Hexagonal Boron Nitride

DOI: 10.1002/cphc.201601063

Articles

Hexagonal Boron Nitride (h-BN) Sheets Decorated with OLi, ONa, and Li2F Molecules for Enhanced Energy Storage Syeda Rabab Naqvi,[a] Gollu Sankar Rao,[d] Wei Luo,[a] Rajeev Ahuja,[a, b] and Tanveer Hussain*[c] First-principles electronic structure calculations were carried out on hexagonal boron nitride (h-BN) sheets functionalized with small molecules, such as OLi, ONa, and Li2F, to study their hydrogen (H2) storage properties. We found that OLi and ONa strongly adsorb on h-BN sheets with reasonably large inter-adsorbent separations, which is desirable for H2 storage. Ab initio molecular dynamics (MD) simulations further confirmed the structural stability of OLi-BN and ONa-BN systems at 400 K. On the other hand, Li2F molecules form clusters over the surface of h-BN at higher temperatures. We performed a Bader charge

investigation to explore the nature of binding between the functionalized molecules and h-BN sheets. The density of states (DOS) revealed that functionalized h-BN sheets become metallic with two-sided coverage of each type of molecules. Hydrogenation of OLi-BN and ONa-BN revealed that the functionalized systems adsorb multiple H2 molecules around the Li and Na atoms, with H2 adsorption energies ranging from 0.20 to 0.28 eV, which is desirable for an efficient H2 storage material.

1. Introduction nancial and safety risk factors.[8, 9] Solid-state materials-based H2 storage is considered to be the most workable option. However, due to the unavailability of suitable materials, the H2 storage targets proposed by the US Department of Energy (USDOE) have not been achieved so far.[10, 19] Carbon-based nanostructures with lightweight, low dimensionality, and large surface-to-volume ratio could be of great advantage in storing large amounts of H2. However, in pristine forms, most of these nanostructures preserve a weak binding affinity for H2 molecules. Thereby, various structural and chemical modifications have to be performed to utilize their potential. Several such nanomaterials like carbon nanotubes,[11–13] fullerenes,[14] graphene,[15–18] and hydrogenated graphene[5] have recently been investigated. Motivated by the potential of carbon-based nanostructures in gas sensing and H2 storage,[20–24] researchers have also started exploring the promise of non-carbon based systems such as boron nitride nanostructures.[25–35] Superior thermal stability ( & 1000 8C) of BN nano-systems compared to carbonaceous nanostructures, which typically oxidize at 600 8C, made them amenable for research perspectives.[36] Hexagonal boron nitride sheets are structural analogs of graphene, with unique electronic, chemical, and mechanical characteristics. Unfortunately, similar to carbon-based nanomaterials, h-BN substrates in pristine forms exhibit a weak binding affinity (0.09 eV) for H2 molecules.[34] Several strategies were employed to improve the binding energy of H2, such as polarization caused by an external electric field[29] and charge induction via defect-creation or functionalization.[27, 33, 37] Functionalization of BN nanostructures with alkali metals (AM),[42, 43] alkaline earth metals (AEM),[44] transition metals,[38–41] and non-metals (NM)[45, 46] has been studied to reach an opti-

The rapid depletion of available fossil reserves and the greenhouse effect have provoked an extensive interest in clean, economical, and renewable alternative energy sources. Electricity generation by photovoltaics, wind, biomass, hydropower, and geothermal resources is currently being utilized for industrial applications.[1, 2] For transportation, the hydrogen (H2) fuel cell is considered as an ideal supplement due to its reversibility, environmental friendliness, and high energy density.[3–5] New strategies for hydrogen (H2) production, such as photoelectric water splitting and nuclear energy are being developed as a CO2-emission-free substitute to the conventional conversion of the natural gas (CNG) technique.[6, 7] However, H2 storage still poses a great challenge for a practical implementation of H2 economy. Storage under compression and liquefaction has fi[a] S. R. Naqvi, Dr. W. Luo, Prof. R. Ahuja Condensed Matter Theory Group Department of Physics and Astronomy, Box 516 Uppsala University, SE-751 20 Uppsala (Sweden) [b] Prof. R. Ahuja Applied Materials Physics Department of Materials and Engineering Royal Institute of Technology (KTH) SE-100 44 Stockholm (Sweden) [c] Dr. T. Hussain Centre for Theoretical and Computational Molecular Science Australian Institute for Bioengineering and Nanotechnology The University of Queensland Brisbane, QLD 4072 (Australia) E-mail: [email protected] [d] Dr. G. S. Rao Department of Physics, University of Basel Basel-4056 (Switzerland) The ORCID identification number(s) for the author(s) of this article can be found under http://dx.doi.org/10.1002/cphc.201601063.

