively low plateau pressures of Na3AlH at low tem-
6
NaAlH4 with Ti-based additives has been thoroughly in-
vestigated, e.g. in Refs. [1–9], since the initial report by
Bogdanovic and Schwickardi in 1997 [10]. Theoretically,
NaAlH4 canreversiblystore5.6 wt.%H2 underpracticaltem-
perature and pressure conditions. The theoretical capacity has
proved difficult to obtain experimentally, however a hydro-
gen storage capacity of 3.5 wt.% has been maintained up to
peratures may represent a problem with regard to potential
applications. To use Na3AlH as a hydrogen storage mate-
6
◦
rial, temperatures above 170 C may be required to obtain a
sufficient plateau pressure.
The LiAlH4 (8.0 wt.%) and Li3AlH (5.6 wt.%) systems
6
have higher theoretical hydrogen storage capacities than
NaAlH4. While some reports exist regarding the desorp-
1
00 hydrogenation/dehydrogenation cycles for Ti-catalyzed
tion reaction of catalyzed LiAlH4 and Li3AlH [12–15], re-
versibility in these two systems has not yet been realized. The
6
◦
NaAlH4 (hydrogenation: 150–120 C, ∼115 bar for 12 h; de-
◦
hydrogenation: 160 C, against 1 bar for 3 h) [11].
mixed alanate Na2LiAlH (3.5 wt.% H2) is however known
6
Despite the high theoretical hydrogen storage capacity,
NaAlH4 with its two decomposition steps is at a disadvantage
to be reversible with a Ti-based additive, and has a lower
plateau pressure than Na3AlH [10] and Li3AlH [16] (both
6
6
relative to the simpler, one-plateau Na3AlH , even though
with addition of a Ti species).
6
the latter has a theoretical storage capacity of only 3.0 wt.%
H2. The dissociation of Ti-enhanced NaAlH4 to Na3AlH6
and Al occurs at very high pressures, so that rehydrogena-
tion pressures above 100 bar are needed [1] for this step at
Claudy et al. [17] synthesized Na2LiAlH by reaction of
6
LiAlH4 with 2 NaH either in toluene or by a solid-state
reaction at elevated temperatures and high H2 pressure. In
the absence of a solvent, pressures > 30 kbar and tempera-
◦
◦
an operating temperature of e.g. 170 C. Rehydrogenation in
tures > 300 C were needed in order for formation of small
˚
amounts of Na2LiAlH to occur. A cubic unit-cell of 7.405 A
6
the Na3AlH system (with Ti-based additive) can however be
6
was obtained for Na2LiAlH from PXD data. Synthesis of
6
∗
Na2LiAlH6 without a solvent is facilitated by ball-milling, as
demonstrated by Huot et al. [18] who obtained Na2LiAlH6
Corresponding author. Tel.: +47 63 80 62 73; fax: +47 63 81 09 20.
E-mail address: anita.fossdal@ife.no (A. Fossdal).
0
925-8388/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2005.01.012

1
36
A. Fossdal et al. / Journal of Alloys and Compounds 397 (2005) 135–139
by ball-milling a mixture of NaH, LiH and NaAlH4 for 40 h.
profile parameters were used and background modelling was
performed by linear interpolation between manually selected
points.
Brinksetal. [19]latersynthesizedNa2LiAlD byball-milling
6
LiAlD4 and 2 NaAlD4, followed by annealing under 30 bar
◦
D2 pressure at 180 C. Na2LiAlD was found by combined
PCT characteristics were measured by an in-house built,
fully automated Sieverts instrument that can be operated up
6
PXD and PND measurements to have an ordered perovskite-
¯
◦
type structure (space group Fm3m) with unit-cell dimension
to 100 bar and 600 C. The pressure in the reference volume
˚
a = 7.38484(5) A.
is monitored by a MKS120 pressure transducer (accuracy
0.08% of reading, supplier’s calibration) below 28 bar and by
a Presens sensor (accuracy ± 4 mbar up to 200 bar, calibration
against a dead-weight tester) based on silicon piezoresistivity
above 28 bar. The volume containing the MKS120 sensor is
closed above its usable pressure range, thereby reducing the
system volume from 50 to 33 ml. The sample is kept in an
autoclave and the sample temperature can be kept constant
Pressure versus composition isotherm (PCT) measure-
ments of alanates are scarce in literature and focus mainly
on the NaAlH4 system [1,8,10,20–23]. Only two of these re-
portsconcernPCTcharacteristicsofLi-substitutedNa3AlH6.
