Enhanced Hydrogen Storage Performance of MgH 2 by the Catalysis of a Novel Intersected Y 2 O 3 /NiO Hybrid

: MgH 2 is one of the most promising hydrogen storage materials due to its high hydrogen storage capacity and favorable reversibility, but it suffers from stable thermodynamics and poor dynamics. In the present work, an intersected Y 2 O 3 /NiO hybrid with spherical hollow structure is synthesized. When introduced to MgH 2 via ball-milling, the Y 2 O 3 /NiO hollow spheres are crushed into ultraﬁne particles, which are homogenously dispersed in MgH 2 , showing a highly effective catalysis. With an optimized addition of 10 wt% of the hybrid, the initial dehydrogenation peak temperature of MgH 2 is reduced to 277 ◦ C, lowered by 109 ◦ C compared with that of the bare MgH 2 , which is further reduced to 261 ◦ C in the second cycle. There is ca. 6.6 wt% H 2 released at 275 ◦ C within 60 min. For the fully dehydrogenation product, hydrogenation initiates at almost room temperature, and a hydrogenation capacity of 5.9 wt% is achieved at 150 ◦ C within 150 min. There is still 5.2 wt% H 2 desorbed after 50 cycles at a moderate cyclic condition, corresponding to the capacity retention of 79.2%. The crystal structure and morphology of the Y 2 O 3 /NiO hybrid is well preserved during cycling, showing long-term catalysis to the hydrogen storage of MgH 2 . The Y 2 O 3 /NiO hybrid also inhibits the agglomeration of MgH 2 particles during cycling, favoring the cyclic stability.


Introduction
Hydrogen is considered to be one of the most promising clean energy carriers, while safe, efficient and low-cost storage of hydrogen is the key bottleneck technology for largescale application of hydrogen energy to date [1][2][3]. MgH 2 has attracted much attention as a hydrogen storage material due to its advantages of good reversibility, high mass hydrogen storage density (7.6 wt%), and low cost [3][4][5][6]. MgH 2 prepared from waste Mg-alloys also show favorable possibility for hydrogen storage [5,6], offering more opportunities to lower the cost. However, the high thermal stability of MgH 2 (∆H = 75 kJ·mol −1 H 2 ) and the reaction barrier (∆E = 161 kJ·mol −1 ) [7] lead to high dehydrogenation temperature, where the peak dehydrogenation temperature is commonly over 360 • C [8], which limit its practical application for hydrogen storage.

Synthesis of the Intersected Y 2 O 3 /NiO Hybrid
The synthesis process of the intersected Y 2 O 3 /NiO hybrid is schematically illustrated in Figure 1, where Y(C 2 H 3 O 2 ) 3 ·4H 2 O (Sinopharm, analytical reagent, Beijing, China) and Ni(C 2 H 3 O 2 ) 2 ·4H 2 O (Sinopharm, analytical reagent, Beijing, China) are used as the sources of Y 2 O 3 and NiO, respectively. A designed molar ratio of 1:1 was used for Y 2 O 3 and NiO. Typically, 50.7 g of Y(C 2 H 3 O 2 ) 3 ·4H 2 O and 37.4 g of Ni(C 2 H 3 O 2 ) 2 ·4H 2 O were added into 1 L of deionized water and mechanically stirred, and then 63 g of citric acid (HOC(COOH)(CH 2 COOH) 2 ·H 2 O, Sinopharm, analytical reagent, Beijing, China) was added as a complexing agent to prevent metal salt hydrolysis and hydroxide precipitation. The obtained mixture is further mechanically stirred to form a solution. Thereafter, the solution was spray dried with inlet and outlet temperatures of 230 and 100 • C, respectively, and a feeding speed of 1 L/h was used. Subsequently, the spray dried precursor was heated to 900 • C at a heating rate of 3 • C/min in an air atmosphere and dwelled for 10 h. After cooling to room temperature, the intersected Y 2 O 3 /NiO hybrid with a hollow spherical structure was obtained.

