D.Pukzhselvn ,K.S.Snhy ,Devrj Rmsmy ,Aliksnr Shul ,Igor Bikin ,Duncn Pul Fgg

a Department of Mechanical Engineering,Nanoengineering Research Group,Centre for Mechanical Technology and Automation (TEMA),University of Aveiro,Aveiro 3810-193,Portugal

b Department of Chemistry,Malankara Catholic College,Kaliyakkavilai,Tamil Nadu 629153,India

c Department of Computational Biology and Bioinformatics,University of Kerala,Thiruvananthapuram,Kerala 695581,India

d UIDM,ESTG,Polytechnic Institute of Viana do Castelo,Viana do Castelo 4900,Portugal

Abstract This study explores how zirconia additive interacts with MgH2 to improve its hydrogen storage performance.Initially it is confirme that the zirconia added MgH2 powder releases hydrogen at a temperature of about 50 °C below that of the additive free MgH2.Subsequent tests by X ray diffraction (XRD) and infrared (IR) spectroscopy techniques reveal that the ZrO2 mixed MgH2 powder contains ZrHx (2 ＜x＞1.5)and MgO secondary phases.This observation is supported by the negative Gibbs free energy values obtained for the formation of ZrH2/MgO from ZrO2/MgH2 powder samples.An X ray photoelectron spectroscopy (XPS) study reveals that apart from Zr4+cations,Zr2+ and zero valent Zr exist in the powder.Atomic force microscopy (AFM) study reveals that the average grain size is 20 nm and the elemental line scan profile further proves the existence of oxygen deficien Zr bearing phase(s).This study strengthens the belief that functional metal oxide additives in fact chemically interact with MgH2 to make active in-situ catalysts in the MgH2 system.

Keywords: Hydrogen storage;Binary hydrides;Metal oxides;Additives;Nanocatalysis.

1.Introduction

The reversible binary hydride MgH2is of considerable interest for vehicular applications,mainly due to its appreciable hydrogen storage capacity (7.6 wt.%) [1].One serious drawback to the commercial MgH2sample is that dehydrogenation requires a high temperature,typically over 350°C,that is much higher than the temperature target (85 °C) for H2fuel cell vehicles [2,3].Nonetheless,it is encouraging that lowering the temperature is possible,by at least a 50 °C,upon employing an appropriate metal oxide additive[4].Oelerich et al.used the oxides of Sc,Ti,V,Cr,Mn,Fe,Cu,Al,and Si and found that a mere 1 wt.% of the best additive (Fe3O4/V2O5)substantially enhances the dehydrogenation kinetics of MgH2at 300 °C [5].Polanski et al.used nanoparticles of Cr2O3,TiO2,Fe3O4,Fe2O3,In2O3and ZnO as additives,and found that except for ZnO all other oxide additives significantl reduce the dehydrogenation activation energy of MgH2[6].Further investigation by a few other researchers has brought to light that Nb2O5is the best additive in the category of metal oxides [7].As per the report of Barkhodarian et al.,Nb2O5additive loaded MgH2releases over 6.5 wt.%hydrogen within 2 min(300°C under vacuum)and reabsorbs the same amount within a minute (300 °C at 8.4 bar pressure) [8].After these observations,there has been considerable research interest for understanding the mechanistic role of Nb2O5on increasing the H sorption performance of MgH2.Although contradictory conclusions have been made,various teams have provided interesting mechanistic insights.Aguey-Zinsou et al.suggested that Nb2O5additive remains intact and the role of Nb2O5is to facilitate cracks and higher active contact surfaces [9].On the other hand,further verificatio explored the existence of metallic Nb spots and reduced Nb oxides in the ball milled powder samples [10-13].These observations confirme that the Nb2O5additive indeed gets reduced by MgH2.Previous studies of the current team explored the existence of Nb dissolved MgO rock salt structure in the Nb2O5loaded MgH2powder and highlighted that these formed phases actively assists hydrogen sorption kinetics [14,15].In similar lines with Nb2O5loaded MgH2system,our consecutive works on TiO2loaded MgH2confirm the formation of Ti dissolved MgO and a significantl enhanced catalytic activity[16,17].Regarding the bond breaking mechanism,it was understood from the theoretical investigation of Sandhya et al.[18] that the Ti dissolved rock salt (MgxTiyOx+y),unlike undoped MgO rock salt,is very reactive to its chemical proximity due to the d orbital free electrons incorporated through the Ti substituent.In agreement with this,another theoretical investigation performed by Zhang et al.[19] for a few transition metals (Ti,V,Fe,Co,Ni and Nb) substituted MgO suggest that MgxTiyOx+y/MgxNbyOx+yrock salts are the best for destabilizing MgH2.The valence charge density analysis performed in this study suggests that the electron clouds from O edges of Ti/ Nb metal substituted MgO strongly engages with the Mg of MgH2which in effect decreases the bond strength between magnesium and hydrogen.Such an Mg-H bond breaking effect was not observed for clean MgO wrapped MgH2configuration Therefore,as for as titania added MgH2is concerned,it is currently understood that reduction of titania and the subsequent formation of Ti dissolved MgO is a key step in the process of catalysis.

