Pengyng Yo, Ying Jing, Yng Liu, Chengzhng Wu,?, Kuo-Chih Chou, To Lyu,Qin Li,,?

aMaterials Genome Institute, Shanghai University, Shanghai 200444, China

b State Key Laboratory of Advanced Special Steel & Shanghai Key Laboratory of Advanced Ferrometallurgy & School of Materials Science and Engineering,Shanghai University, Shanghai 200444, China

Abstract Uniform-dispersed Ni nanoparticles (NPs) anchored on reduced graphene oxide (Ni@rGO) catalyzed MgH2 (MH-Ni@rGO) has been fabricated by mechanical milling.The effects of milling time and Ni loading amount on the hydrogen storage properties of MgH2 have been investigated.The initial hydrogen desorption temperature of MgH2 catalyzed by 10 wt.% Ni4@rGO6 for milling 5 h is significantl decreased from 251 °C to 190 °C.The composite can absorb 5.0 wt.% hydrogen in 20 min at 100 °C, while it can desorb 6.1 wt.% within 15 min at 300 °C.Through the investigation of the phase transformation and dehydrogenation kinetics during hydrogen ab/desorption cycles, we found that the in-situ formed Mg2Ni/Mg2NiH4 exhibited better catalytic effect than Ni.When Ni loading amount is 45 wt.%, the rGO in Ni@rGO catalysts can prevent the reaction of Ni and Mg due to the strong interaction between rGO and Ni NPs.

Keywords: Hydrogen storage materials; Ni@rGO; MgH2; Hydrogenation/dehydrogenation properties; Catalytic mechanism.

1.Introduction

Nowadays, the increasing energy demand and environment pollution have stimulated the research for the utilization of clean energy, in which hydrogen is considered as one of the most potential energy carrier [1].For hydrogen economy, the greatest challenge lies in the hydrogen storage.Compared with gas-state and liquid-state hydrogen storage, solid-state hydrogen storage materials with higher energy density are safer [2].Mg is not only used as a structural material [3,4] or biomedical material [5], but is also regarded as a particularly promising candidate for hydrogen storage, due to its low density, abundant resource and high theoretical hydrogen storage capacity[6,7].Nevertheless,its thermodynamic stability,sluggish reaction kinetics and inherent low thermal conductivity impede the process of practical applications [8-11].In the past decades, many solutions have been developed to overcome these problems, such as nanosizing [12,13], catalyzing[14,15], alloying [16], etc.

Mechanical milling is widely used to disperse different catalysts.Particularly, transition metals (Nb, V, Ti, Co, etc.)[17,18], their oxide [19,20] and non-metal element [21,22],are usually dispersed into MgH2/Mg system.Among the transition metals, Ni-based compounds exhibit excellent catalytic effect on the hydrogen absorption/desorption of MgH2[23-25].When the size is reduced to nanoscale, the catalytic effect will be further enhanced.Liu et al.[23] reported that Mg-Ni nano-composite, prepared by a wet chemical method,could absorb 85% of its maximum hydrogen capacity within 45 s at 125 °C.Furthermore, Ni nanofiber with a diameter of ～50 nm and porous structure could enhance the desorption properties of MgH2[26].For example, the onset temperature (143 °C) and peak temperature (244 °C) of dehydrogenation are much lower for Mg with 4% Ni nanofiber addition, than that mixed with 4% Ni powders (300 °C for onset temperature and 340 °C for peak temperature).In addition, many carbon materials, including graphite [27],single-walled carbon nanotubes (SWNTs) [28], graphene nanosheets (GNs) [29] and activated carbon [30], have been studied to improve the hydrogen storage properties of MgH2.Moreover, recent theoretical calculation has shown that the combination of nanosizing and carbon materials leads to further enhanced hydrogen storage properties [31].Huang et al.[32] revealed that carbon played an important role on inhibiting the aggregation of the catalysts.All in all, among the carbon materials,graphene,with unique 2D nanostructure,excellent electronic and thermal conductivity, and high chemical stability, is an ideal supporter to host disperse nanoparticles(NPs), which exhibits great catalytic effect in the hydrogen storage areas [33,34].The combination of the transition metal and carbon seems to be more efficien on the improvement of hydrogen storage [35,36].Liu et al.[37] synthesized porous Ni@rGO with GO and NiCl2·6H2O as raw materials.Ni NPs loading on the rGO has better catalytic effect on the desorption kinetics of MgH2than that of Ni or rGO alone.Zhang et al.[38] reported that Ni decorated graphene nanoplate(Ni/Gn), which was prepared with Gn and Ni(NO3)2·6H2O,enhanced the sorption properties of MgH2.The sample of Mg@Ni8Gn2absorbed 6.28 wt.% H2in 100 s at 100 °C.To maximize the effect of graphene on preventing the growth and aggregation, a simple solvothermal method was adopted to prepare ordered structure of MgH2NPs with good dispersion on graphene by Xia et al.[39].The hydrogen capacity of the composite did not decay after 100 cycles, which is due to the stable structure.

