BC3 Sheet Functionalized with Lithium-Rich SpeciesEmerging as a Reversible Hydrogen Storage MaterialTanveer Hussain,*[a, b] Sudip Chakraborty,*[a] T. W. Kang,[c] Bçrje Johansson,[b] andRajeev Ahuja[a, b, c]

1. Introduction

Energy inadequacy has always been a critical issue in theglobal economy because of the tremendous increase in energyconsumption over the last few decades. The insufficient availa-bility of fossil fuels to cope with the current and future needsof energy and the environmental issues associated with themdrives us to go for cleaner, cheaper, and renewable alternates.Being pollution free, abundant, and sustainable with the high-est energy density by mass, hydrogen (H2) can replace the cur-rent energy carriers and can also be stored in an efficientway.[1] The search for specific materials that can store H2 hasextensively been going on during the last few years to envis-age a solution for the long-standing storage challenge and fa-cilitate the transition towards a hydrogen economy.[2–8] The USDepartment of Energy (DOE) has set an H2 storage target of5.5 wt %, which has to be achieved by 2017.[9] Therefore, theidentification of a suitable material capable of storing a suffi-cient amount of H2 under feasible conditions demands a pro-found understanding of metal–hydrogen interactions. Hydro-gen can interact with the host material in the atomic formwith significantly high binding energies of 2–4 eV. This requiresa very high desorption temperature, hence making this pro-cess less attractive for mobile applications. On the contrary, H2

interactions with materials in the molecular form involve weakvan der Waals (vdW) forces that permit operation in the low-temperature range.[10] This mechanism would certainly help toattain higher H2 gravimetric densities and fast kinetics, leadingto a desirable candidate for mobile applications.

Along with the fascinating properties and utility in variousfields of technology, the large surface area of carbon-basednanostructures also offer great promise as high capacity H2

storage materials. However, the relatively smaller values of theadsorption energies between the host material’s surface andthe H2 molecules affect the adsorption/desorption mechanismand hence restrict their applications at ambient conditions inthe pristine form.[11–13] To overcome this bottleneck, the interac-tion between the host (carbon-based) materials and H2 mole-cules must be enhanced, which could facilitate the adsorptionand desorption phenomena occurring under feasible physicalconditions. Two promising and extensively used procedures forachieving this goal are the application of an electric field[14–16]

and the functionalization of host materials with various foreignelements.[17–21]

There is an unprecedented interest in novel two-dimensional(2D) materials after the experimental exfoliation of graphene,as they have exceptional physical and chemical properties.[22]

In the field of H2 storage, various graphene-like 2D nanostruc-tures have been the subject of many experimental and theo-retical investigations. Van-der-Waals-induced calculations to in-vestigate the adsorption mechanism of H2 molecules on a calci-um-doped BC7 sheet have been performed recently by Leiet al. The authors reported an average adsorption energy of H2

over Ca-BC7 of 0.26 eV, with a storage capacity of 4.96 wt %.[23]

A graphene-like boron sheet functionalized with light alkalimetals was studied and compared with functionalized gra-phene by Er et al. regarding its H2-storage properties. A strongmetal-to-substrate binding and a feasible adsorption of H2

molecules around the dopants resulted in achieving a high

The decoration of a BC3 monolayer with the polylithiated mol-ecules CLi4 and OLi2 has been extensively investigated to studythe hydrogen-storage efficiency of the materials by first princi-ples electronic structure calculations. The binding energies ofboth lithiated species with the BC3 substrate are much higherthan their respective cohesive energies, which confirms thestability of the doped systems. A significant positive charge on

the Li atom in each of the dopants facilitates the adsorption ofmultiple H2 molecules under the influence of electrostatic andvan der Waals interactions. We observe a high H2-storage ca-pacity of 11.88 and 8.70 wt % for the BC3-CLi4 and BC3-OLi2 sys-tems, respectively, making them promising candidates as effi-cient energy-storage systems.

