Nanolayer-like-shaped MgFe2O4 synthesised via a simple hydrothermal method and its catalytic effect on the hydrogen storage properties of MgH2

In this study, the effect of nanolayer-like-shaped MgFe2O4 that is synthesised via a simple hydrothermal method on the performance of MgH2 for hydrogen storage is studied. MgH2 + 10 wt% MgFe2O4 is prepared by using the ball milling method. The MgFe2O4-doped MgH2 sample started to release H2 at approximately 250 °C, 90 °C and 170 °C lower than the milled and pure MgH2 respectively. At 320 °C, the isothermal desorption kinetic study has shown that the doped sample has desorbed approximately 4.8 wt% H2 in 10 min while the milled MgH2 desorbed less than 1.0 wt% H2. For isothermal absorption kinetics, the doped sample can absorb approximately 5.5 wt% H2 in 10 min at 200 °C. Meanwhile, the undoped sample absorbs only 4.0 wt% H2 in the same condition. The activation energy of 10 wt% MgFe2O4-doped MgH2 composite is 99.9 kJ mol−1, which shows a reduction of 33.1 kJ mol−1 compared to the milled MgH2 (133.0 kJ mol−1). X-ray diffraction spectra display the formation of new species which are Fe and MgO after dehydrogenation, and these new species are believed to act as the real catalyst that plays a crucial role in improving the sorption performance of the MgFe2O4-doped MgH2 system by providing a synergetic catalytic effect.


Introduction
To prepare for the future and ensure global environmental viability, energy systems have to be reliable, clean, low cost, environmentally friendly and exible. Humanity is expected to use 40 TW of power (40 billion of kW) in the future. To satisfy this demand, different sources of renewable energy, such as hydrogen, are needed. Sustainable hydrogen is an ideal clean energy carrier because there is no carbon dioxide or other greenhouse gas emission at the end-user level. Commonly, there are 3 forms of storing hydrogen which is high-pressure gas, cryogenic liquid hydrogen in tanks (stored at 21.2 K) and as solid state hydrogen storage by either reacting with chemical compounds or absorbing. Among these approaches, solid-state hydrogen storage has higher potential for higher hydrogen density and may yield greater utility towards the practical implementation of hydrogen storage. Among the various materials for solid-state hydrogen storage, MgH 2 considered as one of the most potential material due to its high hydrogen storage capacity (7.6 wt%), excellent reversibility and low cost. 1 However, MgH 2 is restricted by the decomposition temperature, which is high with slow sorption kinetics and is excessively stable thermodynamically. 2 Many research have been conducted to overcome these disadvantages by altering the thermodynamics and improve the kinetic properties by producing nanostructures 3,4 and utilizing catalysts such as carbon-based materials, 5,6 metals, 7-10 metal hydrides, 11,12 metal oxides, [13][14][15][16][17][18] metal halides, [19][20][21] and nanosized alloys. [22][23][24] Previous research has proved that catalyst based on ternary metal oxides greatly improved the hydrogen storage performance of MgH 2 . 15, [25][26][27][28][29][30][31][32][33][34][35] Zhang et al. 25 26 This variation paved the way for the debate on how ternary metal oxides, particularly ferrites, work as catalysts in improving the hydrogen sorption performance of MgH 2 . Moreover, the difference in the synthesis method of the catalysts may also provide a different effect in the catalytic role.
Inspired by the role of active species that formed during the heating process in the MgH 2 -ternary metal oxides catalyst system, it is quite interesting to investigate the use of other ferrites (e.g. MgFe 2 O 4 ) as catalysts to improve the hydrogen sorption performance of MgH 2 . Therefore, in this work, MgFe 2 O 4 was synthesised by using a simple hydrothermal method, and its catalytic effects on the hydrogen sorption performance of MgH 2 were systematically studied. To the best of authors' knowledge, this paper is the rst to study the hydrogen sorption performance of MgH 2 catalysed with MgFe 2 O 4 . The possible catalysis mechanisms of MgFe 2 O 4 in the sorption performances of MgH 2 are also discussed in this paper.

Experimental details
The nanolayer-like-shaped MgFe 2 O 4 was synthesised via a hydrothermal method. In a typical synthesis, a stoichiometric amount of Mg(NO 3 ) 2 $6H 2 O (Sigma-Aldrich) and Fe(NO 3 ) 3 $9H 2 O (Sigma-Aldrich) were dissolved in 50 ml distilled water. A total of 10 ml of H 4 N 2 $H 2 O (Sigma-Aldrich) was added dropwise to the above solution to attain the resultant pH of >9. The mixture was then transferred into a sealed Teon lined stainless-steel autoclave (125 ml capacity) and heated for 12 h at 180 C. The nal product was washed several times with deionised water and dried overnight at 60 C under vacuum. A total of 10 wt% of as-prepared MgFe 2 O 4 was mixed with 300 mg of MgH 2 (95% pure; Sigma-Aldrich) and undergo intensive ball milling for 1 h in a planetary ball mill at the rate of 400 rpm. For comparison, pure MgH 2 , and MgH 2 added with 10 wt% Fe (Alfa Aesar) and 10 wt% MgO (R&M Chemicals), respectively were also prepared under the same conditions. All preparations, including loading and weighing, were conducted in an argon atmosphere glove box (MBraun Unilab).