To sum up, carbon supported Ni NPs shows a positive effect on enhancing the hydrogen storage properties of MgH2.However, it is still a challenge to prepare uniformly dispersed Ni NPs on the surface of graphene, and the evolution of Ni during the hydrogenation/dehydrogenation cycles is not clear.In this work, Ni@rGO was successfully synthesized through annealing the Ni(OH)2@rGO with aqueous solution of GO and Ni(NO3)2·6H2O as raw materials, and then it was doped into MgH2by mechanical milling to improve the hydrogen absorption/desorption properties.The effects of milling time and Ni@rGO with different Ni loading amount have been investigated in detail.The evolution of Ni during the hydrogenation/dehydrogenation cycles has been clarified

2.Material and methods

2.1.Synthesis of Ni@rGO

Typically, 25 mL aqueous solution of GO (8 mg/mL,Shenzhen Matterene Technology Company) was diluted to 1 mg/mL, and then 248 mg Ni(NO3)2·6H2O (AR) was added to the solution by ultrasonic vibration for 1 h.Freshly prepared aqueous solution of NaBH4(5 mL, 1 M) was added to prepare composite precursor.After stirring the mixture solution at room temperature for 2 h, the products were collected by centrifuge and washed with deionized water and ethanol several times.Then, it was dried in a blowing dry oven at 100 °C for 12 h.Afterwards, the products was processed with experienced heat treatment at 500 °C for 2 h in a tube under Ar/H2.The mass ratio of Ni and GO was 2:8, 4:6 and 6:4,and the corresponding samples were denoted as Ni2@rGO8,Ni4@rGO6, and Ni6@rGO4.

2.2.Synthesis of MH-Ni@rGO

Magnesium hydride was prepared by reactive ball milling method according to our previous work [40].Firstly, magnesium (purity>99.9%, 400 mesh) was mechanically milled under Hydrogen atmosphere with an initial pressure of ～1 MPa, followed by a long-period hydrogenation at 350 °C.

The as-synthesized Ni2@rGO8catalysts and MgH2powders mixed in a fi ed mass ratio of 10:90.The milling was then performed under the hydrogen pressure of 1 MPa, with the rotational speed of 450 rpm, and the ball-to-powder mass ratio of 40:1.The mixtures were milled for different periods (2 h, 5 h, 10 h, and 20 h).The corresponding samples were denoted as Mg-Ni2@rGO8-2h, Mg-Ni2@rGO8-5h, Mg-Ni2@rGO8-10h, and Mg-Ni2@rGO8-20h.The corresponding samples with different Ni@rGO addition were denoted as MH-Ni2@rGO8, MH-Ni4@rGO6and MH-Ni6@rGO4.

2.3.Characterization

The loading amount of the as-prepared Ni@rGO was quantifie by differential scanning calorimetry combined with thermogravimetry (TG-DSC, STA449 F3, Netzsch).The structure and morphology of the synthesized materials were characterized by powder X-ray diffraction (XRD, D8 Advance, Bruker), scanning electron microscopy (SEM,Sigma 500, Zeiss) and transmission electron microscopy(TEM, F200, FEOL), respectively.