[a] Dr. T. Hussain, Dr. S. Chakraborty, Prof. R. AhujaCondensed Matter Theory GroupDepartment of Physics and AstronomyBox 516, Uppsala University, S-75120 Uppsala (Sweden)E-mail : tanveer.hussain@physics.uu.se

sudip.chakraborty@physics.uu.se

[b] Dr. T. Hussain, Prof. B. Johansson, Prof. R. AhujaApplied Materials PhysicsDepartment of Materials and EngineeringRoyal Institute of Technology (KTH), S-100 44 Stockholm (Sweden)

[c] Prof. T. W. Kang, Prof. R. AhujaQuantum Functional Semiconductor Research Center (QSRC)Dongguk University, 26 Phildong 3ga, Chung guSeoul 100-715 (Republic of Korea)

ChemPhysChem 2015, 16, 634 – 639 � 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim634

ArticlesDOI: 10.1002/cphc.201402696

gravimetric density of 10.7 wt %, which has been seen in thatparticular study.[24] Based on the first principles electronic struc-ture calculations, Hu et al. studied the possibilities of using lay-ered Ti2C, Sc2C and V2C (MXene phases) as efficient H2-storagematerials. The possible modes of hydrogen binding that havebeen studied are mainly confined within chemisorption, physi-sorption, and Kubas interaction. The maximum storage capaci-ty (8.6 wt %) and the adsorption energies (0.272 eV) for H2 byMXene phases reported in this study fall within the desiredrange of reversible H2-storage materials.[25] Besides the metal(alkali, alkaline, transition metals) functionalization, there is an-other fascinating class of materials to be used as dopants forefficient H2 storage called polylithiated compounds. These spe-cies with high density of lithium atoms have been studied andfound stable.[26–28] The prime advantage of decorating 2Dnanostructures with polylithiated compounds is the availabilityof multiple Li atoms capable of adsorbing extra H2 moleculesresulting in a significantly high gravimetric density, which isone of the main requirements for efficient storage materials.Previous studies dealing with the doping of polylithiated spe-cies over graphene, graphane, BN, and BC3 monolayers haveshown promise for such systems as reversible H2-storage mate-rials.[29–31]

The present investigation deals with the functionalization ofa BC3 monolayer with two important members of the polylith-iated family, CLi4 and OLi2. Usually polylithiated molecules arerepresented by the general formula CLin or OLim. They are notrestricted to n = 3–5 and m = 1–4, and molecules with more Liatoms also exist, for example, OLi5, CLi6 and so on(Ref. [26, 28]). But as the density of Li increases from n>4 inCLin and m>2 in OLim, the average charge on Li tends to de-crease due to the metalicity of the coordination shell.[24] Thiscould reduce the ability of Li+ in each polylithiated moleculeto adsorb more H2 molecules through electrostatic interac-tions. Hence, to maximize the H2-storage capacity, the partialcharge on Li should be maximum. In the current manuscript,we have considered OLi2 and CLi4 due to their ability to maxi-mize the charge on Li in both of these species. The stability ofthe dopants over the substrate is also of fundamental impor-tance, which would be compromised by the use of hyper-coor-dinated polylithiated species, OLi4, and CLi5 and so on.

In addition to the stability of the doped systems (BC3-CLi4,BC3-OLi2), we have investigated their H2-storage capacity andthe adsorption mechanism. We have found that with two-sided dopant coverage, the target that has been set by DOEcan easily be achieved indicating the promise of these func-tionalized systems as efficient and reversible H2-storage materi-als.

2. Computational Details

In this study, we have performed all the electronic structurecalculations by using the density functional theory (DFT) for-malism as implemented in the VASP code with the plane wavebasis set having an energy cutoff of 500 eV.[33, 34] We have em-ployed the projector augmented wave method (PAW) thattreats the valence electrons for Li (1s22s1), B (2s22p1), and C

(2s22p2).[35] We have adopted the generalized gradient approxi-mation of the Perdew–Burke–Ernzerhof[36] type exchange corre-lation functional. Since the adsorption of H2 molecules onthese systems involve weak forces (physisorption), the use ofa local density approximation (LDA) has been found useful indealing with these weakly interacting systems in some earlierstudies.[37–39] However, the GGA and LDA functionals usuallyunder- and overestimate the binding energies, respectively,and hence we have incorporated the van der Waals correctionby using the semi-empirical of the DFT-D2 method by Grimmeto have a reasonably better description and reliable adsorptionenergies.[40] We have used the Monkhorst–Pack scheme tosample the Brillouin zone with 7 � 7 � 1 k-point mesh for opti-mization and a thicker 13 � 13 � 1 for obtaining the density ofstates[41] of the individual systems. The energy convergence ofthe starting system was confirmed with respect to the k-pointmesh. The complete structural optimization was performeduntil the forces on each ion became less than 0.001 eV ��1. Toobtain a uniform distribution of dopants over the BC3 mono-layer by preserving an optimum distance between them, weconsidered a 2 � 2 � 1 supercell of the BC3 sheet with 8 boronand 24 carbon atoms. A vacuum space of 15 � was insertedalong the (001) direction to avoid the interaction between therepeating images due to periodic boundary conditions. Thebinding energy (Eb) of the dopants (CLi4, OLi2) over the BC3

sheet was determined using Equation (1):

