**Aluminum-based complex hydrides**

904 Thermodynamics – Interaction Studies – Solids, Liquids and Gases

Fig. 10. Tailoring of the reaction enthalpy by altering the stability of the hydrogenated or dehydrogenated state of the metal hydrides: a) Reduction of total reaction enthalpy by

formation of H0f(Mg2Ni) = -42 kJ/mol. Therefore, compared to pure Mg the dehydrogenated state is stabilised by Hds = -21 kJ/(mol Mg). The enthalpy of formation of Mg2NiH4 is H0f(Mg2NiH4) = -176 kJ/mol (= -88 kJ/(mol Mg)), wherefore the hydride phase is stabilised by Hhd = -10 kJ/(mol Mg) if compared to pure MgH2. In total the hydrogen reaction

is reduced by 11 kJ/mol H2 to aboutH(Mg2Ni-H) = 67 kJ/mol H2. While pure MgH2 exhibits a hydrogen plateau pressure of 1 bar around 300 °C, in case of Mg2NiH4 such a plateau pressure is reached already at around 240 °C and in case of further alloying and substituting Ni by Cu at around 230°C in Mg2Ni0.5Cu0.5 (Klassen et al., 1998). Unfortunately, the gravimetric storage capacity of Mg2NiH4 is reduced to 3.6 wt.% H2 only and thus is less than half the respective value in the MgH2 system. Darnaudery et al. (Darnaudery et al., 1983) were successful to form several quaternary hydrides by hydrogenating Mg2Ni0.75M0.25 with different 3d elements M {V, Cr, Fe, Co and Zn} showing stabilities very similar

Increasing the amount of 3d metals Tsushio et al. (Tsushio et al., 1998) investigated the hydrogenation of MgNi0.86M0.03 with M {Cr, Fe, Co, Mn}. Consequently, they observed a dramatic decrease in hydrogen storage capacity to 0.9 wt.% and in hydrogen reaction enthalpy which amounts to 50 kJ/(mol H2) for MgNi0.86Cr0.03. This reaction enthalpy value is in very good agreement with the value 54 kJ/(mol H2) given by Orimo et al. for amorphous

Lowering even more the content of Mg Terashita et al. (Terashita et al., 2001) found (Mg1-xCax)Ni2 based alloys desorbing hydrogen at room temperature. They determined the hydride formation enthalpy and entropy of (Mg0.68Ca0.32)Ni2 to be H = -37 kJ/(mol H2) and S = - 94 J/(mol H2 K) respectively, which is already quite near to the envisioned target.

Hds or destabilising the hydride phase by

H H HH Mg2Ni-H Mg-H ds hd (20)

Hhd.

Hdd or

b) Increase of total reaction enthalpy by destabilising the dehydrogenated state by

Hhs.

stabilising the dehydrogenated phase by

stabilising the hydrogenated state by

enthalpy of Mg2Ni

Mg2NiH4.

MgNi (Orimo et al., 1998).

As Mg2FeH6 decomposes during hydrogen release into 2 Mg, Fe and 3 H2 NaAlH4 decomposes during hydrogen release in 1/3 Na3AlH6 + 2/3 Al + H2 and finally NaH + Al + 3/2 H2. As written in chapter 2.2 while much lower than those of the Mg-based hydrides the reaction enthalpies of |H|= 37 kJ/(mol H2) and |H|= 47 kJ/(mol H2) are still two high for many applications especially for the usage in combination with low temperature PEM fuel cells. LiAlH4 on the other hand is much less stable. It decomposes in two steps as is the case of the NaAlH4:

$$\text{6LiAlH}\_4 \rightarrow \text{2Li}\_3\text{AlH}\_6 + 4\text{Al} + \text{6H}\_2 \underset{?}{\leftrightarrow} \text{6LiH} + \text{6Al} + \text{9H}\_2.\tag{21}$$

The first reaction step, however, the decomposition of LiAlH4 is found to be exothermic with Hdecomposition = -10 kJ/(mol H2). Since the entropy of decomposition is positive.

