Hydrogen is an energy carrier which holds tremendous promise as a new renewable and cleans energy option [1]

Hydrogen is an energy carrier which holds tremendous promise as a new renewable and cleans energy option 1. The U.S. Department of Energy (DOE) set a target for the ideal hydrogen storage materials should reach 5.5 wt% gravimetric density by the year 2020 2. Hydrogen is a convenient, safe, versatile fuel source that can be easily converted to a desired form of energy without releasing harmful emissions. Hydrogen is recognized as the ideal fuel due to its high utilization efficiency and environmental friendliness 3, 4. However, it is very difficult to store hydrogen under ambient conditions due to very weak intermolecular interactions among hydrogen molecules. Hence, the development of a safe, effective, stable, and cheap hydrogen storage medium has attracted increasing attention in the scientific community. An efficient hydrogen storage material must possess fast sorption kinetics, high volumetric/gravimetric density, relatively low dehydrogenation temperatures for chemical hydrides, 5-8 and the hydrogen adsorption energy should be in the range of -0.2 to -0.7 eV at room temperature, 8. Carbon nanotubes (CNT) and boron nitride nanotubes (BNNT) have attracted much attention as candidates for a hydrogen storage media 5, 9. During the past decade, the carbon-based materials have been considered as promising media for hydrogen storage 10, 11. However, due to the very weak physical adsorption of H2 for most materials including carbon-based materials, attention has been directed at non-carbon nanosystems composed of light elements such as B and N 9. Boron nitride nanotubes (BNNTs) are inorganic analogues of carbon nanotubes (CNTs) 12 and theoretically predicted 13 and then successfully synthesized in 1995 14. BNNTs have attracted considerable attention due to the undisputed fact that in contrast to metallic or semiconducting CNTs: BNNTs are wide-gap semiconductors with almost same band gaps of 5.5 eV, independent of the tube diameter, helicity, and the number of tube walls 15, and they are chemically and thermally more stable 13,16–18. Furthermore, the interaction of hydrogen molecules with material surfaces can be enhanced by heteropolar bonds at surfaces, a feature that is present in BNNTs but absent in CNTs. Given these unique properties, as one of the most interesting non-carbon nanotubes 19, BNNT has high potential practical application in hydrogen storage. In recent years especially, Ma et al. 20 measured the hydrogen storage ability of BN nanotubes and found that multiwall BN nanotubes can uptake 1.8–2.6 wt % hydrogen under about 10 MPa at room temperature and 70 % of adsorbed hydrogen is chemisorbed. In theoretical studies, Yuan and Liew 21 reported that boron nitride impurities will cause a decrease in Young’s moduli of SWCNTs. Moreover, the effect of these impurities in zigzag SWCNTs is more significant because of the linking characteristics of an increase in electrons. In addition, some methods have been shown to improve the efficiency of storage. An increase in the diameter of BNNT can increase the efficiency of hydrogen storage 22. Thus far, several hydrogen storage methods have been suggested. Further, Tang et al. 23 improved the concentration of hydrogen storage to 4.6 wt% by bending the BNNTs.
Of particular interest, BNNTs have excellent mechanical properties, thermal conductivity, and resistance to oxidation at high temperatures, which makes them most valuable in nanodevices working in hazardous and high-temperature environments 24-26. Ju et al. 27 have investigated the effect of uniaxial strain on the electronic properties of (8, 0) zigzag and (5, 5) armchair BNNT. They have found that the two different types of BNNTs show very similar mechanical properties and variations in HOMO–LUMO gaps at different strains. Li et al. 28 demonstrated that the transport property of CNT with a double vacancy is reduced under external force. The stress-strain curve of armchair CNTs shows a step-by-step increasing behaviour, and the C-C bond length varies significantly at specific strain during the tensile process.