The decomposition performance of the composite was measured by DSC.Hydrogen storage properties were determined by using a homemade HPSA-auto apparatus [41], in which the modifie Benedict-Webb-Rubin(MBWR)EOS was applied to calculate the compression factor of H2gas.To reduce the error of the system, two high-accuracy pressure transducers (Keller, ±0.05% FS) and three thermocouple of Pt 100 were used to monitor the pressure and temperature.Absorption-desorption measurements were performed at various temperature with an initial pressure of 3 MPa for hydriding and 0.0004 MPa for dehydriding, respectively.Before ab/desorption measurements, the samples performed dehydrogenation and rehydrogenation at 340 °C.All materials handling was performed in a glove-box fille with purifie argon(99.999%), in which water vapour and oxygen levels were below 1 ppm by a recycling purificatio system to prevent samples from hydroxide formation and/or oxidation.

3.Results and discussion

3.1.Phase structures and morphologies of Ni@rGO

The XRD patterns of the as-prepared Ni2@rGO8,Ni4@rGO6, and Ni6@rGO4are shown in Fig.1.The peak at 26° is contributed by rGO (002).The characteristic diffraction peaks at 44.5° (111) and 51.8° (200) are attributed to face-centered crystalline nickel.Firstly, GO was reduced by the addition of NaBH4, and Ni(OH)2was formed from the Ni2+in the solution and anchored on the rGO (Fig.S1).After annealing in H2/Ar fl w gases at 500 °C,the Ni(OH)2was further reduced to Ni.Based on Scherrer equation,the average grain size of Ni in Ni2@rGO8, Ni4@rGO6, and Ni6@rGO4were estimated to be 8.5 nm, 13.1 nm, and 15.4 nm, respectively.

Fig.1.XRD patterns of Ni2@rGO8, Ni4@rGO6 and Ni6@rGO4.

As-prepared catalysts of Ni2@rGO8, Ni4@rGO6, and Ni6@rGO4were also characterized by SEM.As shown in Fig.2, it can be seen that the Ni NPs are uniformly anchored on the surface of rGO for all samples.The average particle diameter is around 11.5 nm for Ni2@rGO8, 13.5 nm for Ni4@rGO6, and 16.5 nm for Ni6@rGO4, suggesting that the particle size increases with the increasing of Ni loading amount.It is worthy of noticing that, after an exposure in air for a week, Ni NPs doped on the rGO were not oxidized,suggesting that there is a strong interaction between Ni NPs and rGO.

The catalyst of Ni4@rGO6was further characterized by TEM (Fig.3).TEM images show that Ni NPs are uniformly spread over the rGO without aggregation even if the Ni loading amount is up to ～66 wt.%.The diameter of Ni NPs is about 15 nm.Such superior confinemen should be attributed to rGO, which can serve as a support forin-situformation of Ni NPs and prevent the aggregation and growth during the following heat treatment[39].In addition,the highresolution TEM (HRTEM) shown in Fig.3b demonstrates lattice fringes with interplannar distance of 0.205 nm, 0.177 nm and 0.221 nm, which corresponds to the lattice planes of Ni (111), Ni (200) and graphitic carbon (100), respectively.Besides Ni and C, NiO (111) with interplannar distance of 0.241 nm was also observed.The present of NiO may be due to the oxidization of Ni during the TEM sample preparation.It is observed that the rGO appears in the surrounding of Ni and leaves the Ni facets with high activity exposed [42],which can enhance the “synergistic effect”.

To verify Ni loading amount of Ni@rGO,TG measurement was adopted.The as-prepared samples were heated to 900 °C at a rate of 5 °C min?1under high pure air atmosphere.As a result, the rGO would be oxidized to carbon dioxide and released.Ni would be oxidized to NiO and remained as a residual with the increasing temperature.The Ni loading amount in Ni2@rGO8was calculated to be 45 wt.% (Fig.S2), 66 wt.% for Ni4@rGO6, and 77 wt.% for Ni6@rGO4.It is worth noting that the Ni loading amounts are higher than the designed values, which is due to the loss of oxygencontaining functional group, i.e., carboxylic, hydroxyl, and carbonyl during the reduction process.

