NiCo nanoalloy encapsulated in graphene layers for improving hydrogen storage properties of LiAlH4

NiCo nanoalloy (4–6 nm) encapsulated in grapheme layers (NiCo@G) has been prepared by thermolysis of a 3D bimetallic complex CoCo[Ni(EDTA)]2·4H2O and successfully employed as a catalyst to improve the dehydrogenation performances of LiAlH4 by solid ball-milling. NiCo@G presents a superior catalytic effect on the dehydrogenation of LiAlH4. For LiAlH4 doped with 1 wt% NiCo@G (LiAlH4-1 wt% NiCo@G), the onset dehydrogenation temperature of LiAlH4 is as low as 43 °C, which is 109 °C lower than that of pristine LiAlH4. 7.3 wt% of hydrogen can be released from LiAlH4-1 wt% NiCo@G at 150 °C within 60 min. The activation energies of LiAlH4 dehydrogenation are extremely reduced by 1 wt% NiCo@G doping.

Hydrogen storage is one of the most critical issues for fuel cell vehicular applications. Solid hydrogen storage materials such as carbon materials, MOFs, metal hydrides and complex hydrides have received significant attention as the safest and most effective storage media [1][2][3][4][5][6][7][8][9][10] . Among them, lithium alanate (LiAlH 4 ) is considered as one of the most promising hydrogen storage materials due to its high hydrogen storage capacity of 10.5 wt%. In the last two decades, LiAlH 4 has received particular attention aiming at reducing the operation temperature to meet the DOE criteria because it can release a total amount of 7.9 wt% of hydrogen in two steps (eqs 1 and 2) below a relatively low temperature, e.g. 220 °C.
Li AlH 2Al 3H (5 3 wt% H , 150 175 C) Different methods have been explored for decrease of the dehydrogenation temperature of LiAlH 4 , including particle size reduction by ball milling 11 , synthesis of multi-hydride composites 12 , and doping with catalysts 13,14 . Among these methods, doping with catalysts is considered as an effective approach for the dehydrogenation of LiAlH 4 . Various catalysts have been investigated, such as Ti, Fe, Ni, V, Al, Al 3 Ti, TiF 3 , TiCl 4 , TiCl 3 , NiCl 2 , VCl 3 , AlCl 3 , FeCl 3 , TiCl 3 ·1/3AlCl 3 , TiH 2 , NiFe 2 O 4 , carbon nanotube and so on [15][16][17][18][19] . Ni based catalysts have been widely explored. Zheng et al. 20 doped LiAlH 4 with Ni to reduce the temperature of the first dehydrogenation step by approximately 10-15 °C, with a great expense of the hydrogen storage capacity. Yuan et al. 21 prepared 2 mol% NiCo 2 O 4 nanorod doped LiAlH 4 , which can release 6.47 wt% of hydrogen at 150 °C within 150 min. Nevertheless, most of the systems present the disadvantage of loss of the overall hydrogen storage capacity, due to either a hydrogen release during the ball milling process or the large additional weight of the catalysts. Thus, it is desirable to develop an effective catalyst for the dehydrogenation of LiAlH 4 .
Recently, metal/carbon composites has been developed for the dehydrogenation of LiAlH 4 , including Co-decorated MWCNTs 22 , Co@C 23 , and Ni-containing mesoporous carbon scaffold (Ni-MCS) 24 . In the last two decades, bimetallic nanoparticles have received great interest for catalysis and electrocatalysis, due to their higher activity and selectivity than monometallic nanoparticles as a result of "synergistic effects" [25][26][27][28] . In addition, graphene more effectively improves the dehydrogenation behavior of LiAlH 4 than C 60 , carbon nanotubes, and graphite 29 . In this study, we used NiCo nano alloy encapsulated in graphene layers (NiCo@G) as a catalyst and investigated the effects of NiCo@G on the dehydrogenation behavior of LiAlH 4 . Preliminary results showed that excellent dehydrogenation properties of LiAlH 4 were achieved by 1 wt% NiCo@G doping.

