Thermodynamical stabilities of metal-borohydrides

Article abstract of DOI:10.1016/j.jallcom.2007.03.144

Metal-borohydrides M(BH₄)_n (M = Ca, Sc, Ti, V, Cr, Mn, Zn (fourth period in periodic table) and Al; n = 2-4) were synthesized by a mechanical milling process according to the following reaction: MCl_n + nLiBH₄ → M(BH₄)_n + nLiCl. The thermal desorption properties of M(BH₄)_n were investigated by gas-chromatography and mass-spectroscopy combined with thermogravimetry. The results indicate that the hydrogen desorption temperature T_d of M(BH₄)_n correlates with the Pauling electronegativity χ_P of M; that is, T_d decreases with increasing values of χ_P. The components of desorbed gas for M = Ca, Sc, Ti, V and Cr (χ_P ≤ 1.5) are hydrogen only, while those for M = Mn, Zn and Al (χ_P ≥ 1.5) contain borane and hydrogen. Therefore, the appropriate χ_P in M(BH₄)_n for hydrogen storage is expected to be smaller than 1.5, and the resulting predicted relationship between the heat of formation of M(BH₄)_n and χ_P is a useful indicator for estimating the enthalpy change for the desorption reaction for M(BH₄)_n.

Full text of DOI:10.1016/j.jallcom.2007.03.144

Journal of Alloys and Compounds 446–447 (2007) 296–300
Thermodynamical stabilities of metal-borohydrides
a,∗
a
a
b
b
Y. Nakamori , H.-W. Li , K. Kikuchi , M. Aoki , K. Miwa ,
b
a
S. Towata , S. Orimo
a
Institute for Materials Research, Tohoku University, Katahira 2-1-1, Sendai 980-8577, Japan
b
Toyota Central R&D Labs., Inc., Nagakute, Aichi 480-1192, Japan
Received 31 October 2006; received in revised form 17 March 2007; accepted 20 March 2007
Available online 5 April 2007
Abstract
Metal-borohydrides M(BH
4
)
n
(M = Ca, Sc, Ti, V, Cr, Mn, Zn (fourth period in periodic table) and Al; n = 2–4) were synthesized by a mechanical
+ nLiBH + nLiCl. The thermal desorption properties of M(BH were
→ M(BH
milling process according to the following reaction: MCl
investigated by gas-chromatography and mass-spectroscopy combined with thermogravimetry. The results indicate that the hydrogen desorption
temperature T of M(BH correlates with the Pauling electronegativity χ of M; that is, T decreases with increasing values of χ . The components
of desorbed gas for M = Ca, Sc, Ti, V and Cr (χ ≥ 1.5) contain borane and
≤ 1.5) are hydrogen only, while those for M = Mn, Zn and Al (χ
hydrogen. Therefore, the appropriate χ in M(BH for hydrogen storage is expected to be smaller than 1.5, and the resulting predicted relationship
between the heat of formation of M(BH and χ is a useful indicator for estimating the enthalpy change for the desorption reaction for M(BH
2007 Elsevier B.V. All rights reserved.
n
4
4
)
n
4 n
)
d
4
)
n
P
d
P
P
P
P
4 n
)
4
)
n
P
4 n
) .
©
Keywords: Hydrogen absorbing materials; Mechanical alloying; Thermodynamical properties; Thermal analysis
1
. Introduction
orbs hydrogen below 673 K, a lower temperature than for LiBH4
12]. Vajo et al. have reported that mechanically milled mixtures
[
Complex hydrides have attracted growing interest as hydro-
of LiBH4 + (1/2)MgH2 were destabilized than pure LiBH4 and
underwent reversible absorption and desorption reactions [13].
gen storage materials because of their high gravimetric hydrogen
densities [1–2]. Bogdanovi c´ and Schwickardi have reported
that catalyzed sodium alanate (NaAlH4) undergoes reversible
hydrogen desorption and re-absorption reactions at moderate
conditions [3]. Motivated by this report, many attempts have
been made to enhance the hydrogen desorption and absorption
reactions for alkali complex hydrides [4–10].
In order to decrease the hydrogen desorption temperature T and
d
to improve the reversible reaction properties, systematic under-
standings of the thermodynamic stabilities of M(BH4)n are of
great importance.