3.2.Influenc of milling time on the hydrogen storage properties of MH-Ni2@rGO8

Fig.2.SEM images of Ni2@rGO8 (a), Ni4@rGO6 (b) and Ni6@rGO4 (c).

After the preparation of the catalysts of Ni2@rGO8, catalyzed MgH2was then made by reactive ball milling method,which is to mill the catalysts and MgH2together under a hydrogen pressure of 1 MPa.In order to research the effect of milling time on the hydrogen storage properties, various milling time of 2 h, 5 h, 10 h, and 20 h were applied,and the XRD patterns of the MH-Ni2@rGO8with different milling time are shown in Fig.4.The peaks of all samples mainly correspond to MgH2.However, the peaks of Ni phase were not found in the XRD patterns of all samples.In addition, with the increase of milling time, the diffraction peaks become wider and weaker, indicating that the crystallite size decreases.The grain size of MgH2was estimated by Scherrer equation, which shows that the grain size decreases in an order of 16.0 nm, 13.8 nm, 11.9 nm, 9.4 nm with the increase of milling time.Decreased grain size of MgH2would provide more diffusion channel of hydrogen, leading to an improvement of hydrogen sorption kinetics [29].

Fig.3.TEM image of Ni4@rGO6 (a) and HRTEM of Ni4@rGO6 (b).

Fig.4.XRD patterns of MgH2 (a) and MH-Ni2@rGO8 after milling in hydrogen for various periods.MH- Ni2@rGO8-2h (b), MH- Ni2@rGO8-5h (c),MH- Ni2@rGO8-10h (d) and MH-Ni2@rGO8-20h (e).

Fig.5.DSC curves of the MgH2 and as-milled MH-Ni2@rGO8 for various periods with a heating rate of 5 °C min?1.

The morphology of as-milled MH-Ni2@rGO8for different time was further observed by SEM.As shown in Fig.S3, big particles (>2μm) can be clearly observed in the sample of MH-Ni2@rGO8-2h.While the milling time is increased to 5 h and 10 h, the particle size is noticeably decreased.However,when further prolonging the milling time to 20 h, the particle size becomes much larger, and fla y particles appeared.It is noted that the variation trend of particle size observed in SEM images is different from that of grain size calculated by XRD results,which indicates that the grain size could be reduced by increasing the milling time.However, powder particles may be much bigger due to the welding and aggregation during the long milling process.

The dehydrogenation properties of MH-Ni2@rGO8with different milling time were investigated by DSC at a constant heating rate (5 °C min?1).As shown in Fig.5, both the onset temperature (Tonset) and peak temperature (Tpeak) of dehydrogenation for all MH-Ni2@rGO8samples shift to low temperature compared with pure MgH2.For the samples of Ni2@rGO8catalyzed MgH2, The Tpeakof hydrogen desorption is increased in an order of Mg-Ni2@rGO8-5h (280 °C),Mg-Ni2@rGO8-10h (290 °C), Mg-Ni2@rGO8-2h (295 °C),and Mg-Ni2@rGO8-20h (306 °C).It is obvious that 5 h is the best milling time for hydrogen desorption in our experimental condition.The desorption temperature reduces ～98 °C compared to pure MgH2.The hydrogen desorption result is in accordance with the particle size of ball milled samples.When the milling time is short, the particles are not fully ground and still big; when the milling time is too long, the particles will be welded due to the strong mechanical energy input.The reduced particle size leads to the decreasing of diffusion distances for H atom and enhances the hydrogen absorption/desorption kinetics [39].As shown in Fig.S3,the particle size of MH-Ni2@rGO8-2h (about 5μm) is bigger than that of MH-Ni2@rGO8-5h (about 1μm).While the milling time increases to 20 h, large particles (>2μm) can be easily observed.Therefore, with the increase of milling time, the Tpeakof hydrogen desorption decreases firstl and then increases.

Fig.6.Hydrogen absorption curves at 200 °C (a) and desorption curves at 300 °C (b) of the composite MH-Ni2@rGO8 ball milled for different periods.The initial hydrogen pressure is about 3 MPa for absorption and 0.0004 MPa for desorption.The inset to (a) shows the performance of hydrogenation at 100 °C and inset to (b) is the performance of dehydrogenation at 280 °C.

