The influence of HS-AlF3 on the decomposition reaction of MgH2

Article abstract of DOI:10.1016/j.jallcom.2008.12.154

The influence of nanoscopic aluminium fluoride prepared via a sol-gel-synthesis route developed by our group on the thermal decomposition of MgH₂ used as hydrogen storage material has been comprehensively investigated. Combined XRD and MAS NMR experiments have been performed to follow structural changes in ball milled MgH₂/HS-AlF₃ mixtures at thermal decomposition. Although the behaviour of Lewis acidic metal fluorides is discussed being catalytic, our results with HS-AlF₃ clearly give a different picture. We found strong evidence for a decrease of the decomposition temperature in combination with a significant acceleration of the decomposition rate; however, the action of at least AlF₃ is not catalytic in nature. As shown, a fluoride against hydride exchange is taking place, thus transforming AlF₃ irreversibly into AlH₃, which is followed by decomposition into metallic Al (MgAl alloy).

Full text of DOI:10.1016/j.jallcom.2008.12.154

Journal of Alloys and Compounds 479 (2009) 82–86
Contents lists available at ScienceDirect
Journal of Alloys and Compounds
journal homepage: www.elsevier.com/locate/jallcom
The influence of HS-AlF3 on the decomposition reaction of MgH2
Johannes Noack, Gudrun Scholz, Stephan Rüdiger, Michael Feist, Erhard Kemnitz∗
Institute of Chemistry, Humboldt University Berlin, Brook-Taylor-Street 2, 12489, Berlin, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
The influence of nanoscopic aluminium fluoride prepared via a sol–gel-synthesis route developed by our
group on the thermal decomposition of MgH2 used as hydrogen storage material has been comprehen-
sively investigated. Combined XRD and MAS NMR experiments have been performed to follow structural
changes in ball milled MgH2/HS-AlF3 mixtures at thermal decomposition. Although the behaviour of
Lewis acidic metal fluorides is discussed being catalytic, our results with HS-AlF3 clearly give a different
picture. We found strong evidence for a decrease of the decomposition temperature in combination with
a significant acceleration of the decomposition rate; however, the action of at least AlF3 is not catalytic in
nature. As shown, a fluoride against hydride exchange is taking place, thus transforming AlF3 irreversibly
into AlH3, which is followed by decomposition into metallic Al (MgAl alloy).
Received 13 October 2008
Accepted 23 December 2008
Available online 7 February 2009
Keywords:
Hydrogen absorbing materials
Metal hydrides
Sol–gel processes
Gas–solid reactions
MAS NMR
© 2009 Published by Elsevier B.V.
1. Introduction
a fluorination of metal surfaces, accessible by reaction with fluoride
containing aqueous solutions, might act as a protective coating for
Magnesium hydride is considered to be one of the most promis-
ing hydrides studied as hydrogen storage materials, due to its high
gravimetric hydrogen capacity (7.6 wt.%), easy availability and low
cost. However, a commercial use is limited so far, because the high
thermodynamic stability (formation enthalpy ꢀHf = −75.3 kJ/mol)
requires decomposition temperatures above 300 ◦C. In addition the
decomposition is hampered by its slow hydrogen evolution kinetics.
To overcome these limitations a lot of work has focused on devel-
oping catalytic additives to improve the kinetic characteristics of
the involved processes.
The introduction of structural defects and reduction of parti-
cle sizes to nanoscopic range, as obtained upon ball milling of
magnesium hydride, results in enhanced reaction rates of the dehy-
drogenation [1,2]. In recent years, a lot of work has been focused
on the addition of numerous transition metals [3], metal oxides [4]
such as Nb2O5, V2O5, Ta2O5, and metal fluorides such as NbF5, FeF3,
TiF3 and NiF2 [5–9] into the MgH2 matrix to shift the hydrogena-
tion/dehydrogenation reaction to lower temperatures and increase
the rate of the reaction. Until now, the mechanism of this reaction
and the exact influence of the catalytic additive is still under discus-
sion. Unfortunately, Mg0 and MgH2 show high reactivity towards
oxygen and water, so MgH2 particles are usually covered by a MgO-
surface layer. This has a negative influence on the reaction kinetics
by limitation of hydrogen diffusion. Liu et al. [6] demonstrated that
the particles. Fluorinated particle surfaces can act as a barrier to O2,
H2O, CO, or N2 while, on the other side, they may be permeable to
H2 through microcracks of the MgF2 surface [10]. Since ball milling
of magnesium hydride with metal oxides can lead to the formation
of MgO and partial reduction of the oxide, a fluorination of the par-
ticle surfaces should be achieved by milling with metal fluorides
instead. From this point of view, the use of metal fluorides should
have definite advantages.