Recently, we have theoretically investigated a series of
M(BH4)n including (M = Li, Na, K, Cu, Mg, Zn, Sc, Zr, Hf
and Ti) [14,15] theoretically and experimentally. The results
Among the alkali complex hydrides, lithium borohydride
LiBH4) is one of the candidates for use as a hydrogen stor-
n+
(
revealed that the charge compensation by M is a key feature
age material because of its extremely high gravimetric hydrogen
density. LiBH4 desorbs approximately 14 mass% of hydrogen
through the decomposition reaction as follows [11]:
for the stability of M(BH4), and that the hydrogen desorption
temperature T of M(BH4)n decreases with increasing elec-
d
tronegativity χP of M. Moreover, M with a larger χP becomes a
n+
n+
di-, tri- or tetra-valent cation of M , and M combines with n
3
2
LiBH4 → LiH + B + H2
(1)
−
moles of [BH ] to form M(BH ) . This means that M(BH )
4
4 n
4 n
maintain high gravimetric hydrogen density as shown in
Table 1.
However, the hydrogen desorption temperature is higher than the
required temperature for solid state hydrogen storage materials.
Z u¨ ttel et al. have reported that LiBH4 mixed with SiO2 des-
In this study, metal-borohydrides M(BH4)n (M = Ca, Sc, Ti,
V, Cr, Mn, Zn (fourth period in periodic table) and Al; n = 2–4)
are synthesized, and the purposes are (i) to clarify whether our
proposed correlation between the stability of borohydrides and
theelectronegativityofMcanbeextendedalsototheseM(BH4)n
∗
Corresponding author. Tel.: +81 22 215 2093; fax: +81 22 215 2091.
E-mail address: yuko@imr.tohoku.ac.jp (Y. Nakamori).
0
925-8388/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2007.03.144

Y. Nakamori et al. / Journal of Alloys and Compounds 446–447 (2007) 296–300
297
Table 1
Summary of the hydrogen desorption properties of M(BH4)n + nLiCl (M = Ca, Sc, Ti, V, Cr, Mn, Zn and Al)
M(BH4)n
Ca(BH4)2 Sc(BH4)3 Ti(BH4)3 V(BH4)3 Cr(BH4)2 Mn(BH4)2 Zn(BH4)2 Al(BH4)3
Electronegativity χP of M
1.0
11.6
5.2
4.2
1.3
13.5
5.6
4.2
529
H2
1.5
13.1
5.5
1.6
12.7
5.4
1.6
9.9
4.8
0.1
1.5
9.5
4.8
9.1
442
H2, B2H6
1.6
8.5
4.5
13.0
433
1.5
16.9
6.1
23.4
334
Gravimetric hydrogen density (mass%)
Gravimetric hydrogen density with nLiCI (mass%)
Observed weight loss (mass%)
Desorbed temperature Td (first peak temperature in Fig. 4) 747
Desorbed gases
0.1
0.1
a
395
∼400
∼400
b
b
b
H2
H2
H2
H2
H2, B2H6
H2, B2H6
a
The thermal desorption temperature for Al(BH4)3 was defined as the first peak temperature of Fig. 6 (detected by mass-spectroscopy).
The components of desorbed gas are confirmed to be only hydrogen according to the accuracy of our apparatus. However, the desorbed borane is undeniable,
b
because the amount of desorbed gas is very small.
and (ii) to select the appropriate M for the borohydrides as
hydrogen storage materials.
V, Cr, Mn and Zn. These results indicate that the reaction of
Eq. (2 ) did not proceed to completion under the milling con-
ꢀ
ditions in this experiment. In other words, the reaction of Eq.
2
. Experimental
(2) (starting materials of MCl + nLiBH ) proceeds easier than
n
4
ꢀ
the reaction of Eq. (2 ) (starting materials of MCln + nNaBH4)
in the mechanical milling process.
The starting materials were purchased from Aldrich Co., Ltd.; anhydrous
MCln (M = Ca, Sc, Ti, V, Cr, Mn, Zn and Al) with 99.9–99.999% purities,
LiBH4 and NaBH4 with 95–99.9% purities. Then, the mixture of MCln and
nLiBH4/nNaBH4 was mechanically milled under 0.1 MPa argon gas atmosphere
for 5 h in order to synthesize M(BH4)n (M = Ca, Sc, Ti, V, Cr, Mn, Zn and Al).