Thermodynamics of Metal Hydrides: Tailoring Reaction Enthalpies of Hydrogen Storage Materials 907

and the aluminum hydrides the reaction enthalpy of many borohydrides is rather unsuitable for most applications. LiBH4 as one of the most investigated borohydrides with a very high gravimetric hydrogen density of 18.5 wt.% shows an endothermic desorption enthalpy of |DH| = 74 kJ/(mol H2) (Mauron et al., 2008) which is almost the same as in MgH2. Therefore the tailoring of the reaction enthalpy by substitution is a key issue for these materials as well. As in case of the aluminium hydrides there are two different possibilities for substitution in these complex hydrides: cation substitution and anion substitution. Nakamori et al. (Nakamori et al., 2006) reports about a linear relationship between the heat of formation Hboro of M(BH4)n determined by first principle methods and the Pauling

kJ mol BH

shows correlates with the Pauling electronegativity p as well, see Fig. 11.

4 H 248.7 390.8

Aiming to confirm their theoretical results the same group performed hydrogen desorption experiments which show that the experimentally determined desorption temperature Td

Fig. 11. The desorption temperature Td as a function of the Pauling electronegativity P and

Based on these encouraging results several research groups started to investigate the partial substitution of one cation by another studying several bialkali metal borohydrides. The decomposition temperature of the bialkali metal borohydrides like LiK(BH4)2 is approximately the average of the decomposition temperature of the mono alkali borohydrides (Rude et al., 2011). Investigations of Li et al. (Li et al., 2007) and Seballos et al. (Seballos et al., 2009) confirmed that this correlation between desorption enthalpy /

Several experiments are indicating that transition metal fluorides are among the best additives for borohydrides (Bonatto Minella et al., 2011). While for some cases the function of the transition metal part as additive is understood (Bösenberg et al., 2009; Bösenberg et al., 2010; Deprez et al., 2010; Deprez et al., 2011), the function of F so far remained unclear. DFT calculations performed by Yin et al. (Yin et al., 2008) suggest a favourable modification

observed Td holds true for many double cation MM'(BH4)n systems, see Fig. 12.

1 P

(23)

boro

estimated desorption enthalpies Hdes (Nakamori et al., 2007).

electronegativity of the cation p:

Rehydrogenation is not possible at all. The second reaction step, the decomposition of Li3AlH6 is endothermic with Hdecomposition = 25 kJ/(mol H2). The decomposition of LiH itself takes place at much higher temperatures with H = 140 kJ/(mol H2) (Orimo et al., 2007). While the second reaction step, the decomposition of Li3AlH6 and rehydrogenation of LiH + Al shows rather suitable thermodynamic properties, sluggish kinetics prevent this system so far from being used for hydrogen storage.

To increase the storage capacity and tailor the reaction enthalpy of the NaAlH4 system it is a comprehensible approach to replace some of the Na by Li. Indeed Huot et al. (Huot et al., 1999) proved the existence of Na2LiAlH6 and the possible formation by high energy ballmilling of NaH + LiH + NaAlH4. Reversible hydrogen sorption is found to be possible in the Na-Li-Al-H system according to the following reaction:

$$2\text{Na}\_2\text{LiAlH}\_6 \leftrightarrow \text{ 4NaH} + 2\text{LiH} + \text{Al} + 3\text{H}\_2 \tag{22}$$

As in case of the pure Na-Al-H system and the Li-Al-H system kinetics can be improved by the addition of transition metal compounds like metal oxides, chlorides and fluorides, see (Ares Fernandez et al., 2007), (Ma et al., 2005) and (Martinez-Franco et al., 2010). However, due to the lack of any stable compound in the dehydrogenated state and the formation of a rather stable hydride the value of reaction enthalpy isn't decreased but increased if compared to the original single Na and Li based aluminium hydrides. Fossdal et al. (Fossdal et al., 2005) has determined the pressure-composition isotherms of TiF3-doped Na2LiAlH6 in the temperature range of 170 °C – 250 °C. They determined the dissociation enthalpy and the corresponding entropy from the Van't Hoff plot: |DH| = 56 kJ/(mol H2) and S = 138 J/(K mol H2). Therefore, instead of a lowering the heat of reaction the opposite is observed. The heat of reaction of the hexa-hydride phase is increased by about 10 kJ/(mol H2) if compared to the pure Na3AlH6 hydride phase.

