DOI: 10.1002/cphc.201300907

The Influence of N-doped Carbon Materials on SupportedPd: Enhanced Hydrogen Storage and Oxygen ReductionPerformanceXiang-Kai Kong,[a] Qian-Wang Chen,*[a, b] and Zheng-Yan Lun[a]

1. Introduction

Carbon materials have always been a hot research topic andhave attracted increasing attention among scientists—fromthe discovery of carbon nanotubes and fullerene to the discov-ery of graphene and graphyne. On account of their exceptionalphysical and chemical properties, especially for graphenewhich is a two-dimensional single layer of carbon atoms withhoneycomb structure, carbon materials have been widely stud-ied to look for probable applications in hydrogen storage,[1–4]

catalytic reactions,[5–8] energy storage,[9, 10] and so on. Such ma-terials are on a rapid path to commercialization.

Both Pd and Pt are pivotal candidates for hydrogen stor-age[11, 12] and catalytic reactions.[13] Because of their aggregationin the practical usage, researchers have designed to placethem with uniform distribution on porous substrates. Althoughthere are many kinds of materials (including metal oxides andzeolites) that are suitable supports, carbon materials are oftenthe most preferred options,[14] owing to the fact that they arecheap, easy to be obtained and have good biocompatibility aswell as large storage capacity. To investigate the interactionsbetween these carbon materials and their supported transitionmetals, many simulations have been carried out. Corral et al.have studied the hydrogen adsorption on Pd decorated gra-phene and single walled carbon nanotubes, to understand thehydrogen uptake in these systems.[12, 15, 16] Pd clusters from one

single atom to the three-dimensional structure composed offive atoms on graphene have also been investigated systemati-cally, to explore the bonding between metal clusters and theirsupports, as well as the strong Pd–Pd interactions.[17] Pd and Pthave also been found to be the most promising decorationson graphene for hydrogen storage due to their reasonable re-lationship of adsorption energies, which could minimize theoxygen interference.[18]

Recently, growing attention has been given to N-dopedcarbon supports for their interesting and outstanding perfor-mance in many experiments. Xiong et al. reported excellentelectrochemical properties for catalytic methanol oxidation ofPt particles on N-doped graphene supports.[19] Xu et al. foundthat the Pd@CN0.132 catalyst could show very high catalytic ac-tivity in hydrodeoxygenation of vanillin at low hydrogen pres-sure under mild conditions in aqueous media, and they attrib-uted the better performance to the strong interactions be-tween metal particles and carbon surfaces.[14] In addition, it isalso reported that nitrogen-doped carbon materials could im-prove the biocompatibility significantly.[20] Therefore, modifica-tion of carbon materials with N-doping is necessary in mostcases, which could not only change their electronic states forbetter operation of their supported metal particles, but alsoreduce the toxicity, facilitating their practical applications inwide fields.

It has been mentioned that N-doped carbon materials willenhance the combined interaction for their supported metalparticles,[14] improving their performance in catalytic reactionsin experiment. However, as far as we know, there is no theoret-ical investigation concerning on this problem. So we havetried to study this based on density functional theory (DFT) cal-culations here. Graphene, fullerene, and carbon clusters haveall been considered here, which contain both sp2 and sp3 hy-bridized carbon atoms and they have become the most prom-ising materials in the future. Through the analysis of these op-

N-doped graphene has become an important support for Pd inboth hydrogen storage and catalytic reactions. The molecularorbitals of carbon materials (including graphene, fullerene, andsmall carbon clusters) and those of the supported Pd specieswill hybrid much stronger as N dopants are introduced, owingto the increased electrostatic attraction at the interface. Thisenhances the carbon substrates’ catching force for the sup-

ported Pd, preventing its leaching and aggregation in manypractical applications. The better dispersion and stabilization ofPd nanoparticles, which are induced by various carbon sup-ports with N-doping, are pleasing to us and could increasetheir efficiency and facilitate their recycling during various re-action processes in several fields.