The use of Lewis acids for the activation of the H–H bond could
result in higher reaction rates for hydrogen evolution at even lower
temperatures. Computational studies have shown that the donation
of H2 ␴-electron density into the low lying vacant valence orbitals of
Mg2+ and Al3+ as bare ions can result in significant interaction ener-
gies [11]. Unfortunately, it is impossible to establish such bare ions
in real solids. Energies for the interaction of hydrogen molecules
with AlF3 and MgF2 in the gas phase have been calculated to be 11.3
and 11.8 kJ/mol, respectively [12]. High surface (HS) metal fluorides,
e.g. HS-MgF2 and HS-AlF3, should be excellent catalysts to overcome
the aforementioned problems. We have developed a sol–gel synthe-
sis route to highly disordered, X-ray amorphous and high surface
metal fluorides [13].
In this work, HS-AlF3 prepared this way and exhibiting strong
Lewis acidity, has been investigated regarding its impact on the
decomposition of MgH2. In contrast to metal fluorides investigated
so far in this respect, local structures of AlF3 are easily accessible
by 19 F and 27Al MAS NMR. This gives the opportunity to follow
changes and to identify intermediates that occur from high-energy
ball milling and during the thermally induced decomposition reac-
tion. Finally, an approach to the role of HS-AlF3 in these processes
should be possible.
∗
Corresponding author. Tel.: +49 30 2093 7555; fax: +49 30 2093 7277.
E-mail address: erhard.kemnitz@chemie.hu.berlin.de (E. Kemnitz).
0925-8388/$ – see front matter © 2009 Published by Elsevier B.V.
doi:10.1016/j.jallcom.2008.12.154

J. Noack et al. / Journal of Alloys and Compounds 479 (2009) 82–86
83
2. Experimental
points out the formation of MgAl alloys upon decomposition of the
MgH2/HS-AlF3 sample.
All operations in synthesis and milling were performed under argon as inert
atmosphere using Schlenk techniques or a glove box. HS-AlF3 was obtained by
fluorination of Al(OiPr)3 dissolved in i-PrOH with HF/iPrOH in a sol–gel synthe-
sis described elsewhere [14,15]. Samples were prepared by mechanical milling of
MgH2 (Alfa Aesar; 98%) with different ratios of HS-AlF3 (0.5, 3, 17 mol%) using a
planetary mill (Fritsch Pulverisette P7). Each 45 ml syalon vial was filled with 1 g
solid. Milling was performed for 4 h with five syalon balls of 2.9 g each (mass ratio
of balls/sample = 14.5:1) at 600 rpm under inert gas atmosphere.
The observation of chemical reactions by XRD measurements is
limited by the crystallite size of the sample and is therefore not
applicable for amorphous or nanocrystalline particles as obtained
by ball milling. Solid state NMR is a powerful method which also
gives insight into the local structure of X-ray amorphous samples. In
XRD measurements were carried out on a Seiffert XRD 3003 TT diffractometer
(Freiberg, Germany) with CuK␣ radiation. The samples were prepared in a glove box
and were covered with a special X-ray amorphous polystyrene foil to avoid the reac-
tion with air. Phases were identified by comparison with the ICSD powder diffraction
file [16]. 19 F and 27Al MAS NMR spectra have been recorded on a Bruker AVANCE
400 spectrometer at a rotation frequency of 25 kHz using a 2.5 mm MAS probe. The
resonance frequencies have been 376.4 MHz for 19 F and 104.6 MHz for 27Al; chem-
ical shifts are given relative to CFCl3 for 19 F and AlCl3 (1 M in water) for 27Al. 27Al
and 19 F spectra were simulated using dmfit2007 [17]. The thermal dehydrogena-
tion behaviour was studied by using a NETZSCH thermoanalyzer STA 409 C [DTA-TG
sample carrier system; corundum crucibles (baker 0.8 ml); Pt/PtRh10 thermocou-
ples; sample mass 15–30 mg; empty reference crucible; constant purge gas flow of
70 ml/min hydrogen (MESSER-GRIESHEIM 4.8); heating rate 10 K/min for dynamic
treatment, if necessary followed by various isothermal segments set at appropri-
ate temperatures between 250 and 420 ◦C]. The temperature evaluation of the DTA
curves regarding initial (Ti), extrapolated onset (Toenx), and peak (TP) temperatures
was performed following international recommendations [25].
