Enhanced hydrogen storage performance for MgH₂–NaAlH₄ system—the effects of stoichiometry and Nb₂O₅ nanoparticles on cycling behaviour

Abstract

Nowadays, the technological utilization of reactive hydride composites (RHC) as promising hydrogen storage materials is hampered by their reaction kinetics. In the present work, effects of reactant stoichiometry on ensuing hydrogen sorption properties and pathway of the MgH₂–NaAlH₄ (mole ratios 1 : 2, 1 : 1 and 2 : 1) system, both undoped and doped with Nb₂O₅ nanoparticles, were investigated. It was found that the as-prepared reactant stoichiometry of MgH₂/NaAlH₄ system had a profound impact on its dehydrogenation kinetics and reaction mechanism. Variable temperature dehydrogenation data revealed that undoped binary composites possessed enhanced hydrogen desorption properties compared to that of pristine NaAlH₄ and MgH₂. The use of Nb₂O₅ displayed superior catalytic effects in terms of enhancing dehydriding/rehydriding kinetics and reducing the dehydrogenation temperature of MgH₂–NaAlH₄ system. Isothermal volumetric measurements at 300 °C revealed that enhancements arising upon adding Nb₂O₅ were almost double that of undoped MgH₂–NaAlH₄ composites. The apparent activation energies for NaAlH₄, Na₃AlH₆, MgH₂, and NaH relevant decompositions in doped composite were found to be much lower than that for the undoped one. Moreover, Nb₂O₅ doping also markedly enhanced the reversible capacity of MgH₂–NaAlH₄ composites under moderate conditions, persisting well during three de/rehydrogenation cycles. XRD, XPS, and FESEM-EDS analyses demonstrated that reduction of Nb₂O₅ during first desorption was coupled to the migration of reduced niobium oxide species from the bulk to the surface of the material. It was suggested that these finely dispersed oxygen-deficient niobium species might contribute to kinetic improvement by serving as the active sites to facilitate hydrogen diffusion through the diffusion barriers both during dehydrogenation and rehydrogenation.

---

Introduction

The development of new practical hydrogen storage materials with high volumetric and gravimetric hydrogen densities is a key technical challenge to implement the fuel cell technology for transportation applications. According to the U.S. DOE's 2015 targets, a viable hydrogen storage system for fuel cell applications has necessitated a system gravimetric capacity of more than 5.5 wt% with fast desorption kinetics (1.5 wt% min⁻¹). Solid-state hydrogen storage is attractive because it offers a volumetric hydrogen density greater than that of either compressed gas or liquid hydrogen storage, without high pressure containment or cryogenic tanks.^1–5 During the past decade, MgH₂^6–8 and light-element complex hydrides such as alanates,^9–14 amides,^15,16 borohydrides,^17–19 and ammonia borane^20,21 are extensively investigated due to their higher gravimetric and volumetric hydrogen densities. Among the solid-state hydrogen storage materials, magnesium hydride is a good candidate possessing large gravimetric density (7.6 wt% H₂), abundant resources, low cost, and superior reversibility compared to that of conventional transition metal-based hydrides. However, its use is unfortunately hampered by its high thermodynamic stability and its slow kinetics. During the recent years, these disadvantages have been overcome by preparing nanocrystalline MgH₂ produced by high energy milling,^22,23 or doping with various catalysts such as metals,^24–26 metal oxides,^27–29 and metal halides,^30,31etc. Nonetheless, all these efforts have only ameliorated its kinetic performances. The thermodynamic characteristics of the interaction between Mg and hydrogen are hardly modified. During the past decade, another pioneer work in the hydrogen storage field is the use of a mixture of single and complex hydrides with amides based systems.^32–34 Later this approach has been extended to borohydrides systems,^35–37 which are known as Reactive Hydride Composites (RHCs). This approach is aimed at modifying the thermodynamics and kinetics of the hydrogen sorption reaction based on mutual interactions between different hydrides by reducing the enthalpy of dehydrogenation and rehydrogenation reactions. Recent studies have demonstrated that it is possible to reduce the activation energy and decomposition enthalpy of magnesium hydride by mixing it with a second hydride specie, such as LiAlH₄, indicating an effective way to improve the hydrogen storage properties of LiAlH₄ and MgH₂.^38–40 Analogous to LiAlH₄, NaAlH₄ has also been used to modify the de/rehydrogenation thermodynamics of MgH₂, albeit up to now there is limited investigation on sorption reactions of this RHC system by experimental means. Recently, Ismail et al.⁴¹ have investigated the kinetic sorption properties and reaction pathway of a binary MgH₂–NaAlH₄ mixture with 4 : 1 molar stoichiometry prepared by ball milling. Unlike the results obtained for the mixture LiAlH₄/MgH₂, where the destabilization could be caused by the formation of either Al₁₂Mg₁₇ or Li_xMg_y alloys, only the intermetallic Al₁₂Mg₁₇ is ascribed as a destabilizing agent in the system NaAlH₄/MgH₂. The present work intends to contribute to the clarification of the sorption properties, more precisely, indicating the most probable dehydrogenation pathway for each of the various stoichiometries of this system. It is important to properly understand the effect of stoichiometry on such multicomponent system as well as the effect of the dehydrogenation conditions on the reaction paths and sorption properties of the system. Moreover, in our recent studies on LiAlH₄,¹⁰ we have found that Nb₂O₅ nanoparticles are superior to its analogue Cr₂O₃ in enhancing the kinetic and thermodynamic performance of LiAlH₄. Previous studies^29,42 have also indicated that the additive Nb₂O₅ is known to substantially improve the decomposition thermodynamics and kinetics of MgH₂. Motivated by these experimental findings, we have further extended the usage of Nb₂O₅ nanoparticles in the MgH₂–NaAlH₄ system. The addition of Nb₂O₅ is found to result in significantly improved de/rehydrogenation kinetics in the MgH₂–NaAlH₄ system.

To clarify the impact of reactant stoichiometry on the properties of the undoped and doped MgH₂–NaAlH₄ composites, herein we report the synthesis, characterization, hydrogen storage characteristics, and hydrogen desorption pathway for composites with stoichiometric ratios of 1 : 2, 1 : 1, and 2 : 1. Thermogravimetry (TG), differential scanning calorimetry (DSC), and isothermal sorption measurements have been conducted to investigate its thermodynamic and kinetic behavior. On the basis of XRD, XPS, and FESEM-EDS analyses, we suggest a set of consecutive reactions that is consistent with the observed phase evolution and hydrogen desorption behavior. On the basis of these data, we discuss factors that may explain differences between these three sets of related compositions. Furthermore, we have performed a full analysis of the surface and bulk regions in order to investigate the evolution, oxidation state, and the local structure of the Nb species during the different steps of milling, dehydrogenation, and rehydrogenation.

Experimental section

NaAlH₄ (hydrogen storage grade, ≥93% purity) and MgH₂ (hydrogen storage grade) were purchased from Sigma–Aldrich Co. The high purity nano-oxide Nb₂O₅ was provided by SINONANO Co., Ltd (China). All the materials were used as received without any further purification. All material handling (including weighing and loading) was performed in a high purity argon filled glove box, with low oxygen and water vapour content. A hydride mixture was prepared by mixing the as-received MgH₂ and NaAlH₄ together in the mole ratios of 1 : 2, 1 : 1, and 2 : 1. Subsequently, the mixtures were ball milled for 30 min by a high energy Spex mill. All the samples were loaded into the hardened steel vial under an argon atmosphere in a glove box. Steel balls (1 g and 3 g) were added with a ball to powder weight ratio of 15 : 1. Air-cooling of the vial was employed to prevent its heating during the ball-milling process. 2 mol% of Nb₂O₅ nanopowders was also ball milled with 1MgH₂–2NaAlH₄, 1MgH₂–1NaAlH₄, and 2MgH₂–1NaAlH₄ under the same conditions to investigate their catalytic effects. MgH₂ and NaAlH₄ doped with 2 mol% of Nb₂O₅were also prepared under the same conditions for comparison purposes.