In 2007 Yin et al. (Yin et al., 2007) presented DFT calculations about the doping effects of TiF3 on Na3AlH6. Their calculations suggested F- substitution for the H-anion leading to a reduction of the desorption enthalpy and therefore for a favourable thermodynamic modification of the Na3AlH6 system which was experimentally confirmed by Brinks et al. (Brinks et al., 2008) and Eigen et al. (Eigen et al., 2009).

### **Borohydrides**

Only a very few hydrides show a higher gravimetric storage capacity than MgH2. For this they must be composed from very light elements. Knowing that Al-containing compounds can form reversible complex metal hydrides it is a reasonable approach to look for Boroncontaining compounds as reversible hydrogen storage materials with even higher storage capacity. Borohydrides are known since 1940 when Schlesinger and Brown report about the successful synthesis of LiBH4 by reaction of LiEt and diborane (Schlesinger & Brown, 1940). Despite the early patent from Goerrig in 1958 (Goerrig, 1960) direct synthesis from gaseous H2 was not possible for long times. Until in 2004 three different groups from the USA (Vajo et al., 2005), South Korea (Cho et al., 2006) and Germany (Barkhordarian et al., 2007) independently discovered that by using MgB2 instead of pure Boron as starting material formation of the respective borohydrides occurs at rather moderate conditions of 5 MPa H2 pressure. Orimo et al. (Orimo et al., 2005) reports on the rehydrogenation of previously dehydrogenated LiBH4 at 35 MPa H2 pressure at 600 °C. Mauron et al. (Mauron et al., 2008) report that rehydrogenation is also possible at 15 MPa. As in case of the Mg-based alloys

Rehydrogenation is not possible at all. The second reaction step, the decomposition of Li3AlH6 is endothermic with Hdecomposition = 25 kJ/(mol H2). The decomposition of LiH itself takes place at much higher temperatures with H = 140 kJ/(mol H2) (Orimo et al., 2007). While the second reaction step, the decomposition of Li3AlH6 and rehydrogenation of LiH + Al shows rather suitable thermodynamic properties, sluggish kinetics prevent this

To increase the storage capacity and tailor the reaction enthalpy of the NaAlH4 system it is a comprehensible approach to replace some of the Na by Li. Indeed Huot et al. (Huot et al., 1999) proved the existence of Na2LiAlH6 and the possible formation by high energy ballmilling of NaH + LiH + NaAlH4. Reversible hydrogen sorption is found to be possible in the

As in case of the pure Na-Al-H system and the Li-Al-H system kinetics can be improved by the addition of transition metal compounds like metal oxides, chlorides and fluorides, see (Ares Fernandez et al., 2007), (Ma et al., 2005) and (Martinez-Franco et al., 2010). However, due to the lack of any stable compound in the dehydrogenated state and the formation of a rather stable hydride the value of reaction enthalpy isn't decreased but increased if compared to the original single Na and Li based aluminium hydrides. Fossdal et al. (Fossdal et al., 2005) has determined the pressure-composition isotherms of TiF3-doped Na2LiAlH6 in the temperature range of 170 °C – 250 °C. They determined the dissociation enthalpy and the corresponding entropy from the Van't Hoff plot: |DH| = 56 kJ/(mol H2) and S = 138 J/(K mol H2). Therefore, instead of a lowering the heat of reaction the opposite is observed. The heat of reaction of the hexa-hydride phase is increased by about

In 2007 Yin et al. (Yin et al., 2007) presented DFT calculations about the doping effects of TiF3 on Na3AlH6. Their calculations suggested F- substitution for the H-anion leading to a reduction of the desorption enthalpy and therefore for a favourable thermodynamic modification of the Na3AlH6 system which was experimentally confirmed by Brinks et al.