3. Results
Structural analysis was performed before and after decom-
position at 320 ◦C by means of X-ray diffractometry and MAS
NMR spectroscopy. Fig. 1 shows the XRD patterns of MgH2 with
17 mol% HS-AlF3 of the as-milled sample prior (a) and after (b)
decomposition at 320 ◦C. The Bragg-reflections in the X-ray diffrac-
togram of the as-milled sample reveal the presence of tetragonal
␤-MgH2 (PDF 12-0697) as the main component. The existence of
the metastable orthorhombic ␥-MgH2-phase, which has been pre-
viously reported [1] to be formed on milling, could not be observed
under these conditions. HS-AlF3 is a highly disordered, nanoscopic
solid and is therefore not detectable by X-ray diffraction. In addi-
tion, increasing amounts of metallic magnesium (PDF 35-0821) can
be identified as a product of milling-induced decomposition of the
hydride.
After thermal treatment of the sample (pattern b), magnesium
hydride reflections disappear completely and metallic magnesium
as a product of decomposition can be identified. A fluorinating
action of AlF3 during thermal treatment can be deduced from
the reflections of MgF2 in the diffractogram. The reflection at 36◦
Fig. 2. 27Al MAS NMR spectra of (a) a ball milled mixture of MgH2 with 17 mol% HS-
AlF3, (number of accumulations (na): 66688); (b) enlargement of the central signals
of (a) together with their deconvolution: ı27Al1 = −13.3 ppm (78%), ı27Al2 = 12.5 ppm
(12%), ı27Al3 = 67.0 ppm (10%); (c) the same sample thermally treated at 320 ◦C, (na:
72000).
Fig. 1. XRD patterns of a ball milled mixture of MgH2 with 17 mol% HS-AlF3 (a), and
after subsequent thermal decomposition at 320 ◦C (b).

84
J. Noack et al. / Journal of Alloys and Compounds 479 (2009) 82–86
the case of the ball milled samples of MgH2 with different amounts
of HS-AlF3, 19 F and 27Al MAS NMR measurements were applied for
characterisation of the solids. For 1H see supporting information.
The 27Al MAS NMR spectrum (Fig. 2) of an MgH2 sample which
has been milled for 4 h with 17 mol% HS-AlF3 allows the assignment
of at least three different species. Spectrum (a) shows the region
from −3500 to 3500 ppm. At 1638 ppm a signal is observed which
can be assigned to metallic aluminium Al0. The chemical shift val-
ues of Al(III) compounds are usually in the region around 0 ppm.
This region is given as enlargement in Fig. 2b. The major signal is
the typical 27Al signal of HS-AlF3 with ı27Al at −18.9 ppm [15]. In
agreement with Hwang et al. [18] the symmetrical line at 12.5 ppm
can be attributed to AlH3. At 67.0 ppm an additional broad signal
can be interpolated which lies in the region of four-fold coordinated
aluminium sites. The integral proportions of the different species
are given in the caption to Fig. 2.
On decomposition of this sample at 320 ◦C, the HS-AlF3 signal in
the 27Al MAS NMR spectrum (Fig. 2c) disappears almost completely.
The signal belonging to Al0 remains with decreased intensity after
thermal treatment. A new line dominating the spectrum can be
observed at 1265 ppm. Samples with lower HS-AlF3 content show
an emerging line at 1916 ppm after dehydrogenation, which is illus-
trated for the case of 3 mol% HS-AlF3 in Fig. 3b.
Similarly, the 19 F MAS NMR spectrum (Fig. 4a) obtained after
milling of MgH2 with 17 mol% HS-AlF3 for 4 h shows the 19 F sig-
nal of HS-AlF3 as the main component at −168.3 ppm. On the
right hand side of the signal, further species can be interpo-
Fig. 4. 19 F MAS NMR spectra (central lines and their deconvolutions) of (a)
a ball milled mixture of MgH2 with 17 mol% HS-AlF3: ı19F1 = −196.4 ppm (3%),
ı19F2 = −191.1 ppm (4%), ı19F3 = −186.4 ppm (3%), ı19F4 = −168.1 ppm (90%); and
(b) the same sample thermally treated at 320 ◦C: ı19F1 = −197.7 ppm (37%),
ı19F2 = −191.8 ppm (18%), ı19F3 = −185.8 ppm (4%), ı19F4 = −182.2 ppm (41%); *: spin-
ning side band.
lated belonging to additional, so far unidentified species and MgF2
(ı19F = −198.0 ppm). The existence of hydride fluorides such as
AlFxH3−x or MgFxH2−x cannot be ruled out.