Eb ¼ EðBC3 : CLi4=OLi2Þ�EðBC3Þ�EðCLi4=OLi2Þ ð1Þ

where E(BC3: CLi4/OLi2) and E(BC3) are the total energies of theBC3 sheet doped with CLi4/OLi2 and in the pristine form, re-spectively. E(CLi4/OLi2) is the total energy of the lithiated com-pounds. The adsorption energies (Eads) of H2 molecules ad-sorbed by these functionalized systems have been calculatedusing Equation (2):

Eads ¼ ½fEðBC3 : CLi4=OLi2Þ þ n H2g�fEðBC3 : CLi4=OLi2Þþðn�1ÞH2g�fEðH2Þg�

ð2Þ

The first two terms represent the total energies of thedoped systems with n and n�1 H2 molecules, respectively, andE(H2) is the energy of a single hydrogen molecule.

3. Results and Discussion

The experimentally synthesized BC3 monolayer is an analogueof two-dimensional graphene, with each fourth C atom beingreplaced by a B atom to form a honeycomb structure.[42, 43]

Unlike the bulk BC3 compound with adjacent layers resultinginto a metallic structure with promising H2-storage properties,the BC3 monolayer is a semiconducting material.[43] We startwith the optimized structure of a bare BC3 sheet, as shown inFigure 1, with B�C and C�C distances of 1.56 and 1.42 �, re-spectively. It has been observed that each B atom transfers thebulk of its charge (�1.83 e) to the three neighboring C atoms,calculated by Bader charge analysis.[44] After having the opti-mized structure of BC3, one of the important questions to

ChemPhysChem 2015, 16, 634 – 639 www.chemphyschem.org � 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim635

Articles

answer is “can the dopants (CLi4, OLi2) bind strongly to the BC3

sheet without being clustered?” To answer this, we investigat-ed the binding mechanism of both the lithiated species on theBC3 monolayer. The uniform distribution of dopants over thesubstrate, with a binding energy significantly above the bulkenergy, is necessary for a reversible H2-storage material. Other-wise clustering instead of binding will affect the H2-storage ca-pacity and reversibility of the proposed material.

We have considered all thepossible binding sites for thedopants, which include B-top, C-top, two hollow, and two bridgesites. After calculating the bind-ing energies for both dopantsusing Equation (1) for all thepossible configurations, it is ob-served that CLi4 prefers to bindon the C-top whereas OLi2 onthe B-top of the BC3 sheet. Thecorresponding binding energiesof CLi4 and OLi2 over BC3, withtwo-sided coverage, are �4.04and �4.66 eV, respectively. Thedistances of CLi4 and OLi2 to theBC3 sheet are approximated to1.604 and 1.537 �, respectively. Itshould be noted that the report-ed Eb values for CLi4 and OLi2

over the BC3 sheet are muchhigher than the binding energiesover graphene and graphanesheets.[29, 30] The Eb values for CLi4

and OLi2 are also much higherthan their respective dimeriza-tion energies, namely, 2.54 eV/CLi4 and 1.37 eV/OLi2, indicatingthe stability of the functionalizedsystems. In addition to the highEb, the large distances betweenthe dopants CLi4-CLi4 (4.87 �)and OLi2-OLi2 (7.3 �) further nul-lify the possibility of cluster for-mation. The binding of both

dopants does not affect the structure of any of them or theBC3 sheet. The average C�Li and O�Li distances in both CLi4

and OLi2 are 2.0 and 1.77 �, respectively. Thus, the polylithiatedspecies can be distributed uniformly over the BC3 monolayerresulting in stable functionalized materials.

To investigate the bonding mechanism and charge state ofthe doped systems, it will be interesting to see the transfer ofcharge between the substrate, BC3, and the dopants (CLi4,OLi2). We have employed the Bader charge analysis for thispurpose. The higher electronegativities of both C and O com-pared to Li in both CLi4 and OLi2 results in strongly polar C�Liand O�Li bonds with a significant positive charge at the Li.When we introduce CLi4 on BC3 sheets, there is a transfer ofcharge from the sheet to the dopants through C�C bonding.In the case of OLi2, the transfer of charge will be from the BC3

sheet towards the dopants via B�O bonding because of thehigher electronegativity of the O atoms.