Mechanochemical syntheses have been reported for the preparation of Zn(BH4)2
The enthalpy changes of the synthesis reactions for Eqs. (2)
ꢀ
and (2 ), ꢀHsyn, are deduced using our calculated heat of for-
mation for M(BH ) , ꢀH [14] and those reported for MCln,
boro
4
n
e.g., −408 and −411 kJ/mol for LiCl and NaCl, respectively
[
16] and Zr(BH4)4 [17]. The expected reaction during milling is expressed as
[
19]. The results are shown in Fig. 2. The calculated values of
follows:
ꢀ
the enthalpy changes for Eq. (2 ) have more negative values than
MCln + nLiBH4 → M(BH4)n + nLiCl
MCln + nNaBH4 → M(BH4)n + nNaCl
(2)
ꢀ
(2 )
The samples thus prepared were examined by powder X-ray diffraction mea-
surement (PANalytical X’PERT with Cu K␣ radiation), Raman spectroscopy
(Nicolet, Almega-HD, 532 nm laser with back scattering geometry) and ther-
mal desorption spectroscopy detected by gas chromatography (GL Science
GC323, Ar flow rate of 40 ml/min and a heating rate of 5 K/min), and by mass-
spectroscopy (Anelva M-QA200TS, He flow rate of 100 ml/min and heating rate
of 5 K/min) combined with thermogravimetry (Rigaku, TG8120, He flow rate of
1
00 ml/min and a heating rate of 5 K/min). The samples were always handled in
a glove box filled with purified Ar/He (dew point below 183 K) without exposing
the samples to air. The experimental details were described elsewhere [18].
3
. Results and discussion
3
.1. Syntheses of M(BH4)n
Fig. 1(a and b) shows the powder X-ray diffraction pro-
files of the samples after mechanical milling of MCln + nLiBH4
and MCln + nNaBH4, respectively. There are broad diffraction
◦
peaks around 20 in all profiles of Fig. 1(a), originating from
the diffraction of tape covering the samples to avoid (hydro-
)
oxidation by exposure to air. In Fig. 1(a), the diffraction peaks
measured after mechanical milling of MCln + nLiBH4 are iden-
tified as LiCl, although peak shifts were observed, probably due
to the partial substitution of M for Li in LiCl. There are no
diffraction peaks of the starting materials of MCln and LiBH4,
indicating that the reaction in Eq. (2) has progressed to com-
pletion. On the other hand, the starting materials of MCln and
NaBH4 were obviously present confirmed in the diffraction pro-
files for M = Ca, Sc and Al, as shown in Fig. 1(b). Moreover, the
peak positions of NaCl in the profiles prepared by mechani-
cal milling of MCln + nNaBH4 shift to lower angles for M = Ti,
Fig. 1. Powder X-ray diffraction profiles of the samples after mechanical milling
of (a) MCln + nLiBH4 and (b) MCln + nNaBH4 (M = Ca, Sc, Ti, V, Cr, Mn, Zn
and Al).

2
98
Y. Nakamori et al. / Journal of Alloys and Compounds 446–447 (2007) 296–300
Fig. 2. Deduced enthalpy changes of the synthesis reactions for Eqs. (2) and
ꢀ
(
2 ), ꢀHsyn. The heats of formation of M(BH4)n, ꢀHboro, for M = Li, Na, Ca,
Sc, Zn and Al from our previous work were used (−161, −156, −151, −71.7,
1
5.4 and −9.5 kJ/mol BH4, respectively), while the remaining heats of forma-
tion were deduced using the linear relationship ꢀH = 248.7χP − 390.8 in the
units of kJ/mol BH4 [14]. The reported values for LiCl (−408 kJ/mol) and NaCl
(
−411 kJ/mol) [19] were also used for these calculations.