Only a very few hydrides show a higher gravimetric storage capacity than MgH2. For this they must be composed from very light elements. Knowing that Al-containing compounds can form reversible complex metal hydrides it is a reasonable approach to look for Boroncontaining compounds as reversible hydrogen storage materials with even higher storage capacity. Borohydrides are known since 1940 when Schlesinger and Brown report about the successful synthesis of LiBH4 by reaction of LiEt and diborane (Schlesinger & Brown, 1940). Despite the early patent from Goerrig in 1958 (Goerrig, 1960) direct synthesis from gaseous H2 was not possible for long times. Until in 2004 three different groups from the USA (Vajo et al., 2005), South Korea (Cho et al., 2006) and Germany (Barkhordarian et al., 2007) independently discovered that by using MgB2 instead of pure Boron as starting material formation of the respective borohydrides occurs at rather moderate conditions of 5 MPa H2 pressure. Orimo et al. (Orimo et al., 2005) reports on the rehydrogenation of previously dehydrogenated LiBH4 at 35 MPa H2 pressure at 600 °C. Mauron et al. (Mauron et al., 2008) report that rehydrogenation is also possible at 15 MPa. As in case of the Mg-based alloys

2 6 <sup>2</sup> 2Na LiAlH 4NaH 2LiH Al 3H (22)

system so far from being used for hydrogen storage.

Na-Li-Al-H system according to the following reaction:

10 kJ/(mol H2) if compared to the pure Na3AlH6 hydride phase.

(Brinks et al., 2008) and Eigen et al. (Eigen et al., 2009).

**Borohydrides** 

and the aluminum hydrides the reaction enthalpy of many borohydrides is rather unsuitable for most applications. LiBH4 as one of the most investigated borohydrides with a very high gravimetric hydrogen density of 18.5 wt.% shows an endothermic desorption enthalpy of |DH| = 74 kJ/(mol H2) (Mauron et al., 2008) which is almost the same as in MgH2. Therefore the tailoring of the reaction enthalpy by substitution is a key issue for these materials as well. As in case of the aluminium hydrides there are two different possibilities for substitution in these complex hydrides: cation substitution and anion substitution. Nakamori et al. (Nakamori et al., 2006) reports about a linear relationship between the heat of formation Hboro of M(BH4)n determined by first principle methods and the Pauling electronegativity of the cation p:

$$\begin{aligned} \text{4H}\_{\text{boro}} \\ \text{kJ} \text{(mol BH}\_4\text{)}^{-1} \end{aligned} \text{ = 248.7 } \text{\textsuperscript{} } \text{\textsuperscript{} } -390.8 \tag{23}$$

Aiming to confirm their theoretical results the same group performed hydrogen desorption experiments which show that the experimentally determined desorption temperature Td shows correlates with the Pauling electronegativity p as well, see Fig. 11.

Fig. 11. The desorption temperature Td as a function of the Pauling electronegativity P and estimated desorption enthalpies Hdes (Nakamori et al., 2007).

Based on these encouraging results several research groups started to investigate the partial substitution of one cation by another studying several bialkali metal borohydrides. The decomposition temperature of the bialkali metal borohydrides like LiK(BH4)2 is approximately the average of the decomposition temperature of the mono alkali borohydrides (Rude et al., 2011). Investigations of Li et al. (Li et al., 2007) and Seballos et al. (Seballos et al., 2009) confirmed that this correlation between desorption enthalpy / observed Td holds true for many double cation MM'(BH4)n systems, see Fig. 12.

Several experiments are indicating that transition metal fluorides are among the best additives for borohydrides (Bonatto Minella et al., 2011). While for some cases the function of the transition metal part as additive is understood (Bösenberg et al., 2009; Bösenberg et al., 2010; Deprez et al., 2010; Deprez et al., 2011), the function of F so far remained unclear. DFT calculations performed by Yin et al. (Yin et al., 2008) suggest a favourable modification

Thermodynamics of Metal Hydrides: Tailoring Reaction Enthalpies of Hydrogen Storage Materials 909

The system can reversibly store 4.4 wt.% H2. Since the formation enthalpy HForm of Mg17Al12 is -102 kJ/mol the total value of reaction enthalpy of reaction (26) is reduced by ~ 6 kJ/(mol H2) if compared to pure MgH2. An equilibrium pressure of 1 bar is reached at

To further decrease the reaction enthalpy of a Mg-based system a much more stable compound would have to be formed during dehydrogenation. A system investigated by many groups is the MgH2-Si system. Mg2Si has an enthalpy of formation of HForm = - 79 kJ/mol. Due to the formation of Mg2Si the value of reaction enthalpy of MgH2/Si should therefore be reduced by 37 kJ/(mol H2) to about |H| = 41 kJ/(mol H2) (Dornheim, 2010).