ChemPhysChem 2017, 18, 513 – 518

513

T 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Articles 2. Results and Discussions

mum H2 storage efficiency at ambient conditions. Recently, Banerjee et al.[47] reported the superior performance of a Lisubstituted hydrogenated boron nitride sheet for H2 storage. Each Li atom was capable of adsorbing four H2 molecules with average adsorption energy ranging from 0.18 to 0.3 eV. Hussain et al.[31] investigated the stability of h-BN sheet functionalized with polylithiated species (CLi3/OLi2) by MD calculations and further studied H2 storage capacity of the designed systems. They reported a reasonably high H2 storage capacity with adsorption energies lie within acceptable range for practical applications. Peng et al.[48] carried out theoretical investigations of H2 storage in alkali metal coated fullerenes C60(OM)12 (M = Li, Na) and reported the storage capacity of 9.78 % and 8.33 % for C60(OLi)12 and C60(ONa)12, respectively. Wang et al.[49] reported that Li2F coatings on C60 lead to increased binding affinity, and the charge transfer between the host and H2 molecules. Inspired by the prior reports, we expect that small polar molecules with alkali atoms, such as Li and Na, could be prospective candidates for H2 storage. Similarly Li2F, a super alkali with very low ionization energy (3.80 : 0.20 eV), could play a key role in the polarization of H2 molecules. Thus, in the present work, we are interested in exploring the H2 storage properties of h-BN sheets decorated with OLi, ONa, and Li2F. Our results include the investigation of structural stability, bonding mechanism, electronic structure, and charge analysis of functionalized systems.

A h-BN monolayer, sometimes called “white graphene”,[57] is an inorganic analog of graphene with each pair of C atoms in graphene replaced by a B-N couple in the BN sheet. The optimized 2 V 2 V 1 supercell of h-BN sheet is shown in Figure 1. B and N atoms possess different electron affinities that lead to the localization of p electrons around N atoms, and h-BN sheet behaves as a wide band gap material. With the help of Bader analysis, polar nature of B@N bond was confirmed. The average amount of charge equivalent to + 2.134 and @2.137 e was present on B and N atoms, respectively. The structural optimization of pristine h-BN sheet yields the bond length (dBN) and the lattice parameters equivalent to 1.437 and 2.489 a, respectively, which are consistent with previously reported values.[31, 57] The size of BN sheet is large enough to prevent the interaction of the molecules with their periodic images in a-b plane.

Computational Methods First-principles calculations on total energy and structural optimizations were performed within the context of spin-polarized density functional theory (DFT) using the Vienna ab initio simulation package (VASP).[50, 51] Exchange and correlation interactions were considered according to the generalized gradient approximation as implemented in the Perdew–Burke–Ernzerhof method,[52] whereas the ion-electron interactions were investigated using projector augmented wave (PAW) method.[53] We used an energy cut-off around 500 eV throughout the calculations. Our 2 V 2 V 1 super cell of h-BN consists of 9 B and 9 N atoms. The convergence criteria for optimization of structures were set to 0.05 eV a@1. We introduced a vacuum thickness of 15 a along the direction (001) perpendicular to the surface to ensure isolation of periodically repeating images. GGA is well known for accurate geometrical optimization; however, it underestimates the binding energies. For an accurate description of our system, where van der Waals interactions could be a dominant mode of interactions (between H2 molecule and the host material), we included a van der Waals correction to our calculations as proposed by Grimme.[54] The adsorption of OLi, ONa, and Li2F molecules on host material lead to charge redistributions, which were investigated using Bader charge analysis.[55] For Brillouin zone sampling, we used Monkhorst pack scheme with 5 V 5 V 1 mesh for structural relaxation. A denser k-point mesh 11 V 11 V 1 was employed for calculations of the density of states according to tetrahedron method.[56]