Preparation of the Y 2 O 3 /NiO Introduced MgH 2 Systems
MgH 2 used in the present study was synthesized from commercially purchased Mg powder (Macklin, 99.8%, 20~100 mesh, Shanghai, China) by hydrogenating at 340 • C under 50 bar of hydrogen pressure (Jingong Gases, 99.999%, Hangzhou, China,) for 12 h. Thereafter, the as-prepared Y 2 O 3 /NiO hybrid was introduced into MgH 2 with different amounts by ball milling under 50 bar of hydrogen at 500 rpm for 5 h, where tungsten carbide balls and a ball-to-power ratio of 180:1, which was a conventional and optimized parameter obtained in our group, were used. The obtained systems are denoted as MgH 2 -x wt% Y 2 O 3 /NiO systems (x = 2, 5, 7, 10, 12). For comparison, the commercial Y 2 O 3 (Rhawn, 99.99%, 0.5 µm, Shanghai, China) and NiO (Rhawn, 99.5%, 30 nm, Shanghai, China) was also introduced into MgH 2 using the same ball-milling parameters. All MgH 2 -related operations were performed in a glove box (Etelux, Lab 2000, Beijing, China) filled with argon (Jingong Gases, 99.999%, Hangzhou, China), with the concentrations of water and oxygen both below 1 PPM.

Structure and Morphology Characterizations
The phase structures of samples were identified by X-ray diffraction (XRD, Miniflex600, Tokyo, Japan) with the radiation source of Cu Kα (40 kV, 15 mA), where the samples were sealed in a custom-designed container covered with a Scotch tape to prevent air and moisture contamination during operation. Scanning electron microscope (SEM, Hitachi-4800, Tokyo, Japan) attached with energy dispersive X-ray spectroscopy (EDS) was used to observe the surface morphologies and element distributions of the samples. Morphologies were also analyzed by transmission electron microscope (TEM, Tecnai G2 F20 S-Twin, Hillsboro, OR, USA). The surface compositions of the samples and the chemical states of the corresponding elements were analyzed by X-ray photoelectron spectroscopy (XPS, Escalab 250 Xi, Waltham, MA, USA) with a monochromatic Al Kα X-ray source under a base pressure of 3.7 × 10 −10 Torr. The XPS data were calibrated using the adventitious C 1s signal at 284.8 eV as the reference, and binding energy spectra were fitted using XPSPEAK41 software.

Hydrogen Storage Property Measurements
The hydrogen storage property measurements of the systems were carried out on a home-made Sieverts type apparatus. For non-isothermal dehydrogenation and hydrogenation tests, samples were heated to designed temperatures with a heating rate of 2 • C/min under a primary vacuum and 5 MPa H 2 pressure, respectively. For isothermal tests, samples are heated to preset temperatures with a heating rate of 10 • C/min. Isothermal dehydrogenation was also conducted under a primary vacuum. For isothermal hydrogenation, a primary vacuum was used until the preset temperatures were reached, and then 5 MPa of H 2 was loaded instead of the vacuum. The cyclic stability was measured with an isothermal regime of dehydrogenation at 300 • C for 40 min in an initial vacuum and hydrogenation at 300 • C for 20 min in 5 MPa H 2 , where the periods only denote the isothermal ones. The thermodynamic properties of the systems were studied by differential scanning calorimetry (DSC, Netzsch, Selb, Germany). Thermodynamic enthalpies of the systems were calculated by integrating the peak areas of DSC curves through Proteus Analysis Software provided by Netzsch. The kinetic performance was evaluated based on the dehydrogenation apparent activation energy (E a ), which was obtained according to the Kissinger method [48], as shown in Equation (1).