In the study by Bogdanovic and Schwickardi [10], a PCT
◦
curve was measured at 211 C for a Na2LiAlH material
6
with 2 mol% of a Ti-based additive. In that measurement, the
middle of the desorption plateau was found at approximately
◦
within ± 0.2 C. Apart from a volume of 4 ml heated to the
◦
1
3 bar. Zaluski et al. [23] investigated a more Li-rich mate-
sample temperature, the remaining volume is kept at 40.00 C
rial, Na1.7Li1.3AlH , and found a desorption plateau pressure
at 220 C of approximately 9 bar, confirming the trend that
Li-substitution increases the stability of the alanate relative
to the parent Na3AlH6.
in two heating cabinets with temperature stabilities of ± 0.02
6
◦
◦
and ± 0.05 C. The stability criterion during the PCT mea-
surements was set at a pressure change with time of 8 (at
◦
◦
170, 200, 230 and 250 C) or 15 mbar/h (at 210 C), corre-
sponding to a change in the hydrogen content per formula
The aim of this study is to synthesize Na2LiAlH by ball-
6
milling and obtain PCT characteristics and thermodynamic
information for this mixed alanate with addition of 2 mol%
TiF3.
unit of Na2LiAlH (H/f.u.) of 0.007 and 0.013 per hour, re-
spectively. The H2 compressibilities of Hemmes et al. [25]
were used.
6
2
. Experimental
3. Results and discussion
Na2LiAlH was synthesized by ball-milling (Pulverisette
PXD confirmed formation of Na LiAlH6 and Al in
6
2
7
) a 2:1 molar mixture of NaAlH4 (techn. > 90%, Sigma
the as-prepared sample, corresponding to the expected
Aldrich) and LiH (Sigma Aldrich) for 3 h at a speed of
00 rpm. The milling vial and milling balls were made of
stainless steel, and the weight ratio of balls to powder was
0:1. An initial hydrogen pressure of 80 bar was then ap-
reaction:
7
2
NaAlH4 + LiH → Na2LiAlH + Al + (3/2)H2
(R1)
6
2
In addition, 0.7 mol% of Na3AlH was observed. In the
cycled sample, the Na3AlH phase had completely disap-
peared. No evidence of Ti- or F-containing phases could be
found prior to cycling. The measured unit-cell parameter of
Na2LiAlH , a = 7.40654(15) A agrees well with the 7.405 A
6
plied to the milled powder in a constant-volume system and
6
◦
the sample was allowed to equilibrate overnight at 180 C.
The annealed powder was further ball milled with 2 mol%
of TiF3 (Alfa Aesar) for 20 min at 350 rpm. All opera-
tions were performed under argon with <1 ppm of O2 and
H2O.
˚
˚
6
measured by Claudy et al. [17] and is, as expected, larger than
that found for Na2LiAlD (a = 7.38484 A) [19]. The fit from
˚
6
Powder X-ray diffraction (PXD) data at 295 K was col-
lected at the Swiss-Norwegian beam line (station BM01B)
at the European Synchrotron Radiation Facility (ESRF) in
Grenoble, France. Measurements were performed after addi-
tion of TiF3 to the annealed powder (termed the as-prepared
sample) and after cycling in the PCT experiments (termed
the cycled sample). The latter had been desorbed stepwise
the Rietveld refinement of the as-prepared sample is shown
in Fig. 1.
Assuming that Ti3+ from the TiF3 additive is reduced to ze-
rovalent by reaction with Na2LiAlH during the ball-milling,
6
some reversible storage capacity is lost. This capacity loss can
occur in several ways, depending on whether formation of
LiF is preferred over formation of NaF, as would be thermo-
dynamically favorable in a simpler system, or not. Assum-
◦
◦
at 250 C and then reabsorbed in one step at 170 C under
8 bar of H2. The samples were contained in rotating 0.5 mm
boron–silica–glass capillaries. Data was collected between
5
ing no solid solubility between Na2LiAlH and Na3AlH ,
6
6
the reactions with minimum and maximum decomposition
◦ ◦ ◦ ◦
2
θ = 6.0 and 30.0 in steps of ꢀ(2θ) = 0.003 or 0.005 .
of Na2LiAlH can be represented, respectively, as
6
˚
The wavelength was 0.51979 A, obtained from a channel-cut
Si(1 1 1) monochromator.