After dehydrogenation of this sample at 320 ◦C (Fig. 4b), the 19
F
signal belonging to HS-AlF3 signal vanishes almost completely. The
signals at −186, −192 and −198 ppm, respectively, remain under
these conditions (cf. Fig. 4b).
The thermal decomposition of the as-milled MgH2 can be fol-
lowedvia theDTAcurvesupon heatingat 10 K/minin pure hydrogen
atmosphere (Fig. 5). The untreated MgH2 sample starts to decom-
pose at about 420 ◦C which causes a sharp endothermal DTA signal
with an extrapolated onset temperature (Toenx) of 436 ◦C and a peak
temperature (TP) of 480 ◦C under these experimental conditions.
Ball milling of this MgH2 sample results in particle size reduc-
tion as can be seen from the broadening of the X-ray reflections
ex
(not shown). As a consequence, T for the decomposition reaction
on
decreased to 350 ◦C (TP 384 ◦C). A further shift to lower tempera-
tures is observed for the dissociation of samples containing 3 and
17 mol% of HS-AlF3. Addition of 0.3 mol% HS-AlF3 has a very little
effect on the decomposition reaction, but 3 mol% and more has a
direct impact on the thermal behaviour of MgH2. It is noteworthy
mentioning that both samples (3 and 17 mol% HS-AlF3) show almost
Fig. 3. 27Al MAS NMR spectra of (a) a ball milled mixture of MgH2 with 3 mol% HS-
AlF3 (na: 68277); (b) the same sample after thermal decomposition at 320 ◦C (na:
72000).
ex
the same DTA signal upon heating (T 290, TP 360 ◦C). This means
on
that further increasing the amount of added HS-AlF3 (more than

J. Noack et al. / Journal of Alloys and Compounds 479 (2009) 82–86
85
at 1638 ppm in the 27Al MAS NMR experiment (Fig. 2a).
milling
AIH3 −→ Al0 + 3/2H2
(3a)
(3b)
milling
Mg(AlH4)2 −→ MgH2 + 2Al0 + 3H2
Both XRD and 27Al MAS NMR prove that Mg and Al form an
alloy upon thermal decomposition. Depending on the content of
HS-AlF3 in the sample, the stoichiometry of the MgxAly alloy varies.
The alloy obtained after decomposition of MgH2 with 17 mol% HS-
AlF3 as presented in Fig. 2c induces a 27Al NMR chemical shift of
1265 ppm. According to Bastow and Smith [21] this NMR shift could
be assigned to an Al-rich alloy with the stoichiometry Mg17 Al12
.
With lower Al contents, the formation of an Mg-rich alloy with
6 wt.% Al is observed, which is illustrated in Fig. 3b for the case
of 3 mol% HS-AlF3 content in the MgH2 sample.
Moreover, the fluoride transfer according to Eq. (1) is enhanced
during the decomposition of the MgH2/HS-AlF3 sample and the
content of HS-AlF3 is dramatically reduced.
Fig. 5. DTA curves of MgH2 samples milled with different amounts of HS-AlF3 in
The DTA results shown in Fig. 5 give clear evidence for the
effect of the addition of strongly Lewis acidic nanoscopic AlF3 as
the decomposition temperature is significantly decreased by about
25 K. Obviously amounts of 3 mol% AlF3 are sufficient enough to
induce this effect.
pure hydrogen, normalised on hydrogen content.
3 mol%) does not lead to a stronger decrease of the decomposition
temperature.
We have also performed kinetic measurements for the
decomposition (see supporting information) of pure MgH2 and
MgH2/HS-AlF3 mixtures and have calculated the activation ener-
gies (Ea) from the rate constants at different temperatures applying
a nucleation and growth model [22]. These values confirm the effect
of milling and HS-AlF3 addition, respectively, as already displayed
above: the Ea values are 184 kJ/mol for pure MgH2, 154 kJ/mol for
milled MgH2 and 124 kJ/mol for MgH2/AlF3. Data for the activa-
tion of the MgH2 decomposition given in the literature deviate
from 91 to 156 kJ/mol [1,23]. Hence, the obtained values clearly
have a strong dependence on the detailed procedure and the
kinetic model applied. However, since these kinetic investigations
were performed under the same conditions, the data clearly give
evidence that addition of strongly Lewis acidic AlF3 reduces the
activation energy and, consequently, significantly improves the
decomposition rate. This effect might have two different origins.