where β is the heating rate, T m is the dehydrogenation peak temperature at corresponding heating rate, R is the gas constant, and C is a constant which shows no specific meaning in the calculation of apparent activation energy. In this work, a home-made temperature programmed desorption (TPD) apparatus was used, which can intuitively obtain the onset, peak, and ending dehydrogenation temperatures at different heating rates. Figure 2 shows the XRD patterns of the as-prepared Y 2 O 3 /NiO hybrid, the raw MgH 2 , and a representative MgH 2 -x wt% Y 2 O 3 /NiO system with x of 10. It seems that, for the as-prepared Y 2 O 3 /NiO hybrid, only sharp peaks from Y 2 O 3 (PDF-# 43-0661) and NiO (PDF-# 44-1159) are detected, indicating that the synthesized product is composed of Y 2 O 3 and NiO with high crystallinity. For the as-milled MgH 2 -10 wt% Y 2 O 3 /NiO system, there are still Y 2 O 3 and NiO detected, besides the phase of MgH 2 . The result suggests that there is no chemical reaction occurring during ball-milling process. However, the diffraction peak intensities of MgH 2 , Y 2 O 3 , and NiO are all obviously weakened after ball milling compared with those of their as-prepared states, indicating reduced particle size and/or lowered crystallinity.  Figure 3a shows the SEM image of the Y 2 O 3 /NiO hybrid. It is seen that the hybrid has a spherical structure with coarse surface in size of micron and sub-micron level. A hollow structure can be observed from the crushed particle. EDS element mapping results of one selected nanosphere, as presented in Figure 3b-e, show that Y, Ni, and O elements are uniformly distributed in the whole nanosphere. TEM observation (Figure 3f) confirms that the hybrid spheres have internal hollows. Moreover, the high-resolution TEM (HRTEM) analysis ( Figure 3g) clearly shows that there are crystal grains with interplanar spacing of 0.306 nm, corresponding to lattice fringes of the (222) plane for Y 2 O 3 , and interplanar spacing of 0.208 nm, corresponding to the lattice fringes of the (012) plane of NiO, which are interactively grown, forming an intersected structure. Combining the EDS and HRTEM analyses, it is thus obtained that the prepared Y 2 O 3 /NiO hybrid is composed of interactively grown Y 2 O 3 and NiO crystal grains. SEM observation of the MgH 2 -10 wt% Y 2 O 3 /NiO system as a representative is also conducted, as shown in Figure 3h. It is seen that the particle size of the system is not evidently different from that of the bare MgH 2 , which is performed for a same ball-milling regime ( Figure S1). The MgH 2 -10 wt% Y 2 O 3 /NiO system has a size range of 300-900 nm with most particles distributing between 500-700 nm. The spherical structure of the assynthesized Y 2 O 3 /NiO hybrid is not observed after the ball milling. Further TEM analysis, as presented in Figure 3i, shows image of dark-gray fragments in size of nanometers dispersing in light-gray matrix, which indicates that the fine dark-gray fragments in size around 100-200 nm are the Y 2 O 3 /NiO hybrid, and the matrix is MgH 2 . It is obtained that the hollow nanospheres of Y 2 O 3 /NiO are crushed into fine fragments and are well dispersed in the MgH 2 matrix. Figure 4a shows the TPD curves of the MgH 2 -x wt% Y 2 O 3 /NiO systems with that of the pristine MgH 2 for comparison. It is seen that a small amount addition of the Y 2 O 3 /NiO hybrid can significantly reduce the onset and the main dehydrogenation temperatures of MgH 2 . All the composite systems show a two-step dehydrogenation process, where the first step in the relatively low temperature range is the major one, which shows sharp peaks, and the subsequent step in the comparatively high temperature range is a minor one, which is partially overlapped with the major one, occurring as a shoulder peak. The onset peak dehydrogenation temperatures of the composite systems are around 220 • C, and the peak temperatures of the first step are in the range of 280-310 • C, where a higher addition of the hybrid of Y 2 O 3 /NiO in the range of 2-10 wt% corresponds to a lower peak temperature. Further increasing the addition of Y 2 O 3 /NiO to 12 wt% only results in a little reduction in the peak temperature. The second-step dehydrogenation occurs in the temperature range of 320-350 • C. In contrast, the bare MgH 2 releases H 2 with onset and peak dehydrogenation temperatures as high as 350 and 386 • C, respectively. As a representative, the MgH 2 -10 wt% Y 2 O 3 /NiO system shows low onset and peak dehydrogenation temperatures of ca. 220 and 279 • C, respectively, which are lowered by 130 and 107 • C, compared with those of the pristine MgH 2 (350 and 386 • C, respectively). The dehydrogenation peak temperature of the MgH 2 -10 wt% Y 2 O 3 /NiO system is ca. 144 • C and 156 • C lower than that of Y 2 O 3 and Y-doped MgH 2 systems [43], and the dehydrogenation kinetics is better than that of the Pt-doped MgH 2 system which releases less than 2 wt% of hydrogen at 275 • C for 200 min, while the amount of hydrogen released of the present system reaches 6.2 wt% under the same condition [49]. Figure 4b shows the non-isothermal dehydrogenation curves of the MgH 2 -x wt% Y 2 O 3 /NiO systems. A two-step dehydrogenation process is also shown, which is consistent with the result from the TPD result, where the first step is the main one and the second one is the minor one. The occurrence of the second step is supposed due to the less efficient catalytic effect on the Y 2 O 3 /NiO hybrid to MgH 2 , which may be due to