Na2LiAlH + 0.02 TiF3
6
Rietveld refinements were carried out with the program
Fullprof (version 2.50) [24]. The X-ray form factor coef-
ficients were taken from the Fullprof library. Pseudo-Voigt
→
0.98 Na2LiAlH + 0.04 NaF + 0.02 LiF + 0.02 Al
6
+ 0.02 Ti + 0.06 H2
(R2)

A. Fossdal et al. / Journal of Alloys and Compounds 397 (2005) 135–139
137
Fig. 1. PXD pattern for Na2LiAlH6 at 295 K showing observed (circles),
calculated (upper line) and difference (bottom line) plots. The positions of
the Bragg reflections are shown for Al (upper), Na2LiAlH6 (middle) and
Na3AlH6 (lower).
Fig. 2. Pressure–composition isotherms for Na LiAlH with addition of
2
6
2
mol% TiF3. The term H/f.u. equals the number of hydrogen atoms per
formula unit of Na2LiAlH6. Filled and open symbols represent absorption
and desorption measurements, respectively.
and
Na2LiAlH + 0.02 TiF3
6
show only one plateau, in accordance with the reversible re-
action:
→
0.94 Na2LiAlH + 0.04 Na3AlH + 0.06 LiF
6
6
+
0.02 Al + 0.02 Ti + 0.06 H2
(R3)
Na2LiAlH ↔ 2 NaH + LiH + Al + (3/2)H2
(R5)
6
Crystalline NaF or LiF were not detected by PXD in the as-
prepared or cycled samples. Any Bragg reflections from LiF,
however, would overlap strongly with the reflections from
Al due to similar crystal structures and unit-cell parameters,
and presence of LiF can thus not be excluded. The presence
UpondrawingthePCTcurves, completionof(R5)isassumed
in the desorption measurements, i.e. H/f.u. = 3 at the end of
the desorption and start of the absorption measurements. Fur-
thermore, H/f.u. corresponding to the amount of Na2LiAlH6
remaining after reaction with TiF3 (R2) is given.
of Na3AlH in the as-prepared sample could suggest that
6
The PCT curves were measured in the following order:
(R3) occurs during the reaction with TiF3, however the dis-
◦
2
10, 200, 170, 230 C (full desorption and absorption cycle)
appearance of Na3AlH after cycling is counterindicative.
◦
6
and finally 250 C (desorption only). The desorption plateau
pressures, defined at the center of the plateau, were deter-
Furthermore, Na3AlD was also observed upon preparation
6
of Na2LiAlD without additives [19]. As the presence of NaF
6
mined to be (in order of increasing temperature) 2.95, 9.35,
in an X-ray amorphous state cannot be ruled out, (R2) is thus
considered to be more likely than (R3). The assumption of
◦
1
2.75, 22.15 and 36.95 bar. The plateau pressure at 210 C
is in excellent agreement with the ∼13 bar found by Bog-
constant composition Na2LiAlH is however an oversimpli-
◦
6
danovic and Schwickardi [10] at 211 C.
fication. Taking into account that the Na:Li ratio in the mixed
alanate has been observed to be variable [17,23], a varia-
tion on (R3) is also possible, wherein Li is leached from the
alanate to form LiF and thus shifting the alanate composition
according to
The plateau in the absorption isotherms is not as well
defined as in the desorption measurements. The onset of
the plateau was abrupt, and in some cases an overpressure
of more than 1 bar was needed in order to nucleate the
Na2LiAlH phase. On the high-pressure side, the end of the
6
plateau was more gradual, extending over a relatively wide
pressure–composition region. The difference between the ab-
sorption and desorption plateau was 2.8, 3.4, 7.8 and 4.0 bar
Na2LiAlH + 0.02 TiF3
6
→
0.98 Na2.04Li0 AlH + 0.06 LiF + 0.02Al
.96
6
◦
at 170, 200, 210 and 230 C, respectively.
+
0.02 Ti + 0.06 H2
(R4)
◦
At 200 C, the measured width of the plateau corresponds
Further studies are needed to unequivocally determine
the reaction path for reduction of the TiF3 additive. In
the following, all calculations are performed assuming
to 2.8 wt.% H2. This is slightly lower than the theoretical ca-
pacity (3.5 wt.%), even after correction for reaction with the
TiF3 additive during ball-milling. The reason why the hydro-
gen storage capacity is lower than the theoretical value is not
clear. The width of the plateau in the desorption isotherms
appears to increase with increasing temperature. The largest
additive-induced decomposition of Na2LiAlH according to
6
(
R2).
