**Mg-based hydrides**

902 Thermodynamics – Interaction Studies – Solids, Liquids and Gases

of Siebtechnik GmbH Eigen et al. (Eigen et al., 2007; Eigen et al., 2008) showed that upscaling of material synthesis is possible: After only 30 min milling under optimised process conditions in such a tube vibration mill in kg scale, fast absorption and desorption kinetics with charging/discharging times of less than 10 min can be obtained. The operation temperatures of this complex hydride are much lower than compared to MgH2 and other light weight hydrides. Fast kinetics is achieved at 100 °C to 150 °C which is much less than what is required in case of MgH2, however, still significantly higher than in case of the conventional hydrides which show only a very limited storage capacity. Such hydride working temperatures offer the possibility for combinations of metal hydride tanks based on these complex hydrides with e. g. combustion engines, high temperature PEM fuel cells or other medium to high temperature fuel cells. However, compared to MgH2 the gravimetric hydrogen storage capacity is significantly reduced. Having a maximum theoretical storage capacity of about 5.6 wt. % NaAlH4 exhibits a long term practical storage capacity of 3.5-4.5 wt. % H2 only. Furthermore, in difference to MgH2 NaAlH4 decomposes in two reaction steps upon dehydrogenation which implies two different pressure plateaus instead of just

 NaAlH4 1/3 Na3AlH6 + 2/3 Al + H2(g) NaH + Al +3/2 H2(g) (19) The first decomposition step has an equilibrium pressure of 0.1 MPa at 30 °C, the second step at about 100 °C (Schüth et al., 2004). A maximum of 3.7 wt.% H2 can be released during the first desorption step, 5.6 wt.% in total. The remaining hydrogen bonded to Na is

While the reaction kinetics was optimized significantly, the desorption enthalpy of NaAlH4 of 37 kJ/molH2 and Na3AlH6 of 47 kJ/mol H2 respectively remains a challenge. For many

While there are plenty of known hydrides with suitable thermodynamics for hydrogen uptake and release at ambient conditions (several bar equilibrium pressure at or nearby room temperature) currently no hydride is known which combines suitable thermodynamics and kinetics with such a high gravimetric storage capacity that a hydrogen storage tank based on such a material could compete with a 700 bar compressed composite vessel in regard to weight. Depending on the working temperature and pressure as well as the reversible gravimetric storage capacity of the selected hydride the achievable capacity of a metal hydride based storage tank is usually better than half of the capacity of the metal hydride bed itself (Buchner & Povel, 1982). Since modern composite pressurized gas tanks meanwhile show gravimetric hydrogen storage capacities of around 4 wt.% according to conservative extrapolations the possible choice of hydrides should be limited to those having the ability to reversibly store at least 6 wt.%H2. All currently known high capacity hydrides, however, show either too small values of the respective reaction enthalpy and are therefore not reversible or would require several thousand bar hydrogen pressure or alternatively electrochemical loading or on the other hand are too stable and have an equilibrium pressure which around room temperature is much below the required pressures. The value of reaction enthalpy aimed at is between 20 and 30 kJ/mol H2. Fig. 9 shows the potentially available hydrogen content of some well known hydrides plotted

technically not exploitable due to the high stability of the respective hydride.

applications even this value which is much below that of MgH2 is still too large.

**3. Tailoring thermodynamics of light weight metal hydrides** 

against their hydrogen reaction enthalpies.

one:

One of the first examples using this approach for tuning the thermodynamic properties of light weight metal hydrides was the discovery of the Mg-Ni –system as potential hydrogen storage system by Reilly and Wiswall (Reilly & Wiswall, 1968). Mg2Ni has a negative heat of

Thermodynamics of Metal Hydrides: Tailoring Reaction Enthalpies of Hydrogen Storage Materials 905

Unfortunately, with lowering the Mg content the hydrogen storage capacity dropped down

On the other side, as schematically shown in Fig. 10b the absolute value of reaction enthalpy can be increased by either stabilising the hydride phase or destabilising the dehydrogenated phase. In case of Mg-based hydrogen absorbing alloys this is not at all of interest for hydrogen storage itself since MgH2 is too stable for most hydrogen storage applications , however, this is of interest for other applications like the storage of thermal energy (Dornheim & Klassen, 2009). Mg2FeH6 is an example of such materials with increased amount of reaction enthalpy. Furthermore, it is the one with the highest known volumetric hydrogen density which amounts to 150 kg m-3. This enormously high hydrogen density is more than double the value found in case of liquid hydrogen at 20 K and moderate pressures of up to 20 bar (Klell, 2010). The gravimetric storage capacity is 5.6 wt.% and thus still rather high. Since Mg and Fe are immiscible the dehydrogenated state is destabilised compared to pure Mg: Hdd > 0 kJ/(mol H2). Accordingly the hydride phase is more difficult to be synthesised and reversibility as well as long term stability is more difficult to