The hydrogen absorption/desorption kinetics of composite MH-Ni2@rGO8ball milled for various times was then studied by volumetric method.The initial hydrogen pressure during absorption and desorption is 3 MPa and 0.0004 MPa,respectively.The hydrogenation curves at 200 °C are plotted in Fig.6a, and the dehydrogenation curves at 300 °C are plotted in Fig.6b.The results show that the composite of MH-Ni2@rGO8-5h has the fast hydriding/dehydriding rate.It can absorb 5.6 wt.% H2within 60 s at 200 °C and 4.8 wt.%H2within 30 min at 100 °C (inset of Fig.6a).For the desorption properties, MH-Ni@rGO-5h can desorb 6.2 wt.% H2in 20 min at 300 °C and 6 wt.% H2in 60 min even at 280 °C(inset of Fig.6b).When the temperature reaches to 300 °C,all the Ni@rGO-containing samples reach 95% of their maximum dehydriding capacity in 20 min, indicating that faster hydrogen desorption kinetics are obtained at higher temperature.The hydrogen desorption behavior measured by volumetric method is the same as that by DSC.Fig.6 shows that the increase of the ball milling time would lead to a negative influenc on the hydrogen storage kinetics when the milling time is longer than 5 h.The composite of MH-Ni2@rGO8-20h can only absorb 3.9 wt.% H2within 30 min at 100 °C.

3.3.Influenc of Ni loading amount on the hydrogen storage properties of MH-Ni@rGO

Above results showed that the composite of MH-Ni@rGO with a milling time of 5 h had the best hydrogen storage properties.Thus,the milling time of 5 h was used for further study about the catalytic effect of Ni@rGO with different Ni loading amount on the hydrogen storage performance of MgH2.Fig.7 presents the DSC profile of the as-milled composite of MH-Ni2@rGO8, MH-Ni4@rGO6and MH-Ni6@rGO4.The MH-Ni4@rGO6displays the lowest desorption peak temperature (Tpeak= 259 °C) in comparison of MH-Ni2@rGO8and MH-Ni6@rGO4.Furthermore, it is worth to note that the onset temperature of MH-Ni4@rGO6(Tonset= 190 °C) is much lower than that of MH-Ni2@rGO8(Tonset=240°C)and MHNi6@rGO4(Tonset= 270 °C).In addition, MH-Ni6@rGO4shows two desorption peaks at higher temperature, one is at 285 °C and the other is at 320 °C, which indicates that the excessive Ni NPs may aggregate on the rGO and weaken the interaction between Mg and rGO.Obviously, too much Ni loading would degrade the catalytic effect.

Fig.7.DSC curves of the as-milled MH-Ni2@rGO8, MH-Ni4@rGO6 and MH-Ni6@rGO4 with a heating rate of 5 °C min?1.

Fig.8 gives the hydrogen absorption and desorption kinetics curves of MH-Ni2@rGO8, MH-Ni4@rGO6, and MHNi6@rGO4.The MH-Ni4@rGO6can absorb 3.7 wt.% H2at 100 °C in 10 min, higher than that of MH-Ni6@rGO4(3.0 wt.%) and MH-Ni2@rGO8(2.7 wt.%).The dehydrogenation kinetics is further improved in the sample of MHNi4@rGO6.At 300 °C, the MH-Ni4@rGO6can desorb 6.1 wt.%H2in15 min,while the MH-Ni2@rGO8needs 20 min to reach the same capacity.Moreover,the MH-Ni6@rGO4can only desorb 5.7 wt.% H2, which is due to the higher Tpeak(320 °C).

Fig.8.Hydrogen absorption curves at 100 °C (a) and desorption curves at 300 °C (b) of the MH-Ni@rGO.The initial hydrogen pressure is about 3 MPa for absorption and 0.0004 MPa for desorption.

Table 1Summary of hydrogen absorption/desorption data for the MgH2 system with metal/carbon catalysts.