Pressure–composition isotherms of Na2LiAlH6 with
mol% TiF3 additive are shown in Fig. 2. The PCT diagrams
◦
2
increase in width occurs between 210 and 230 C. This ap-

1
38
A. Fossdal et al. / Journal of Alloys and Compounds 397 (2005) 135–139
parent increase in capacity with temperature may be a result
of repeated cycling rather than originating from the material
properties of Na2LiAlH6.
The PCT curves below the plateau appear, both in ab-
sorption and desorption mode, to be nearly vertical, whereas
in the high-pressure section of the isotherms, particularly
◦
above 200 C, the slope is lower. The change of slope in the
◦
isotherm above the plateau pressure from ≤200 to >200 C
is likely to originate from intrinsic material properties. The
plateau pressure of Na3AlH (and Li3AlH ) is higher than for
6
6
Na2LiAlH and neither Na nor Li hexa-alanates should there-
6
fore be formed in these PCT measurements. Even the final
◦
reabsorption at 170 C and 58 bar, which is much above the
plateau pressure of Na3AlH [1], does not result in formation
6
of Na3AlH . The initial Na3AlH impurity could therefore
6
6
◦
only affect the isotherm at 210 C, as this was measured first.
However, the impurity phase should be fully desorbed prior
to the start of the measurement series (the sample was al-
Fig. 4. Time required to cross the desorption plateau of Na2LiAlH6 as a
function of temperature.
◦
lowed to equilibrate at 27 bar and 210 C), since the plateau
It is also worth mentioning that measurements of
pressure–composition isotherms at near equilibrium condi-
tions are much more time-consuming than cycling with a
large chemical driving force. As an example, almost complete
pressure of Ti-enhanced Na3AlH is higher than 30 bar at this
temperature [10].
The equilibrium desorption pressures, taken from the mid-
dle of the plateau, are shown as a van’t Hoff plot in Fig. 3. The
thermodynamic properties for the desorption of Na2LiAlH6
6
rehydrogenation of Na2LiAlH could be achieved within
6
1
4
–2 h when a large overpressure was used, whereas typically
0 h would be required to reach equilibrium for each point at
(R5) are extracted using the van’t Hoff equation. A dissoci-
ation enthalpy of 60.7 ± 0.9 kJ/mol H2 and a corresponding
the plateau when measuring the PCT curve. The time needed
entropy of 146.5 ± 1.9 J/K mol H2 is found for Na2LiAlH
6
to cross the Na2LiAlH desorption plateau, defined as the to-
6
using data for all five measured temperatures. However, a
better fit to the data is found by excluding the data point mea-
tal time spent waiting for the equilibrium criterion to be ful-
filled, is given as a function of sample temperature in Fig. 4.
The figure clearly shows that reduction of the temperature
leads to a significant increase in the required measurement
time. Only a small amount of time is gained by increasing
◦
sured at 170 C, which lies significantly below a linear fit to
◦
the four points measured at 200–250 C. This gives a dissoci-
ation enthalpy of 56.4 ± 0.4 kJ/mol H2 and a corresponding
entropy of 137.9 ± 0.7 J/K mol H2. The dissociation enthalpy
is lower than the 62.8 kJ/mol measured by Claudy et al. [17]
by DSC, but is slightly more positive than the 47 kJ/mol H2
◦
the temperature from 230 to 250 C. The measurement at
2
◦
10 C was carried out with a less restrictive equilibrium
condition (15 mbar/h instead of 8 mbar/h), hence the point
for this temperature lies significantly below the trend from
the other measurements. These results show that even with
the strict equilibrium criterion of 8 mbar/h (0.007 H/f.u. per
hour), the low curvature of P(t) at the plateau leads to un-
derestimation of the plateau pressures. Note also that even
with addition of a Ti catalyst, total measurement times for
a complete desorption–absorption measurement can reach 3
weeks or more at low temperatures, whereas approximately
reported for Na3AlH [1], in accordance with the higher sta-
6
bility of Na2LiAlH relative to Na3AlH .
6
6
2
weeks are needed at higher temperatures.
Acknowledgements
Financial support from project “Hydrogen Storage in Hy-
drides for Safe Energy Systems—HYSTORY” (contract no.