Nevertheless, hydrogenation is possible at hydrogen pressures of at least 90 bar and temperatures of at least 450 °C (Selvam & Yvon, 1991). Bogdanovic et al. (Bogdanovic et al., 2002) achieved very good reversibility and cycling stability with the hydrogen storage capacities remaining unchanged throughout 550-600 cycles at a level of 5-5.2 wt.% H2. The reaction enthalpy value is reported to be in between 77 kJ/(mol H2) and 98 kJ/(mol H2) (Bogdanovic et al., 2002), (Konstanchuk et al, 1987), (Puszkiel et al., 2008), (Didisheim et al.,

The large reaction enthalpies of MgH2 and Mg2FeH6 lead to weight and volume related heat storage densities in the temperature range of 500 °C which are many times higher than that of the possible sensible or latent heat storage materials (Bogdanovic et al., 2002). The calculated and experimental heat storage densities to weight given by Bogdanovic et al. are 2814 kJ/kg and 2204 kJ/kg for the MgH2-Mg system and 2106 and 1921 kJ/kg for the Mg2FeH6 – 2Mg+Fe system respectively. The corresponding calculated and experimental values for the volumetric thermal energy storage density are 3996 kJ/dm³ and 1763 kJ/dm³ for the MgH2-Mg system and 5758 kJ/dm³ and 2344 kJ/dm³ respectively (Bogdanovic et al., 2002). These thermal energy densities ought not to be mistaken with the energy stored in the

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

> 4 36 <sup>2</sup> <sup>2</sup> ? 6 LiAlH 2Li AlH 4Al 6 H 6 LiH 6 Al 9 H . (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.

hydrogen (lower heating value) which is more than a factor of three larger.

**Aluminum-based complex hydrides** 

case of the NaAlH4:

to 1.4 wt.% only.

be accomplished.

1984).

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 stabilising the dehydrogenated phase by Hds or destabilising the hydride phase by Hhd. b) Increase of total reaction enthalpy by destabilising the dehydrogenated state by Hdd or stabilising the hydrogenated state by Hhs.

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 H0 f(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 enthalpy of Mg2Ni

$$\left| \Delta \mathbf{H}\_{\mathrm{Mg}\_2\mathrm{Ni-H}} \right| \, = \left| \Delta \mathbf{H}\_{\mathrm{Mg-H}} \right| + \Delta \mathbf{H}\_{\mathrm{ds}} - \Delta \mathbf{H}\_{\mathrm{hd}} \tag{20}$$

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 Mg2NiH4.

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 MgNi (Orimo et al., 1998).

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. Unfortunately, with lowering the Mg content the hydrogen storage capacity dropped down to 1.4 wt.% only.

On the other side, as schematically shown in Fig. 10b the absolute value of reaction enthalpy can be increased by either stabilising the hydride phase or destabilising the dehydrogenated phase. In case of Mg-based hydrogen absorbing alloys this is not at all of interest for hydrogen storage itself since MgH2 is too stable for most hydrogen storage applications , however, this is of interest for other applications like the storage of thermal energy (Dornheim & Klassen, 2009). Mg2FeH6 is an example of such materials with increased amount of reaction enthalpy. Furthermore, it is the one with the highest known volumetric hydrogen density which amounts to 150 kg m-3. This enormously high hydrogen density is more than double the value found in case of liquid hydrogen at 20 K and moderate pressures of up to 20 bar (Klell, 2010). The gravimetric storage capacity is 5.6 wt.% and thus still rather high. Since Mg and Fe are immiscible the dehydrogenated state is destabilised compared to pure Mg: Hdd > 0 kJ/(mol H2). Accordingly the hydride phase is more difficult to be synthesised and reversibility as well as long term stability is more difficult to be accomplished.

Nevertheless, hydrogenation is possible at hydrogen pressures of at least 90 bar and temperatures of at least 450 °C (Selvam & Yvon, 1991). Bogdanovic et al. (Bogdanovic et al., 2002) achieved very good reversibility and cycling stability with the hydrogen storage capacities remaining unchanged throughout 550-600 cycles at a level of 5-5.2 wt.% H2. The reaction enthalpy value is reported to be in between 77 kJ/(mol H2) and 98 kJ/(mol H2) (Bogdanovic et al., 2002), (Konstanchuk et al, 1987), (Puszkiel et al., 2008), (Didisheim et al., 1984).

The large reaction enthalpies of MgH2 and Mg2FeH6 lead to weight and volume related heat storage densities in the temperature range of 500 °C which are many times higher than that of the possible sensible or latent heat storage materials (Bogdanovic et al., 2002). The calculated and experimental heat storage densities to weight given by Bogdanovic et al. are 2814 kJ/kg and 2204 kJ/kg for the MgH2-Mg system and 2106 and 1921 kJ/kg for the Mg2FeH6 – 2Mg+Fe system respectively. The corresponding calculated and experimental values for the volumetric thermal energy storage density are 3996 kJ/dm³ and 1763 kJ/dm³ for the MgH2-Mg system and 5758 kJ/dm³ and 2344 kJ/dm³ respectively (Bogdanovic et al., 2002). These thermal energy densities ought not to be mistaken with the energy stored in the hydrogen (lower heating value) which is more than a factor of three larger.