ChemPhysChem 2017, 18, 513 – 518

www.chemphyschem.org

Figure 1. Optimized (h-BN) sheet: a) Top and b) side view. The dotted rhombus represents a unit cell, which consists of two atoms. The lattice dimensions are: a = b = 2.489 a, c = 15 a, and g = 608. Boron (B) and Nitrogen (N) atoms are indicated as green and white balls, respectively.

Previous studies suggested that a uniform distribution of adsorbed molecules over the surface of the pristine sheet is necessary to accommodate a large number of H2.[58, 31] We functionalized BN sheets considering all the possible adsorption sites, for instance, hollow of BN hexagon, B-top, N-top and B-N bridge sites. For the structural stability of the functionalized systems, binding between the dopant molecules and the sheet is of utmost importance. The binding energies (Eb) were obtained using the following relation [Eq. (1)]: E b ¼ Eðh-BN : XÞ@Eðh-BNÞ@EðXÞ X ¼ OLi, ONa, and Li2 F

ð1Þ

Here E(h-BN:X), E(h-BN), and E(X) represent the total energies of functionalized h-BN sheet, pristine h-BN sheet, and coating molecules, respectively. The binding energies and the structural parameters for each adsorbent are listed in Table 1. 514


Articles Table 1. Binding energy [eV] for the most stable configuration, moleculeto-sheet distance (D), molecule–molecule distance (d), and final charge on each molecule. System

Binding energy per molecule, Eb [eV]

Average molecule–sheet distance, D [a]

Minimum intermolecular distance, d [a]

Charge per molecule [e]

BN:(OLi)2

@5.43

1.443

3.597

BN:(ONa)2

@5.45

1.436

3.904

BN:(Li2F)2

@1.12

2.212

4.616

[email protected] Li! + 0.877 [email protected] Na! + 0.878 Li! + 0.796 [email protected]

Both OLi and ONa molecules were strongly adsorbed on the sheet with a molecule–molecule distance of 3.597 and 3.904 a, respectively. The binding energies of OLi and ONa on h-BN reported here are much stronger than those on graphene and CNT.[58, 48] Strong adsorption energies and high enough intermolecular separation ensures that the molecular coalescence is suppressed. Electronic structure analysis confirms the presence of a strong hybridization of each of the dopant with nearby B and N atoms in the BN sheet (Figure 4). Bader analysis depicts that O atoms in both OLi and ONa molecules gain a bulk of electronic charge (@1.168 e) from nearby B atoms and the metal (Li/Na) centers as expected due to the difference of their electronegativities. Results obtained from Bader charge analysis are presented in Table 1. In a free molecular state, the O@Li and O@Na bond lengths were determined as 1.714 and 2.101 a, respectively. After their adsorption on the h-BN sheet, the bond lengths of O@Li and O@Na increased to 1.836 and 2.180 a, respectively. Bond elongation can be attributed to the transfer of electronic charge from h-BN sheet to the adsorbent molecules. Both molecules seem to be tilted towards BN bridge sites making angles of ]BOLi & 1078 and ]BONa & 112.88. Minimum distances between the adsorbent and the sheet are 1.443 and 1.436 a for OLi and ONa, respectively. Top and side views of functionalized h-BN sheets are represented in Figure 2. Unlike OLi and ONa, the adsorption of a Li2F molecule follows a different trend. Having tried all the possible binding sites, the average adsorption energy of a Li2F molecule for the most preferential configuration is @1.12 eV, which suggests relatively weaker interactions. The adsorption energy is much weaker than its value of 3.172 eV on the C60 cluster.[49] Each Li2F molecule binds to h-BN with its Li atoms attached to hollow sites. Calculations depict that the bond length (Li@F) and bond angle (]LiFLi) for a free Li2F molecule are 1.708 a and 100.108, respectively. Upon adsorption, the average molecular bond length and bond angle change to 1.752 a and 137.68, respectively. The average minimum distance between a h-BN sheet and the Li2F molecule is 2.212 a, which is relatively large compared to other adsorbents (OLi, ONa). Similarly, the cohesive energy of Li2F (1.902 eV)[49] is larger than its adsorption energy on h-BN. Thus, the molecules are prone to clustering. ChemPhysChem 2017, 18, 513 – 518