After discussing the stability of polylithiated functionalizedBC3, we now turn our attention to the electronic properties ofthese doped systems. For this, we have plotted the total andpartial density of states for the pure and doped BC3 monolay-ers. A semiconducting behavior is observed with a band gap

Figure 1. Optimized structure of a BC3 monolayer. The brown and black ballsrepresent B and C atoms, respectively.

Figure 2. Optimized top and side views of bare and hydrogenated BC3-CLi4 and BC3-OLi2. The brown, black, blue,red, and green balls represent B, C, O, Li and H atoms, respectively.

ChemPhysChem 2015, 16, 634 – 639 www.chemphyschem.org � 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim636

Articles

of 0.54 eV in the case of the pure BC3 monolayer, as shown inFigure 3. There is a strong hybridization between the p orbitalsof both B and C atoms. As the B atoms donate the bulk oftheir valence charge to the C atoms, due to the higher electro-negativity of C, there is a significant ionic character involved intheir bonding. Upon introduction of CLi4, there is a transitionfrom a semiconductor to a metallic system, as shown inFigure 4. The states arising in the band gap result from the p

orbitals of the C atoms of the dopant (CLi4). The lower panel ofFigure 4 shows the partial densities of state, showing a hybridi-zation between the C atoms of BC3, denoted by C1, and the Catoms of CLi4, denoted by C2.

In the case of BC3:OLi2, both the O atoms of OLi2 and the Batoms of BC3 strongly hybridize in the band gap, as shown inFigure 5. A strong contribution from the p orbitals of the OLi2

molecules just before the Fermi level is an indication of thecharge transfer between the monolayer and the dopant.

Finally, we describe the mechanism of H2 adsorption (hydro-genation) on the BC3�CLi4 and BC3�OLi2 systems. As men-

tioned earlier, a significant amount of positive charge accumu-lates on each Li+ in both dopants CLi4 and OLi2. With the two-sided doping of the lithiated molecules, CLi4 has eight andOLi2 has four Li+ ions available for the adsorption of H2 in theBC3 monolayers. We start with the adsorption of one H2 mole-cule on each Li+ on either side of the functionalized monolay-ers. As the H2 molecule approaches the Li+ species in the BC3-CLi4/BC3-OLi2 system, it is held by electrostatic interactions. It isobserved that Li+ also polarizes the H2 molecules, whichmeans that apart from electrostatic interactions, polarization isalso responsible for the adsorption of H2 around Li+ . The weakvan der Waals interaction plays a vital role in describing thiskind of attraction. Thus, to approximate the adsorption ener-gies of H2 molecules with the functionalized BC3 monolayerwith accuracy, the inclusion of a van der Waals correction onthe top of the DFT energy is extremely important. The adsorp-tion energies of the H2 molecules were calculated using Equa-tion (2) with the help of both LDA and van-der-Waals-inducedGGA-PBE. To maximize the H2-storage capacity, we kept in-creasing the number of H2 molecules around each Li+ speciesin a stepwise manner while maintaining a sufficient distance tothe Li+ as well as among H2 molecules. Each time after the in-troduction of H2, the system was fully relaxed. It is also worthmentioning that the H2 molecules adsorbed by CLi4/OLi2 retainthe molecular behavior. The first H2 molecule on CLi4 (OLi2)had an adsorption energy (Eads) of �0.423 eV (�0.437 eV). Amaximum of three H2 molecules can be adsorbed around eachLi+ in CLi4 with an average Eads of �0.213 eV and a significantlyhigh storage capacity of 11.88 wt %. In the case of OLi2, four H2

molecules can be held around each Li+ with an average Eads of�0.223 eV that lies within the desired range for practical appli-cations. The storage capacity 8.70 wt % of BC3-OLi2 easily sur-passes the target set by the DOE to be achieved by 2017.[45]

Further addition of H2 beyond this limit on both systems isthought to be hindered by the electrostatic repulsion amongthe adsorbed molecules. The complete results, showing thevariation of adsorption energies with increasing number of H2

molecules, are shown in Figure 6. The optimized structures ofhydrogenated BC3�CLi4 and BC3�OLi2 are shown in Figure 2.