To facilitate comparison, representative hydrogen absorption/desorption data for MgH2system (with metal/carbon catalysts, prepared by ball milling method) are summarized in Table 1 [29,32,37,43,44,45].Obviously, the Ni4@rGO6shows outstanding catalytic efficien y in enhancing the ab/dehydrogenation kinetics of MgH2.In comparison with other reported systems, the sample of MH-Ni4@rGO6is competitive in the hydrogen absorption at low temperature(100 °C) and hydrogen desorption at 300 °C.

3.4.Thermodynamics and kinetics investigation of MH-Ni4@rGO6

The composite of MH-Ni4@rGO6exhibits excellent hydrogen absorption and desorption properties.The PCT measurements of hydrogen absorption and desorption for the sample were performed at 300 °C, 320 °C, and 340 °C in a hydrogen pressure range from 0.05 to 2 MPa.Particularly, to ensure equilibrium, each absorption/desorption stage lasts at least 2 h.As shown in Fig.9a, the plateau pressure was measured as 0.205, 0.320, and 0.499 MPa for absorption and 0.162, 0.270, 0.420 MPa for desorption at 300, 320, 340 °C,respectively.The corresponding van’t Hoff plots (Eq.1)for both hydrogen absorption and desorption are shown in Fig.9b.

WherePis the H2pressure,Tis the temperature,ΔHis the enthalpy,Ris the gas constant (8.3145 J mol?1K?1), andCis a constant.TheCvalue equals toΔS/R, in whichΔSrefers to entropy.

According to the fittin result, the hydride formation and decomposition reaction enthalpy value (ΔH) are calculated to be ?64.8 and 69.6 kJ mol?1, respectively.The values are lower than the theoretical values of MgH2[38].The results indicate that the addition of Ni@rGO can destabilize the MgH2,which might be a reason for lower onset dehydrogenation temperature shown in the DSC curves.Therefore, it can be concluded that the Ni@rGO can not only improve the absorption/desorption kinetics, but also change the thermodynamics.

Fig.9.Hydrogen absorption/desorption PCT curves measured at 300, 320, and 340 °C (a) and van’t Hoff plots (b) for the MH-Ni4@rGO6 sample.

A number of kinetic models for the gas-solid reaction were adopted to analyze the evolution of kinetics, such as Johnson-Mehl-Avrami-Kolmogorov (JMAK) model [46,47],Chou model [48,49], etc.The classical JMAK model can well describe the hydrogenation and dehydrogenation of nucleation-growth-impingement mode.Thus, the improved kinetics of hydrogenation and dehydrogenation of MHNi4@rGO6was further determined by the JMAK model (see Eq.2) through fittin the absorption and desorption curves of MH-Ni4@rGO6and the activation energy (Ea) can be calculated according to the Arrhenius equation (Eq.3).

Wherekis the reaction rate constant,nis the Avrami exponent of the reaction order,αis the fraction transformed at timet, andAis the temperature-independent coefficient The reacted fraction of 0.2< α <0.8 was used in this study.Fig.10 shows the JMAK plots for the absorption of the MH-Ni4@rGO6at the temperature of 100, 150,200 °C, and desorption at 280, 300, 320 °C.Generally, the rate-limiting process, growth dimensionality and nucleation behavior of the hydrides can affect the reaction order n.Thenvalues of hydrogenation (0.91, 1.03 and 1.09) are close to 1 (Fig.10a), indicating that the hydriding reaction of the sample follows a diffusion-controlled mechanism[50,51].There are numerous growth and nucleation scenarios consistent with a value ofn= 1, including nucleation and growth along one-dimensional (1D) dislocation lines and thickening of cylinders, needles and plates.Jeon [52] pointed out that hydrogen atoms rapidly nucleate and accumulate along the defects and form a metal hydride layer in one dimension, followed by subsequent growth and thickening from the metallic core.Similarly, thenvalues are in close proximity to 1.5 for decomposition of MH-Ni4@rGO6at 300 and 320 °C (Fig.10b).Thus, the phase transformation from MgH2to Mg in this case exhibits a zero nucleation rate,which is consistent with previous report [28].Besides, theEaof MH-Ni4@rGO6is calculated to be 47.6 ± 3.4 kJ mol?1for hydrogenation (Fig.10c) and 117.8 ± 3.4 kJ mol?1for dehydrogenation (Fig.10d).The value for hydrogenation is lower than that of MgH2-5 wt.% GNs (Eaof hydrogenation:78.4 kJ mol?1) [53], and the dehydrogenation value is also lower than pure MgH2(157 kJ mol?1) [37].