In continuation to our previous studies on Nb2O5and TiO2additives incorporated MgH2,our current interest is ZrO2incorporated MgH2.In the literature very few published works are available for ZrO2loaded MgH2for hydrogen storage.Zhang et al.reported that an MgH2sample milled using zirconia vial/balls provide better results as compared to a sample prepared using steel milling medium under identical sample processing conditions.It was concluded that the zirconia impurities incorporated through the milling medium were responsible for the observed improved performance[20].Another study performed by Chen et al.agrees with this conclusion and further suggests that zirconia improves the surface interaction by acting as a size refinin agent for MgH2nanocrystals [21].In another study,Hwang et al.showed that ZrO2behaves as an excellent co-catalyst with single walled carbon nanotubes for lowering the hydrogenation activation energy of magnesium nanocrystals [22].Although these studies reveal the possibility of using ZrO2as another useful additive for MgH2,mechanistic details were not adequately discussed.Therefore,owing to the increasing interest for exploring how metal oxides interact with MgH2,our current research is focused on the chemical interaction between MgH2and ZrO2.Generally,after studying a few metal oxide additives,many researchers highlight only the best performing oxide and limit the subsequent mechanistic studies to only the best additives.In our opinion,in order to fully explore what makes an oxide additive better over other oxides it is necessary to understand the reason(s) pertaining to why different oxides offer different levels of performance.Moreover,there are chances that a moderately active oxide additive can undergo slow chemical variations over several cycles,while,in a certain stage,converts to exhibit a surprising behavioral change as a promising catalyst [23].Therefore it is necessary to make detailed studies regarding the interaction of a wide variety of oxide additives with MgH2.In this context,the authors of the current article make clear that our aim is not to suggest ZrO2as peak performing additive over that of the existing ones,but instead to provide valuable insights to assist understanding of the mechanism of oxides incorporated MgH2.In this respect,for a detailed overview regarding a general mechanism of oxide additives loaded hydrogen storage systems the reader is redirected to our recent chapter published elsewhere [4].

In this work,through XRD investigation we reveal that the MgH2+xZrO2powder prepared under strong mechanical milling conditions contains zirconium hydride.It is further proved that ZrH2containing powder samples improve the hydrogen sorption performance of MgH2over that of a pure ZrO2additive.The samples were tested by differential scanning calorimetry,X ray diffraction,Infrared spectroscopy,X ray photoelectron spectroscopy,scanning electron microscopy and atomic force microscopy techniques.Appropriate discussions are presented to consolidate the current observations with the literature observations for exploring a general hydrogen storage mechanism for metal oxide additives added MgH2system.

2.Experimental

Fig.1.Differential scanning calorimetry profile of 10 wt.% ZrO2 mixed MgH2 and the additive free ball milled MgH2 sample.