Figure 2. Side (left) and top (right) views of functionalized h-BN sheets: a) BN:(OLi)2, b) BN:(ONa)2, and c) BN:(Li2F)2. Green, red, yellow, and blue balls correspond to lithium (Li), oxygen(O), sodium (Na), and fluorine (F), respectively.

The stability of OLi, ONa, and Li2F functionalized h-BN sheets were further investigated using ab initio molecular dynamics (MD) simulations. Nose Thermostat algorithm was employed at 400 K with a time interval of 1 ps. The MD analysis and the resulting structures for OLi and ONa suggest that these systems preserve high stability even at 400 K. However, BN-(Li2F)2 system behaved differently. One Li2F molecule detached and moved closer to other Li2F molecule on opposite side of the sheet. The total energy variations could well explain the detachment of Li2F molecule from one side of the sheet and formation of cluster (Figure 3, lower panel). We found that the

Figure 3. Ab initio molecular dynamics (MD) simulations. The plots represent the variation of energy with time at 400 K, according to the Nose Thermostat algorithm.

515


Articles clustered molecules stay weakly intact with BN surface. However, the interactions between the surface and the molecule are weaker, which lead to the molecular coalescence.

2.2. Hydrogenation of OLi and ONa Functionalized Systems Functionalization with different kinds of coating molecules induces charge redistribution over the surface of h-BN, which gives rise to an enhancement of the binding affinity for H2 molecules. As previously stated, the Li atom in the OLi molecule donates its electronic charge to the neighboring highly electronegative O atom and acquires a partial positive charge (+ 0.877 e). The Li cation (Li + ) polarizes the approaching H2 molecules and holds them via electrostatic and Van der Waals interactions. The hybridization between H2 s-bond and Li(s) is responsible for this kind of interaction.[59] Similar to Li, Na cation (+ 0.878 e) polarizes the H2 molecules and behaves as an attraction-center for approaching H2 molecules. H2 s-bond and Na s-orbital interactions contribute to the adsorption of H2 molecules. The number of H2 was gradually increased while maintaining a reasonable distance to avoid unwanted electrostatic repulsion among them. The adsorption energies of H2 on functionalized h-BN sheets were calculated using the following equation [Eq. (2)]:

2.1. Electronic Density of States The electronic structures of pure h-BN, BN:(OLi)2, BN:(ONa)2, and BN:(Li2F)2 were investigated by plotting the partial density of states (PDOS) as shown in Figure 4 (a–d). In a pure BN sheet, the p-electrons localize around the N atoms due to their higher electronegativity compared to the neighboring B atoms. The absence of electronic states above the Fermi level suggests that the pure h-BN sheet is a wide-band-gap semiconductor (Figure 4 (a)). Doping with foreign species introduces electronic states above the Fermi level. It is evident from the plots that BN:(OLi)2, BN:(ONa)2, and BN:(Li2F)2 sheets exhibit a metallic character (Figure 4 (b–d)). The dotted line in each plot represents the Fermi level. We found a significant hybridization between N(p) and O(p) states close to the Fermi level at @0.5 eV. It is evident that N(p) and O(p) states are highly localized above the Fermi level owing to their higher electron affinities compared to B atoms (Figure 4 (b,c)). In the case of Li2F, we find a small hybridization between Li(s) and N(p) on the right side of Fermi level. A strong contribution of B(p) can be seen at the top of Fermi with little involvement of Li(s) and N(p). Fluorine atom seems to be slightly contributing to the binding process. Negligible hybridization of F with B or N is observed (Figure 4 (d)).