Figure 3. Total and partial density of states of a pure BC3 monolayer.

Figure 4. Total and partial density of states of a CLi4-doped BC3 monolayer.

Figure 5. Total and partial density of states of an OLi2-doped BC3 monolayer.

ChemPhysChem 2015, 16, 634 – 639 www.chemphyschem.org � 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim637

Articles

4. Conclusion

We have performed systematic first principles calculations,based on density functional theory, to design polylithiatedfunctionalized BC3 monolayers and investigate their H2-storageefficiencies. Two important members of the polylithiatedfamily, namely, CLi4 and OLi2, are found to be strongly bondedover the C- and B-top of the BC3 sheet, respectively. The calcu-lated binding energies for both dopants, CLi4 and OLi2, overthe BC3 monolayer are much higher than their correspondingdimerization energies, ensuring the stabilities of the BC3-CLi4

and BC3-OLi2 systems. A large distance between the dopantsensures their uniform distribution over the substrate andmakes the possibility of clustering negligibly small. Because ofthe difference in electronegativities, a significant amount ofcharge has been transferred from Li to C and O in both CLi4

and OLi2, resulting in the formation of a Li+ ion. The adsorp-tion of H2 molecules around Li+ results from electrostatic andweak van der Waals interactions. The H2-storage capacities forthe two-sided coverage of CLi4 and OLi2 are found to be 11.88and 8.70 wt %, respectively, with optimum adsorption energiesfor the practical H2-storage systems. Thus, we have presented

computationally engineered stable systems for the high-ca-pacity storage of hydrogen at ambient conditions, which cancertainly be beneficial for practical energy applications.

Acknowledgements

We would like to acknowledge the Carl Tryggers Stiftelse for Ve-tenskaplig Forskning (CTS), the Swedish Research Council (VR),and the Swedish Energy Agency for financial support. SNIC andUPPMAX are acknowledged for providing computing time. Thisresearch was supported by Leading Foreign Research Institute Re-cruitment Program through the National Research Foundation ofKorea(NRF) funded by the Ministry of Education, Science andTechnology(MEST) (No.2014-039452)

Keywords: energy materials · first principles calculations ·functionalization · hydrogen storage · lithium

[1] L. Schlapbach, Nature 2009, 460, 809.[2] J. C. Yang, Int. J. Hydrogen Energy 2008, 33, 4424.[3] B. Sakintuna, F. L-Darkrim, M. Hirscher, Int. J. Hydrogen Energy 2007, 32,

1121.[4] C.-S. Tsao, Y. Liu, M. Li, Y. Zhang, J. B. Leao, H.-W. Chang, M.-S. Yu, S.-H.

Chen, J. Phys. Chem. Lett. 2010, 1, 1569.[5] T. Hussain, T. Kaewmaraya, S. Chakraborty, R. Ahuja, Phys. Chem. Chem.

Phys. 2013, 15, 18900.[6] Q. Sun, Q. Wang, P. Jena, Appl. Phys. Lett. 2009, 94, 013111.[7] A. Blomqvist, C. M. Araujo, P. Srepusharawoot, R. Ahuja, Proc. Natl. Acad.

Sci. USA 2007, 104, 20173.[8] A. R. Muniz, M. Meyyappan, D. Maroudas, Appl. Phys. Lett. 2009, 95,

163111.[9] DOE: US Department of Energy. Website: http://www.doe.gov.

[10] P. Jena, J. Phys. Chem. Lett. 2011, 2, 206.[11] C. Ataca, E. Akturk, S. Ciraci, H. Ustunel, Appl. Phys. Lett. 2008, 93,

043123.[12] H. Lee, J. Ihm, M. L. Cohen, S. G. Louie, Nano Lett. 2010, 10, 793.[13] R. Lu, D. Rao, Z. Meng, X. Zhang, G. Xu, Y. Liu, E. Kan, C. Xiao, K. Deng,

Phys. Chem. Chem. Phys. 2013, 15, 16120.[14] W. Liu, Y. H. Zhao, J. Nguyen, Y. Li, Q. Jiang, E. J. Lavernia, Carbon 2009,

47, 3452.[15] J. Zhou, Q. Wang, Q. Sun, P. Jena, X. S. Chen, Proc. Natl. Acad. Sci. USA

2010, 107, 2801.[16] Z. Zhang, J.-Y. Hwang, M. Ning, X. Li, Int. J. Hydrogen Energy 2012, 37,

16018.[17] A. Blomqvist, C. M. Araujo, P. Srepusharawoot, R. Ahuja, Proc. Natl. Acad.