All sample precursors were procured from Merck chemicals ltd.MgH2+xZrO2samples (x,as per the required concentration) were ball milled using a planetary type mechanical milling facility,model:Retsch PM200.A stainless steel milling medium with a ball to powder weight ratio of 75:1 was employed.The ball milling time was fi ed between 5 and 20 h with the speed of 200 and 300 rotations per minute(rpm) as required for the specifi experiment (explained in detail in the results section).The powder processing was always performed inside an Ar fille glove box.The ball milled samples were immediately tested by X ray diffraction (XRD)technique using a Rigaku X-ray diffractometer operating with CuKαradiation,λ=1.541 °A.All the XRD samples were sealed by kapton tape inside the glove box for preventing air induced contamination during measurements.Differential scanning calorimetry (DSC) tests were conducted by using NETZSCH STA 449 F3 thermal analyzer.Pure alumina was used as the reference sample for calorimetry and Ar with a fl w rate of 50 ml/min was used as the carrier gas during measurements.Oxidation state analysis for Zr was performed by X ray photoelectron spectroscopy (XPS) technique.XPS spectra were acquired under ultra-high vacuum (2 × 10-10mbar) chamber which is attached with a hemispherical electron energy analyzer (SPECS Phoibos 150),a delay-line detector and a monochromatic AlKα(1486.74 eV)X-ray source.High resolution spectra were recorded at normal emission take-off angle with a pass-energy of 20 eV,which provides an overall instrumental peak broadening of 0.5 eV.Fourier transform infrared (FT IR) spectra were acquired (KBr method)with a Shimadzu FTIR 8400 spectrometer operating with an excitation laser wavelength of 314 nm.Hydrogen desorption/absorption measurements were performed by volumetric technique using an automated Sieverts apparatus built in the laboratory.The Gibbs free energy values were calculated with the help of HSC Chemistry 10 thermodynamic data calculator package.Microstructural characterization was performed by using a tabletop scanning electron microscope (SEM) facility,Hitachi TM4000.Scanning was performed with 15 kV fiel energy at the back scattered electron (BSE) imaging mode.The elemental chemical mapping (line scanning) was performed with the chemical analyzer (Bruker Energy Dispersive X Ray Spectrometer (EDS):QUANTAX 75/ 80) attached with the SEM facility.For morphological studies,we used an atomic force microscope (AFM) facility,Nanoscope IV MMAFM-2.For AFM study,samples were prepared by dropping and drying a sample dispersed solution on a clean mica base.The solution was prepared by dispersing a small amount of sample in pure ethanol solution with a sonication treatment for 10 min.

3.Results and discussion

The differential scanning calorimetry(DSC)profile shown in Fig.1 illustrate the effect of incorporation of a small amount of zirconia with MgH2.The sample was prepared by mixing a 10 wt.% of ZrO2additive with MgH2for 5 h milling time at the milling speed of 200 rpm (sample “a”).A reference sample,which is an additive free clean MgH2,was ball milled for 5 h at 200 rpm under similar experimental conditions and the corresponding DSC profil is indexed as sample “b”.As seen,with respect to sample “b”,the endothermic dehydrogenation occurs at about 50 °C lower temperature in the case of sample “a”.This suggests that ZrO2is a good additive for promoting the low temperature dehydrogenation of MgH2.After confirmin this effect,our interest turns to explore the interaction between MgH2and zirconia.Making a detailed study for understanding this interaction is vital,not only for exploring the catalytic dehydrogenation mechanism of zirconia loaded MgH2,but also for consolidating a general catalytic mechanism for metal oxide additives loaded MgH2systems.

Fig.2.(a) X ray diffraction profile correspond to (bottom to top),10 wt.% ZrO2 mixed as-prepared MgH2 sample (blue),after dehydrogenation (pink),after firs cycle hydrogenation (black) and after 5th cycle hydrogenation (red).Inset:The maximum intensity peak regions for MgH2 and ZrO2,(b) Gibbs free energy profil calculated for reaction (2) over room temperature to 600 °C (For interpretation of the references to color in this figur legend,the reader is referred to the web version of this article).