ENK6-CT-2002-00600) under the European Commission’s
Fifth Framework Program and the NANOMAT program un-
der the Research Council of Norway, as well as the skilful as-
sistance from the project team at the Swiss–Norwegian beam
line, ESRF is gratefully acknowledged.
Fig. 3. van’t Hoff plot showing equilibrium desorption pressures as a func-
tion of temperature for 2 mol% TiF3-catalyzed Na2LiAlH6.

A. Fossdal et al. / Journal of Alloys and Compounds 397 (2005) 135–139
139
References
[13] D. Blanchard, H.W. Brinks, B.C. Hauback, P. Norby, Mater. Sci.
Eng. B 108 (2004) 54.
[
[
[
[
[
[
[
[
14] H.W. Brinks, B.C. Hauback, P. Norby, H. Fjellv a˚ g, J. Alloys Compd.
[
[
[
[
[
[
[
[
[
1] B. Bogdanovic, R.A. Brand, A. Marjanovic, M. Schwickardi, J.
T o¨ lle, J. Alloys Compd. 302 (2000) 36.
2] H.W. Brinks, C.M. Jensen, S.S. Srinivasan, B.C. Hauback, D. Blan-
chard, K. Murphy, J. Alloys Compd. 376 (2004) 215.
3] M. Fichtner, O. Fuhr, O. Kircher, J. Rothe, Nanotechnology 14
351 (2003) 222.
15] J. Chen, N. Kuriyama, Q. Xu, H.T. Takeshita, T. Sakai, J. Phys.
Chem. B 105 (2001) 11214.
16] H.W. Brinks, A. Fossdal, J.E. Fonneløp, B.C. Hauback, J. Alloys
Compd. 397 (2005) 291.
(2003) 778.
17] P. Claudy, B. Bonnetot, J.-P. Bastide, J.-M. Letoffe, Mater. Res. Bull.
4] C.M. Jensen, R. Zidan, N. Mariels, A. Hee, C. Hagen, Int. J. Hy-
drogen Energy 24 (1999) 461.
5] G.P. Meisner, G.G. Tibbets, F.E. Pinkerton, C.H. Olk, M.P. Balogh,
J. Alloys Compd. 337 (2002) 254.
1
7 (1982) 1499.
18] J. Huot, S. Boily, V. G u¨ ther, R. Schulz, J. Alloys Compd. 283 (1999)
04.
3
19] H.W. Brinks, B.C. Hauback, C.M. Jensen, R. Zidan, J. Alloys
Compd. 392 (2005) 27–30.
20] K.J. Gross, G.J. Thomas, C.M. Jensen, J. Alloys Compd. 330–332
6] G. Sandrock, K. Gross, G. Thomas, J. Alloys Compd. 339 (2002)
2
99.
7] G. Sandrock, K. Gross, G. Thomas, C. Jensen, D. Meeker, S. Takara,
J. Alloys Compd. 330–332 (2002) 696.
8] A. Zaluska, L. Zaluski, J.O. Str o¨ m-Olsen, J. Alloys Compd. 298
(2002) 683.
21] K.J. Gross, G.J. Thomas, G. Sandrock, in: Proceedings of the U.S.
DOE Hydrogen Program Review, Hydride Development for Hydro-
gen Storage, San Ramon, CA, 2000.
(
2000) 125.
9] R.A. Zidan, S. Takara, A.G. Hee, C.M. Jensen, J. Alloys Compd.
85 (1999) 119.
10] B. Bogdanovic, M. Schwickardi, J. Alloys Compd. 253/254 (1997)
[
22] G.J. Thomas, S.E. Guthrie, K. Gross, in: Proceedings of the U.S.
DOE Hydrogen Program Review, NREL/CP-570-26938, 452, Hy-
dride Development for Hydrogen Storage, Denver, CO, 1999.
23] L. Zaluski, A. Zaluska, J.O. Str o¨ m-Olsen, J. Alloys Compd. 290
2
[
[
[
1
.
[
11] S.S. Srinivasan, H.W. Brinks, B.C. Hauback, D. Sun, C.M. Jensen,
J. Alloys Compd. 377 (2004) 283.
12] V.P. Balema, V.K. Pecharsky, K.W. Dennis, J. Alloys Compd. 313
(1999) 71.
[
24] J. Rodr ´ı guez-Carvajal, Physica B 192 (1993) 55.
[25] H. Hemmes, A. Driessen, R. Griessen, J. Phys. C 19 (1986) 3571.
(2000) 69.