the existence of large Y 2 O 3 /NiO particles that are heterogeneously mixed with MgH 2 . The main dehydrogenation process is completed at a close temperature range of ca. 260-300 • C for the systems with the addition of Y 2 O 3 /NiO. In addition, with the increasing amount of the Y 2 O 3 /NiO hybrid introduced, the system shows lower main ending dehydrogenation temperatures. As an overall, dehydrogenation is mostly completed upon heating to 350 • C for all the composite systems. In comparison, completing the dehydrogenation needs a high temperature over 400 • C for bare MgH 2 . Moreover, the main dehydrogenation curves move to lower temperature ranges as overall with the increasing amounts of the Y 2 O 3 /NiO hybrid introduced, except that the two of the systems with 10 and 12 wt% Y 2 O 3 /NiO additions are almost overlapped. Nevertheless, the former shows a slightly higher dehydrogenation capacity in the followed second step. Among the composite systems, the MgH 2 -10 wt% Y 2 O 3 /NiO one shows the highest hydrogen capacity of 5.2 wt% upon heating to 300 • C. Further increasing the dehydrogenation temperature results in more hydrogen release for the systems with less amounts of Y 2 O 3 /NiO addition as their higher theoretical hydrogen storge capacity. The system with 5 wt% Y 2 O 3 /NiO addition can release more amount of hydrogen compared with the others when the dehydrogenation temperature is over 320 • C, and a capacity of 6.8 wt% H 2 is achieved upon heating to 350 • C. The peak dehydrogenation temperatures of the systems obtained by making the differentiation of the dehydrogenation curves is highly consistent with the TPD results as shown in Figure S2a. The onset dehydrogenation temperature of the present MgH 2 doped with 10 wt% Y 2 O 3 /NiO is comparable with that of a Mn-doped MgH 2 system [50] and is 35 and 50 • C lower than those of the FeCoNi@GS doped [51] and HfCl 4 doped [52] MgH 2 systems, respectively. Moreover, the amount of hydrogen released upon heating to 300 • C of the present work (5.2 wt% H 2 ) is also much higher than that of the two works, where the FeCoNi@GS doped MgH 2 system only shows a dehydrogenation capacity less than 3 wt% at the same temperature. Figure 4c shows the hydrogenation curves of the systems which are pre-dehydrogenated to 410 • C, as noted in Figure 4b. It is seen that the hydrogenation temperatures of the composite systems are significantly lowered by the introduction of the Y 2 O 3 /NiO hybrid. For the composite systems, the hydrogen absorption takes place at temperatures lower than 50 • C. Especially for the systems with 10 and 12 wt% Y 2 O 3 /NiO additions, the hydrogen absorption takes place at room temperature, whereas for the bare MgH 2 , hydrogen absorption takes place at ca. 150 • C. The systems added with 5-10 wt% Y 2 O 3 /NiO can absorb almost the same amount of H 2 , 6.3-6.5 wt%, in the hydrogenation temperature range of 25-235 • C. The system with 2 wt% Y 2 O 3 /NiO shows somewhat higher capacity but, accordingly, a higher hydrogenation temperature. In addition, the system with 12 wt% Y 2 O 3 /NiO addition turns out to have a slightly lower hydrogenation temperature compared with the one with 10 wt% Y 2 O 3 /NiO, but the hydrogen absorption capacity of the former is lower than that of the latter. As a representative, a saturated value of 6.3 wt% H 2 can be absorbed for the MgH 2 -10 wt% Y 2 O 3 /NiO system when the temperature reaches 180 • C. Figure 4d shows the dehydrogenation curves of the hydrogenated systems as noted in Figure 4c. It is found that, overall, a higher amount of Y 2 O 3 /NiO introduced results in a lower dehydrogenation temperature for the system, which shows similar regulation as the initial dehydrogenation process. Moreover, the main dehydrogenation for the systems with different additions of Y 2 O 3 /NiO is in a larger temperature range compared with that of the initial one. The dehydrogenation curves of the systems with 10 and 12 wt% Y 2 O 3 /NiO almost overlap in the main dehydrogenation process. However, less amount of H 2 can be released in the following high temperature range for the latter. For the MgH 2 -10 wt% Y 2 O 3 /NiO system, a capacity of 5.0 wt% is obtained upon heating to 280 • C, where the main dehydrogenation is almost complete. The capacity is significantly higher than that of the first around dehydrogenation at 280 • C, where a capacity of only 3.1 wt% is obtained (Figure 4b). There is evident decrease in the onset, main, and ending dehydrogenation temperatures for the system with 10 wt% Y 2 O 3 /NiO for the second dehydrogenation compared with the data in the initial one, with onset, peak, and ending temperatures of ca. 215, 261, and 327 • C, respectively, as obtained from the differentiation of the second cycle dehydrogenation curve ( Figure S2b), lowered by 5, 16, and 19 • C compared with those of the initial dehydrogenation. Differences in the dehydrogenation temperature between the second and initial dehydrogenation are also observed for the other systems. The dehydrogenation temperature of the system with 5 wt% Y 2 O 3 /NiO addition for the second cycle is a little lower than that of the first cycle, while the capacity is much lower than that of the initial cycle. In addition, the second dehydrogenation temperature of the system with 2 wt% Y 2 O 3 /NiO is even a little higher than that of the first cycle. The reason for these differences on the dehydrogenation temperatures of the first and second cycles will be discussed in the following section.