E ads ¼ Eðh-BN : X þ n H2 Þ@Eðh-BN : X þ ðn@1ÞH2 Þ@EðH2 Þ X ¼ OLi and ONa

ð2Þ

Here, E(h-BN:X + n H2) is the total energy of the functionalized h-BN sheet with n H2. The second term, E(h-BN:X + (n@1)H2), gives the energy of the functionalized h-BN sheet with (n@1) H2, and the third term E(H2) is the energy of the H2

Figure 4. Partial density of states (PDOS) plots for: a) Pure h-BN; b) BN:(OLi)2 ; c) BN:(ONa)2 ; and d) BN:(Li2F)2 sheets. The dotted line shows the Fermi level.

ChemPhysChem 2017, 18, 513 – 518


516


Articles molecule. Our findings indicate that both OLi-BN and ONa-BN systems could adsorb 6 H2 each with binding energies ranging from 0.28–0.21 eV and 0.23–0.20 eV, respectively. Here, the adsorption behavior of OLi on the h-BN sheet is a bit different to OLi functionalized graphene,[58] where 3 H2 were adsorbed around each OLi molecule with relatively lower adsorption energies (0.10–0.12 eV). The reason for higher adsorption energy per H2 in the present case could be due to a large polarization caused by an excessive charge present on each Li + in OLi-BN as compared to that of OLi-graphene system. In the present work, the adsorption energies for both OLiBN and ONa-BN lie in the desirable energy window for practical H2 storage. The complete set of the adsorption energies is given in Figure 5, which shows that the adsorption energies decrease as the number of adsorbed H2 increases. Each H2 binds to AM atom at average physisorption distance of 2.057 and 2.380 a, whereas the average H@H bond lengths were calculated as 0.755 and 0.754 a for OLi and ONa functionalized systems, respectively. Top and side views of the hydrogenated functionalized OLi and ONa systems are shown in Figure 6.

Figure 6. Top (left) and side (right) views of hydrogenated functionalized sheets: a) BN:(OLi)2 ; b) BN:(ONa)2. Green, red, yellow, and purple balls correspond to lithium (Li), oxygen(O), sodium (Na), and hydrogen (H) atoms.

cules through electrostatic and Van der Waals interaction. Both OLi-BN and ONa-BN sheets could adsorb multiple H2 molecules with binding energies that are desirable for ambient conditions. Thus, h-BN sheets functionalized with small molecules like OLi and ONa hold the promise to perform as an efficient H2 storage material.

Acknowledgements The Swedish Research Council (VR), StandUp, Swedish Energy Agency and Swedish Institute are acknowledged for financial support. Authors are grateful to SNIC and UPPMAX for provided computing time. T. Hussain is indebted to the resources at NCI National Facility systems at the Australian National University through National Computational Merit Allocation Scheme supported by the Australian Government and the University of Queensland Research Computing Centre. Figure 5. Van der Waals corrected adsorption energies of H2 on BN:(OLi)2 and BN:(ONa)2 sheets.

Keywords: electronic properties · functionalization hydrogenation · nanosheets · structural stability

