Temperature Programmed Desorption (TPD) measurements were done on several ball milled mixtures of MgH₂ with 4 mol% of Nb₂O₅, ZrO₂, CeO₂ and Fe₃O₄. The samples were ball milled for 10 h and their thermal desorption performance was studied to determine the temperature at which hydrogen was released from the sample mixtures. By so doing, we can understand the effect of each catalyst on the hydrogen desorption properties of MgH₂. It can be seen from the desorption curves shown in Figure 1 that MgH₂ has the highest onset temperature of about 310 °C and that the catalyzed samples have lower desorption temperatures. The mixture of MgH₂ + Fe₃O₄ has the lowest onset desorption temperature of about 200 °C. The onset temperature of the reacting mixtures are in the order: pure MgH₂ > CeO₂ > ZrO₂ > Nb₂O₅ > Fe₃O₄. The plot also revealed that all of the samples released less than 6 wt% of hydrogen. The reduction in hydrogen weight percentage is most likely due to the partial oxidation of the Mg in the alloy caused by the presence of oxide in all the transition metal oxide catalysts. These results confirm that the addition of transition metal oxide catalysts is effective in reducing the desorption temperature of MgH₂, although accompanied with a small weight penalty. The values of the onset temperatures are summarized in Table 1.

Pressure Composition Isotherms were constructed for MgH₂ and the catalyzed MgH₂ mixtures. Figure 2 shows the desorption isotherms for these samples at 400 °C. It can be seen from the curves that the plateau pressures are about the same for all the samples. The data from these isotherms were used to construct the Van’t Hoff plots shown in Figure 3. The reaction enthalpies for the mixtures were determined from the slopes of these plots and the values are summarized in Table 1. It is evident from the ΔH values that the thermodynamic stability of MgH₂ decreases with the addition of transition metal oxide catalysts.

In addition to having a low desorption temperature, it is also important that samples have fast reaction rates. Desorption kinetics experiments were carried out on each sample at 400 °C to determine the catalytic effect of transition metal oxides on hydrogen desorption rates from MgH₂. A novel concept of constant pressure thermodynamic driving force was used to achieve these desorption kinetics measurements. This was done by adjusting the hydrogen pressure in the reactor to a slightly higher value than that of the mid-plateau pressure (p_m), to ensure that only the hydrogen rich phase was present initially and sealing off the reactor. The pressure in the remaining system (p_op) was then adjusted to a value such that the ratio between the mid-plateau pressure and the opposing pressure (p_m/p_op) was a small whole number. This ratio is defined as the N-value. In these experiments, the N-value was set at 5 for all the sample mixtures. The theoretical basis for using constant pressure thermodynamic forces is that the pressure ratio is proportional to ΔG (the Gibbs free energy). The proportionality can be expressed as: ΔG ≈ RT ln(p_m/p_op). If the pressure ratio is the same then ΔG will be the same for all the reactions. Figure 4 contains plots of the reacted fraction versus the time for hydrogen desorption from the MgH₂ mixtures with and without catalysts. It can be seen that the uncatalyzed MgH₂ sample has the slowest hydrogen desorption rate. The addition of transition metal oxides improved the kinetics of MgH₂ with Nb₂O₅ and Fe₃O₄ having the fastest desorption reaction kinetics. The times required for all of these reactions to reach 90 percent completion (T₉₀) are summarized in Table 1.

Smith and Goudy [15], while performing kinetics modeling studies on LaNi_5−xCo_x hydride system (x = 0, 1, 2 and 3), reported two theoretical equations based on the shrinking core model that were successfully used to model the system. The equations are shown below:

    (1)

where

    (2)

where , t is the time at a specific point in the reaction, X_B is the fraction of the metal reacted. R is the initial radius of the hydride particles, b is a stoichiometric coefficient of the metal, C_Ag is the gas phase concentration of reactant, D_e is the effective diffusivity of hydrogen atoms in the hydride, ρ_B is the density of the metal hydride and k_s is a rate constant. It was found that a model based on Equation (1) will have chemical reaction at the phase boundary controlling the reaction rate. This is called the shrinking particle model (SPM). A model based on Equation (2) is one in which the overall reaction rate is controlled by diffusion. Both models were applied to the current study to determine which kinetic model best describes the reactions. Equations (1) and (2) were fitted to the kinetic data for each of the reaction sample mixtures. Figure 5, Figure 6, Figure 7, Figure 8, Figure 9 each contain three curves. One is an experimental curve taken from the desorption kinetics curve shown in Figure 4, a second curve was calculated from the SCM with diffusion controlling the overall reaction and a third curve was calculated with chemical reaction at the phase boundary controlling the rate. In order to determine the theoretical curves, it was first necessary to determine a value for τ. It was not necessary to know the values of all the physical parameters in Equations (1) and (2) in order to do this. The determination of τ was accomplished through a series of statistical data analyses in which the value of τ necessary to minimize the standard deviation between the experimental and theoretical data was calculated. Thus τ was a fitting parameter in these analyses As shown in Figure 5, Figure 6, Figure 7, Figure 8, Figure 9, data generated from the SPM with chemical reaction at the phase boundary controlling the overall rate fits the experimental data better than the data generated from the SCM with diffusion controlling the overall reaction rate. Therefore we can say that chemical reaction at the phase boundary is the most likely mechanism for all the reactions in this study.

To further understand the effects of the transition metal oxide catalysts on the dehydrogenation of MgH₂, the activation energy of dehydrogenation for the systems studied were investigated using an isoconversion method based on the Kissinger Equation stated below.

    (3)

where T_max is the temperature at maximum reaction rate, β the heating rate, E_a the activation energy, α the fraction of transformation, F_KAS(α) a function of the fraction of transformation, and R is the gas constant. Figure 10 shows the DTA curves for one of the samples (MgH₂ + 4 mol% Nb₂O₅) at different heating rates from 1 to 15 °C per minute. The figure shows that the endothermic peak corresponding to the maximum rate of dehydrogenation shifts to higher temperatures as the heating rate is increased. The same trend was observed in all the other samples. The plots based on the Kissinger equation are shown in Figure 11. A good linear relationship between ln (β/T²_max) versus 1/T_max existed for all the samples. Activation energies of dehydrogenation were calculated from the slopes of the straight lines. The calculated activation energies are summarized in Table 1. From the values it is clear that the addition of transition metal oxide catalysts to MgH₂ helped lower its activation energy. The calculated activation energies correlate well with the times required for 90% of the hydrogen to desorb from the samples (T₉₀). Lower activation energies correspond to faster desorption kinetics. There was also a slight correlation between the activation energy and desorption temperatures as well. Samples with high activation energies tended to have high thermal stabilities.