Hydrogen Storage Performance of the MgH 2 -xwt%Y 2 O 3 /NiO Systems
Considering that the MgH 2 -10 wt% Y 2 O 3 /NiO system shows an overall favorable performance of low dehydrogenation and hydrogenation temperatures, less capacity loss and high reversibility, further kinetics and cyclic studies are performed on this system. For comparison, parallel tests are also performed for the bare MgH 2 , the results of which are also shown. Figure 5a shows the isothermal dehydrogenation curves of the MgH 2 -10 wt% Y 2 O 3 /NiO system as well as the bare MgH 2 at different temperatures. The record of the amount of H 2 released starts when the temperature reaches the designed one. It is seen that the hydrogen desorption rate of MgH 2 is greatly enhanced by the introduction of the Y 2 O 3 /NiO hybrid. When the temperature reaches 275 • C, 3.7 wt% H 2 is already released, and following a 15-min dwelling, a capacity of 6.0 wt% is reached. This isothermal property is comparable to that of Mn-doped MgH 2 system [50]. In addition, a total amount of 6.6 wt% H 2 is released when the isothermal period is further extended to 60 min. In contrast, there is almost no H 2 released for the bare MgH 2 upon heating to 275 • C, and there is only 1.2 wt% hydrogen released even after the isothermal period is extended to 160 min. At a lower isothermal temperature of 250 • C, a dehydrogenation capacity of 6.3 wt% H 2 is obtained for an isothermal period of 225 min for the MgH 2 -10 wt% Y 2 O 3 /NiO system, whereas there is less than 1 wt% H 2 released for the bare MgH 2 in the same isothermal period. Since much lower isothermal dehydrogenation temperatures are used for the present system than those used in the FeCoNi@GS doped [51] and HfCl 4 doped [52] systems, the reaction kinetics cannot be quantitatively compared, but the dehydrogenation kinetics of the present work is superior to that of the MgH 2 system catalyzed with Pt [49] and is also superior to that of the system catalyzed by TiC x @C [53] in terms of the same test temperatures of 275 and 250 • C, respectively.
Isothermal hydrogenation tests at different temperatures with a H 2 pressure of 5 MPa are also performed for the MgH 2 -10 wt% Y 2 O 3 /NiO system as well as the bare MgH 2 , which are pre-dehydrogenated upon heating to 410 • C as noted in Figure 4b. H 2 is loaded when the temperature reaches the designed one. The results are shown in Figure 5b. It is seen that the hydrogen absorption kinetics of MgH 2 is also significantly improved by the introduction of the Y 2 O 3 /NiO hybrid. At an isothermal temperature of 150 • C, a hydrogen absorption capacity of 4.2 wt% is achieved within 50 min for the MgH 2 -10 wt% Y 2 O 3 /NiO system, and a capacity of 5.9 wt% is achieved when further extending the isothermal period to 150 min. In addition, further extending the isothermal period to 200 min, there will be a slightly higher amount of H 2 absorbed, whereas over 200 min is needed for the bare MgH 2 to achieve a hydrogenation capacity of 4.4 wt%. Moreover, even at a low temperature of 100 • C, a hydrogen absorption capacity of 3.0 wt% can be reached for the MgH 2 -10 wt% Y 2 O 3 /NiO system within 240 min, whereas no obvious hydrogen uptake is observed for the bare MgH 2 . It is obtained that the intersected Y 2 O 3 /NiO hybrid exhibits highly effective bidirectional catalysis on both dehydrogenation and hydrogenation kinetics of MgH 2 . Figure 6 shows the cyclic dehydrogenation and hydrogenation capacities of the MgH 2 -10 wt% Y 2 O 3 /NiO system for 50 cycles. The initial dehydrogenation capacity is 6.5 wt%. A dehydrogenation capacity of 5.2 wt% is maintained for the MgH 2 -10 wt% Y 2 O 3 /NiO system after 50 cycles, corresponding to capacity retention of 79.2%. Close values for hydrogenation capacity are obtained for the same cycle, indicating high reversibility. The average dehydrogenation capacity decay for each cycle in the 50 cycles is only 0.026 wt%. The result demonstrates that the introduction of Y 2 O 3 /NiO also significantly improves the cyclic stability of MgH 2 , of which the hydrogen desorption capacity retention after 50 cycles was only 10.9% as shown in Figure S7. As different cyclic regimes are used in different research groups, the cyclic properties from different groups cannot be quantitatively compared. It is still necessary to present some data, such as the Mn-doped MgH 2 system shows the capacity retention of 89% after 20 cycles, maintaining a capacity of 5.72 wt% [49] at a cyclic regime of isothermal dehydrogenation and hydrogenation (3 MPa) at 275 • C. A capacity of 6.02 wt% remained after 20 cycles for the FeB/CNTs-doped MgH 2 system [54], corresponding to the capacity retention of 93% with a cyclic regime of isothermal dehydrogenation at 300 • C (vacuum) and isothermal hydrogenation at 150 • C (5 MPa of H 2 ).