[a] Dr. X.-K. Kong, Prof. Q.-W. Chen, Z.-Y. LunHefei National Laboratory forPhysical Sciences at the MicroscaleDepartment of Materials Science & Engineering& Collaborative Innovation Center ofSuzhou Nano Science and TechnologyUniversity of Science and Technology of China, Hefei (China)E-mail : cqw@ustc.edu.cn

[b] Prof. Q.-W. ChenHigh Magnetic Field LaboratoryHefei Institutes of Physical ScienceChinese Academy of Sciences, Hefei (China)

� 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 2014, 15, 344 – 350 344

CHEMPHYSCHEMARTICLES

timized structures and calculated total electron densities, it isfound that N doping could really strengthen the interactionbetween the carbon substrates and their supported Pd metalboth in hydrogen storage and oxygen reduction reaction(ORR), which will reduce the aggregation and leaching of Pd inthe application process. Via calculating of the partial density ofstates (PDOS), we have tried to study their enhancementmechanism.

2. Methods

All quantum chemical calculations were performed using the Gaus-sian 09 software.[21] The graphene model discussed herein con-tained 42 carbon atoms and 16 hydrogen atoms and was takenfrom Sidik’s structure. This model has been used well in our previ-ous work.[8, 22–24] The structure of fullerene was taken from the clas-sical buckyball model, which was composed of 60 carbon atoms,and for the carbon cluster, the C6 cluster was selected here as wedid before.[25] All calculations were performed with a (75, 302)pruned grid. The ground-state geometries of all were optimized byDFT with the B3PW91 functional until the gradient forces werelower than a threshold of 0.00045 a.u. There were no imaginary forall complexes, ensuring that all of the systems for each structurewere stable and reasonable. Because of the existing electronega-tive atoms such as the doped N atoms, the 6-31G** basis set wasemployed in all the below calculations, supplementing polarizationfunctions to C, H, N and O atoms. The selected functional andbasis set were suitable for large systems and depicting the weaklyinteracting complexes, and they have also been used to study thedoped graphene by us before.[8, 23, 26]

In the case of fullerene (as this is the first time that we study thissystem), we decided to discuss first which functional and basis setwould be appropriate for the carbon atoms. As shown in Table 1,the B3PW91 functional and D95 V basis set gives the most reason-able configuration based on the well-known values of 1.46 and1.39 � for the C�C bonds of the pentagons and hexagons of fuller-ene. In the meantime, B3PW91 has been taken for other complexesand is better than B3LYP to depict weak interactions at the inter-face.Considering the computational cost and our large system, one Pdatom has been used to simulate the supported transition metal,which has also been selected to study the Pd-decorated carbonmaterials in previous investigations.[15, 17, 27, 28] The effect core pseu-dopotential with (8s7p5d)/[6s5p3d] was chosen to describe the Pdatom.[29–31] In our method, the calculated Pd–Pd bonding length forthe free Pd dimer was 0.50 �, which was close to the reported0.52 �.[17, 32] The PDOS and energy levels of occupied and unoccu-pied molecular orbitals of the models were analyzed with GAUSS-SUM software.[33]

3. Results and Discussion

3.1. N-doped Graphene Support

First, we will discuss the effect of a N-doped graphene sub-strate on hydrogen storage. As mentioned above, one Pdatom was placed in the center of a graphene sheet to elimi-nate the edge influence of the planar sheet. The supported Pdon graphene was taken as a new hydrogen-storage media, soits combined interaction at the interface should be discussed.Their optimized configurations are shown in Figure 1 with thedata listed in Table 2. The binding energy is calculated as: Eb =

Egraphene-Pd�Egraphene�EPd, where Egraphene-Pd, Egraphene, and EPd standfor the ground-state energies of the complex of Pd supportedon graphene, isolated graphene and a free Pd atom, respec-tively. There are three sites for the Pd adsorbed: the hollowsite with Pd at the center of a hexagon, the bridge site withPd at the midpoint of a carbon-carbon bond, and a top sitewith Pd directly above a carbon atom of graphene.

In our results, the Pd atom is located at the top site on theC atom, surrounded by the N atoms, for all of the dopedmodels, whereas it is located at the bridge site for pristine gra-phene. As the N-doping level increases, the catch force for Pdgets stronger with a larger combined energy at the interface.In the meantime, the distance between the graphene surfaceand the Pd atom becomes shorter, reflecting the stronger in-teraction. In addition, it could be seen that the graphene could

Table 1. Pentagon and hexagon C�C bond lengths for the free fullerenemodel, calculated with different functionals and basis sets.