Non-isothermal dehydrogenation performances were investigated by thermogravimetry (TG) and differential scanning calorimetry (DSC). The DSC and TG analyses were conducted using NETZSCH STA 449C. All measurements were carried out under a flow (50 ml min⁻¹) of high purity argon (99.999%). Sample mass was typically 5 mg. Heating runs were performed at different rates (4, 7 and 10 °C min⁻¹) from 35 to 500 °C.

The isothermal de/re-hydrogenation kinetics were measured using a pressure composition–temperature (PCT) apparatus. The details of the apparatus are given in our previous reports.^9,10,43 The apparatus can be operated up to a maximum pressure of 10 MPa and 600 °C. About 0.5 g of sample was loaded into the sample vessel. The isothermal dehydrogenation measurements for the undoped and doped samples were performed at 280 °C and 300 °C under a controlled vacuum atmosphere. Following the first complete dehydrogenation, the samples were subjected to rehydrogenation studies at 300 °C under 9.5 MPa for 1 h. The pressure drop with time in the closed system testified the rehydrogenation of the samples. Subsequently, the rehydrogenated samples were dehydrogenated at similar temperature.

The phase structure of the sample following the ball milling, dehydrogenation, and rehydrogenation was determined by a MXP21VAHF X-ray diffractometer (XRD with Cu-Kα radiation) at room temperature. XRD was done at a tube voltage of 40 kV and a tube current of 200 mA. The samples were covered with paraffin film to prevent the oxidation during the XRD test. X-ray photoelectron spectroscopy (XPS) was performed with a PHI-5300 XPS spectrometer.

The as-received, doped, dehydrogenated, and rehydrogenated samples were examined by a field emission scanning electron microscope (FESEM-6301F) coupled with energy dispersive spectroscopy (EDS). Sample preparation for the FESEM measurement was carried out inside the glove box and, moreover, the samples were transferred to the SEM chamber by means of a device maintaining an Ar overpressure.

Results and discussion

The TG profiles in Fig. 1 depict the non-isothermal dehydrogenation performance of the milled MgH₂–NaAlH₄ composites with different mole ratios (1 : 2, 1 : 1 and 2 : 1) undoped and doped with 2 mol% Nb₂O₅. The manometric desorption profiles recorded on the unary components NaAlH₄ and MgH₂, with and without 2 mol% Nb₂O₅, are also reported in Fig. 1 for comparison. In the case of pristine NaAlH₄, the hydrogen desorption starts at 178 °C. After heating to 450 °C, this material has released a total hydrogen capacity of 6.3 wt%. For the bare MgH₂, the hydrogen release starts at 345 °C, and the weight loss is about 7.2 wt% after heating to 450 °C. It is evident that both the undoped and doped 1MgH₂–2NaAlH₄ and 1MgH₂–1NaAlH₄ systems display three significant stages of dehydrogenation occurring during the heating process. On the contrary, the undoped 2MgH₂–1NaAlH₄ composite has exhibited four-stage hydrogen release. The relative proportion of hydrogen evolved during these three or four stages is of course dependent upon the ratio of the as prepared samples. The first stage for all the undoped and doped composite samples seems to be the two-step decomposition of NaAlH₄ as indicated in eqn (1) and (2), because both the dehydrogenation temperatures and the hydrogen liberation amounts correspond to the decomposition of NaAlH₄.

3NaAlH₄ → Na₃AlH₆ + 2Al + 3H₂ (1)

Na₃AlH₆ → 3NaH + Al + 3/2H₂ (2)

Fig. 1 Comparison of the TG profiles of neat MgH₂ and NaAlH₄, and MgH₂–NaAlH₄ composites (in mole ratios of 1 : 2, 1 : 1 and 2 : 1), (i) without Nb₂O₅ nanoparticles and (ii) with Nb₂O₅ nanoparticles. The ramping rate is 4 °C min⁻¹.

Although the first step dehydrogenation temperature (175 °C and 170 °C) in the MgH₂–NaAlH₄ (molar ratios, 1 : 2 and 1 : 1) samples has not exhibited the significant reduction compared to that of the pristine NaAlH₄ sample, faster dehydrogenation kinetics have been observed for both samples. The first stage of the decomposition of 1MgH₂–2NaAlH₄ and 1MgH₂–1NaAlH₄ samples terminates at 255 and 248 °C, respectively, which are 45 °C and 52 °C lower than that for the pure NaAlH₄ sample. The 1MgH₂–2NaAlH₄ composite starts to decompose at 278 °C and terminates at 326 °C for the second stage, whereas the 1MgH₂–1NaAlH₄ sample starts the second decomposition at 275 °C and concludes at 335 °C. Obviously, The MgH₂ decomposition temperature in 1MgH₂–1NaAlH₄ composite has reduced by 70 °C compared to that of the pristine MgH₂ sample. Further heating of the 1MgH₂–1NaAlH₄ sample leads to the third decomposition, starting at about 360 °C, corresponding to the decomposition of NaH according to eqn (3), which also occurs at 50 °C lower than that of the pure NaAlH₄ sample.

NaH → Na + 1/2H₂ (3)

For the undoped 2MgH₂–1NaAlH₄ system, the first and second steps (Fig. 1(i)) occur in the temperature ranges of 152–225 °C and 240–300 °C while the onset temperatures of the third and fourth desorption steps are at 320 °C and 360 °C, respectively. The lowered dehydrogenation temperatures of 2MgH₂–1NaAlH₄ system compared to that for 1MgH₂–2NaAlH₄ and 1MgH₂–1NaAlH₄ systems suggest that higher amount of MgH₂ enhances the decomposition of NaAlH₄. After heating to 390 °C the total desorption capacity from the 2MgH₂–1NaAlH₄ mixture is about 6.9 wt%, which is higher than that of both MgH₂ and NaAlH₄ alone at the same temperature. These results indicate that the mixing between NaAlH₄ and MgH₂ can decrease the onset desorption temperature compared to that of NaAlH₄ (178 °C) and MgH₂ (345 °C). Therefore, the dehydrogenation properties of the MgH₂–NaAlH₄ system are improved compared to those of its unary components (NaAlH₄ and MgH₂), suggesting that a mutual destabilization has occurred in the binary MgH₂–NaAlH₄ system. Obviously, a greater reduction in the decomposition temperatures has occurred for the composites with higher concentrations of MgH₂. This kinetic enhancement in the undoped 2 : 1 system can be partially ascribed to the more refinement of powders in these composites compared to that in other mixtures (Fig. 14).

In order to further clarify the mechanistic effects, three undoped binary composites have been also heated at much lower heating rate of 1 °C min⁻¹. The results are displayed in Fig. S1 of the ESI.† Evidently, all the decomposition steps still exist, albeit the dehydrogenation temperatures have been further reduced.