Fig.3.X ray diffraction profile correspond to MgH2+xZrO2 (x=0.125,0.17,0.25 and 0.5) ball milled for 20 h at 200 and 350 rpm milling speed.Bottom set/Cyan-blue profile correspond to 200 rpm samples and top set/red profile correspond to 350 rpm samples.XRD of ZrO2 is given for comparison (For interpretation of the references to color in this figur legend,the reader is referred to the web version of this article).

The 10 wt.%zirconia mixed MgH2test sample was further characterized by XRD technique (generally a little amount of additive,typically less than 5 wt.%,is preferred for hydrogen storage studies because the added weight should be small enough for maintaining the higher gravimetric hydrogen density of the system.On the other hand,as far as the characterization studies are concerned,with small amounts of additive concentration it is difficul to explore the existence ofin situgenerated catalytic phases.This is particularly true when the reacted portion is much smaller as compared to the major phases existing in the powder sample).The XRD profil of the 10 wt.% ZrO2mixed MgH2sample (as prepared)is shown in Fig.2a (blue profile) This sample was dehydrogenated at 300 °C/1 bar and the subsequent XRD profil was recorded (pink profile) The dehydrogenated sample was then hydrogenated at 300 °C/5-6 bar and XRD profile were recorded for two different cycles,cycle 1 (black profile and cycle 5 (red profile as illustrated in Fig.2a.(The maximum intensity peak positions for MgH2and ZrO2are very close,therefore for clarity the zoomed max intensity peak regions are provided in the inset).In these samples,along with the expected Mg/MgH2peaks,peaks corresponding to chemically unchanged zirconia were also observed and no clear evidence regarding the existence of any other Zr bearing phases can be identified This is in contrast to our previous XRD study for TiO2added MgH2which suggested that TiO2reduces by interacting with MgH2during dehydrogenation at 300 °C [16].This is understandable because zirconia is known as a stable oxide as compared to titania,as is proven by experiments under various circumstances in heterogeneous catalysis [24-26].Nonetheless,we cannot ignore the possibility that if the reacted portion of the additive is too small and the phases incorporated through additive-hydride interaction are too little with respect to the resolution limit of our testing facility,catalytically active Zr containing traces other than ZrO2may exist in the powder but remain unidentified This possibility gains ground as the Gibbs free energy values for a potentially possible chemical reaction shown in reaction (1) for a wide range of ′n′values is negative (see Fig.2b and Table 1).

Table 1 Summary of the calculated Gibbs free energy and enthalpy values at two different temperatures,100 °C and 300 °C for the reaction,nMgH2+ZrO2→2MgO+ZrH2+(n-1)H2+(n-2)Mg (where n=2,4,6 and 8).

As for example,1/n=0.02 closely correspond to MgH2with 10 wt.% of ZrO2.In this case,the derived chemical reaction is,

The Gibbs free energy profil obtained for reaction (2)within the temperature range,30-600 °C is provided in Fig.2b.The observed negative values clearly suggest that formation of zirconium hydride and MgO through the chemical interaction between MgH2and ZrO2is indeed possible.In view of this,since a further deeper XRD test is necessary,we prepared the further set of samples,MgH2+xZrO2(x=0.125,0.17,0.25 and 0.5).For preparing these samples we increased the milling time to 20 h instead of 5 h,while retaining the milling speed,200 rpm.