Functional Basis set Lpentagon [�] Lhexagon [�]

B3LYP D95V 1.4638 1.4046B3PW91 D95V 1.4591 1.4011B3LYP 6-31G** 1.4536 1.3957B3PW91 6-31G** 1.4495 1.3934

Figure 1. Optimized structures for: a) pristine graphene–Pd (PGPd), b) N2-doped graphene–Pd (N2GPd), c) N4-doped graphene–Pd (N4GPd), andd) N6-doped graphene–Pd (N6GPd). The insets are the side-view configura-tions for each of the above complexes. The gray, blue, glaucous and whiteballs stand for C, N, Pd, and H atoms, respectively.

Table 2. Calculated binding energies and distances between grapheneand Pd according to Figure 1.

Eb [Hartree] G�Pd [�]

PGPd �0.0287 2.0809N2GPd �0.0323 2.0253N4GPd �0.0570 1.9984N6GPd �0.0779 1.9958

� 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 2014, 15, 344 – 350 345

CHEMPHYSCHEMARTICLES www.chemphyschem.org

maintain the planar configuration, except for the six-N-atom-doped case, which twists apparently in the entire surface, asshown in Figure 1 (d). This is unwanted in practical usage asthe heavy doping may break the electronic and chemical prop-erties, as well as the stability of the original graphene. So wefound that a four-N-atoms doping for our graphene modelcould give an excellent adsorption enhancement for the sup-ported Pd atom, with a binding energy of �0.057 Hartree anda separated distance of 1.9984 �. Therefore, we will focus ourattention mainly on the four-N-atom-doped case in the follow-ing investigations.

One hydrogen molecule was placed near the Pd atom tostudy its usage in hydrogen storage. The optimized structuresfor the pristine and N4-doped graphene with one H2 moleculeare displayed in Figure 2. PdH2 clusters all prefer to combinewith the planar supports at the top sites of the graphenesheets. This is consistent with the results reported by Corralet al. ,[15] where a top-site adsorp-tion of PdH2 on pristine gra-phene gives the largest bindingenergy.

From the side views (inset inFigure 1) it could be found thatafter PdH2 combination, the orig-inal graphene sheet still retainsthe same planar geometry asbefore, although the centralcarbon atom of N4GPdH2 hasbeen dragged out from the sur-face, displacing toward Pd. Thecombined energies here are cal-culated between the two partsof graphene substrates and Pd-adsorbed molecules, to depictand study the influence of N-doping on the supported Pd inpractical applications. The four-N-atom-doped graphene couldapparently enhance the catchforce for Pd, showing a strongbinding energy of �0.0562 Har-tree, which is nearly twice theone obtained for pristine gra-phene (�0.0277 Hartree). Theseparation distance becomesclearly shorter, changing from2.1734 to 2.0872 �, and reflect-ing a stronger interaction (seeTable 3). At the same time, theH�H bond gets shorter. Our cal-culated free hydrogen moleculebond is 0.7433 �, which is closeto the well-known 0.75 �value.[15] When one hydrogenmolecule is adsorbed on thetransition metals, the H�H bondelongates, giving a dissociated

state. In our system, both H2 molecule and graphene supportincline to draw Pd, and the draw force from the H2 moleculewill impair the combined interaction between Pd and thecarbon support. This competitive process has been depicted asthe an inset in Table 3. Compared with the pristine graphenesystem (PGPdH2), it is found that although the H�H bond be-comes 1.2 % shorter for N4GPdH2, it is still in the dissociatedstate and much longer than 0.7433 �. This result is pleasingand it will increase the value of graphene substrates in practi-cal applications. In addition, the Mulliken charge distributionsat the Pd atom have also been calculated and are displayed inTable 3. The local electron density increases after introducingdopants, and this is attributed to the increasing Fermi leveldue to the electron-rich graphitic N-doping method, whichcould facilitate electron escaping.