Fig. 2 and 3 present the evolution of the XRD patterns of the 1MgH₂–1NaAlH₄ and 2MgH₂–1NaAlH₄ samples upon ball-milling and heating to different temperatures, successively. The as-milled samples are detected as the physical mixtures of MgH₂ and NaAlH₄. Fig. 2(c) and 3(b) characterize the NaH, Al and MgH₂ phases in the XRD patterns of the 1MgH₂–1NaAlH₄ and 2MgH₂–1NaAlH₄ systems after dehydrogenation at 250 °C and 230 °C, respectively. It excludes the possibility of MgH₂-relevant reactions, demonstrating that only self-decomposition of NaAlH₄ dominates this stage. For the 1 : 1 composite, after dehydrogenation at 350 °C, the Al phase has disappeared, and the peaks pertaining to Al₁₂Mg₁₇ and Mg are observed clearly from the XRD pattern in Fig. 2(d). These results indicate that the hydrogen release in the temperature range of 250–350 °C is mainly from the reaction of Al with MgH₂ and the decomposition of MgH₂ in accordance with the following reactions:

12Al + 17MgH₂ → Al₁₂Mg₁₇ + 17H₂ (4)

MgH₂ → Mg + H₂ (5)

Fig. 2 XRD patterns for the MgH₂–NaAlH₄ (1 : 1) composite at different states: (a) before dehydrogenation; (b) after dehydrogenation at 225 °C; (c) after dehydrogenation at 250 °C; (d) after dehydrogenation at 350 °C; (e) after dehydrogenation at 400 °C.

Fig. 3 XRD patterns for the MgH₂–NaAlH₄ (2 : 1) composite at different states: (a) before dehydrogenation; (b) after dehydrogenation at 230 °C; (c) after dehydrogenation at 310 °C; (d) after dehydrogenation at 360 °C; (e) after dehydrogenation at 390 °C.

After further heating to 400 °C, the NaH phase has disappeared with the evolution of Na phase, suggesting that the hydrogen release in the temperature range of 350–400 °C is due to the decomposition of NaH as indicated in eqn(3). These results demonstrate that the three stages of dehydrogenation in the 1MgH₂–1NaAlH₄ composite correspond to the sequential decomposition of the NaAlH₄, MgH₂ and NaH phases (NaH resulting from the decomposition of NaAlH₄).

The XRD pattern of 2MgH₂–1NaAlH₄ sample, heated to 310 C, shows the formation of the NaMgH₃ phase together with Al₁₂Mg₁₇ in the temperature range of 240–300 °C (Fig.3c). The presence of NaMgH₃, in agreement with the TG analysis of Fig. 1 (the third decomposition at about 320 °C), strongly suggests that apart from the reaction between Al and MgH₂, a partial reaction between Mg/MgH₂ and NaH has also taken place in the temperature range of 240–300 °C. The formation of NaMgH₃ during the 2MgH₂–1NaAlH₄ decomposition reaction demonstrates that the desorption reaction occurs via the formation of an intermediate phase. On the contrary, the 1MgH₂–1NaAlH₄ sample has not exhibited the formation of NaMgH₃. This corroborates that NaMgH₃ can be formed while the ratio of MgH₂/NaAH₄ is relatively high. Otherwise, NaH and Mg are formed preferentially without the formation of NaMgH₃. Ismail et al.⁴¹ have also identified the formation of NaMgH₃ for 4MgH₂–1NaAlH₄ composite. XRD measurements of 2MgH₂–1NaAlH₄ composite at 360 °C indicate that the decomposition of NaMgH₃ has accomplished below that temperature according to the following reaction.

NaMgH₃ → NaH + Mg + H₂ (6)

Moreover, the XRD pattern in Fig. 3(e) confirms that the hydrogen release in the temperature range of 360–390 °C is induced by the decomposition of NaH. These experimental results, described above, point out the fact that various stoichiometric ratios of the MgH₂/NaAlH₄ system can lead to significant variations of the reaction mechanism and the sorption properties, even if the same final decomposition products (Al₁₂Mg₁₇, Mg and Na) are obtained. It can be surmised that the more MgH₂-rich compositions facilitate the favorable formation of NaMgH₃ phase. It has already been hypothesized that the formation of a mixed compound such as NaMgH₃ is strongly associated with the microstructure of the starting materials, which should contain interfaces between the reacting phases.³⁶ It is believed that higher concentrations of MgH₂ favour the formation of NaMgH₃ by creating large amounts of such interfaces. Moreover, the binary NaAlH₄–MgH₂ systems displayed superior hydrogen storage properties compared to that of the unary components NaAlH₄ and MgH₂. This phenomenon may be attributed to the mutual interactions among the two hydrides.

The desorption curves in Fig.1(ii) clearly depict that the use of nanometric Nb₂O₅ additions has rendered quite striking effects not only on the dehydrogenation characteristics of the binary composites but also on the unary components. In order to comprehend the catalytic role of Nb₂O₅ on the binary system, the samples of NaAlH₄ and MgH₂ doped with 2 mol% Nb₂O₅ nanopowders are prepared and investigated in comparison with the undoped binary samples as well as the undoped and doped unary components. It can be seen that nanosized Nb₂O₅ is effective in ameliorating the dehydrogenation properties of the unary components NaAlH₄ and MgH₂ as well. In the case of the Nb₂O₅-doped NaAlH₄ and MgH₂, the onset dehydrogenation temperatures are decreased to 100 °C and 275 °C, which are 78 °C and 70 °C lower than that of pure NaAlH₄ (178 °C) and MgH₂ (345 °C) samples. In addition, the dehydrogenation temperatures of the NaAlH₄ and MgH₂ in all the MgH₂–NaAlH₄ systems are also lower than that of the pure and doped NaAlH₄ and MgH₂ samples. The onset temperatures of hydrogen desorption for the milled 2 mol% Nb₂O₅ added MgH₂–NaAlH₄ with molar ratios of 2 : 1, 1 : 1, and 1 : 2 composites appear around 70, 80, and 95 °C, respectively, which are 82, 90, and 80 °C lower than the corresponding decomposition temperatures of the milled MgH₂–NaAlH₄ composite without Nb₂O₅ addition. Further heating leads to the decomposition of MgH₂ near 213, 220 and 235 °C (Fig. 1(ii)), which are also lower than the corresponding decomposition temperatures at 240, 275 and 278 °C in Fig. 1(i). In particular for the doped samples with molar ratios of 1 : 1 and 2 : 1, the dehydrogenation temperature ranges have lowered to 80–355 °C and 70–350 °C due to the additive Nb₂O₅. Obviously, the addition of Nb₂O₅ nanoparticles plays a key role in decreasing the decomposition temperature of the MgH₂–NaAlH₄ system. Moreover, the nanometric Nb₂O₅-doped MgH₂–NaAlH₄ samples have shown faster desorption rates compared to undoped ones during the heating process. For example, the hydrogen desorption capacities of 6.6, 6.4 and 5.9 wt% are reached on heating the doped MgH₂–NaAlH₄ (molar ratios of 2 : 1, 1 : 1 and 1 : 2) samples to 350 °C. In contrast, the MgH₂–NaAlH₄ composites (molar ratios of 2 : 1, 1 : 1 and 1 : 2) without additive have exhibited hydrogen release capacities of 5.7, 5.8 and 5.6 wt% hydrogen, respectively, by 350 °C. In addition, the total hydrogen release contents from 2MgH₂–1NaAlH₄ doped samples are around 6.6 wt%. It indicates that a significant reduction in the decomposition temperature arising from the addition of a Nb₂O₅ catalyst is achieved without much penalty in the practical capacity of the materials. The above results clearly indicate that the dehydrogenation properties of the NaAlH₄–MgH₂ composites have been enhanced by the addition of Nb₂O₅ nanopowders.