Results and Discussion
Preparation and characterization of NiCo@G. Briefly, a 3D bimetallic complex CoCo[Ni(EDTA)] 2 ·4H 2 O, a precursor for NiCo@G, was firstly synthesized through solvothermal method. CoCo[Ni(EDTA)] 2 ·4H 2 O precursor was thermally decomposed to NiCo@G as the final product at 500 °C in an argon flow. The Powder X-ray diffraction peaks (Fig. 1a) of the as-synthesized precursor CoCo[Ni(EDTA)] 2 ·4H 2 O match well with the simulated pattern on the basis of the single crystal structure reported by Sapiña et al. (Supplementary Fig. S1) 30 . The Ni/Co molar ratio in NiCo@G is 1:1 as that in the precursor complex, which is confirmed by SEM-EDS data ( Supplementary Fig. S2). As shown in Fig. 1b, the XRD pattern of NiCo@G shows peaks at 2θ = 44.48, 51.69 and 76.25°, matching those reported for (111), (200) and (220) planes of Ni 50 Co 50 alloy with a fcc structure 31,32 . Their positions are slightly higher than those of pure fcc Co (44.22, 51.52 and 75.86°) and slightly lower than those of pure fcc Ni (44.51, 51.85 and 76.37°). Broadness of the characteristic diffraction peaks for NiCo alloy is due to the formation of nanosized NiCo particles. Furthermore, the characteristic reflections corresponding to hexagonal close packed (hcp) metallic Co can not be observed. The broad peak around 2θ = 26° is the characteristic reflection for carbon. X-ray photoelectron spectroscopy (XPS) was used to examine the species present in the particles. The spectra of Ni 2p and Co 2p energy ranges were recorded ( Supplementary Fig. S3). The positions of the 2p peaks are respectively 852.8 and 870.1 eV for Ni 2p, 778.3 and 793.3 eV for Co 2p, demonstrating Ni and Co in their zero-valent states [33][34][35] . TEM images are shown in Fig. 1c,d, indicating that the NiCo nanoalloy is encapsulated in multilayered graphene shells (NiCo@G). The NiCo nanoalloy are spherical in shape, with a highly Dehydrogenation performances. Figure 2a as shows the non-isothermal dehydrogenation performances of as-received LiAlH 4 , as-milled LiAlH 4 , and LiAlH 4 doped with 1 wt%, 5 wt% and 10 wt% NiCo@G. Compared to as-received LiAlH 4 , as-milled LiAlH 4 exhibits a similar dehydrogenation behavior. The as-received LiAlH 4 starts to decompose at 152 °C, while the as-milled LiAlH 4 exhibits a slight decrease of 4 °C. It is obvious that addition of NiCo@G extremely improves the onset dehydrogenation temperature of LiAlH 4 . The onset dehydrogenation temperature and the amount of hydrogen released of all samples are shown in Fig. 2b. The onset desorption temperature decreases with the increasing NiCo@G percent. LiAlH 4 -1 wt% NiCo@G starts to decompose at 43 °C, which is 109 °C lower than as-received LiAlH 4 . For 5 wt% and 10 wt% NiCo@G doped samples, the onset dehydrogenation temperature is as low as 36 °C, which is 116 °C lower than as-received LiAlH 4 . However, raising the NiCo@G percent results in a decrease of the amount of hydrogen released. Only 5.9 wt% and 3.7 wt% of hydrogen are respectively released for LiAlH 4 -5 wt% NiCo@G and LiAlH 4 -10 wt% NiCo@G, due to the increasing catalyst percent and the premature dehydrogenation during the ball milling process. It is noteworthy that the amount of hydrogen released for LiAlH 4 -1 wt% NiCo@G reaches up to 7.3 wt%, which is identical to that of as-milled LiAlH 4 . This phenomenon is attributed to the small NiCo@G percent and a good preservation of hydrogen during the ball milling process. Compared with the performance of other additives or catalysts, NiCo@G developed in this work exhibits high catalytic activity (Table S1 † ). For graphene, Hsu 29 and Jiang 36 suggested that the interaction between electronegative carbon and Li + , high electronic conductivity promoting electron exchange between metal and [AlH 4 ] − , and delocalized π bonds facilitates hydrogen release. Furthermore, in NiCo@G, graphene is the shell preventing NiCo nanoparticles aggregation, leading to NiCo nanoalloy with a uniform size distribution ranging from 4 to 6 nm. The beneficial effect of catalyst size on dehydrogenation behaviors has been confirmed in previous literatures. Figure 3 shows the isothermal dehydrogenation kinetics measurements of as-milled LiAlH 4 and LiAlH 4 -1 wt% NiCo@G at 150 °C. For as-milled LiAlH 4 , only 1.6 wt% of hydrogen releases within 10 min. However, for-LiAlH 4 -1 wt% NiCo@G, the dehydrogenation goes on rapidly with 5.8 wt% of hydrogen released within 10 min. Furthermore, total 7.3 wt% of hydrogen can be thoroughly released within 60 min for LiAlH 4 -1 wt% NiCo@G while 350 min for as-milled LiAlH 4 . This result confirms that dehydrogenation kinetics are significantly improved by addition of NiCo@G.