Based on the above observations, another hydrogen mole-cule has been introduced into the system to study the dop-

Figure 2. Optimized structures for: a) pristine graphene–PdH2 (PGPdH2), b) N4-doped graphene–PdH2 (N4GPdH2),c) pristine graphene–PdH4 (PGPdH4), d) N4-doped graphene–PdH4 (N4GPdH4), e) pristine graphene–PdOOH(PGPdOOH), and f) N4-doped graphene–PdOOH (N4GPdOOH). The inset figures are the side-view configurationsfor each of the corresponding complexes. The red balls stand for O atoms.

Table 3. Calculated binding energies, distances between graphene and Pd, bonding lengths of the hydrogenmolecule for the two-hydrogen-molecules adsorption systems, and Mulliken charge distributions at the Pdatom.

Eb [Hartree] G�Pd [�] H�H [�] Q [e]

PGPdH2 �0.0277 2.1734 0.8493 0.028N4GPdH2 �0.0562 2.0872 0.8389 �0.229

PGPdH4 �0.0017 2.4127 0.8328 �0.011

N4GPdH4 �0.0278 2.1750 0.8677 �0.138PGPdOOH -0.0300 2.1457N4GPdOOH �0.0735 2.0575

� 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 2014, 15, 344 – 350 346

CHEMPHYSCHEMARTICLES www.chemphyschem.org

ants’ influence more systematically. Unlike the single H2

system, here the adsorption method for pristine graphenewith two hydrogen molecules (H4) changes to the bridge site,while it remains at the top site for the N4-doped complex.Thus, the length of the G�Pd bond for the former should becalculated by the distance of Pd to the graphene surface inthe vertical direction. At the same time, the H�H bond is takenby the average length of the two H�H bonds of these ad-sorbed hydrogen molecules. The optimized structures are alsoshown in Figure 2 and the obtained configuration data are dis-played in Table 3. It is clearly found that the combined energychanges from �0.0017 to �0.0278 Hartree and the G�Pd dis-tance decreases for the doping system (by about 9.9 %) where-as the H�H bond becomes longer, indicating a more dissociat-ed state. This is not only positive to the supported Pd, due tothe stronger adsorption, but is also helpful to the spillovermechanism of hydrogen storage. We have also tried to intro-duce the third hydrogen molecule, however, it could notadsorb on the supported Pd metal.

In addition to the hydrogen-storage case, we have also stud-ied the applications of these results to the ORR. On the basisof our previous investigations and the work carried out byother groups,[6, 26] we conclude that oxygen adsorption andOOH cluster (because of the acid condition) adsorption playa decisive role in the ORR process. If an OOH molecule cannotbind with the catalyst, the following reactions will cease. As il-lustrated in Figures 2 (e) and 2 (f), OOH can bind with two sup-ported Pd atoms. Interestingly, the distance between Pd andthe graphene support (G�Pd bond) changes from 2.1457 to2.0575 � for the four-N-atom-doped complex. This is in accord-ance with the change in the calculated binding energy, show-ing a stronger interaction at the interface (of �0.0735 Hartree)and demonstrating that N-doping can also have a positive in-fluence on the ORR because of the stronger binding force forthe Pd catalyst coming from the doped graphene support, thisleads to its dispersion and stabilization in the reaction process.

3.2. N-doped Fullerene Support

Although graphene is one of the hottest materials-relatedtopics nowadays, fullerene is still an important carbon materialwith 60 carbon atoms in our model. Its p orbital is no longerbased on pure p electrons, but also contains a certain amountof s electrons. All optimized structures can be found inFigure 3, and their configuration data are listed in Table 4. Onaccount of the above discussions, we only consider the four-N-atom-doped case here for convenience. It could be seen thatfor all the complexes, the Pd atoms prefer to stand at thebridge sites, so the separated distance between fullerene andPd (F�Pd bond) should be calculated by the average length ofthe two shortest distances between the Pd atom and its twosupported C atoms.