Discussion
To investigate the mechanism of the improved hydrogen storage performance of MgH 2 by the introduction of the intersected Y 2 O 3 /NiO hybrid, the dehydrogenation apparent activation energy (E a ) and the dehydrogenation reaction enthalpy (∆H des ) of the MgH 2 -10 wt% Y 2 O 3 /NiO system are evaluated based on the Kissinger's method and DSC measurement, respectively. Figure 7a shows the Kissinger's plot obtained by fitting the TPD-MS curves of different heating rates, where the E a value of the MgH 2 -10 wt% Y 2 O 3 /NiO system is calculated to be 86 ± 5 kJ·mol −1 , which is 46% lower than 160 ± 10 kJ·mol −1 for the bare MgH 2 . The decreased E a value indicates a lowered dehydrogenation energy barrier of MgH 2 by the introduction of the Y 2 O 3 /NiO hybrid, contributing to the improvement of dehydrogenation kinetics. Figure 7b shows the Kissinger's fitting plot of the MgH 2 -10 wt% Y 2 O 3 /NiO system after 50 cycles. It is seen that the E a value of the composite system only shows a slight increase of 3 kJ·mol −1 after cycling, which is extremely small and smaller than 6 kJ·mol −1 for the bare MgH 2 . The result indicates a sustainable catalysis of the Y 2 O 3 /NiO hybrid on the kinetics of MgH 2 .  Figure 7c shows the DSC curve of the MgH 2 -10 wt% Y 2 O 3 /NiO system. There is a strong endothermic peak observed at 278 • C, ascribed to the dehydrogenation of MgH 2 . By integrating the peak area, the corresponding ∆H des value is calculated to be 71.9 kJ·mol −1 H 2 , which is close to 73.2 kJ·mol −1 H 2 of the bare MgH 2 . This indicates that the introduction of the Y 2 O 3 /NiO hybrid almost does not change the thermodynamic behavior of MgH 2 . It is concluded that the improvement of the hydrogen storage performance of the catalytic system is mainly attributed to kinetic factor.