The onset decomposition temperature and sorption kinetic measurement for doped and undoped samples were characterised by using Sievert-type pressure-composition-temperature apparatus (Advanced Materials Corporation). For onset decomposition temperature measurement, the samples were heated from room temperature to 450 C at a heating rate of 5 C min À1 in vacuum chamber. Meanwhile, the sorption kinetics was conducted under 1.0 atm at 320 C for desorption kinetic measurement and under 33.0 atm at 200 C for absorption kinetic measurement. The thermal properties of the doped and undoped samples were performed using differential scanning calorimeter (DSC)/thermogravimetric analysis from Metler Toledo. With a ow of 50 ml min À1 argon, the samples were heated with 15, 20, 25 and 30 C min À1 heating rate from 25 C to 500 C.
The phase composition of the samples was analysed by XRD via a Rigaku MiniFlex X-ray diffraction apparatus equipped with Cu Ka radiation. Data were collected in the 2q range 20 to 80 at 2 min À1 . The morphologies of the samples were observed by scanning electron microscopy (SEM) (JEOL JSM-6350LA). Fourier transform infrared (FTIR) spectrometry was recorded on an IR Shimadzu Tracer-100 between 400 and 2000 cm À1 . Raman spectra were recorded on Renishaw Raman spectroscopy (532 nm radiation) extended with 0.1% power laser measurement at room temperature.

Results and discussion
Before milling with MgH 2 , the phase structure of MgFe 2 O 4 was conrmed by XRD, as shown in Fig. 1 where L is the average crystallite size (nm), q is the angle of diffraction, k is Scherrer's constant (k ¼ 0.94), l is the X-ray wavelength (0.15405 nm) and B is the full width at half maximum of the diffraction peak in radian (FWHM). The SEM image ( Fig. 1(b)) reveals that the MgFe 2 O 4 forms a large layer with a nanosized thickness. From the FTIR spectrum ( Fig. 1(c)), two typical peaks of MgFe 2 O 4 were observed at the low wavenumber, thus indicating the formation of spinel ferrite structure. 36,37 The peak at 417 cm À1 can be ascribed to the Fe-O vibration in the octahedral site, and the peak at 538 cm À1 can be assigned to the Fe-O vibration in the tetrahedral and octahedral sites. Furthermore, the Raman peaks ( Fig. 1(d)) at 475 and 694 cm À1 can be assigned to the typical characteristic peaks of MgFe 2 O 4 , 38 whereas the peak at 283 cm À1 corresponding to the stretching vibration of the Mg-O chemical bond. 39 The XRD, FTIR and Raman spectroscopy results conrm that pure MgFe 2 O 4 was successfully synthesised by the hydrothermal method. Fig. 2(a) shows the onset decomposition temperature results for the pure MgH 2 , milled MgH 2 and MgH 2 with 10 wt% MgFe 2 O 4 . Before milling, pure MgH 2 started to desorb hydrogen at approximately 420 C. The total amount of hydrogen desorbed is approximately 7.0 wt%. Aer milling for 1 h, the onset decomposition temperature of MgH 2 was decreased to approximately 340 C. This phenomenon demonstrate that the sorption performance of MgH 2 also inuenced by the milling process. From the curve, it can be seen that aer milling, the amount of hydrogen desorb of MgH 2 slightly decreases. This might be ascribed to the hydrogen released from MgH 2 during the milling process. Aer doping with 10 wt% of MgFe 2 O 4 , it is clear that the onset decomposition temperature of the MgH 2 was dramatically reduced to 250 C, 90 C and 170 C lower than that for the milled and pure MgH 2 , respectively. However, the hydrogen desorption capacity decrease slightly to approximately 6.5 wt% because the dopant used in this study, namely, MgFe 2 O 4 , does not contain hydrogen. 40 From Fig. 2(a), it can be concluded that MgFe 2 O 4 additive plays a positive role in decreasing the decomposition temperature of MgH 2 .