N-atom doping on fullerene could also enhance the com-bined interaction for the supported Pd. As for the complexesof four-N-atom-doped fullerene with one (N4FPdH2) and two(N4FPdH4) hydrogen molecules, their combined energies bothare stronger than the corresponding pristine fullerene cases,

with the separation distances (F�Pd bonds) changing from2.1299 to 2.0828 � and from 2.1301 to 2.0862 �, respectively,confirming the stronger combination of the supported catalyt-ic metal with the N-doped fullerene (compared to the corre-sponding pristine fullerene case) in hydrogen storage. Al-though the H�H bonds display shortening during the doping,the change is small. Taking the H4 adsorbed model, for exam-ple, the change is just 0.7 % for the H�H bond, which is0.7995 �, obviously longer than 0.7433 � of the free hydrogenmolecule, still suggesting an elongated state. This also provesa positive effect of N-doping on the Pd metal supported byfullerene. Moreover, the ORR case has also been consideredhere. The calculated combined energy changes from �0.0533to �0.0844 Hartree, and the separation distance changes from2.1355 to 2.0177 � when N atoms are introduced, showinga stronger interaction caused by the N dopants. The variationsin the Mulliken charge density are similar to those observedfor the graphene case because the N atoms introduce moreelectrons to fullerene so that its work function decreases andthe number of electrons transferred from the Pd atom to thefullerene support is reduced.

Figure 3. Optimized structures of : a) pristine fullerene–PdH2 (PFPdH2), b) N4-doped fullerene–PdH2 (N4FPdH2), c) pristine fullerene–PdH4 (PFPdH4), d) N4-doped fullerene–PdH4 (N4FPdH4), e) pristine fullerene–PdOOH (PFPdOOH),and f) N4-doped fullerene–PdOOH (N4FPdOOH).

Table 4. Calculated combined energies, distances between fullerene andPd, bonding lengths of the adsorbed hydrogen molecules, and Mullikencharge distributions at the Pd atom.

Eb [Hartree] F�Pd [�] H�H [�] Q [e]

PFPdH2 -0.0608 2.1299 0.8143 0.360N4FPdH2 -0.0826 2.0828 0.8003 0.240PFPdH4 -0.0503 2.1301 0.8049 0.340N4FPdH4 -0.0726 2.0862 0.7995 0.223PFPdOOH -0.0533 2.1355N4FPdOOH -0.0844 2.0177

� 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 2014, 15, 344 – 350 347

CHEMPHYSCHEMARTICLES www.chemphyschem.org

3.3. N-doped C6 Cluster Support

Apart from graphene and fullerene, carbon materials at a small-er scale, such as carbon clusters and carbon dots, have also at-tracted recent attention. Now, we will discuss the N-doped C6cluster, based on our previous investigation,[25] to further provethe wide suitability of N-doped carbon supports. As mentionedabove, both hydrogen storage and ORR have been simulated,and the optimized configurations are shown in Figure 4.

The separation distance between Pd and C6 is calculated bythe average length of the shortest two Pd�C bonds, and thedistances are marked directly in Figure 4.

It is obvious that as an N atom is introduced into the smallcarbon cluster, it catches the supported Pd atom more tightly,both in the hydrogen-storage and ORR processes. This is in ac-cordance with the graphene and fullerene cases. In addition,the binding-energy calculations can confirm this again; as an Natom is introduced, it changes from �0.0927 to �0.0931 Har-tree and from �0.0753 to �0.077 Hartree for the one- andtwo-hydrogen-molecules adsorbed models, respectively, andfrom �0.0912 to �0.1154 Hartree for the OOH adsorption case.Therefore, based on the above discussions, it is concluded thatN-doping can really enhance the interaction between Pd andthe carbon support material. This will prevent the leaching,flopping, and aggregation of the supported transition metalsduring practical applications, increasing their efficiency. Then,we tried to investigate the enhancement mechanism behindthese observations

3.4. PDOS and Electronic Cloud Surfaces

Although it has been accepted that N-doping may introducelocally high electron and spin densities (which benefits the ad-sorption of some small molecules), this is not suitable to ex-plain the supported enhancement phenomenon.

The PDOS of six representative complexes were studied andare displayed in Figure 5. Two results could be obtained: 1) thedegree of overlapping between the carbon support orbitaland the transition metal orbital becomes larger for all of the N-doped models compared to their corresponding pristine com-plexes; and 2) at the same time, the energy gap between the

Figure 4. Optimized structures of : a) C6PdH2, b) N1C5PdH2, c) C6PdH4,d) N1C5PdH4, e) C6PdOOH, and f) N1C5PdOOH. The separation distances aremarked in each figure with � units.