With the aim of structurally elucidating the catalytic mechanism of Nb₂O₅ nanoparticles, XRD measurements of the doped composites, before and after being subjected to the dehydrogenation at different temperatures, are displayed in Fig. 4. It is clear that the Nb₂O₅ phase can be detected in the XRD pattern of the as milled materials. It suggests that the Nb₂O₅ nanocrystalline particles remain stable with the composite matrix during ball-milling under the high-energy impact mode. Although, the dehydrogenation products of the doped 1 : 1 composite are analogous to that observed for undoped one, nevertheless, some by-products due to the decomposition/reaction of the dopant can be revealed in the discharged mixture of the doped sample. Fig. 4a clearly depicts that the doped 1 : 1 sample, at the beginning, consists mainly of Nb₂O₅ and changes significantly during the heating and desorption process. It is evident that the heating of the sample up to 200 °C has induced the disappearance of the crystalline Nb₂O₅, coupled to a parallel growth of the newly reduced niobium species with different oxidation states similar to NbO₂ (+4) and NbO/NbH (+2, +1). During further heating, the Mg liberated from the hydrogen induces a fast reduction of the niobium oxide due to its very low redox potential, leading to the evolution of new peaks corresponding to pure MgO. Obviously, the reaction between the reduced niobium oxide species and the Mg/MgH₂ has also yielded the formation of a ternary oxide phase MgNb₂O_3.67. It has been previously documented⁴² that NbO and MgO exhibit similar crystal structure. Moreover, both have similar lattice parameters with comparable bond lengths between metal and oxygen atoms. Therefore, a reaction between the niobium oxide and the Mg/MgH₂ with a formation of a ternary oxide Mg_xNb_yO is possible. Hence, the evolution of the additive to the real active species has taken place in the form of mixed MgNb₂O_3.67 phase. In the literature, the formation of ternary Mg_xNb_yO has already been reported for the MgH₂ doped with Nb₂O₅. It has been documented that these ternary Mg–Nb oxides facilitate the hydrogen diffusion by acting as pathways by the formation of metastable niobium hydrides.⁴² When the desorption process is finished, a steady state is reached consisting mainly of phases with the oxidation states rich in +1(NbH) and +2(NbO).

Fig. 4 XRD patterns for the Nb₂O₅-doped MgH₂–NaAlH₄ (1 : 1) composite at different states: (a) before dehydrogenation; (b) after dehydrogenation at 200 °C; (c) after dehydrogenation at 300 °C; (d) after dehydrogenation at 350 °C; (e) XRD pattern of the Nb₂O₅-doped 2MgH₂–1NaAlH₄ composite upon heating to 300 °C.

Fig. 4e presents the evolution of the XRD pattern of the Nb₂O₅-doped 2MgH₂–1NaAlH₄ composite upon heating to 300 °C. Interestingly, no diffraction peaks corresponding to the NaMgH₃ phase have been detected for the 2 : 1 doped mixtures (in the sensitivity limits of the XRD), suggesting that the presence of Nb₂O₅ alters the desorption pathway by inhibiting the formation of NaMgH₃. These results coincide well with those obtained in Fig. 1, indicating that the four-stage decomposition for the undoped 2MgH₂–1NaAlH₄ sample has transformed to three- stage owing to the additive Nb₂O₅. Moreover, these results are also in contrast with that reported by Milanese et al.,⁴⁴ pointing to the fact that the dopants hamper the formation of NaMgH₃ phase.

The thermal decomposition behavior of the undoped and doped MgH₂–NaAlH₄ (molar ratios of 2 : 1, 1 : 1 and 1 : 2) composites, as well as of undoped unary components NaAlH₄ and MgH₂, is further investigated by DSC, as presented in Fig. 5, 6, and 7. Fig. 5 illustrates the DSC results of the as-received MgH₂ and alanate samples at various heating rates (4 °C min⁻¹, 7 °C min⁻¹ and 10 °C min⁻¹, respectively). the DSC results of Nb₂O₅-doped NaAlH₄ at various heating rates (4 °C min⁻¹, 7 °C min⁻¹ and 10 °C min⁻¹, respectively) have been reported in Fig. S2 of the ESI.† The calorimetric profiles, recorded on the binary 1 : 1, 1 : 2 and 2 : 1 mixtures, are reported in Fig. 6. Five endothermic events in the 1 : 1 and 1 : 2 mixtures are assigned to the melting of NaAlH₄, the decomposition of molten NaAlH₄ to Na₃AlH₆, the decomposition of Na₃AlH₆ to NaH and Al, the decomposition of MgH₂, and the decomposition of NaH, respectively, which is comparable with the three-step dehydrogenation in the TG results. Concomitant with the aforementioned XRD and TG results, the first four endothermic processes in 2 : 1 sample are similar to that in other composites, while the fifth and sixth events are attributed to the decompositions of NaMgH₃ and NaH, respectively. The resulting peak temperatures, measured in Fig. 6(i), are small compared to that of pure NaAlH₄ and MgH₂ samples. For instance, the peak temperatures for 2 : 1 composite for the first and second dehydrogenation steps of NaAlH₄ are 191 °C and 213 °C, respectively, which are 49 °C and 58 °C lower than those of the bare alanate sample. It is also evident that the peak temperature for pure MgH₂ is 406 °C, which is 97 °C higher than that for the MgH₂-relevant decomposition in 1MgH₂–1NaAlH₄ sample. These results further testify the mutual destabilization between NaAlH₄ and MgH₂. In contrast with the calorimetric profiles of the undoped MgH₂/NaAlH₄ composites, the features of all the doped samples are strikingly different, displaying only four endothermic peaks [Fig. 7]. The results indicate that NaAlH₄ decomposes at a much lower temperatures without melting with Nb₂O₅ catalysis. Obviously, the addition of Nb₂O₅ nanopowders has rendered a significant reduction in the decomposition temperatures associated with the MgH₂–NaAlH₄ system.

Fig. 5 The DSC profiles at various heating rates (4 °C min⁻¹, 7 °C min⁻¹ and 10 °C min⁻¹) for pristine (i) MgH₂ and (ii) NaAlH₄.

Fig. 6 The DSC profiles for (i) MgH₂–NaAlH₄ composites (in mole ratio of 1 : 2, 1 : 1 and 2 : 1) at heating rate of 4 °C min⁻¹ and (ii) MgH₂–NaAlH₄ composite (in mole ratio of 1 : 1) at various heating rates (4 °C min⁻¹, 7 °C min⁻¹ and 10 °C min⁻¹).

Fig. 7 The DSC profiles for (i) Nb₂O₅-doped MgH₂–NaAlH₄ composites (in mole ratios of 1 : 2, 1 : 1 and 2 : 1) at heating rate of 4 °C min⁻¹ and (ii) Nb₂O₅-doped MgH₂–NaAlH₄ composite (in mole ratio of 1 : 1) at various heating rates (4 °C min⁻¹, 7 °C min⁻¹ and 10 °C min⁻¹).