To further examine the sorption properties of the MgFe 2 O 4doped MgH 2 sample, the isothermal absorption kinetic was studied. The amount of hydrogen absorbed from the milled MgH 2 and the MgFe 2 O 4 -doped MgH 2 sample was measured under 33.0 atm H 2 and at constant temperature of 200 C, as shown in Fig. 2 For further studies on the catalytic effect of MgFe 2 O 4 on the sorption kinetic of MgH 2 , isothermal desorption kinetic was performed under 1.0 atm at 320 C. As shown in Fig. 2(c) Fig. 3(a) presents the isothermal absorption kinetics of the 10 wt% MgFe 2 O 4 doped with MgH 2 at 320 C under a hydrogen pressure of 33.0 atm over 10 cycles. From the result, it can be seen that aer the ten cycles, the absorption kinetics show a small reduction in the hydrogen capacity. Aer completing the 10 th cycle, the system is able to absorb 5.6 wt% of hydrogen in 60 minutes. The result shows that the doped system displays good absorption properties even aer 10 cycles. As for the desorption kinetics, Fig. 3(b) shows the isothermal desorption kinetics for 10 cycles that was carried out at 320 C and under 1.0 atm of pressure. Like the absorption kinetics, a small hydrogen capacity degradation is shown aer completing the 10 th cycle. The doped system possesses a good performance aer completing the 10 th cycle as it is able to desorb about 5.5 wt% of hydrogen within 60 minutes. These results demonstrated that MgFe 2 O 4 plays a vital catalytic role for the cycle life of MgH 2 .
The thermal properties of the 10 wt% MgFe 2 O 4 -doped MgH 2 and undoped MgH 2 sample were further studied by DSC at heating rate of 30 C min À1 and under a ow of 50 ml min À1 argon (Fig. 4). Obviously, the DSC trace for the pure MgH 2 showed one endothermic peak at approximately 482.9 C. This strong endothermic peak related to the released of hydrogen  from the MgH 2 . Similar to the pure MgH 2 , DSC traces of the milled MgH 2 and MgFe 2 O 4 -doped MgH 2 showed only one strong endothermic peak at 438.8 C and 393.3 C respectively. The peaks correlated to the decomposition of MgH 2 but at lower temperatures.
The improvement in desorption behaviour is correlated with the kinetic barrier of the hydrogen desorbed from the MgH 2 . By doping MgH 2 with MgFe 2 O 4 , low value of activation energy for released hydrogen is obtained. Kissinger analysis 41 (eqn (2)) was conducted to determine the activation energy of doped and undoped MgH 2 samples.
where b is the heating rate, E A is the activation energy, R is the gas constant, T p is the peak temperature of DSC curves and A is the linear constant. Fig. 5 Fig. 6 presents the microstructures of the pure and milled MgH 2 , and MgFe 2 O 4 -doped MgH 2 . From the images, it can be seen clearly that the particle size of the pure MgH 2 is around 50-100 mm (Fig. 6(a)). Fig. 6(b) shows the image of the MgH 2 aer 1 h ball milling. The size of the milled MgH 2 was decreased dramatically compared to the pure MgH 2 . However, the image shows agglomeration and inconsistent particle sizes. Fig. 6(c) shows that the particle size of 10 wt% MgFe 2 O 4 -doped MgH 2 was the smallest and had less agglomeration than the pure and milled MgH 2 . Smallest particle size gives a larger region of contact to the MgH 2 , thus resulting in the higher rate of reaction of MgH 2 .
To investigate the phase structure, XRD measurement was performed on the 10 wt% MgFe 2 O 4 -doped MgH 2 sample, as shown in Fig. 7. From Fig. 7(a), it can be observed that the MgH 2 and MgFe 2 O 4 phases are present in the as-milled MgFe 2 O 4doped MgH 2 sample. No additional peaks were found from the spectra. Aer dehydrogenation at 450 C ( Fig. 7(b)), the XRD pattern showed that the MgH 2 was completely dehydrogenated to Mg. This result demonstrates that the decomposition of MgH 2 was completed aer heating for up to 450 C. Furthermore, a small peak of MgO and Fe formed aer the desorption process, thus demonstrate that the partial reaction of MgH 2 with MgFe 2 O 4 may occur during the heating process as follows: The standard Gibbs Free energy, DG f , of MgH 2 , MgFe 2 O 4 and MgO are À35.9824, À1317.1232 and À569.024 kJ mol À1 , respectively. 42 The total change DG correlated with the reaction in eqn (3) is À815.0168 kJ mol À1 . These values can conrm the possibility of the reaction in eqn (3) from thermodynamic potentials. Fig. 7(c) shows the XRD patterns for the rehydrogenated sample under 33.0 atm H 2 at 320 C. The result illustrates that the phase of Mg was fully converted into MgH 2 . Furthermore, the peak of Fe and MgO still remained aer undergo rehydrogenation.