Metal-functionalization has been found to be a very useful scheme to improve or induce some unique properties of nanotubes 29. Both experimental 30-33 and theoretical 34, 35 studies were reported on the interaction of transition metal atoms with “perfect” BNNTs (BNNTs without intrinsic defects). The defective 36-39, carbon-doped 40, 41, Si-doped 38, Ti-doped 42, Ni-doped 43. Pt-doped 35 BNNTs have been examined, and these results show that the hydrogen storage capacity is significantly enhanced with respect to pristine BNNT and BN clusters 20,44-46. The adsorption of Ni onto single-walled BNNTs with intrinsic defects has been studied using DFT calculations, and the results of that study were reported by Zhao et al. 47. They found that the existence of defects in BNNTs clearly improves the chemical reactivity associated with Ni adsorption. In addition, the charge transfer, band gap, and density of states of Ni adsorbed onto BNNTs have been reported. Y. Liu et al. 48 investigated the hydrogen storage of Na-decorated single- and double-sided graphyne and their BN analogues. They found that the Na decorated double-sided graphyne and BN analogue the hydrogen storage capacities could reach to 5.98 and 5.84 wt%, with the average adsorption energies of 0.25 and 0.17 eV/H2, respectively. W. Lei et al. 49 measured the hydrogen storage ability of oxygen doped boron nitride (BN) nanosheets with 2–6 atomic layers, synthesized by a facile sol-gel method and found that a storage capacity of 5.7 wt% under 5 MPa at room temperature and 89% of the stored hydrogen can be released when the hydrogen pressure is reduced to ambient conditions.
However, the phenomenon of the mechanical activation of the material for hydrogen storage, especially from the point of view of the US Department of Energy (DOE) requirements, has been little explored. The US DOE has established an ultimate set of technical targets for onboard hydrogen storage systems to represent the desired end state of the technology department 50. In practical, bending nanotubes represent one of a variety of different shapes extending BNNTs into a variety of BNNT-based nanostructures, including finite BNNTs. These nanostructures have diverse morphologies as well as unique and distinct physical and chemical properties so as to extend the potential applications of BNNTs.
This study explores the usefulness of the bending-deformed (8,0) BNNT as hydrogen storage material. With reference to the ultimate targets of the DOE for physisorption, and thermodynamic properties. We have therefore investigated the hydrogen storage properties of the undeformed and bending-deformed Ni metal functionalized (8, 0) BNNT with special attention paid to the characterization and the thermodynamic capabilities of the hydrogen storage capacity reaction 4H2 + Ni- BNNT-?, (?=0, 15, 30, 45). The characterization is carried out in terms of density of states (DOS), pairwise and non-pairwise additivity, infrared (IR), Raman(R), electrophilicity , molecular electrostatic potentials (MEPs) and statistical thermodynamics.

Computational details
First-principle calculations have been performed for the interactions between Ni and single-walled zigzag BNNT within the DFT by using the Gaussian 09 program 51. The DFT calculations were performed by simultaneously using Becke’s three-parameter exchange functional (B3) with Lee–Young–Parr (LYP) correlation functional 52-55. B3LYP correctly reproduces the thermochemistry of many compounds including TM atoms 57–59. We chose the finite cluster (8,0) zigzag BNNT with the length of 12 Å including totally 40 boron, 40 nitrogen as the studied systems. Full geometry optimizations were carried out for Ni-BNNT, nH2-Ni-BNNT -? = 0, 15, 30 and 45 (n=1-5) complexes by using the larger B3LYP/6?31g (d, p) level of theory. In addition, NBO analysis implemented in Gaussian 09 program 51 was applied through a series of intermolecular interactions under the above system to evaluate the natural bond orbital (NBO) charges. The optimal geometries were visualized by using the corresponding Gauss View 5.0 software. The density of states (DOS) and Fermi levels were calculated by using Gauss Sum 2.2.5 which is a post-processing of Gaussian 09 code 61.

3. Results and discussion
3.1. Bending deformation and hydrogen storage
We constructed a zigzag BNNTs (8,0) with length 12 Å, B–N bond length 1.45 Å, consisting of 40 B and 40 N atoms and defined the bending angle in degrees (?=0) for the undeformed BNNT, and (?=15,30, and 45) for the deformed BNNT. The frozen geometries of BNNT-?=0, BNNT-?=15, 30, and 45 are shown in Figure 1. At first, we analyzed the interaction between Ni atom and (8, 0) BNNTs -? (?=0, 15, 30, and 45). Several various adsorption positions were selected for the Ni atom on the BNNTs as presented in Figure 2: (1) directly above a boron atom (B), (2) above a nitrogen atom (N), (3) over an axial BN bond (BN), (4) above a center of a hexagon (h).