To acquire detailed information about the kinetics of the reactions, the apparent activation energy related to the NaAlH₄, Na₃AlH₄, MgH₂, and NaH decompositions in both the undoped and doped 1MgH₂–1NaAlH₄ composites, are calculated by using the nonisothermal Kissinger method⁴⁵ which is based on the shift in peak temperatures with the heating rates of 4, 7, and 10 °C min⁻¹. The activation energies of pristine NaAlH₄ and MgH₂, and Nb₂O₅-doped NaAlH₄, for hydrogen desorption, are also reported. The derived values of the activation energies are listed in Tables 1 and 2. For comparison, the activation energies of NaAlH₄ and MgH₂, undoped and doped with other additives, are also recapitulated in Tables 1 and 2.^{6,25–27,46–49,50–55} The calculated apparent activation energies for the first, second, and third decomposition steps for the pure alanate are 116 kJ mol⁻¹, 149 kJ mol⁻¹, and 180 kJ mol⁻¹, respectively. It is evident in Table 1 that the activation energy values for the first and second stages obtained in our work, agrees very well with that reported by Fan et al.⁴⁷ The apparent activation energy calculated is about 153 kJ mol⁻¹ for pure MgH₂, which is consistent with the values of 150, 156, 158, and 160 kJ mol⁻¹ H₂ for un-milled MgH₂ in Table 2. With the mixing of two hydrides, the activation energy has lowered to 109, 124, 120, and 164 kJ mol⁻¹ for NaAlH₄, Na₃AlH₆, MgH₂, and NaH relevant decompositions, respectively, which exhibits an enhancement in kinetics. This result clearly indicates that the mixing of two hydrides reduces the energy barriers for all the four stages but is more efficient for the last three stages, i.e., the dehydrogenation of Na₃AlH₆, MgH₂, and NaH, respectively. For the uncatalyzed MgH₂–NaAlH₄ composite, the reduction of activation energy may be attributed to the particle size reduction and the occurrence of mutual interactions between two species. The derived apparent activation energies corresponding to NaAlH₄, Na₃AlH₆, MgH₂, and NaH decompositions for the MgH₂–NaAlH₄ composite, catalyzed with Nb₂O₅ nanoparticles, are 68, 81, 75, and 130 KJ mol⁻¹, respectively. It implies that the activation energies of four stages are reduced by around 41, 43, 44, and 34 kJ mol⁻¹, compared with the activation energies associated with the undoped 1MgH₂–1NaAlH₄ sample, respectively. Table 1 and 2 clearly demonstrate that these values are slightly higher than those reported for TiO₂-doped NaAlH₄ and MgH₂ doped with TiH₂, but smaller than that reported for NaAlH₄ and MgH₂ doped with all other additives. Thereby, it is surmised that the presence of the Nb₂O₅ nanoparticles has essentially rendered the diminution of activation energy barrier for the dehydrogenation of the MgH₂–NaAlH₄ samples.

Table 1 Comparison of Activation energies (KJ mole⁻¹) of pure and doped NaAlH₄

Sample	Determination method	NaAlH4	NaAlH3	References
Pristine NaAlH4	Kissinger Method	114.2,132	156.8,268	47,46
Pristine NaAlH4	Isothermal	118.1	120.7	48
Pristine NaAlH4	Kissinger Method	116	149	Present work
NaAlH4 + MgH2	Kissinger Method	109	124	Present work
NaAlH4 + MgH2 + Nb2O5	Kissinger Method	68	81	Present work
Nb2O5-doped	Kissinger Method	65	86	Present work
CeCl3-doped	Kissinger Method	80.8	97.2	47
CeAl4-doped	Kissinger Method	80.9	98.9	47
TiCl3-doped	Isothermal	80	97.5	48
Ti-doped	Kissinger Method	77		49
TiO2-doped	Kissinger Method	67		49
Ti/Til3•6THF-doped	Kissinger Method	96.2	197.2	50
Ti/TiCl3-doped	Kissinger Method	139.5		50
CeAl-doped	Kissinger Method	72.3	98.9	51
LaCl3-doped	Kissinger Method	86.4	96.1	52
La3Al11-doped	Kissinger Method	92.9	99.2	52
SmCl3-doped	Kissinger Method	89	96.7	52
SmAl3-doped	Kissinger Method	91.9	98.9	52
TiF3-doped	Kissinger Method	98	130	46
SiO2-doped	Kissinger Method	127	138	46
TiF3 + SiO2-doped	Kissinger Method	99	122	46

---

Table 2 Comparison of Activation energies (KJ mole⁻¹) of pure and doped MgH₂

Sample	Determination method	MgH2	References
Pristine MgH2	Kissinger Method	120,135,150,191,191.3	53,6,27,26,54
Pristine MgH2	Isothermal	158.5,156,160	25
Pristine MgH2	Kissinger Method	153.5	Present work
NaAlH4 + MgH2	Kissinger Method	120	Present work
NaAlH4 + MgH2 + Nb2O5	Kissinger Method	75	Present work
TiC-doped	Kissinger Method	144.6	54
NbF5-doped	Kissinger Method	90,88	53
Cr2O3-doped	Kissinger Method	84	27
Fe2O3-doped	Kissinger Method	124	27
Ni-doped	Kissinger Method	81	26
TiO2-doped	Kissinger Method	94	27
Fe3O4-doped	Kissinger Method	115	27
In2O3-doped	Kissinger Method	122	27
ZnO-doped	Kissinger Method	147	27
TiH2-doped	Kissinger Method	58.4	6
Nb2O5-doped	Kissinger Method	71	55
CoCl2-doped	Kissinger Method	121.3	25
NiCl2-doped	Kissinger Method	102.6	25

---

The pronounced effect of Nb₂O₅ nanopowders in promoting dehydrogenation reactions of the Na–Mg–Al–H system is further demonstrated in the isothermal volumetric measurements. Fig. 8 presents the dehydriding kinetics at 280 °C and 300 °C for the samples with and without the Nb₂O₅ additive, respectively. For comparison, the dehydrogenation kinetics of the pure MgH₂, decomposed at 280 °C and 300 °C, are also included. The desorption rate is much slower for the pristine MgH₂ both at 280 °C and 300 °C. The desorption kinetics exhibit an enhancement after mixing with NaAlH₄. However, the enhancements arising upon the addition of Nb₂O₅ to the MgH₂/NaAlH₄ sample are much more significant. Within 60 min, the pure MgH₂ has only desorbed 0.7 wt% and 1.1 wt% hydrogen at 280 °C and 300 °C, respectively. The 1MgH₂–1NaAlH₄ composite can release 4.5 wt% and 5.2 wt% hydrogen in 60 min at 280 °C and 300 °C, respectively. Under the same conditions, the doped sample has yielded 5.9 wt% and 6.2 wt% of hydrogen. Moreover, the MgH₂–NaAlH₄–Nb₂O₅ sample has desorbed about 5.2 wt% hydrogen after 25 min at 300 °C, which is higher than for the MgH₂–NaAlH₄ (3.2 wt%) and much higher than for the pure MgH₂ (0.5 wt%) at same temperature. In contrast, 60 min is required for the undoped MgH₂–NaAlH₄ sample to release 5.2 wt% hydrogen at 300 °C. These results indicate that the dehydrogenation kinetics of MgH₂ is significantly improved by combining with NaAlH₄, and also further enhanced by the addition of Nb₂O₅. In conclusion, the Nb₂O₅-doped MgH₂/NaAlH₄ mixture exhibits superior dehydrogenation properties than that of the pure MgH₂ as well as the uncatalyzed MgH₂/NaAlH₄ samples.