3.5.Evolution and catalytic mechanism of Ni@rGO

Above results show that Ni4@rGO6is an excellent catalyst to improve the hydrogen storage properties of MgH2.Generally, the uniformly distributed ultrafin Ni NPs could be beneficia to the decomposition of H2and recombination of atomic hydrogen during hydrogen absorption/desorption[54].Moreover, the graphene can also provide more nucleation sites for the alloy or hydride and hydrogen diffusion channel [39], exhibiting a “synergistic effect” with Ni.To gain a further insight into the catalytic mechanism, XRD measurement was carried on the samples of MH-Ni4@rGO6at following fi e states: as-milled MH-Ni4@rGO6, after firs dehydrogenation, after firs rehydrogenation and after 8th de/rehydrogenation.As shown in Fig.11, XRD pattern of as-milled MH-Ni4@rGO6sample exhibits diffraction peaks corresponding to MgH2and Ni (Fig.11a), indicating that Ni has not reacted with Mg during ball milling process.After firs dehydrogenation, only two phases can be detected, Mg and Mg2Ni (Fig.11b), implying that Ni has already reacted with Mg and transformed to Mg2Ni.During the next rehydrogenation and dehydrogenation cycles, except for the main phase transformation of Mg/MgH2, the phase transformation of Mg2Ni and Mg2NiH4occurred (Fig.11c-e).

It is well acknowledged that the XRD investigation can not give the information of trace amount of sample or amorphous sample.Thus, the evolution of Ni during the dehydrogenation process of MH-Ni4@rGO6was also detected by TEM.As shown in Fig.12a,the size of dehydrogenated MH-Ni4@rGO6particles is around 200-600 nm, and some small catalyst particles are anchored on the surface of matrix.Further HRTEM analysis (Fig.12c-d) shows that MH-Ni4@rGO6sample was fully dehydrogenated to Mg and Mg2Ni with a crystallize size of smaller than 10 nm.Moreover, Ni NPs can’t be found in the HRTEM images, implying that the Ni NPs were totally reacted with Mg and yielded Mg2Ni during the hydrogenation of MH-Ni4@rGO6, which agrees with the result of XRD(Fig.11).Moreover, it is worthy of note that, a metastable alloy of Mg6Ni was also identifie by HRTEM, and evidenced by selected area electron diffraction (SAED) pattern(Fig.12b).It has been reported that Mg6Ni alloy can be formed in a Mg83Ni17alloy due to the solute accumulation in solidification and decomposes into Mg and Mg2Ni with a low velocity at the temperature of 300-350 °C [45].In this study, the as-milled MH-Ni4@rGO6sample was composed of MgH2and Ni rather than Mg-Ni alloys (Fig.11).During the following dehydrogenation process, Mg6Ni and Mg2Ni alloys might be formed due to the entering of Ni atoms into the lattice of Mg.At high temperature, Mg6Ni alloy is not stable and transforms into Mg and Mg2Ni again.In addition,as shown in Fig.12d, Mg2Ni and rGO grains were clearly dispread in the surrounding of Mg, which indicates that the co-catalyst of Mg2Ni and graphene nanosheet may have a“synergetic effect” on the hydrogen storage properties of Mg.

Fig.10.JMAK plots for the absorption (a), desorption (c) of MH-Ni4@rGO6 and Arrhenius plots for the absorption (b), desorption (d).

Fig.11.XRD patterns of as-milled MH-Ni4@rGO6 (a), MH-Ni4@rGO6 after firs dehydrogenation (b), MH-Ni4@rGO6 after rehydrogenation (c), MHNi4@rGO6 after 8th dehydrogenation (d) and MH-Ni4@rGO6 after 8th rehydrogenation (e).