At first, due to extraordinary high Lewis acidity, hydrogen atoms
might be transferred to HS-AlF3 sites on the particle surfaces, which
then combine to hydrogen gas. Second, metallic magnesium, that
is formed on dissociation of the hydride, is thermodynamically
stabilised by alloying. As a consequence of this, the correspond-
ing MgH2 is destabilised and the decomposition occurs at a lower
temperature [24].
4. Discussion
Due to the small crystallite sizes, the identification of chemi-
cal reactions or phase transformations induced by high-energy ball
milling only by XRD analysis of the as-milled samples is not pos-
sible. Together with 19 F and 27Al MAS NMR experiments however,
clear evidence can be given for the existence of Mg0, MgH2, AlF3,
Al0, AlH3, MgF2 and further so far unidentified 19 F species already
in the milled samples.
From the 19 F and 27Al MAS NMR spectra it can be deduced, that
a partial fluoride–hydride exchange reaction is already taking place
during ball milling of MgH2 with HS-AlF3, clearly shown by the for-
mation of MgF2 and AlH3 (Figs. 2b and 4a). Taking the enthalpies of
formation (ꢀH◦ ) for MgH2, MgF2, AlF3 and AlH3 into account, which
f
are −75.3, −1124.2, −1510.4 and −46.0 kJ/mol, respectively [19], a
total reaction enthalpy, ꢀH◦, of −217.9 kJ/mol can be estimated for
the reaction given in Eq. (1).
milling
3MgH2 + 2HS-AlF3 −→ 3MgF2 + 2AlH3
(1)
Bearing in mind that surface energies of nanoscopic solids might
be significantly higher than those of crystalline compounds, this
reaction should be possible from the thermodynamic point of view.
Although this reaction is thermodynamically favoured it is far from
being quantitative. Even after milling, HS-AlF3 is the main fluoride
component in the matrix (see Figs. 2a,b, 3a, 4a).
Apart from AlH3, which is formed by milling of HS-AlF3 with
MgH2, the broad signal at 67.0 ppm in the 27Al MAS NMR spectrum
(Fig. 2b) can be possibly assigned to Mg(AlH4)2. It could either be
formed from the binary hydrides during the milling procedure (Eq.
(2a)) or in a direct metathesis reaction (Eq. (2b)) which is known
from milling of AlCl3 with MgH2 [20].
5. Conclusion
Combined XRD and MAS NMR experiments allowed to follow
structural changes in ball milled MgH2/HS-AlF3 mixtures at thermal
decomposition.
Although the addition of Lewis acidic metal fluorides in litera-
ture is discussed as being catalytic, our results with HS-AlF3 clearly
give a different picture. We found strong evidence for a decrease
of the decomposition temperature in combination with a signifi-
cant acceleration of the decomposition rate; however, the action
of – at least AlF3 – is not catalytic in nature. As shown, a fluo-
ride against hydride exchange is taking place, thus transforming
AlF3 irreversibly into AlH3, which is followed by decomposition
into metallic Al (MgAl alloy). Fortunately this reaction is far from
being stoichiometric, therefore several cycles might be performed
for hydrogen storage but it is running into a “dead end”. We suspect
that other metal fluorides will show very similar behaviour and so
their action should be investigated in more detail too in order to
milling
MgH2 + 2AlH3 −→ Mg(AlH4)2
(2a)
(2b)
milling
4MgH2 + 2HS-AlH3 −→ Mg(AlH4)2 + 3MgF2
As a consequence of the H/F substitution, the resulting alane or
alanate can be easilydecomposed tometallicaluminiumandhydro-
gen by mechanical impact (Eq. (3a)) or by thermal treatment (Eq.
(3b)). This reaction is indicated by the signal of metallic aluminium

86
J. Noack et al. / Journal of Alloys and Compounds 479 (2009) 82–86
gain a better understanding of the rule of Lewis acid addition in
general but especially in case of metal fluorides.
[9] N. Recham, V.V. Bhat, M. Kandavel, L. Aymard, J.-M. Tarascon, A. Rougier, J. Alloys
Compd. 464 (2008) 377–382.