ꢀ
those of Eq. (2). Therefore, the reaction in Eq. (2 ) is expected
to proceed easier, while the reaction of Eq. (2) proceeds read-
ily in this experiment. One of the reasons for the progression
of the reaction of Eq. (2) seems to be faster kinetics due to
+
n+
the similar ionic radii of Li and M in the solid–solid cation
+
exchange reaction; the ionic radius of Li (0.076 nm) is closer
3+
3+
3+
Fig. 3. Raman spectra of the samples measured after mechanical milling of
MCln + nLiBH4 (M = Ca, Sc, Ti, V, Cr, Mn, Zn and Al). The Raman spectrum
of LiBH4 is also shown for reference.
to those of Sc (0.075 nm), Ti (0.067 nm), V (0.079 nm),
Cr (0.073 nm), Mn2+ (0.067 nm) Zn (0.074 nm) and Al
2
+
2+
3+
+
(
0.054 nm) than that of Na (0.102 nm) [20].
There are no X-ray diffraction peaks of M(BH4)n in the pro-
M(BH4)n, because LiCl decomposes at temperatures higher than
files of the samples prepared by mechanical milling, probably
due to the absence of any long range order in the structure.
Therefore, the atomic vibrations were investigated by Raman
spectroscopy in order to confirm the existence of M(BH4)n. The
Ramanspectraofthesamplesmeasuredaftermechanicalmilling
of MCln + nLiBH4 are shown in Fig. 3. The spectrum of LiBH4
is also shown as reference [18]. The Raman peaks were observed
8
78 K. The hydrogen desorption peaks of M(BH4)n decrease
with increasing values of χP of M. Here, the desorption tem-
perature T was defined as the temperature of the first peak,
d
and T as a function of χP was plotted in Fig. 5. A linear rela-
d
tionship was observed between T and χP, indicating that χP
d
−
1
at around 2200–2500 and 1100–1400 cm for M = Li, Ca, Sc,
Ti, Cr, Mn, Zn and Al. These results agree fairly well with the-
−
oretical predictions for free [BH4] anion; the normal mode
−
1
−1
frequencies are 2225 cm and 2236 cm for B H stretching
−
1
−1
modes and 1029 cm and 1154 cm for B H bending modes
21]. Raman spectroscopy measurements for M = V were not
[
successful because the sample was charred by laser irradiation.
The products of Eq. (2), LiCl and M(BH4)n, were confirmed
in the samples by powder X-ray diffraction measurements and
Raman spectroscopy, respectively, except for M = V.
3
.2. Thermal desorption properties of M(BH4)n
Fig. 4 shows the thermal desorption profiles of the mix-
ture of MCln + nLiBH4 (M = Ca, Sc, Ti, V, Cr, Mn, and Zn)
after mechanical milling as detected by gas-chromatography.
The previous results for M = Li, Na and K are shown for refer-
ence [14]. Although the samples contain not only M(BH4)n but
also LiCl, the thermal desorption profiles originate only from
Fig. 4. Thermal desorption profiles of the samples of MCln + nLiBH4 (M = Ca,
Sc, Ti, V, Cr, Mn, and Zn) after mechanical milling detected by gas-
chromatography. The results of MBH4 (M = Li, Na and K) are also shown as
references. Pauling electronegativity χp is shown below each M. The decompo-
sition reaction for M = Al cannot be studied because its boiling temperature is
317 K.

Y. Nakamori et al. / Journal of Alloys and Compounds 446–447 (2007) 296–300
299
Fig. 5. The desorption temperature Td as a function of the Pauling electroneg-
ativity χp. The previous results for M(BH4)n (M = Li, Na, Mg and Zr) are also
shownasreferences[14]. (ThethermaldesorptiontemperatureforAl(BH4)3 was
the first peak temperature of Fig. 6). The inset shows the correlation between Td
and estimated ꢀHdes for the desorption reaction.
is a good indicator for approximate estimations of the T of
d
M(BH4)n (M = elements of fourth period in periodic table and
Al), approximately.
Since M(BH4)n desorbs hydrogen accompanying the forma-
tion of hydrides and/or borides, the stabilities of the decomposed
products should be taken into account. We assume that M(BH4)n
Fig. 6. Thermal desorption profiles of the samples of MCln + nLiBH4 (M = Ca,
Sc, Ti, V, Cr, Mn, Zn and Al) after mechanical milling detected by (a) mass-
spectroscopy and (b) thermogravimetry. In (b), the right vertical axis is for
M = Mn, Zn Al. The differences in the peak features and peak temperatures
between the thermal desorption profiles detected by gas-chromatography (Fig. 4)
and those detected by mass-spectroscopy (a) seems to be due to different condi-
tions, such as heating atmospheres, gas-flow rates, temperatures and the length
of gas flow routes.