Dehydrogenation mechanism.
To obtain insight on the catalytic mechanism of NiCo@G for the LiAlH 4 dehydrogenation, morphologies and intergrain dispersion of both as-milled LiAlH 4 and LiAlH 4 -1 wt% NiCo@G are investigated by SEM, as shown in Fig. 4. Compared to as-milled LiAlH 4 , the particle size significantly decreases after doping with 1 wt% NiCo@G, leading to more grain boundaries and larger surface area. This important observation suggests that NiCo@G readily influences the LiAlH 4 texture at room temperature during the ball milling process, by preliminarily breaking their particle aggregation. Graphene has been confirmed as an effective grinding agent to reduce the crystal size of LiAlH 4 owing to its high mechanical strength 29,36 . At that stage it is not yet clear whether the consequent decrease of the dehydrogenation temperature is due to the smaller LiAlH 4 particles generated after ball milling or to another effect of the catalyst on the mechanism governing this decomposition. The strong catalytic effect of such a small percent of NiCo@G was further investigated by combining XRD and DSC. Figure 5 shows the XRD patterns of NiCo@G, as-received LiAlH 4 , as-milled LiAlH 4 and LiAlH 4 -1 wt% NiCo@G. The diffraction peaks of as-milled LiAlH 4 match well with those of the as-received LiAlH 4 , demonstrating a high stability of LiAlH 4 during the ball milling process. NiCo@G can not be distinguished in LiAlH 4 -1 wt% NiCo@G, due to the exceptionally small concentration (1 wt%) of NiCo@G. The weak peaks of Li 3 AlH 6 (2θ~21.9°, 31.6°) and Al (111) (2θ~38.4°) appear in LiAlH 4 -1 wt% NiCo@G sample, indicating partial dehydrogenation of LiAlH 4 during ball milling process (R1), in agreement with the small hydrogen capacity loss observed in the TGA visualization of the dehydrogenation process (Fig. 2). Furthermore, the diffraction peaks of LiAlH 4 in the 1 wt% NiCo@G doped sample become broader than those of as-milled LiAlH 4 , indicating smaller particle size of LiAlH 4 .
DSC measurements were conducted to further verify the effect of NiCo@G on the dehydrogenation of LiAlH 4 , as shown in Fig. 6a. Compared to as-received LiAlH 4 , as-milled LiAlH 4 presents a similar DSC profile including the melting peak of LiAlH 4 , indicating that the ball milling process does not alter its intrinsic properties. Surprisingly, the DSC profile of LiAlH 4 -1 wt% NiCo@G is totally different and shows three distinct endothermic peaks. In order to understand the phase changes at different stages of LiAlH 4 -1 wt% NiCo@G, we stopped the dehydrogenation of LiAlH 4 -1 wt% at temperatures (110 °C, 170 °C and 210 °C ) corresponding to three dehydrogenation stages in DSC profile and investigated the samples by XRD. As shown in Fig. 6b, Li 3 AlH 6 , Al and small amount of retained LiAlH 4 are present in the sample which was stopped dehydrogenation at 110 °C. Thus, the first endo peak in the DSC profile of LiAlH 4 -1 wt% NiCo@G is attributed to the decomposition of solid LiAlH 4 (eqn.1). This decomposition (eqn. 1) apparently ends around 110 °C, which explains the absence of any LiAlH 4 melting. For the sample heated up to 170 °C corresponding to the second stage, Al, LiH, small amount of Li 3 AlH 6  Activation energies. To understand the dehydrogenation kinetics, the apparent activation energy (E a ) at each stage of LiAlH 4 -1 wt% NiCo@G dehydrogenation was calculated using the Kissinger equation (eqn. 3), considering the three endothermic peaks in DSC profiles at heating rates of 2, 5, 10 and 20 °C min −1 , as shown in Fig. 7.