3. Conclusions

[1] National Renewable Energy Laboratory (NREL). Website. http:// www.nrel.gov/news/press/2016/40763, 2015. [2] M. S. Dresselhaus, I. L. Thomas, Nature 2001, 414, 332. [3] S. Pacala, R. Socolow, Science 2004, 305, 968. [4] P. L. Bora, R. Ahmad, A. K. Singh, Int. J. Hydrogen Energy 2015, 40, 1054. [5] T. Hussain, B. Pathak, M. Ramzan, T. A. Maark, R. Ahuja, Appl. Phys. Lett. 2012, 100, 183902. [6] J. Nowotny, T. Hoshino, J. Dodson, A. J. Atanacio, M. Ionescu, V. Peterson, K. E. Prince, M. Yamawaki, T. Bak, W. Sigmund, T. N. Veziroglu, M. A. Alim, Int. J. Hydrogen Energy 2016, 41, 12812. [7] J. Nowotny, T. Bak, D. Chu, S. Fiechter, G. E. Murch, T. N. Veziroglu, Int. J. Hydrogen Energy 2014, 39, 4151. [8] L. Schlapbach, Nature 2009, 460, 809. [9] P. Jena, J. Phys. Chem. Lett. 2011, 2, 206. [10] U.S. Department of Energy’s Office of Renewable Energy Efficiency and Renewable Energy. Website. https://www1.eere.energy.gov/hydrogenandfuelcells/storage/current technology.html, 2010. [11] N. Faginas-Lago, D. Yeni, F. Huarte, Y. Wang, M. Alcami, F. Martin, J. Phys. Chem. A 2016, 120, 6451.

In this study, we investigated the stability, electronic structure, and H2 storage capability of OLi-, ONa-, and Li2F-functionalized h-BN systems. Our findings indicate that Li2F molecules weakly bind to h-BN and tend to form molecular clusters at high doping concentrations. The uniform and stable distribution of Li2F molecules on h-BN sheet could not be attained, which hinders the practicability of BN:(Li2F)2 system for H2 storage. On the other hand, OLi and ONa molecules strongly bind to the sheet with adequate adsorption energy for H2 storage. Ab initio MD simulations at a high enough temperature of 400 K further confirmed the structural stabilities of both OLi-BN and ONa-BN monolayers. The difference of electronegativities causes the alkali atoms in OLi-BN and ONa-BN in partial positive states and facilitates the adsorption of incident H2 moleChemPhysChem 2017, 18, 513 – 518


·

517


Articles [12] S.-J. Hwang, Y.-S. Chuang, J. Alloys Compd. 2016, 664, 284. [13] W. Liu, Y. H. Zhao, Y. Li, Q. Jiang, E. J. Lavernia, J. Phys. Chem. C 2009, 113, 2028. [14] D. J. Durbin, N. L. Allan, C. M-. Jugroot, Int. J. Hydrogen Energy 2016, 41, 13116. [15] C. M. Ramos-Castillo, J. U. Reveles, M. E. Cifuentes-Quintal, R. R. Zope, R. de Coss, J. Phys. Chem. C 2016, 120, 5001. [16] E. S. Cho, A. M. Ruminski, S. Aloni, Y.-S. Liu, J. Guo, J. J. Urban, Nat. Commun. 2016, 7, 10804. [17] C. Zhou, J. A. Szpunar, X. Cui, ACS Appl. Mater. Interfaces 2016, 8, 15232. [18] C. M. Ramos-Castillo, J. U. Reveles, R. R. Zope, R. de Coss, J. Phys. Chem. C 2015, 119, 8402. [19] D. J. Durbin, C. Malardier-Jugroot, Int. J. Hydrogen Energy 2013, 38, 14595. [20] T. Hussain, M. Hankel, D. J. Searles, J. Phys. Chem. C 2016, 120, 25180. [21] D. Umadevi, S. Panigrahi, G. N. Sastry, Acc. Chem. Res. 2014, 47, 2574. [22] T. Hussain, P. Panigrahi, R. Ahuja, Phys. Chem. Chem. Phys. 2014, 16, 8100. [23] D. Umadevi, G. N. Sastry, ChemPhysChem 2013, 14, 2570. [24] D. Umadevi, G. N. Sastry, Curr. Sci. 2014, 106, 1224. [25] Q. Sun, Q. Wang, P. Jena, Nano Lett. 2005, 5, 1273. [26] M. Chen, Y. J. Zhao, J. H. Liao, X. B. Yang, Phys. Rev. B 2012, 86, 045459. [27] S. A. Shevlin, Z. X. Gue, Phys. Rev. B 2007, 76, 024104. [28] R. N. Muthu, S. Rajashabala, R. Kannan, Renewable Energy 2016, 85, 387. [29] J. Zhou, Q. Wang, Q. Sun, P. Jena, X. S. Chen, Proc. Natl. Acad. Sci. USA 2010, 107, 2801. [30] Q. H. Weng, X. B. Wang, C. Y. Zhi, Y. Bando, D. Golberg, ACS Nano 2013, 7, 1558. [31] T. Hussain, A. D. Sarkar, T. W. Kang, R. Ahuja, Sci. Adv. Mater. 2013, 5, 1960. [32] R. Ma, Y. Bando, H. Zhu, T. Sato, C. Xu, D. Wu, J. Am. Chem. Soc. 2002, 124, 7672. [33] T. Hussain, D. J. Searles, K. Takahashi, J. Phys. Chem. A 2016, 120, 2009. [34] S.-H. Jhi, Y.-K. Kwon, Phys. Rev. B 2004, 69, 245407. [35] Q. Weng, X. Wang, Y. Bando, D. Golberg, Adv. Energy Mater. 2014, 4, 1301525. [36] S. Nigam, C. Majumder, ACS Nano 2008, 2, 1422.