The XRD patterns of the MgH 2 -10 wt% Y 2 O 3 /NiO system at different dehydrogenation and hydrogenation stages are shown in Figure 8. It is found that MgH 2 and Mg phases interconvert accompanying with the hydrogenation and dehydrogenation processes, while the phases of Y 2 O 3 and NiO are always detected in the products either after the first dehydrogenation or after 50th dehydrogenation with no other new phase detected. The result indicates that Y 2 O 3 and NiO almost maintain stable in crystal structure during cycling instead of being reduced, which is consistent with the results presented of previous reports [47,55]. Moreover, chemical state analyses of the elements of Y and Ni in the dehydrogenation and hydrogenation stages of the system identified by XPS ( Figure S3) show that the chemical states of Y and Ni also remain as Y 2 O 3 and NiO, respectively, for their as-milled sates and the dehydrogenation/hydrogenation states of the first and second cycles. Further HRTEM analysis of the dehydrogenation product of the system performed by the regime as noted in Figure 4b shows that there are lattice planes of (200) plane of Y 2 O 3 and (101) plane of NiO vaguely observed, which still shows the intersected structure ( Figure S4). Although the image is not in a high resolution, as the Y 2 O 3 /NiO hybrid is high-energy ball-milled and tightly connected with the dehydrogenation product of MgH 2 (Mg), it still can be obtained that the intersected structure is well inherited to the dehydrogenation product. Combining with the results of the aforementioned XRD analysis, it is obtained that the intersected Y 2 O 3 /NiO hybrid does not participate in the dehydrogenation and hydrogenation reactions of MgH 2 , maintaining the states of Y 2 O 3 and NiO. Comparison of the particles size of the systems with low (2 wt%) and high (12 wt%) amount of Y 2 O 3 /NiO additions as well as the bare MgH 2 at as-milled and initial hydrogenation sates by SEM observation ( Figure S5) demonstrates that there is severe particle growth and agglomeration for the bare MgH 2 and for the composite system with low amount (2 wt%) Y 2 O 3 /NiO addition, whereas there is almost no particle growth for the system introduced with a high (12 wt%) amount of Y 2 O 3 /NiO, and it seems that there is even a slight particle size decrease. It is reported that cracks are generated in the oxide of MgO during annealing and hydrogenation processes, providing more surface area and active sites and resulting in more pathways for hydrogen diffusion. As a result, the second dehydrogenation temperature is reduced compared with the initial one [56]. In the present work, it is possibly that there are some cracks generated in the Y 2 O 3 /NiO particles, especially for those with large size. Agglomeration and particle growth commonly occur for bare MgH 2 after initial dehydrogenation, thus degrading the dehydrogenation performance. The low amount of addition of the Y 2 O 3 /NiO hybrid is insufficient in preventing the agglomeration and particle growth. As a result, the second dehydrogenation temperature is higher than the initial one. Adding a high enough amount of the Y 2 O 3 /NiO hybrid can effectively prevent the agglomeration and particle growth of MgH 2 , with the extra increased surface area, active sites, and therefore, the second dehydrogenation temperature is reduced. Figure 9 shows the SEM images of the MgH 2 -10 wt% Y 2 O 3 /NiO system at the dehydrogenation states of the initial cycle (a) and after 50 cycles (b). No obvious change in particle size is observed after the initial dehydrogenation ( Figure 9a) compared with its as-milled sate (Figure 3h), whereas it is common that the particles of bare MgH 2 are severely agglomerated after dehydrogenation and hydrogenation cycles, as reported in literature [57,58]. This is also one of the main reasons for the capacity decay of MgH 2 during cycling. Figure 9c shows the TEM images of the dehydrogenation product of the MgH 2 -10wt% Y 2 O 3 /NiO system after 50 cycles. It is seen that the dark-grey particles, which correspond to the Y 2 O 3 /NiO hybrid, are still uniformly distributed on the brightgrey matrix (Mg). It is obtained that either the spatial structure or the crystal structure of the Y 2 O 3 /NiO particles is mostly preserved during cycling. The Y 2 O 3 /NiO particles act as a barrier to inhibit the growth and agglomeration of MgH 2 /Mg particles during cycling, favoring the cyclic stability. To further understand the role of the intersected Y 2 O 3 /NiO hybrid on the hydrogen storage of MgH 2 , the dehydrogenation and hydrogenation properties of the MgH 2 -10 wt% Y 2 O 3 /NiO system are further compared with two reference systems, which are added with 10 wt% of commercial Y 2 O 3 and NiO, respectively, prepared by the same ball-milling regime, which are performed for non-isothermal dehydrogenation and hydrogenation. The result ( Figure S6) shows that the MgH 2 -10 wt% Y 2 O 3 /NiO system has superior dehydrogenation and hydrogenation properties to the two reference systems, demonstrating that there is a synergistic catalytic effect between Y 2 O 3 and NiO on the hydrogen storage of MgH 2 .
To sum up, the introduction of the intersected Y 2 O 3 /NiO hybrid into MgH 2 significantly reduces the de-/hydrogenation temperatures and enhances the hydrogen sorption rate. As the intrinsic high hardness of the metal oxides of Y 2 O 3 and NiO, and also the hollow structure of the hybrid with thin thickness of the shell, the hybrid serves as auxiliary abrasive agents reducing the particle size of MgH 2 during ball milling after it is crushed at the initial milling process. Thus, the particle size of MgH 2 is significantly reduced by the introduction of Y 2 O 3 /NiO, which is much smaller than that of the bare MgH 2 performed by the same milling process. The crushed Y 2 O 3 /NiO fragments are further ground to nanosized particles and uniformly dispersed in the MgH 2 matrix, acting as highly effective bidirectional catalyst to both dehydrogenation and hydrogenation of MgH 2 .