where β is the heating rate, T p (K) is the DSC peak temperature, A is the pre-exponential factor, and R is the gas constant. E a , for each dehydrogenation stage of LiAlH 4 -1 wt% NiCo@G, was evaluated to be 54.8 ± 6 kJ mol −1 , 80.1 ± 1.3 kJ mol −1 and 119.7 ± 2.8 kJ mol −1 respectively. The activation energy E a1 for R1 involved in the catalytic dehydrogenation of LiAlH 4 is comparable to the lowest values of those reported catalysts.
Regarding the superior efficiency of NiCo@G, many factors obviously play a role. Such efficiency can either be attributed to the support (carbon) 29 or to a combination of properties associated with two different nanosized metals 37,38 or to both 19,39,40 . The mechanism need further investigations for the particular NiCo bimetallic nanoparticles encapsulated in graphene used here.
As a summary, NiCo nanoalloy (4-6 nm) encapsulated in grapheme layers (NiCo@G) was prepared and introduced into LiAlH 4 by solid-state ball milling. A tremendous improvement in the dehydrogenation properties of LiAlH 4 was achieved. When 1 wt% NiCo@G was doped with LiAlH 4 , the onset dehydrogenation temperature is decreased to 43 °C with 7.3 wt% of hydrogen released below 200 °C. For LiAlH 4 doped with 10 wt% NiCo@G, the onset dehydrogenation temperature is as low as 36 °C, which is 116 °C lower than that of pristine LiAlH 4 . Ea of LiAlH 4 -1 wt% NiCo@G for the first dehydrogenation step decreased to 54.8 kJ mol −1 . The significant catalytic effect makes NiCo@G a promising candidate for LiAlH 4 dehydrogenation. A more in depth study of the effect of NiCo@G on LiAlH 4 dehydrogenation, in particular regarding the critical roles of Ni/Co and catalyst/LiAlH 4 ratios, is in process.   Preparation of LiAlH 4 -NiCo@G samples. LiAlH 4 (97%) was purchased from Alfa Aesar, and used without further purification. Typically, 0.5 g powder mixture composed of LiAlH 4 and NiCo@G was loaded into a stainless milling pot with 10 steel balls (10 mm in diameter). Ball milling was carried out on a QM-1SP2 planetary under an argon atmosphere at 300 rpm for 30 min. All sample handlings were performed in a glove box filled with argon to avoid oxidation and moisture.
Characterizations. Powder X-ray diffraction (XRD) measurements were conducted on a PANalytical X'pert diffractometer operated at 40 kV and 40 mA with a Cu K α radiation (λ = 1.5418 nm). The samples were covered by Mylar film in glove box to avoid oxidation and moisture. Scanning electron microscopy (SEM) images were obtained by using JSM-6360LV SEM (JEOL, Japan). Transmission electron microscopy (TEM) studies were performed on a FEI Tecnai F30 microscope and a G 2 microscope operated at 300 kV. The powders were dropped on an ultrathin carbon film supported on a copper grid by using ethanol as a dispersant. X-ray photoelectron spectroscopy (XPS) was recorded using a Thermo ESCALAB 250Xi instrument with Al Kα X-rays (1486.6 eV). Thermogravimetric analysis (TGA) was carried out on a Cahn Thermax 500 with a heating rate of 2 °C min −1 in an argon flow. The isothermal dehydrogenation kinetics were measured using a Sieverts-type apparatus (Advanced Materials Corporation, USA) at 150 °C under an initial pressure of 10 −5 MPa. Differential scanning calorimetry (DSC) data was collected from a TA Q1000 in a constant argon flow (50 mL min −1 ) at different heating rates (2 K min −1 , 5 K min −1 ,10 K min −1 , 20 K min −1 ).