The thermodynamic data indicate a very favourable equilibrium pressure of about 1 bar at 20 °C and 50 bar at 120 °C (Vajo, 2004). While so far rehydrogenation of Mg2Si was not shown to be possible the system LiH-Si turned out to be reversible. The enthalpy of dehydrogenation of LiH being 190 kJ/(mol H2) an equilibrium H2 pressure of 1 bar is reached at 910 °C (Sangster, 2000; Dornheim, 2010). LiH reversibly reacts with Si via a two step reaction with the equilibrium pressure being more than 104 times higher and the

This approach has recently also been applied to borohydrides. According to Cho et al. (Cho et al., 2006) the decomposition temperature of pure LiBH4 is determined by CALPHAD to 1 bar H2 pressure at 403 °C while the corresponding equilibrium temperature for the reaction

is reduced to 188 °C. Kang et al. (Kang et al., 2007) and Jin et al. (Jin et al., 2008) could show

The only disadvantage of this approach is that the total reversible storage capacity per weight is reduced if something is added to the hydrogen storing material which contains no

The problem of reduced hydrogen capacity by using reactive additives has recently overcome by the approach of the Reactive Hydride Composites (Dornheim, 2006). Thereby, different high capacity hydrogen storage materials are combined which react exothermically

One of the first examples of such a system is the LiNH2-LiH system which was discovered

However, the value of reaction enthalpy is |H| = 80 kJ/(mol H2) and therefore for most

 Mg(NH2)2 + 2 LiH ↔ Li2Mg(NH)2 + 2H2 (30) shows a much more suitable desorption enthalpy of |H|~40 kJ/(mol H2) with an expected equilibrium pressure of 1 bar at approximately 90 °C (Xiong et al., 2005; Dornheim,

dehydrogenation enthalpy being reduced by 70 kJ/(mol H2) (Vajo, 2004).

that this system indeed is reversible if suitable additives are used.

with each other during decomposition, see Fig. 13.

applications still much to high. In contrast the system

by Chen et al. (Chen et al., 2002):

2 22 2 MgH Si M g Si 4 H (27)

<sup>4</sup> 2 2 2 LiBH Al 2 LiH AlB 3 H (28)

<sup>2</sup> <sup>2</sup> <sup>2</sup> <sup>3</sup> H2 LiNH 2 LiH Li NH LiH H Li N 2 (29)

around 240 °C again.

hydrogen.

2010).

Theoretically 5 wt.% H2 can be stored via the reaction

of hydrogen reaction enthalpy in the LiBH4 system by substitution of the H- -ion with the F- ion. However, no clear indicative experimental results for F- -substitution in borohydrides are found yet. In contrast to the F the heavier and larger halides Cl, Br, I are found to readily substitute in some borohydrides for the BH4- -ion and form solid solutions or stoichiometric compounds and are so far reported to stabilize the hydride phase leading to an increase of the desorption enthalpy |H| (Rude et al., 2011).

Fig. 12. Decomposition temperatures, Tdec for metal borohydrides plotted as a function of the electronegativity of the metal, M'. (Rude et al., 2011)

### **3.2 Thermodynamic tuning using multicomponent systems: reactive additives and reactive hydride composites**

In 1967 Reilly and Wiswall (Reilly & Wiswall, 1967) found another promising approach to tailor reaction enthalpies of hydrides (MHx) by mixing them with suitable reactants (A):

$$\text{MH}\_{\text{x}} + \text{yA} \leftrightarrow \text{MA}\_{\text{y}} + \,\,\_{2}^{\text{x}}\text{H}\_{2} \tag{24}$$

They investigated the system MgH2/MgCu2 which reversibly reacts with hydrogen according to:

$$2\text{ 3 MgH}\_2 + \text{MgCu}\_2 \leftrightarrow 2\text{ Mg}\_2\text{Cu} + \text{3H}\_2\tag{25}$$

The formation of MgCu2 from Mg2Cu and Cu is exothermic and thus counteracts the endothermic release of hydrogen. Thereby, the total amount of hydrogen reaction enthalpy is reduced to roughly |H| = 73 kJ/(mol H2) (Wiswall, 1978). The equilibrium temperature for 1 bar hydrogen pressure is reduced to about 240 °C. In spite of the lower driving force for rehydrogenation, Mg2Cu is much more easily hydrogenated than pure Mg. A fact found in many other systems like the Reactive Hydride Composites as well.