[10] F.-J. Liu, S. Suda, J. Alloys Compd. 231 (1995) 742–750.
[11] R.C. Lochan, M. Head-Gordon, Phys. Chem. Chem. Phys. 8 (2006) 1357–1370.
[12] C.A. Nicolaides, E.D. Simandiras, Chem. Phys. Lett. 196 (1992) 213–219.
[13] St. Rüdiger, E. Kemnitz, Dalton Trans. 9 (2008) 1117–1127.
[14] E. Kemnitz, U. Groß, St. Rüdiger, S. Chandra Shekar, Angew. Chem. 115 (2003)
4383–4386;
Acknowledgements
The authors would like to thank Mrs. S. Bäßler for performing
the thermoanalytical measurements.
E. Kemnitz, U. Groß, St. Rüdiger, S. Chandra Shekar, Angew. Chem. Int. Ed. 42
(2003) 4251–4254.
[15] St. Rüdiger, G. Eltanany, U. Groß, E. Kemnitz, J. Sol–Gel. Sci. Technol. 41 (2007)
299–311.
Appendix A. Supplementary data
[16] JCPDS-ICDD – International Centre for Diffraction Data: PDF-2 Database (Sets
1-51 plus 70-89), PA 19073-3273 U.S.A., Release 2001, – PCPDFWIN Version 2.2.
[17] D. Massiot, F. Fayon, M. Capron, I. King, S. Le Calve, B. Alonso, J.O. Durand, B.
Bujoli, Z.H. Gan, G. Hoatson, Magn. Reson. Chem. 40 (2002) 70–76.
[18] S.-J. Hwang, R.C. Bowman, J. Graetz, J.J. Reilly, W. Langley, C.M. Jensen, J. Alloys
Compd. 446–447 (2007) 290–295;
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jallcom.2008.12.154.
References
S.-J. Hwang, R.C. Bowman, J. Graetz, J.J. Reilly, in: J.C.F. Wang, W. Tumas, A.
Rougier, M. J. Heben, E. Akiba, (Eds.), Hydrogen Storage Materials, Mater. Res.
Soc. Symp. Proc. Warrendale, PA, 2006, vol. 927E, 0927.
[1] J. Huot, G. Liang, S. Boily, A. Van Neste, R. Schulz, J. Alloys Compd. 293 (1999)
495–500.
[19] D.R. Lide (Ed.), Handbook of Chemistry and Physics, 76th ed., CRC Press, Boca
Raton, FL, USA, 1995.
[20] T.N. Dymova, N.N. Mal’tseva, V.N. Konoplev, A.I. Golovanova, D.P. Aleksandrov,
A.S. Sizareva, Russ. J. Coord. Chem. 29 (2003) 385–389.
[21] T.J. Bastow, M.E. Smith, J. Phys. Condens. Matter. 7 (1995) 4929–4937.
[22] J.F. Fernández, C.R. Sánchez, J. Alloys Compd. 340 (2002) 189–198.
[23] H. Kawano, Y. Zhu, A. Tanaka, S. Sugimoto, Thermochim. Acta 371 (2001)
155–161.
[24] J.J. Vajo, T.T. Salguero, A.F. Gross, S.L. Skeith, G.L. Olson, J. Alloys Compd. 446–447
(2007) 409–414.
[2] A. Zaluska, L. Zaluski, J.O. Ström-Olsen, J. Alloys Compd. 288 (1999) 217–225.
[3] G. Liang, J. Huot, S. Boily, A. Van Neste, R. Schulz, J. Alloys Compd. 292 (1999)
247–252.
[4] G. Barkhordarian, T. Klassen, R. Bormann, J. Phys. Chem.
11020–11024.
B 110 (2006)
[5] S.-A. Jin, J.-H. Shima, Y.W. Choa, K.-W. Yi, J. Power Sources 172 (2007) 859–
862.
[6] F.-J. Liu, S. Suda, J. Alloys Compd. 231 (1995) 742–750.
[7] J.F.R. de Castro, A.R. Yavari, A. LeMoulec, T.T. Ishikawa, W.J. Botta, J. Alloys Compd.
389 (2005) 270–274.
[25] J. O. Hill (Ed.), For Better Thermal Analysis III, Special edition of the International
Confederation for Thermal Analysis (ICTA), 1991.
[8] S. Deledda, A. Borissova, C. Poinsignon, W.J. Botta, M. Dornheim, T. Klassen, J.
Alloys Compd. 404–406 (2005) 409–412.