Fig. 8 Comparison of the isothermal dehydriding curves for neat MgH₂ at (a) 280 °C and (b) at 300 °C, undoped MgH₂–NaAlH₄ composite (in mole ratio of 1 : 1) at (c) 280 °C and (d) at 300 °C, and Nb₂O₅-doped MgH₂–NaAlH₄ composite (in mole ratio of 1 : 1) at (e) 280 °C and (f) at 300 °C.

In order to further analyze the catalytic efficiency of Nb₂O₅ nanoparticles for the rehydrogenation/dehydrogenation cycles, the reversibility and cyclic properties of the Nb₂O₅-doped binary composite system have been investigated. The rehydrogenation of the samples has been conducted under 9.5 MPa of H₂ at 300 °C after complete dehydrogenation. The undoped binary system has also been examined for comparison. Fig. 9 displays the isothermal rehydrogenation kinetics for the first three hydrogenation cycles of the Nb₂O₅-doped 1 : 1 system and the first rehydrogenation cycle of the undoped 1 : 1 sample. It is evident that the rehydriding capacity for the undoped sample (3.3 wt%) is smaller than that of the Nb₂O₅-doped sample (4.4 wt%), indicating that the reversibility is facilitated by the Nb₂O₅ dopant. Moreover, the recycled MgH₂–NaAlH₄–Nb₂O₅ sample exhibits well maintained kinetics and some capacity loss compared to the performance in the first cycle. It decreases from 4.4 wt% (first rehydrogenation) to 4.2 wt% (second rehydrogenation) and 4 wt% (third rehydrogenation). The XRD profiles of both the undoped and doped samples following the rehydrogenation, shown in Fig. 10b and c, indicate the reformation of MgH₂ and NaH. This suggests that reaction (3) is reversible. However, no Al₃Mg₂ alloy is found in the rehydrogenated samples, indicating that full recovery of MgH₂ from Al₁₂Mg₁₇ alloy has been acquired according to the following equations.

Mg₁₇ Al₁₂ + (17 − 2y)H₂ → yMg₂Al₃ + (17 − 2y)MgH₂ + (12 − 3y)Al (7)

Mg₂Al₃ + 2H₂ → 2MgH₂ + 3Al (8)

Fig. 9 Comparison of the rehydriding curves for the undoped MgH₂–NaAlH₄ composite (in mole ratio of 1 : 1) and Nb₂O₅-doped MgH₂–NaAlH₄ composite (in mole ratio of 1 : 1) at 300 °C under 9.5 MPa.

In addition, the Na₃AlH₆ phase has been detected for the rehydrogenated doped sample (Fig. 10c), which is not found for the rehydrogenated undoped sample (Fig. 10b). The formation of Na₃AlH₆ may be accounted for the enhanced rehydriding capacity of the doped sample engendered by the Nb₂O₅ dopant, indicating that the Nb₂O₅ component in the MgH₂–NaAlH₄–Nb₂O₅ sample plays a catalytic role through the formation of Nb-containing catalytic species.

Fig. 10 XRD spectra for (a) undoped MgH₂–NaAlH₄ (1 : 1) composite after complete dehydrogenation, (b) undoped MgH₂–NaAlH₄ (1 : 1) composite after rehydrogenation, and (c) Nb₂O₅-doped MgH₂–NaAlH₄ (1 : 1) composite after rehydrogenation.

To further analyze the milling and cycling process, the chemical characterization of the Nb-based additive has been also investigated by XPS in the surface and sub surface regions of the milled 1MgH₂–1NaAlH₄ + 2 mol% Nb₂O₅ sample before and after hydrogen cycling. Fig. 11 depicts the Nb 3d photoelectron peak region for the as prepared sample as compared to the fully desorbed and fully reabsorbed samples. According to Fig. 11a for the as prepared sample, no Nb catalyst has been detected at the surface. During desorption, the reduced Nb species migrate from the bulk to the surface.⁵⁶ This is consistent with the results obtained from XRD. It is hard to specify whether the inception of the reduction of Nb₂O₅ takes place during milling or not. XRD results for the milled sample have not revealed the existence of any reduced Nb species. During the first hydrogen desorption, the reduction of Nb₂O₅ commences and the originating products disperse in the sample and emerge to the surface. This fact is further strengthened by the XPS spectra from the Mg 2p regions of the doped sample (Fig. 12b), exhibiting the occurrence of MgO after first dehydrogenation. The Na₃AlH₆ photoelectron peak in Fig. 13(a–d) for the rehydrogenated samples, located at 64.92 eV,⁵⁷ further proves the reformation of Na₃AlH₆ during the rehydrogenation process. These observations are in accordance with the results of X-ray diffraction, showing that the enhanced rehydriding capacity of the doped sample is induced by the reformation of Na₃AlH₆.

Fig. 11 Narrow scan XPS Nb 3d spectra for Nb₂O₅-doped MgH₂–NaAlH₄ (1 : 1) composite before dehydrogenation, after dehydrogenation, and after 3rd rehydrogenation.

Fig. 12 Narrow scan XPS Mg 2p spectra for (a) pristine MgH₂ and (b) Nb₂O₅-doped MgH₂–NaAlH₄ (1 : 1) composite before dehydrogenation, after dehydrogenation, and after 3rd rehydrogenation.

Fig. 13 Narrow scan XPS Na 2s spectra for pristine NaAlH₄ and Nb₂O₅-doped MgH₂–NaAlH₄ (1 : 1) composite before dehydrogenation, after dehydrogenation, and after 3rd rehydrogenation.

The FESEM qualitative and quantitative microstructural investigations have been employed to examine the particle size and morphology of the as-milled and cycled samples. Fig. 14a and b indicates that the pure MgH₂ and NaAlH₄ particles are irregularly shaped with a mean particle size of more than 30 μm. Fig. 14c,d brings out the microstructural features of the as-milled undoped 1MgH₂–1NaAlH₄ and 2MgH₂–1NaAlH₄ powders. Remarkably, the stoichiometry has exhibited an evident influence on the particle size distribution reached after the milling process. This noticeable difference may be ascribed to the higher brittleness of the MgH₂ powders which, indeed, improves the efficiency of the ball milling procedure. The FESEM analysis carried out on the cross-section of 1MgH₂–1NaAlH₄ + 2 mol% Nb₂O₅ composite has revealed the substantial refinement induced by the Nb₂O₅ nanoparticles (Fig. 14e). The size of most of the particles is less than 5 μm for the sample with Nb₂O₅ dopant, indicating the significant diminution of particle size compared to that of undoped unary as well as binary samples. It is vividly discernible in the inset of Fig. 14e that the oxide particles are embedded heterogeneously in the composite matrix, resulting in the substantial surface modifications exhibiting many deformed and disordered surface regions. Moreover, the hardness of Nb₂O₅ is much higher than that of MgH₂ and NaAlH₄. Therefore, the fraction of embedded Nb₂O₅ nanoparticles will enhance the hardness and brittleness of hydride particles that will ultimately shift the balance between fracturing and agglomeration to smaller particle sizes. Moreover, Nb₂O₅ is well known for its good lubricant and dispersive properties that prevent the agglomeration and cold welding of hydride particles and thereby facilitates the refinement of matrix particles during milling. Fig. 14f brings out the microstructural features of the doped materials after third rehydrogenation. The small pores are observed at the matrix surface are imputable to the repeated volume shrinkage and expansion during the sorption cycles. Fig. 15(a–b) is the surface EDS spectra collected at the positions circled in Fig. 14e,f. The variation in the Nb peak positions following the cycling (Fig. 15b) further corroborates that Nb₂O₅ has been reduced during cycling.