Aluminum is another example of a reactive additive for MgH2. The reaction occurs via two steps (Bouaricha et al., 2000):

$$17\text{ MgH}\_2 + 12\text{ Al} \leftrightarrow 9\text{ MgH}\_2 + 4\text{ Mg}\_2\text{Al}\_3 + 8\text{ H}\_2 \leftrightarrow \text{Mg}\_{17}\text{Al}\_{12} + 17\text{ H}\_2 \tag{26}$$

ion. However, no clear indicative experimental results for F- -substitution in borohydrides are found yet. In contrast to the F the heavier and larger halides Cl, Br, I are found to readily

compounds and are so far reported to stabilize the hydride phase leading to an increase of


Fig. 12. Decomposition temperatures, Tdec for metal borohydrides plotted as a function of

**3.2 Thermodynamic tuning using multicomponent systems: reactive additives and** 

In 1967 Reilly and Wiswall (Reilly & Wiswall, 1967) found another promising approach to tailor reaction enthalpies of hydrides (MHx) by mixing them with suitable reactants (A):

They investigated the system MgH2/MgCu2 which reversibly reacts with hydrogen

The formation of MgCu2 from Mg2Cu and Cu is exothermic and thus counteracts the endothermic release of hydrogen. Thereby, the total amount of hydrogen reaction enthalpy is reduced to roughly |H| = 73 kJ/(mol H2) (Wiswall, 1978). The equilibrium temperature for 1 bar hydrogen pressure is reduced to about 240 °C. In spite of the lower driving force for rehydrogenation, Mg2Cu is much more easily hydrogenated than pure Mg. A fact found

Aluminum is another example of a reactive additive for MgH2. The reaction occurs via two

<sup>2</sup> <sup>2</sup> 23 2 17 12 <sup>2</sup> 17 MgH 12 Al 9 MgH 4 Mg Al 8 H Mg Al 17 H (26)

in many other systems like the Reactive Hydride Composites as well.

<sup>x</sup> MH x yA MA H <sup>y</sup> <sup>2</sup> 2 (24)

2 2 22 3 MgH M gCu 2 M g Cu 3H (25)




of hydrogen reaction enthalpy in the LiBH4 system by substitution of the H-

substitute in some borohydrides for the BH4

the desorption enthalpy |H| (Rude et al., 2011).

the electronegativity of the metal, M'. (Rude et al., 2011)

**reactive hydride composites** 

steps (Bouaricha et al., 2000):

according to:

The system can reversibly store 4.4 wt.% H2. Since the formation enthalpy HForm of Mg17Al12 is -102 kJ/mol the total value of reaction enthalpy of reaction (26) is reduced by ~ 6 kJ/(mol H2) if compared to pure MgH2. An equilibrium pressure of 1 bar is reached at around 240 °C again.

To further decrease the reaction enthalpy of a Mg-based system a much more stable compound would have to be formed during dehydrogenation. A system investigated by many groups is the MgH2-Si system. Mg2Si has an enthalpy of formation of HForm = - 79 kJ/mol. Due to the formation of Mg2Si the value of reaction enthalpy of MgH2/Si should therefore be reduced by 37 kJ/(mol H2) to about |H| = 41 kJ/(mol H2) (Dornheim, 2010). Theoretically 5 wt.% H2 can be stored via the reaction

$$2\text{ MgH}\_2 + \text{Si} \rightarrow \text{Mg}\_2\text{Si} + 4\text{ H}\_2\tag{27}$$

The thermodynamic data indicate a very favourable equilibrium pressure of about 1 bar at 20 °C and 50 bar at 120 °C (Vajo, 2004). While so far rehydrogenation of Mg2Si was not shown to be possible the system LiH-Si turned out to be reversible. The enthalpy of dehydrogenation of LiH being 190 kJ/(mol H2) an equilibrium H2 pressure of 1 bar is reached at 910 °C (Sangster, 2000; Dornheim, 2010). LiH reversibly reacts with Si via a two step reaction with the equilibrium pressure being more than 104 times higher and the dehydrogenation enthalpy being reduced by 70 kJ/(mol H2) (Vajo, 2004).