The recorded XRD profile corresponding to the samples,MgH2+0.125ZrO2,MgH2+0.17ZrO2,MgH2+0.25ZrO2and MgH2+0.5ZrO2(all,20 h at 200 rpm)are shown in Fig.3(indexed as “b”,“c”,“d” and “e”,respectively.The profil “a”corresponds to as-received pure ZrO2).Among these samples,MgH2+0.125ZrO2,MgH2+0.17ZrO2,MgH2+0.25ZrO2exhibit nearly identical phase structural features and there is no clear evidence regarding the formation of any other Zr bearing phase(s).On the other hand,the sample MgH2+0.5ZrO2(profil “e”) shows a broad shoulder at 32.2° (highlighted by pink rectangle box),signifying that there may bein-situformation of a new nanophase in this sample.The occurrenceof this phase only in the MgH2+0.5ZrO2sample may be highlighting the fact that higher direct contact surfaces between MgH2and ZrO2is important for facilitating a chemical interaction between MgH2and ZrO2.We reached this viewpoint because the main factor that differentiates sample MgH2+0.5ZrO2from the other samples is the abundance of zirconia that can facilitate larger contact surfaces with MgH2.In this context,increasing the milling intensity seems to be a good strategy for promoting interaction between ZrO2and MgH2for all samples because higher fragmentation and dispersion of particles can be achieved by intense milling [27].With this understanding,we increased the milling speed from 200 to 350 rpm for the MgH2+xZrO2(x=0.125,0.17,0.25 and 0.5) samples for 20 h and again tested the phase structural features by the XRD technique.The observed diffraction patterns corresponding to the powder samples MgH2+xZrO2(x=0.125,0.17,0.25 and 0.5) (20 h,350 rpm) are indexed as patterns,“f”,“g”,“h” and “i” in the figure It is interesting that the broad shoulder that we observed for profil “e” at 32.2° develops as a prominent peak in all samples.Moreover,in all these samples we can notice MgO,which strengthens our belief that the reduction reaction of ZrO2by interaction with MgH2may be the most important event in these samples.Careful peak profil analysis with all the known Mg/Zr/O/H bearing phases suggest that the phase exhibiting the maximum intensity peak at 32.2 in profile “e”,“f”,”g” and “h”corresponds to cubic ZrH1.5(a=b=c=4.76 °A,space group:m-3 m,DB No.04-008-1383).On the other hand,the zirconium hydride existing in the sample indexed by “i”correspond to a tetragonal structured ZrH1.81with lattice parameter,a=b=3.491 °A andc=4.506 °A (space group:I4/mmm,DB No.00-036-1340).Upon comparing profil “e”with “i” one can understand that,not only hydrogenation of zirconium occurs,but also a phase transformation of hydrogenated zirconium occurs depend upon the concentration of hydrogen.The best identifie transformation in this study is from cubic ZrH1.5to tetragonal ZrH1.81.However,we believe that existence of other ZrHxtraces with hydrogen concentration between 1.5 and 2 per Zr atom cannot be ruled out because a number of ZrHxphases ranging fromx=1.5 to 2 is known to give reflection with maximum intensity peak centered at～32.2°[28-31].