Hydrogen desorption kinetics of as-milled MH-Ni4@rGO6and rehydrogenated MH-Ni4@rGO6are presented in Fig.13.It can be clearly seen that the hydrogen desorption rate of rehydrogenated samples are faster than that of as-milled sample.According to the above XRD and HRTEM results, the only difference between as-milled and rehydrogenated sample is chemical surrounding of Ni.We believed that the change of dehydrogenation kinetics of MgH2in the sample of MHNi4@rGO6is due to the change of Ni: elementary Ni for as-milled sample and Mg2NiH4for rehydrogenated samples.The result indicates that thein-situformed Mg2Ni/Mg2NiH4may have better catalytic effect than Ni.

Fig.12.TEM images (a, c), SAED (b) and HRTEM (d) of MH-Ni4@rGO6 in the dehydrogenation state of the firs cycle.

Fig.13.Hydrogen desorption curves of as-milled and rehydrogenated MHNi4@rGO6 at 300 °C.The initial hydrogen pressure during desorption is about 0.0004 MPa.The 1st means the as-milled samples.

Interestingly, we found that Ni NPs in the sample of MHNi2@rGO8seems to be more stable than Ni NPs in the sample of MH-Ni4@rGO6.As shown in Fig.S4, although Ni is not detectable in the as-milled MH-Ni2@rGO8, the diffraction peaks of Ni still present in the rehydrogenated samples even after 8 cycles.And Mg2Ni/ Mg2NiH4is not appeared in all MH-Ni2@rGO8samples, implying no reaction between Ni and Mg.For the sample of MH-Ni2@rGO8, the hydrogen desorption kinetics of as-milled sample is almost the same as that of rehydrogenated sample (Fig.S5).The results further proved that thein-situformed Mg2Ni has more effective catalysis than Ni.It is well acknowledged that, Mg2NiH4is easier to release hydrogen compared with MgH2.Therefore,thein-situformed and uniformly dispersed Mg2NiH4on the surface of rGO can serve as a “hydrogen pump” to enhance the dehydrogenation kinetics [18,41].The rGO could provide more active “catalytic sites” and H “diffusion channels” to reduce the dehydrogenation temperature and enhance the dehydrogenation kinetics [36], leading to a “synergetic effect”with Mg2NiH4.According to the results, it is assumed that there is a strong interaction between Ni NPs and rGO.When the amount of Ni is low, the binding may be strong enough to prevent the reaction between Ni and Mg/MgH2.However,if the amount of Ni is high, the interaction between Ni NPs and rGO will be weakened, therefore, the Ni NPs can react with Mg/MgH2more easily.It is worth noting that too much Ni loading will weaken the interaction between Mg and rGO and degrade the catalytic effect.

4.Conclusions

(1) Ni@rGO with different loading amounts was synthesized by wet chemical method, and the average crystallites size of Ni for Ni2@rGO8, Ni4@rGO6, Ni6@rGO4were calculated to be 8.5 nm, 13.1 nm, and 15.4 nm,respectively.

(2) The MH-Ni4@rGO6composite absorbs 5 wt.% hydrogen in 20 min at 100 °C.And the composite shows enhanced dehydrogenation rate: it can release 6.1 wt.%hydrogen within 15 min at 300 °C.The activation energy for the rehydrogenation of MH-Ni4@rGO6is 47.6 ± 3.4 kJ mol?1.Hydride formation and decomposition reaction enthalpy (ΔH) are determined to be?64.8 and 69.6 kJ mol?1, respectively, indicating a little thermodynamic change for the composite.

(3) We found that thein-situformed Mg2Ni/ Mg2NiH4exhibits better catalytic effect than Ni.Ni couldn’t react with Mg due to the strong interaction between rGO and Ni NPs when the loading amount of Ni is low.

Declaration of Competing Interest

The authors declare that they have no known competing financia interests or personal relationships that could have appeared to influenc the work reported in this paper.

Acknowledgements

This work was supported by the National Natural Science Foundation of China(Grant No.51671118),the research grant(No.16520721800 and No.19ZR1418400) from Science and Technology Commission of Shanghai Municipality.The authors gratefully acknowledge support for materials analysis and research from Instrumental Analysis and Research Center of Shanghai University.

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.jma.2019.06.006.