From the above analyses, the improvements in the sorption properties of MgH 2 doped with 10 wt% MgFe 2 O 4 may be resulted from the formations of Fe and MgO. Fe and MgO may act as the real catalysts that play a vital role in the improvements of MgH 2 sorptions. Therefore, to verify the effect of Fe and MgO on MgH 2 , samples of 10 wt% MgO-doped MgH 2 and 10 wt% Fedoped MgH 2 were prepared and the TPD proles for the dehydrogenations were shown as in Fig. 8. It is clearly seen that the decomposition temperature of MgH 2 are reduced with the addition of MgO or Fe as compared to the pure and milled MgH 2 . However, the performance of these MgO and Fe are not signicant as the sample of 10 wt% of MgFe 2 O 4 doped with MgH 2 . This demonstrated that the in situ generated MgO and Fe from the reaction of MgH 2 + 10 wt% of MgFe 2 O 4 may play a signicant role that introduce a synergetic catalytic effect that cause a signicant improvement on the dehydrogenation performances of MgH 2 doped with 10 wt% of MgFe 2 O 4 . In addition, the in situ formed Fe and MgO in the MgH 2 + 10 wt% of MgFe 2 O 4 sample are speculated to have a higher degree of dispersion and more compact phase segregation as compared to the milled MgH 2 + 10 wt% Fe and milled MgH 2 + 10 wt% MgO. This would be likely to lead the improvement of the sorption kinetics.
From the result obtained, we postulate that formation of fresh and ne MgO and Fe species which resulted from the reaction between MgH 2 and MgFe 2 O 4 during the dehydrogenation process may play signicant role in improving the sorption performances of MgH 2 . Numerous studies have claimed that the newly active species formed during the de/ absorption process may play as a real catalyst in the enhancement of MgH 2 sorption. 43,44 Many works have proven that Fe is  an excellent catalyst for sorption performance in MgH 2 . 7-9 It is believed that the presence of fresh and ne Fe could interact with H 2 molecules, thus possibly leading to the dissociation of H 2 molecules and the improvement of the de/rehydrogenation kinetic. Furthermore, previous studies have shown that the sorption performance in MgH 2 can be enhanced with the addition of MgO. Ares-Fernández and Aguey-Zinsou 45 claimed that the addition of MgO to MgH 2 during the milling process led to an enhancement of sorption kinetics because of the high electronegativity MgO. In another study, the same group also claimed that during the milling process, MgO may act as a process control agent that can lead to the reduction of the particle agglomeration of MgH 2 by an optimal breakage rate, thus aiding the high stability of these particles and evading the use of cold welding. 46 Shan et al. 15 also revealed that MgO has a great catalytic effect on the MgH 2 sorption performance. Their study showed that during the heating process in CoFe 2 O 4 -doped MgH 2 composite system, MgO is formed. The catalytic effect of MgO could work together with the catalytic role of the Fe metal to create a synergetic effect. Therefore, the in situ active species of Fe and MgO may actually act as real catalysts and further enhance the hydrogen sorption performance of MgH 2 . However, further studies are needed to elucidate more details on the exact role of MgFe 2 O 4 addition in MgH 2 .

Conclusion
In this study, nanolayer-like-shaped MgFe 2 O 4 was successfully synthesised through a rapid, simple hydrothermal method. The addition of 10 wt% as-synthesised MgFe 2 O 4 to MgH 2 reduces the onset decomposition temperature and enhances sorption kinetics. The MgFe 2 O 4 -doped MgH 2 sample has started to release H 2 at approximately 250 C, 90 C and 170 C lower than milled and pure MgH 2 respectively. In a duration of 10 min, the isothermal desorption kinetic study showed that the doped sample can release approximately 4.8 wt% H 2 at 320 C while the milled MgH 2 only desorbed less than 1.0 wt% H 2 under the same condition. For isothermal absorption kinetics, the doped sample can absorb approximately 5.5 wt% H 2 in 10 min at 200 C. By contrast, the milled MgH 2 sample absorbed only 4.0 wt% H 2 in the same condition. From the Kissinger analysis, the apparent activation energies, E A , for the MgFe 2 O 4 -doped MgH 2 sample were calculated to be 99.9 kJ mol À1 , which is decreased by 33.1 kJ mol À1 compared with the milled MgH 2 (133.0 kJ mol À1 ). The XRD exploration displays the formation of new species of Fe and MgO aer the dehydrogenation process, and these species remained unchanged aer rehydrogenation. It is believed that the new species of Fe and MgO play a synergistic role in enhancing the hydrogen storage performances of MgH 2 .

Conflicts of interest
There are no conicts to declare.