From the present experiments, the amount of 10 wt% is suggested to be an optimized value for the addition of the intersected Y 2 O 3 /NiO hybrid, as obtained from Figure 4, for the dehydrogenation and hydrogenation of MgH 2 -x wt% Y 2 O 3 /NiO systems. The catalytic effect is insufficient for the less amount of Y 2 O 3 /NiO introduction, and too much amount of Y 2 O 3 /NiO introduction results in obvious decrease in hydrogen storage capacity. The crystal structure and morphology of the Y 2 O 3 /NiO hybrid are almost maintained during ball milling process and after long-term cycling (for 50 cycles, as shown in Figure 6). The stable structure of the Y 2 O 3 /NiO hybrid provides a long-term catalytic effect on the hydrogen storage of MgH 2 . In terms of kinetic properties, as reported [40], the dissociation of the molecular hydrogen adsorbed on the surface of the metal oxide catalysts can be facilitated by the high-density defects introduced on the surface of hard oxides by ball milling, thus proving superior catalysis.

Conclusions
A novel intersected Y 2 O 3 /NiO hybrid with the morphology of hollow nanospheres is prepared, which shows bidirectional catalysis for MgH 2 hydrogen storage. The dehydrogenation temperature is significantly reduced by the introduction of the hybrid, where an increasing reduction is obtained with the increasing addition of the hybrid from 2 to 10 wt%, but the reduction becomes limited when the addition is further increased to 12 wt%. With an addition of 10 wt% Y 2 O 3 /NiO hybrid, MgH 2 shows an overall optimized hydrogen storage performance. It initiates the dehydrogenation at 220 • C and reaches the peak at 277 • C for the first dehydrogenation, and notably, the second desorption peak temperature is further reduced to 261 • C, which is lower by 125 • C compared with that of the bare MgH 2 , and 6.6 wt% H 2 is desorbed at 275 • C within 60 min. The hydrogenation takes place at almost room temperature, and 4.2 wt% H 2 can be absorbed at 150 • C within 50 min. The apparent activation energy of dehydrogenation of MgH 2 is significantly lowered by the addition of the hybrid, which contributes to the improved reaction kinetics. A dehydrogenation capacity of 5.2 wt% is maintained after 50 cycles at a moderate cyclic condition, corresponding to the capacity retention of 79.2%. The crystal structure and distribution of the Y 2 O 3 /NiO hybrid are almost maintained during cycling, and the hybrid inhibits the agglomeration of MgH 2 particles, which all contribute to the favorable cyclic stability of MgH 2 .
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/pr9050892/s1, Figure S1: SEM image of the as-milled bare MgH 2 . Figure S2: The differentiation of the dehydrogenation curves of the MgH 2 -10 wt% Y 2 O 3 /NiO system of (a) the first dehydrogenation and (b) the second dehydrogenation. Figure S3: XPS spectra of Y 3d and Ni 2p of the products of the ball-milled (BM) sate (a,b), the 1st dehydrogenation (c,d), the 1st hydrogenation (e,f), and the 2nd dehydrogenation (g,h) of the MgH 2 -10 wt% Y 2 O 3 /NiO system. The products are by the regime as noted for Figure 4. Figure S4: HR-TEM image of the dehydrogenated product of the MgH 2 -10wt% Y 2 O 3 /NiO system, performed as Figure 4b. Figure S5: SEM images of the MgH 2 -12 wt% Y 2 O 3 /NiO system of as-milled (a) and hydrogenated (b); the MgH 2 -2 wt% Y 2 O 3 /NiO system of as-milled (c) and hydrogenated (d); the bare MgH 2 of as-milled (e) and hydrogenated (f). Dehydrogenation and hydrogenation are performed as noted in Figure 4. Figure S6: Comparison of the hydrogen desorption (a) and absorption (b) curves of the MgH 2 -10 wt%Y 2 O 3 /NiO system with MgH 2 + 10 wt%Y 2 O 3 and MgH 2 + 10 wt%NiO systems. Figure S7: Cyclic dehydrogenation and hydrogenation capacities of the pure MgH 2 with a cyclic regime of dehydrogenation at 320 • C for 60 min and hydrogenation at 320 • C and 5 MPa H 2 for 120 min.