Sci. USA 2007, 104, 20173.[18] H. Lee, J. Ihm, M. L. Cohen, S. Louie, Phys. Rev. B 2009, 80, 115412.[19] T. Hussain, T. A. Maark, A. De Sarkar, R. Ahuja, Appl. Phys. Lett. 2012, 101,

243902.[20] R. Zacharia, K. Y. Kim, A. K. M. F. Kibria, K. S. Nahm, Chem. Phys. Lett.

2005, 412, 369.[21] T. Hussain, B. Pathak, M. Ramzan, T. A. Maark, R. Ahuja, Appl. Phys. Lett.

2012, 100, 183902.[22] K. S. Novoselov, A. K. Geim, S. V. Morozov, G. Jiang, Y. Zhang, S. V. Bubo-

nos, I. V. Grigorieva, A. A. Firsov, Science 2004, 306, 666.[23] X. L. Lei, G. Liu, M. S. Wu, B. Xu, C. Y. Ouyang, B. C. Pan, Int. J. Hydrogen

Energy 2014, 39, 2142.[24] S. Er, G. A. de Wijs, G. Brocks, J. Phys. Chem. C 2009, 113, 18962.[25] Q. Hu, D. Sun, Q. Wu, H. Wang, L. Wang, B. Liu, A. Zhou, J. He, J. Phys.

Chem. A 2013, 117, 14253.[26] P. V. R. Schleyer, E. U. Wuerthwein, E. Kaufmann, T. Clark, J. A. Pople, J.

Am. Chem. Soc. 1983, 105, 5930.[27] C. H. Wu, H. Kudoa, H. R. Ihle, J. Chem. Phys. 1979, 70, 1815.[28] H. Kudo, Nature 1992, 355, 432.[29] S. Er, G. A. de Wijs, G. Brocks, J. Phys. Chem. C 2009, 113, 8997.

Figure 6. Adsorption energy per H2 molecule for the BC3�CLi4 and BC3�OLi2

systems calculated by LDA and vdW-induced calculations.

ChemPhysChem 2015, 16, 634 – 639 www.chemphyschem.org � 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim638

Articles

[30] T. Hussain, A. De Sarkar, R. Ahuja, Int. J. Hydrogen Energy 2014, 39, 2560.[31] T. Hussain, A. De Sarkar, T. W. Kang, R. Ahuja, Sci. Adv. Mater. 2013, 5, 1.[32] P. Hohenberg, W. Kohn, Phys. Rev. 1964, 136, B864.[33] G. Kresse, J. Hafner, Phys. Rev. B 1993, 47, 558.[34] P. E. Blçchl, Phys. Rev. B 1994, 50, 17953.[35] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865.[36] J. P. Perdew, K. Burke, M Ernzerhof, J. Chem. Phys. 1996, 105, 9982.[37] Z. M. Ao, F. M. Peeters, Phys. Rev. B 2010, 81, 205406.[38] Z. Yang, J. Ni, Appl. Phys. Lett. 2012, 100, 183109.[39] S. Grimme, J. Comput. Chem. 2006, 27, 1787.[40] H. J. Monkhorst, J. D. Pack, Phys. Rev. B 1976, 13, 5188.[41] H. Sun, F. J. Ribeiro, J.-L. Li, D. Roundy, M. L. Cohen, S. G. Louie, Phys.

Rev. B 2004, 69, 024110.

[42] H. Tanaka, Y. Kawamata, H. Simizu, T. Fujita, H. Yanagisawa, S. Otani, C.Oshima, Solid State Commun. 2005, 136, 22.

[43] X. Sha, A. C. Cooper, W. H. Bailey, H. Cheng, J. Phys. Chem. C 2010, 114,3260.

[44] R. F. W. Bader, Atoms in Molecules-A Quantum Theory, Oxford UniversityPress, Oxford, 1990.

[45] US. DOE. EERE, Fuel Cell Technologies Program Multi-Year Research, De-velopment and Demonstration Plan, Section 3.3.

Received: October 8, 2014

Published online on November 25, 2014

ChemPhysChem 2015, 16, 634 – 639 www.chemphyschem.org � 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim639

Articles