(M = Sc, Ti and V) decompose into hydrides in the first desorp-
tion reaction as follows:
1
2
M(BH4) → MHm + nB + (4n − m)H2
(3)
n
For M = Cr, Mn, Zn and Al, decomposition into elements was
assumed due to the instability of MHm. The decomposition reac-
tion for Ca(BH4)2 is described in literature [22]; Ca(BH4)2
decomposes to 2/3CaH2, 1/3CaB and 10/3H2. The enthalpy
6
theoretical values, as shown in Table 1. On the other hand, the
measured weight losses for M = Ti, V and Cr (χP = 1.5 and 1.6)
are significantly smaller than the theoretical values, while those
for M = Ca and Sc correspond to the theoretical values as sum-
marized in Table 1. Therefore, M(BH4)n (χP < 1.5) is expected
to be a candidate hydrogen storage material with appropriate
stability and without desorbing borane. The adjustment of the
desorption temperature using two kind of mixed cations is also
under investigation [24].
change for the desorption reaction, ꢀHdes, was estimated using
the predicted ꢀHboro for M(BH4)n [14] and the reported exper-
imental data for product e.g., ꢀHhyd [1,2,23]. The estimated
ꢀ
Hdes as a function of observed T is depicted in the inset of
d
Fig. 5. A good correlation between ꢀHdes and T was observed,
d
indicating that ꢀHdes can be approximately estimated using the
reported ꢀHboro [14]. However, there is a deviation from linear
in the relationship between ꢀHdes and T for M = Mn, Cr and
d
Al (χP ≥ 1.5). This difference might be due to the weak ionic
n+
−
bonding between M and [BH4] because of their unclosed and
delocalized 3d electron shell. Another possibility for the cause
of the deviation is that the assumption of desorption reaction
4. Summary
is incorrect, or that kinetics influence the observed T . In order
In order to clarify the correlation between the stability of
borohydrides and the cation electronegativity χP of M, M(BH4)n
(M = Ca, Sc, Ti, V, Cr, Mn, Zr (elements at fourth period in peri-
odic table) and Al; n = 2–4) were synthesized by mechanical
milling and the thermal desorption properties were investi-
gated. The mixture of MCln and nLiBH4 changes into M(BH4)n
with nLiCl by mechanical milling, which was confirmed by
Raman spectroscopy and powder X-ray diffraction measure-
ment, respectively. The results of thermal desorption reactions
d
to discuss ꢀHdes in detail for the desorption reaction, we are
currently investigating the actual hydrogen desorption equation
using well crystallized M(BH4)n synthesized in solution.
Finally, mass-spectroscopy and thermogravimetry have been
carriedoutinordertoinvestigatethecomponentsofthedesorbed
gas. From the results of mass-spectroscopy (Fig. 6(a)), the com-
ponents of desorbed gas for M = Ca, Sc, Ti, V and Cr (χP ≤ 1.6)
are confirmed to be only hydrogen within the accuracy of our
apparatus. However, those for M = Mn, Zn and Al (χP ≥ 1.5)
contain borane in addition to hydrogen. Weight loss for M = Mn,
Zn and Al as observed by thermogravimetry are larger than the
indicated that T decreases with increasing values of χP of
d
M(BH4)n (M = element in the fourth period in the periodic table
or Al). The estimated ꢀHdes for the desorption reaction corre-

3
00
Y. Nakamori et al. / Journal of Alloys and Compounds 446–447 (2007) 296–300
lates well with T , whereas a deviation from a linear relationship
[5] L. Zaluski, A. Zaluska, J.O. Str o¨ m-Olsen, J. Alloys Compd. 290 (1999)
1.
d
7
appears in M(BH4)n with χP ≥ 1.5. Moreover, the components
of desorbed gas from M = Mn, Zn and Al (χP ≥ 1.5) contain
borane, and a small amount of hydrogen is confirmed for M = Ti,
V, Cr (χP ≥ 1.5). Therefore, the appropriate χP of M(BH4)n for
hydrogen storage is expected to be smaller than 1.5, and the pre-
dicted relationship between ꢀHboro and χP is a useful indicator
for estimating ꢀHdes for M(BH4)n.