The Fourier transform infrared (FTIR) spectra corresponding to the samples,MgH2+0.5ZrO2reacted at 200 and 350 rpm (codes:“a1” and “a2”);MgH2+0.25ZrO2reacted at 200 rpm and 350 rpm (codes:“b1” and “b2”);MgH2+0.175ZrO2reacted at 200 and 350 rpm (codes:“c1”and “c2”);MgH2+0.125ZrO2reacted at 200 and 350 rpm(codes:“d1” and “d2”) are shown in Fig.4.In these samples one can see a series of bands in the low frequency region,typically between 425 and 1200 cm-1.This is because several vibration modes from ZrO2and MgO show peak signatures at this frequency range.For instance,ZrO2is known to exhibit peak positions at 445,410,500,572,740,1104 and 1187 cm-1[32,33] whereas MgO exhibits peaks at 424,652 and 836 cm-1due to stretching vibrations [34-36].Among the ZrO2peaks,the well discernible intense peaks positioned at 572 cm-1is due to the Zr-O stretching vibration and the peak at 740 cm-1is due to Zr-O2-Zr bond stretching vibrations.The broad peak at 3430 cm-1and the one at 1632 cm-1can be attributed to the existence of adsorbed water.Adsorption of water,CO2(peak at 2350 cm-1) and other minor organic impurities (e.g.C-H vibrations [37] observed at 2849 and 2918 cm-1) are known to be due to the sample processing conditions by KBr method.However,the work of Coutant et al.[38],Quignard et al.[39] and Andrews et al.[40]brings evidence that Zr-H stretching vibrations also show a peak at 1632 cm-1.Proceeding with these observations,in order to know further information regarding the chemical states of Zr in the ZrO2/MgH2powder,two test samples,ZrO2+0.5MgH2and ZrO2+0.25MgH2(ball milled at 350 rpm,20 h) were studied by X-ray photoelectron spectroscopy (XPS) technique.The XPS Zr 3d5/2 and 3d3/2 peak profile corresponding to these samples are compared with the clean ZrO2sample in Fig.5 (profil “a” ZrO2,profil “b”ZrO2+0.25MgH2and profil “c” ZrO2+0.5MgH2).For simplifying the identificatio of the oxidation states,the standard 3d5/2 position corresponding to Zr0,Zr2+and Zr4+is highlighted by vertical red lines in the figure It is clear that in samples,ZrO2+0.25MgH2and ZrO2+0.25MgH2,apart from Zr4+cations,clear indication for the existence of Zr in lower valence states can be identified Interestingly,apart from the expected Zr2+cations,evidence for the zero valent Zr can also be identifie in both the samples.Especially in the case of ZrO2+0.25MgH2sample,a clear XPS signal representing the existence of metallic Zr can be identified This is in good agreement with the ZrHxphases observed by the XRD study.The existence of oxygen deficien and/or oxygen free Zr bearing regions in the samples,MgH2+0.25ZrO2and MgH2+0.5ZrO2(both samples,20 h/350 rpm),is further tested by recording the EDS line scan profiles The line scan profile corresponding to MgH2+0.25ZrO2,MgH2+0.5ZrO2and a reference sample ZrO2are shown in Fig.6.In these samples,the line scan performed for a surface cross sectional distance up to 62 micrometers is indicated by the yellow horizontal line.The three observed elements correspond to:magnesium (green),zirconium (violet) and oxygen (red).Any spikes in the Zr profile with respect to oxygen profile over the scanned length represents the existence of oxygen deficien or oxygen free Zr regions in the sample.As seen,a number of such regions are notable for both the samples MgH2+0.25ZrO2and MgH2+0.5ZrO2(e.g.,the corresponding particles embedded in the positions towards the line distance,19-20,38 and 42 for MgH2+0.25ZrO2and 9-12,17.5,27.5,45-48 and 51-53 for MgH2+0.5ZrO2.All the provided values are at positions noted from left to write,in μm scale).

Fig.4.FT-IR profile corresponding to samples MgH2+xZrO2 (x=0.125,0.17,0.25 and 0.5).Samples reacted at 200 rpm are given by the index,d1,c1,b1 and a1 and the ones reacted at 350 rpm are given the index,d2,c2,b2 and a2.

Fig.5.X ray photoelectron spectroscopy profile correspond to two test samples,MgH2+0.50ZrO2 and MgH2+0.25ZrO2,both milled at the speed of 350 rpm for 20 h,compared with ZrO2 starting sample.The standard positions for Zr0,Zr2+ and Zr4+ cations are highlighted by vertical red lines (For interpretation of the references to color in this figur legend,the reader is referred to the web version of this article).

Fig.6.EDS line scan profile obtained for the samples MgH2+0.25ZrO2 (bottom) MgH2+0.5ZrO2 (middle).The line scan profile correspond to zirconia(top) is provided for reference.

In order to visualize the existence of small sized particles in our samples,we performed an atomic force microscopy study (AFM) for two test samples (i.e.ZrO2+0.5MgH2samples milled at 200 and 350 rpm).The observed AFM pictures with both 3D and 2D projections are shown in Fig.7.The profil indexed as “a” corresponds to ZrO2+0.5MgH2milled at 200 rpm and “b” corresponds to ZrO2+0.5MgH2milled at 350 rpm.For comparison,the pure ZrO2starting sample was also studied and the corresponding AFM profil is indexed as “c”.As seen,the size of ZrO2is over 200 nm but its interaction with MgH2leads to creation of considerably smaller particles.The AFM 3D and the corresponding 2D images shown as “a” suggest that after milling the sample ZrO2+0.5MgH2at 200 rpm for 20 h,the size of particles significantl drops to the range of 25-100 nm.On the other hand,AFM profile of the powder milled at 350 rpm proves the existence of particles of much smaller size,roughly within a range of 10-50 nm.This size evolution is interesting in the context that the mechanochemically activated interaction between ZrO2and MgH2also results in the creation of higher contact surfaces between additive and MgH2.