Fig. 14 Field emission scanning electron microscopy (FESEM) images of (a) as-received MgH₂, (b) as-received NaAlH₄, (c) undoped MgH₂–NaAlH₄ (1 : 1) composite, (d) undoped MgH₂–NaAlH₄ (2 : 1) composite, (e) Nb₂O₅-doped MgH₂–NaAlH₄ (1 : 1) composite after ball milling, and (f) Nb₂O₅-doped MgH₂–NaAlH₄ (1 : 1) composite after third rehydrogenation.

Fig. 15 Energy-dispersive spectroscopy (EDS) results of Nb₂O₅-doped MgH₂–NaAlH₄ (1 : 1) composite (a) after ball milling (area circled in Fig. 14e) and (b) after third rehydrogenation (area circled in Fig. 14f).

The above experimental results suggest that all the stoichiometries follow the same initial low-temperature reaction pathway (decomposition of NaAlH₄). Nevertheless, the differences in the respective pathways begin to emerge at elevated temperatures (the formation of NaMgH₃), depending upon the proportion of the potential reactant (MgH₂). The improvement in dehydrogenation kinetics of binary composite is believed to be due to the interaction between MgH₂ and alanate that changes the thermodynamics of the reactions by either lowering the enthalpy of the dehydrogenation reaction or leads the balance of the reaction towards one direction. Moreover, the smaller powder agglomerates are obtained for the 2 : 1 composites (Fig. 14), which accounts for their enhanced dehydrogenation kinetics compared to that of other samples. The results of the present investigation also reveal that additive, i.e. Nb₂O₅ nanopowders can lead to the generation of surface defects by inducing a substantial refinement of powder particles during the high-energy milling, which are expected to further ameliorate the kinetics as well as the hydrogen sorption capability. It is documented in our previous work^9,10,43 that the hydrogen desorption/absorption is closely associated with the surface defects and the refinement, and the hydrogen sorption capability increases with the increase in the defects in the nanostructures. In addition, the lubricant, dispersive, and hardness properties of Nb₂O₅ can facilitate a further reduction of hydride particles by preventing the agglomeration of matrix particles during milling, which has to be considered also an important factor for enhancing the reaction kinetics. Nonetheless, other factors, such as the nature and local electronic structure of the added oxide as well as its reduction during heating also have to be taken into account. Since the Nb₂O₅ is reduced during the heating, the possible candidates for catalysis remain in the non-stoichiometric magnesium–niobium oxide or even other possible niobium phases with lower oxidation states formed during cycling. Similarly, in previous studies by ourselves and others,^10,42 it is suggested that the partially reduced Nb species with a wide range of valence might play a major role as a catalyst. Thereby, the finely dispersed oxygen-deficient Nb species may contribute to kinetic improvement by facilitating the diffusion of hydrogen through the diffusion barriers both in dehydrogenation and rehydrogenation processes.

Conclusions

In conclusion, we have examined the hydrogen storage properties and reaction pathways of three distinct stoichiometries within the MgH₂–NaAlH₄ binary composite system both undoped and doped with Nb₂O₅ nanopowders. The premilled reactant stoichiometry of the MgH₂/NaAlH₄ system has a profound impact on the reaction kinetics and the sorption properties because of the reactant availability. It has also been demonstrated that the dehydrogenation kinetics of MgH₂ and NaAlH₄ can be improved by combining them with each other. Apart from the existence of the mutual interactions between two hydrides, the crystallite size of the undoped binary composites is also found to be smaller than that of the pristine components, which explains for the superior hydrogen performance of the mixed samples. Significant improvements in the dehydrogenation/rehydrogenation properties of the MgH₂–NaAlH₄ system have been achieved by adding a small amount of Nb₂O₅ nanoparticles. The onset as well as the peak temperatures of hydrogen desorption shift to lower temperatures. The kinetics of hydrogen desorption with Nb₂O₅ additions are found to be 2–6 times as fast as the undoped MgH₂ and MgH₂–NaAlH₄ samples. The rehydrogenation properties of the MgH₂–NaAlH₄–Nb₂O₅ composite are also improved significantly as compared to the MgH₂–NaAlH₄ composite. The reversible capacity of 4.4 wt% is acquired for the MgH₂–NaAlH₄–Nb₂O₅ composite, which is larger than that of MgH₂–NaAlH₄ (3.3 wt%). Moreover, this catalytically enhanced rehydrogenation capacity persists well during three de/rehydrogenation cycles. XRD, XPS, and FESEM-EDS analyses suggest that hydrogen cycling has induced the reduction of Nb₂O₅, which leads to the formation of oxygen-deficient reduced niobium oxide species. Accordingly, it is believed that the fine dispersion of these oxygen-deficient niobium oxide nanoparticles facilitates the dehydrogenation process by serving as the active sites for nucleation and growth of the dehydrogenated product associated with the shortening of the diffusion paths among the reaction ions and thus reducing the kinetic barriers and rendering the amelioration of dehydrogenation kinetics. Besides, the nanocrystalline reduced niobium oxide phases present in the mixture can also provide the active catalytic sites for the hydrogen dissociation and surface adsorption, which improves the kinetics of the rehydrogenation as well. Moreover, the lubricant, dispersive, and hardness properties of Nb₂O₅ increase the surface defects and grain boundaries by a large reduction in the particle size, creating a larger surface area for hydrogen to interact, thereby decreasing the temperature for decomposition.

Acknowledgements

This work is financially supported by the University of Science and Technology Beijing (USTB). The authors also thank the Higher Education Commission (HEC) of Pakistan for the financial support to Rafi-ud-din.