This approach has recently also been applied to borohydrides. According to Cho et al. (Cho et al., 2006) the decomposition temperature of pure LiBH4 is determined by CALPHAD to 1 bar H2 pressure at 403 °C while the corresponding equilibrium temperature for the reaction

$$2\text{ LiBH}\_4 + \text{Al} \leftrightarrow \text{ 2LiH} + \text{AlB}\_2 + \text{3H}\_2 \tag{28}$$

is reduced to 188 °C. Kang et al. (Kang et al., 2007) and Jin et al. (Jin et al., 2008) could show that this system indeed is reversible if suitable additives are used.

The only disadvantage of this approach is that the total reversible storage capacity per weight is reduced if something is added to the hydrogen storing material which contains no hydrogen.

The problem of reduced hydrogen capacity by using reactive additives has recently overcome by the approach of the Reactive Hydride Composites (Dornheim, 2006). Thereby, different high capacity hydrogen storage materials are combined which react exothermically with each other during decomposition, see Fig. 13.

One of the first examples of such a system is the LiNH2-LiH system which was discovered by Chen et al. (Chen et al., 2002):

$$\text{LiNH}\_2 + 2\,\text{LiH} \leftrightarrow \text{Li}\_2\text{NH} + \text{LiH} + \text{H}\_2 \leftrightarrow \text{Li}\_3\text{N} + 2\,\text{H}\_2\tag{29}$$

However, the value of reaction enthalpy is |H| = 80 kJ/(mol H2) and therefore for most applications still much to high. In contrast the system

$$\text{Mg(NH}\_2\text{)}\_2 + 2\text{ LiH} \leftrightarrow \text{Li}\_2\text{Mg(NH}\_2\text{)}\_2 + 2\text{H}\_2\tag{30}$$

shows a much more suitable desorption enthalpy of |H|~40 kJ/(mol H2) with an expected equilibrium pressure of 1 bar at approximately 90 °C (Xiong et al., 2005; Dornheim, 2010).

Thermodynamics of Metal Hydrides: Tailoring Reaction Enthalpies of Hydrogen Storage Materials 911

equilibrium pressure at given temperature and, therefore, are important parameters to be taken into account. Optimised system integration for a given application is not possible without selecting a hydride with suitable thermodynamic properties. To achieve highest possible energy efficiencies the heat of reaction and temperature of operation of the metal hydride should be adapted to the waste heat and temperature of operation of the fuel cell / fuel combustion system. It has been found that the thermodynamic properties of metal hydrides can be tailored in a wide range. Unfortunately, so far all the known conventional metal hydrides with more or less ideal reaction enthalpies and hydrogen equilibrium pressures above 5 bar at room temperature suffer from a rather limited reversible hydrogen storage capacity of less than 2.5 wt.%. With such a material it is not possible to realise a solid storage hydrogen tank with a total hydrogen storage density of more than 1.8 wt.% H2. Such tank systems still have advantages for the storage of small quantities of hydrogen for larger quantities, however, modern high pressure composite tank shells have a clear advantage in respect of gravimetric storage density. To realise a solid storage tank for hydrogen with a comparable gravimetric storage density it is required that novel hydrogen storage materials based on light weight elements are developed. There are several promising systems with high gravimetric storage densities in the range of 8 – 12 wt.% H2. For the applications of these novel material systems it is important to further adapt thermodynamic properties as well as the temperatures of operation towards the practical requirements of the system. The discovery of the approach of combining different hydrides which react with each other during hydrogen release by forming a stable compound, the so-called Reactive Hydride Composites, show a great promise for the development of novel suitable hydrogen storage material systems with elevated gravimetric storage densities. However, so far, the ideal storage material with low reaction temperatures, a reaction heat in the range of |H| = 20- 30 kJ/(mol H2) and a on-board reversible hydrogen storage density of more than 6 wt.% H2