[
6] B. Bogdanovi c´ , R.A. Brand, A. Marjanvi c´ , M. Schwickardi, J. T o¨ lle, J.
Alloys Compd. 302 (2000) 36.
[
[
7] C.M. Jensen, K.J. Gross, Appl. Phys. A 72 (2001) 213.
8] D. Sun, T. Kiyobayashi, H. Takeshita, N. Kuriyama, C.M. Jensen, J. Alloys
Compd. 337 (2002) L8.
[
9] G. Sandrock, K. Gross, G. Thomas, J. Alloys Compd. 339 (2002) 299.
[
10] H. Morioka, K. Kakizaki, S.C. Chung, A. Yamada, J. Alloys Compd. 353
2003) 310.
(
[
[
11] D.S. Stasinevich, G.A. Egorenko, Russ. J. Inorg. Chem. 13 (1968) 341.
12] A. Z u¨ ttel, P. Wenger, S. Rentsch, P. Sudan, Ph. Mauron, Cd. Emmenegger,
J. Power Sources 5194 (2003) 1.
Acknowledgements
[
[
13] J.J. Vajo, S.L. Skeith, F. Martens, J. Phys. Chem. 109 (2005) 3719.
14] Y. Nakamori, K. Miwa, A. Ninomiya, H.-W. Li, N. Ohba, S. Towata, A.
Z u¨ ttel, S. Orimo, Phys. Rev. B 74 (2006) 045126.
The authors would like to thank T. Noritake and S. Hyodo
for valuable discussion, and N. Warifune for her techni-
cal support. This study was partially supported by the New
Energy and Industrial Technology Development Organization
[
15] Y. Nakamori, H.-W. Li, K. Miwa, S. Towata, S. Orimo, Mater. Trans. 47
(
2006) 1898.
(
NEDO), International Joint Research under the “Development
for Safe Utilization and Infrastructure of hydrogen” Project
2005–2006), by the Ministry of Education, Science, Sports and
Culture, “Grant-in-Aid for Encouragement of Young Scientists
B), #17760555” and for Scientific Research (A), #18206073”.
[16] V.E. Wiberg, W. Henle, Z. Naturforschg. 76 (1952) 579.
[17] V.V. Volkov, K.G. Myakishev, S.I. Yugov, J. Appl. Chem. 48 (1975) 2184.
[
18] S. Oirmo, Y. Nakamori, A. Z u¨ ttel, Mater. Sci. Eng. B 108 (2005) 51.
[19] D.R. Lide, CRC Handbook of Chemistry and Physics, 83rd ed., CRC press,
(
(
2002) 5-5-5-25.
(
[
[
[
20] L. Pauling, The Nature of Chemical Bonds, third ed., Cornell University
Press, 1960.
21] K. Miwa, N. Ohba, S. Towata, Y. Nakamori, S. Orimo, Phys. Rev. B 69
References
(
2004) 245120.
22] K. Miwa, M. Aoki, T. Noritake, N. Ohba, Y. Nakamori, S. Towata, A.
Z u¨ ttel, S. Orimo, Phys. Rev. B 74 (2006), 155122.
[
[
[
[
1] R. Schlapbach, A. Z u¨ ttel, Nature 414 (2001) 353.
2] W. Grochala, P.P. Edwards, Chem. Rev. 104 (2004) 1283.
3] B. Bogdanovi c´ , M. Schwickardi, J. Alloys Compd. 253 (1997) 1.
4] R.A. Zidan, S. Takara, A.G. Hee, C.M. Jensen, J. Alloys Compd. 285 (1999)
[
[
23] http://hydpark.ca.sandia.gov/MaterialsFrame2.html.
24] H.-W. Li, S. Orimo, Y. Nakamori, K. Miwa, N. Ohba, S. Towata, A. Z u¨ ttel,
J. Alloys Compd. 446–447 (2007) 315.
1
19.