The metal adsorption energy of the functionalized BNNT is defined as
?E_(ads.)=E_((BNNT+Ni))-(E_((BNNT) )+E_((Ni) ) ) (1)
where E_((BNNT+Ni)) is the total energy of the fully relaxed Ni-BNNT, and E_((BNNT) ) and E_((Ni)) are the energies of the isolated systems. The negative value of ?E_(ads.)corresponds to exothermic adsorption. For these sites, our calculations show that the adsorption of Ni is all exothermic and the most stable site is N site BNNT-?=0,15,30 structures while for BNNT-?=45 the most stable site is BN site (Figure 2).The binding energies for the most stable configurations BNNT-?=0,15,30,45 structures are summarized in Table 1. In the case of a Ni-BNNT-?=0 system, the adsorption energy at the N site is -3.033 eV with the bond length of B-Ni and N-Ni are 2.210 and 1.808 Å, respectively. This bond length is consistent with the results suggested by 34, 47, 62 and the value predicted by Auwarter et al. 63. For the Ni-BNNT-?=15 system, the adsorption energy at the N site is -3.872 eV with the bond length of B-Ni and N-Ni are 2.375 and 1.839Å, respectively. The adsorption energy of the Ni-BNNT-?=30 is -6.980 eV with the bond length of B-Ni and N-Ni are 2.189 and 1.835 Å, respectively. While for the Ni-BNNT-?=45 system, the adsorption energy is -5.153 eV with bond lengths of Ni-B and Ni-N being 1.873 and 1.763 Å respectively. According to the Mulliken charge analysis of single Ni atom adsorbed on BNNT-?=0,15,30,45, Ni atom donate electrons to the neighboring boron and nitrogen atoms on the BNNT-?=0,15,30,45, and this charge transfer decreases for boron and nitrogen atoms far away from the metal atom. The d orbital’s of Ni atom overlap with the sp and sp3orbitals of the Ni-B and Ni-N bond to form mixed (spd and sp3d) hybridization. This charge transfer behavior leads to Ni atom in cationic form and renders extensive heteropolar bonding between the Ni atom and the nearest boron and nitrogen atoms. As a consequence, extra dipole moments are formed, thus resulting in an increase in the H2 molecule uptake.
Full geometry optimizations without symmetry constraints were carried out for 4H2 and Ni, while the BNNT was kept frozen, at the B3LYP level of theory using the 6-31g (d,p) basis set. The adsorption of four hydrogen molecules on the stable configuration of Ni- BNNT-(? = 0, 15, 30, 45) system was studied. This was then followed by full geometry optimization of the adsorbates (4H2+Ni) over the outer surface of the frozen core BNNT (Figure 3). Bending beyond 45° was avoided because of tube fracture. The calculated binding energies, structural and other physical quantities are summarized in Table 2. As shown in Table 2 the H2 bond length of the first adsorbed H2 molecule in the four complexes (0.942, 0.818, 0.808, and 0.847Å) are increased relative to the experimental bond length of 0.74 Å 64. When more than one H2 molecule is adsorbed on the complexes BNNT-(?=15, 30, 45) the average H2 internuclear distances (0.729 – 0.744 Å) do not change relative to the experimental bond length of 0.74 Å. The average binding energy for hydrogen over the Ni-BNNT is defined as
?E_ads (H_2 )=E(4H_2-Ni-BNNT)-E(Ni-BNNT)-4E(H_2 ) /4 (2)
where E(4H_2-Ni-BNNT) is the total energy of the fully relaxed 4H_2-Ni fragment deposited on frozen BNNT, E(H_2 ) is the energy of an isolated relaxed hydrogen molecule, E(Ni-BNNT) is the total energy of the fully relaxed Ni atom deposited on frozen BNNT and (4) is the number of hydrogen molecules. By definition,
Eads. ; 0 corresponds to an exothermic procedure for the adsorption of Ni atoms.
The average adsorption energies per H2 of the undeformed 4H2-Ni-BNNT-?=0(-0.478 eV) are greater than the deformed 4H2-Ni-BNNT-?=15, 30 and 45 (-0.255, -0.328, -0.288 eV), respectively. The adsorption energies of the deformed 4H2-Ni-BNNT-? = 30 are stronger than those with the deformed 4H2-Ni-BNNT-?=15, and 45). This implies that the adsorption property of the deformed 4H2-Ni-BNNT is either correlated with the values of the bending angle which increases in the direction of ?=0 ; ?=30 ; ?=