Ares Fernandez, J.R.; Aguey-Zinsou, F.; Elsaesser, M.; Ma, X.Z.; Dornheim, M.; Klassen, T.;

Barkhordarian, G.; Klassen, T.; Bormann, R. (2006). Kinetic investigation of the effect of

Barkhordarian, G.; Klassen, T.; Dornheim, M.; Bormann, R. (2007). Unexpected kinetic effect

*of Alloys and Compounds*, Vol. 440, No. 1-2, pp. (L18-L21), ISSN: 0925-8388 Barkhordarian, G.; Jensen, T.R.; Doppiu, S.; Bösenberg, U.; Borgschulte, A. ; Gremaud, R.;

Bösenberg, U.; Vainio, U.; Pranzas, P.K.; Bellosta von Colbe, J.M.; Goerigk, G.; Welter, E.;

*Chemistry C*, Vol. 112, No. 7, pp. (2743-2749), ISSN: 1932-7447

Bormann, R. (2007). Mechanical and thermal decomposition of LiAlH4 with metal halides*. International Journal of Hydrogen Energy*, Vol. 32, No. 8, pp. (1033-1040),

milling time on the hydrogen sortpion reaction of magnesium catalyzed with different Nb2O5 contents. *Journal of Alloys and Compounds*, Vol. 407, No. 1-2, pp.

of MgB2 in reactive hydride composites containing complex borohydrides. *Journal* 

Cerenius, Y.; Dornheim, M., Klassen, T.; Bormann, R. (2008). Formation of Ca(BH4)2 from Hydrogenation of CaH2+MgB2 Composite. *Journal of Physical* 

Dornheim, M.; Schreyer, A.; Bormann, R. (2009). On the chemical state and

has not been found.

ISSN: 0360-3199

(249-255), ISSN: 0925-8388

**5. References** 

Fig. 13. Schematic of the reaction mechanism in Reactive Hydride Composite.

In 2004 Vajo et al. (Vajo et al., 2005) , Cho et al. (Cho et al., 2006) and Barkhordarian et al. (Barkordarian et al., 2007) independently discovered that the usage of borides especially MgB2 as a starting material facilitates the formation of different borohydrides. This finding initiated the development and investigation of several new reversible systems with high storage capacities of 8 – 12 wt.% H2 and improved thermodynamic and kinetic properties such as 2 LiBH4+MgH2 (Bösenberg et al., 2009; 2010; 2010b), 2 NaBH4+MgH2 (Garroni et al., 2010; Pistidda et al., 2010; 2011; Pottmaier et al., 2011), Ca(BH4)2+MgH2 (Barkhordarian et al., 2008), 6 LiBH4+CeH2, 6 LiBH4+CaH2 (Jin et al., 2008b), LiBH4/Ca(BH4)2 (Lee et al., 2009) . One of the most intensely studied systems hereof is the 2 LiBH4 + MgH2 system. The indended reaction pathway is:

$$2\,\text{LiBH}\_4 + \text{MgH}\_2 \leftrightarrow 2\,\text{LiH} + \text{MgB}\_2 + 4\,\text{H}\_2\tag{31}$$

However, several other reaction pathways are possible leading to products such as LiB2, amorphous B, Li2B12H12 or Li2B10H10. Bösenberg et al. (Bösenberg et al., 2010b) could show that due to a higher thermodynamic driving force for the favoured reaction the competing reactions can be suppressed by applying a hydrogen back pressure and limiting the dehydrogenation temperature. Nevertheless, since long-range diffusion of metal atoms containing species is required, see Fig. 13, in bulk ball-milled samples dehydrogenation so far occurs only at temperatures higher than 350 °C, hydrogenation at temperatures higher than 250 °C.

The dehydrogenation temperatures of this Reactive Hydride Composite, however, can be significantly reduced by using nanoconfined 2 LiBH4 + MgH2 stabilised in inert nanoporous aerogel scaffold materials whereby long-range phase separation is hindered and thus the diffusion path length reduced (Gosalawit-Utke, 2011).