ChemPhysChem 2017, 18, 513 – 518


[37] Y. Lin, T. V. Williams, W. Cao, H. E. Elsayed-Ali, J. W. Connell, J. Phys. Chem. C 2010, 114, 17434. [38] T. Yildirim, S. Ciraci, Phys. Rev. Lett. 2005, 94, 175501. [39] A. S. Rad, K. Ayub, Int. J. Hydrogen Energy 2016, 41, 22182. [40] M. Mananghaya, D. Yu, G. N. Santos, Int. J. Hydrogen Energy 2016, 41, 13531. [41] N. S. Venkataramanan, M. Khazaei, R. Sahara, H. Mizuseki, Y. Kawazoe, Chem. Phys. 2009, 359, 173. [42] S. Bhattacharya, A. Bhattacharya, G. P. Das, J. Phys. Chem. C 2012, 116, 3840. [43] M. Khazaei, M. S. Bahramy, N. S. Venkataramanan, H. Mizuseki, Y. Kawazoe, J. Appl. Phys. 2009, 106, 094303. [44] S.-H. Wen, W. Q. Deng, K. L. Han, J. Phys. Chem. C 2008, 112, 12195. [45] R. J. Baierlo, P. Piquini, T. M. Schmidt, A. Fazzio, J. Phys. Chem. B 2006, 110, 21184. [46] F. Shayeganfar, R. Shahsavari, Langmuir 2016, 10, 23. [47] P. Banerjee, B. Pathak, R. Ahuja, G. P. Das, Int. J. Hydrogen Energy 2016, 41, 14437. [48] Q. Peng, G. Chen, H. Mizuseki, Y. Kawazoe, J. Chem. Phys. 2009, 131, 214505. [49] K. Wang, Z. Liu, X. Wang, X. Cui, Int. J. Hydrogen Energy 2014, 39, 15639. [50] G. Kresse, J. Hafner, Phys. Rev. B 1993, 47, 558. [51] G. Kresse, J. Hafner, Phys. Rev. B 1993, 49, 14251. [52] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865. [53] P. E. Blochl, Phys. Rev. B 1994, 50, 17953. [54] S. Grimme, J. Comput. Chem. 2006, 27, 1787. [55] R. F. W. Bader, Atoms in Molecules—a Quantum Theory, Oxford University Press, Oxford, 1990. [56] H. J. Monkhorst, J. D. Pack, Phys. Rev. B 1976, 13, 5188. [57] H. Zeng, C. Zhi, Z. Zhang, X. Wei, X. Wang, W. Guo, Y. Bando, D. Golberg, Nano Lett. 2010, 10, 5049. [58] E. Seleyman, G. A. de Wijs, G. Brocks, J. Phys. Chem. C 2009, 113, 8997. [59] R. Davy, E. Skoumbourdis, T. Kompanchenko, Mol. Phys. 1999, 97, 1263. Manuscript received: September 29, 2016 Revised: December 14, 2016 Final Article published: January 18, 2017

518