Fig.7.the atomic force microscopy images corresponding to MgH2+0.5ZrO2 sample ball milled for 20 h at the speed of (a) 200 rpm and (b) 350 rpm.Profil (c) correspond to the ZrO2 starting sample.

Fig.8.Comparison of DSC profile of MgH2 containing a 3 wt.% of additives,such as (x) 20 h milled MgH2+0.5ZrO2 at the speed of 350 rpm and (y)clean ZrO2.

In order to understand the significanc of interaction between MgH2and ZrO2on the low temperature hydrogen storage behavior of zirconia loaded MgH2,we decided to use a small amount (3 wt.%) of MgH2+0.5ZrO2powder (reacted at 350 rpm/20 h) as additive for MgH2.The additive was mixed with MgH2by ball milling for 5 h at the speed of 200 rpm.The DSC profil recorded for this sample,indexed as “x” is compared with the DSC profil corresponding to 3 wt.% of clean ZrO2mixed MgH2(indexed as “y”) in Fig.8.As we can see,the dehydrogenation occurs at slightly lower temperature in “x” as compared to “y” (the reader is referred to these articles [41,42] for further reading regarding the existence of duel endothermic peak maxima in MgH2samples).After observing this information,a dehydrogenation and rehydrogenation kinetics test was performed for both the samples“x” and “y”.For comparison,H-de/ab-sorption kinetics measurement for an additive free ball milled sample was also made (sample indexed as “z”) and compared in Fig.9a and b.The hydrogen desorption kinetics profile were recorded at 305 °C (Fig.9a) and the re-hydrogenation kinetics tests were performed at the same temperature with 5-6 bar pure hydrogen pressure (Fig.9b).As seen in the figure the observed kinetics in both desorption and absorption stages reaffir the trend demonstrated in Figs.1 and 8.This observation is in good agreement with a recent work published by Chen et al.[43] which suggests that nano ZrH2is a highly efficien catalyst for MgH2.

Fig.9.(a) and (b) Dehydrogenation and Rehydrogenation profiles “x”:MgH2 with 3 wt.% 20 h milled MgH2+0.5ZrO2 at the speed of 350 rpm,“y”:MgH2+3 wt.% ZrO2,and “z”:ball milled MgH2.

The above set of experiments suggest that zirconia interacts with MgH2and makes ZrHxand MgOin-situproducts.Our previous studies on TiO2and Nb2O5mixed MgH2suggested that metal(Ti/Nb)dissolved MgO rock salt is the main product of interaction between the additive and MgH2[14-17].On the other hand,in this study,although it is clear that MgO is one of the by-products in the ZrO2added MgH2powder,no clear evidence for the dissolution of Zr in MgO lattice can be obtained.Moreover,the existence of ZrHxsecondary phase further suggest that dominant existence of Zr dissolved MgO is unlikely,in contrast to the case of TiO2/Nb2O5incorporated MgH2system.This anomaly suggests that not every oxide behaves the same way for MgH2and that different mechanistic paths exist for improving its hydrogen sorption performance.In this context,our view is that more oxides need to be tested for providing a generalized mechanism of oxides loaded MgH2system for hydrogen storage.

4.Conclusions

It is clear that ZrO2added MgH2performs better than additive free MgH2for reversible hydrogen storage applications.When the ZrO2added MgH2power is prepared under mild milling conditions there may be surface chemical interaction between MgH2and ZrO2but the product(s) cannot be identifie by XRD technique possibly due to their existence in very small quantities.Nonetheless,when the powder is prepared under strong milling conditions,it can be clearly explored that ZrO2and MgH2interact to form ZrHxand MgO and this powder is catalytically more active than the neat ZrO2.It is also clear that after chemical interaction between MgH2and ZrO2the size of particles in the product powder is very small (～20 nm) and the particles are also highly dispersed.Moreover the chemical state of Zr is a mix of Zr4+,Zr2+and Zr0which seems beneficia than when Zr exists in Zr4+state alone.