References

1. R. A. Varin, T. Czujko and Z. S. Wronski, Nanomaterials for Solid State Hydrogen Storage; Springer Science + Business Media, New York, 2009.
2. J. Yang, A. Sudik, C. Wolverton and D. J. Siegel, Chem. Soc. Rev., 2010, 39, 656–675.
3. A. W. C. Berg and C. O. Arean, Chem. Commun., 2008, 668–681.
4. S.-I. Orimo, Y. Nakamori, J. R. Eliseo, A. Züttel and C. M. Jensen, Chem. Rev., 2007, 107, 4111–4132.
5. T. Umegaki, J. M. Yan, X. B. Zhang, H. Shioyama, N. Kuriyama and Q. Xu, Int. J. Hydrogen Energy, 2009, 34, 2303.
6. L. Jun, J. C. Young, Z. F. Zhigang, Y. S. Hong and R. J. Ewa, AM. CHEM. SOC., 2009, 131, 15843–15852.
7. J. F. Mao, Z. Wu, T. J. Chen, B. C. Weng, N. X. Xu and T. S. Huang, J Phys Chem C, 2007, 111, 12495–12498.
8. M. David, B. J. Daniel, S. Toyoto, N. Dag, K. Daisuke, S. Tetsuo, K. Naoyuki and Y. Hitoshi, J. Mater. Chem., 2009, 19, 8150–8161.
9. Rafi-ud-din, L. Zhang, L. Ping and Q. Xuanhui, J. Alloys Compds., 2010, 508, 119–129.
10. Rafi-ud-din, Q. Xuanhui, L. Ping, L. Zhang and M. Ahmad, J Phys Chem C, 2011, 115, 13088–13099.
11. L. Li, F. Qiu, Y. Wang, Y. Wang, G. Liu, C. Yan and C. An, J. Mater. Chem., 2012, 22, 3127–3132.
12. J. Chen, N. Kuriyama, Q. Xu, H. T. Takeshita and T. Sakai, J. Phys. Chem. B, 2001, 105, 11214–11220.
13. F. Schuth, B. Bogdanovic and M. Felderhoff, Chem. Commun., 2008, 668–681.
14. J. F. Mao, X. B. Yu, Z. P. Guo, C. K. Poh, H. K. Liu, Z. Wu and J. Ni, J. Phys. Chem. C, 2009, 113, 10813–10818.
15. P. Chen, Z. T. Xiong, L. F. Yang, G. T. Wu and W. F. Luo, J Phys Chem B, 2006, 110, 14221–14225.
16. H. Y. Leng, T. Ichikawa, S. Hino, N. Hanada, S. Isobe and H. Fujii, J Power Sources, 2006, 156, 166–170.
17. M. Au, A. R. Jurgensen, W. A. Spencer, D. L. Anton, F. E. Pinkerton and S. J. Hwang, J Phys Chem C, 2008, 112, 18661–18671.
18. X. B. Yu, G. L. Xia, Z. P. Guo and H. K. Liu, J Mater Res, 2009, 24, 2720–2727.
19. Y. H. Guo, X. B. Yu, L. Gao, G. L. Xia, Z. P. Guo and H. K. Liu, Energy Environ. Sci., 2010, 3, 465–470.
20. X. D. Kang, Z. Z. Fang, L. Y. Kong, H. Cheng, X. D. Yao, G. Lu and P. Wang, Adv. Mater., 2008, 20, 2756–2759.
21. J. Zhao, J. Shi, X. Zhang, F. Cheng, J. Liang and Z. Tao, Adv. Mater, 2010, 22, 394–397.
22. B. Peng, J. Liang, Z. Tao and J. Chen, J. Mater. Chem., 2009, 19, 2877–288.
23. F. Zeppelin, H. Reule and M. Hirscher, J. Alloys Compds., 2002, 330, 723–726.
24. C. X. Shang, M. Bououdina1, Y. Song and Z. X. Guo, Int. J. Hydrogen Energy, 2004, 29, 73–80.
25. J. F. Mao, Z. Guo, X. Yu, H. Liu, Z. Wu and J. Ni, Int. J. Hydrogen Energy, 2010, 35, 4569–4575.
26. L. Xie, Y. Liu, X. Zhang, J. Qu, Y. Wang and X. Li, J. Alloys Compds., 2009, 482, 388–392.
27. M. Polanski and J. Bystrzycki, J. Alloys Compds., 2009, 486, 697–701.
28. D. L. Croston, D. M. Grant and G. S. Walker, J. Alloys Compds., 2010, 492, 251–258.
29. O. Friedrichs, T. Klassen, J. C. Sanchez-Lopez, R. Bormann and A. Fernandez, J. scripta mat., 2006, 54, 1293–1297.
30. L. P. Ma, P. Wang and H. M. Cheng, Int. J. Hydrogen Energy, 2010, 35, 3046–3050.
31. N. Recham, V. V. Bhat, M. Kandavel, L. Aymard, J. M. Tarascon and A. Rougier, J. Alloys Compds., 2008, 464, 377–382.
32. J. Lu and Z. G. Z. Fang, J. Phys. Chem., 2005, 109, 20830–20834.
33. Z. T. Xiong, G. T. Wu, J. J. Hu and P. Chen, J. Power Sources, 2006, 159, 167–170.
34. P. Wang, L. P. Ma, Z. Z. Fang, X. D. Kang and P. Wang, Energy Environ. Sci., 2009, 2, 120–123.
35. G. S. Walker, D. M. Grant, T. C. Price, X. Yu and V. Legrand, J. Power Sources, 2009, 194, 1128–1134.
36. D. Pottmaier, C. Pistidda, E. Groppo, S. Bordiga and G. Spoto, Int. J. Hydrogen Energy, 2011, 36, 7891–7896.
37. J. F. Mao, Z. Guo, H. Leng, Z. Wu, Y. Guo, X. Yu and H. Liu, J. Phys. Chem. C, 2010, 114, 11643–11649.
38. J. F. Mao, Z. Guo, X. Yu, M. Ismail and H. Liu, Int. J. Hydrogen Energy, 2011, 36, 5369–5374.
39. Y. Zhang, T. Qi-Feng, L. Shu-Sheng and S. Li-Xian, J. Power Sources, 2008, 185, 1514–1518.
40. M. Ismail, Y. Zhao, X. B. Yu and S. X. Dou, RSC Advances, 2011, 1, 408–415.
41. M. Ismail, Y. Zhao, X. B. Yu, J. F. Mao and S. X. Dou, Int. J. Hydrogen Energy, 2011, 36, 9045–9050.
42. O. Friedrichs, F. Aguey-Zinsou, J. R. Ares Fernandez, J. C. Sanchez-Lopez, A. Justo, T. Klassen, R. Bormann and A. Fernandez, Acta mat., 2006, 54, 105–110.
43. A. Mashkoor, Rafi-ud-Din, P. Caofeng and Z. Jing, J. Phys. Chem. C, 2010, 114, 2560.
44. C. Milanese, S. Garroni, A. Girella, G. Mulas, V. Berbenni, G. Bruni, S. Surinach, M. D. Baro and A. Marini, J. Phys. Chem. C, 2011, 115, 3151–3162.
45. H. E. Kissinger, Anal. Chem., 1958, 29, 1702.
46. S. Zheng, Y. Li, F. Fang, G. Zhou, X. Yu, G. Chen, D. Sun, L. Ouyang and M. Zhu, J. Mater. Res., 2010, 25, 2047–2053.
47. X. Fan, X. Xiao, L. Chen, S. Li and Q. Wang, J. Alloys Compds., 2011, 509, 386–389.
48. G. Sandrock, K. Gross and G. Thomas, J. Alloys Compd., 2002, 339, 299–308.
49. D. Pukazhselvan, M. Hudson, A. Sinha and O. Srivastava, Energy, 2010, 35, 5037–5042.
50. O. Kircher and M. Fichtner, J. Alloys Compd., 2005, 404, 339–342.
51. X. Fan, X. Xiao, L. Chen, S. Li, H. Ge and Q. Wang, J. Phys. Chem. C, 2011, 115, 2537–2543.
52. X. Fan, X. Xiao, L. Chen, L. Han, S. Li, H. Ge and Q.Wang, Int. J. Hydrogen Energy, 2011, 36, 10861–10869.
53. Y. Luo, P. Wang, L. Ma and H. Cheng, J. Alloys Compd., 2008, 453, 138–142.
54. M. Fan, S. Liu, Y. Zhang, J. Zhang, L. Sun and F. Xu, Energy, 2010, 35, 3417–3421.
55. N. Hanada, T. Ichikawa, S. Hino and H. Fujii, J. Alloys Compd., 2006, 420, 46–49.
56. http://srdata.nist.gov/xpx/ .
57. C. Wan, X. Ju, Y. Qi, S. Wang, X. Liu and L. Jiang, J. Alloys Compd., 2009, 486, 436–441.

---

Footnote

† Electronic Supplementary Information (ESI) available. See DOI: 10.1039/c2ra20518a/

---