Figure 5. Calculated PDOS for the six complexes: a) pristine graphene with Pd, b) N4-doped graphene with Pd, c) pristine fullerene with Pd, d) N4-doped full-erene with Pd, e) C6 cluster with Pd, and f) N1-doped C6 cluster with Pd. The left and right vertical lines stand for HOMO and LUMO, respectively.

� 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 2014, 15, 344 – 350 348

CHEMPHYSCHEMARTICLES www.chemphyschem.org

highest occupied molecule orbital (HOMO) and the lowest un-occupied molecule orbital (LUMO) decreases. This behavior canbe attributed to changes in the electronic states of the carbonmaterials, owing to the doped N atoms. Thus, we concludethat the doping N atoms can increase the hybridized orbitalsof Pd and various carbon materials, enhancing their combinedinteractions.

We have shown the electronic distributional surfaces for gra-phene and fullerene complexes (see Figure 6). At the isoval-ue = 0.08 a.u. , the electronic distributional surface is broken atthe interface between pristine graphene and Pd, whereas forthe doped model, the electronic cloud is still continuous. Thesame situation also exists in the fullerene case. At the isoval-ue = 0.1 a.u. , the electron cloud is disconnected at the inter-face of pristine fullerene and Pd; however, the electronic distri-butional surface of the N-doped fullerene and Pd is linked to-gether all the same. This demonstrates that the combined in-teractions of Pd induced by N-doped carbon materials arestronger than those due to pristine carbon materials without Ndecoration. This is in accordance with the above results.

The active site of a carbon atom for combination is adjacentto the doping N atoms. Thus, it will be positively chargedowing to the larger electronegativity of these doped N atoms.As mentioned above, electron-rich graphitic N-doping can in-crease the Fermi level, facilitating electron escaping, so thatthe supported Pd atom is more negatively charged after N-doping. In this case, the electrostatic attraction between Pdand the N-doped carbon substrates increases, which can beused to explain the N-doping enhancement.

4. Conclusion

In summary, we have systematically studied N-doped carbonmaterials (including graphene, fullerene, and carbon clusters)and investigated the doping influence on supported Pd spe-cies for applications in hydrogen storage and ORR. We con-clude that the doped N atoms can change the electronicstates of the carbon materials, leading to stronger attractive in-teractions with the supported Pd. The hybridized PDOS of

these carbon supports and Pd become more intensive whenthey are doped with N atoms, owing to the enhanced electro-static attraction at the interface. The electronic-density overlap-ping at the interface is also stronger for the N complexes. Thiscould prevent the leaching and aggregation of supported tran-sition metals on N-doped carbon materials, making them suita-ble for a wide range of applications.

Acknowledgements

This work is supported by the National Natural Science Founda-tion (NSFC, 21071137 and U1232211). The calculations have beencompleted on the Supercomputing system in the SupercomputingCenter of USTC.

Keywords: carbon · fullerene · graphene · hydrogen storage ·oxygen reduction reaction

[1] Z. Yufeng, K. Yong-Hyun, A. C. Dillon, M. J. Heben, S. B. Zhang, Phys. Rev.Lett. 2005, 94, 155504.

[2] Y. Mina, Y. Shenyuan, W. Enge, Z. Zhenyu, Nano Lett. 2007, 7, 2578 –2583.

[3] S. Dag, Y. Ozturk, S. Ciraci, T. Yildirim, Phys. Rev. B 2005, 72, 155404.[4] Q. Sun, Q. Wang, P. Jena, Y. Kawazoe, J. Am. Chem. Soc. 2005, 127,

14582 – 14583.[5] Y. Gao, D. Ma, C. Wang, J. Guan, X. Bao, Chem. Commun. 2011, 47,

2432 – 2434.[6] L. Zhang, Z. Xia, J. Phys. Chem. C 2011, 115, 11170 – 11176.[7] Y. Sun, C. Li, G. Shi, J. Mater. Chem. 2012, 22, 12810 – 12816.[8] X. Kong, Z. Sun, Q. Chen, Phys. Chem. Chem. Phys. 2012, 14, 13564 –

13568.[9] K. Srinivasu, S. K. Ghosh, J. Phys. Chem. C 2012, 116, 5951 – 5956.

[10] X. Kong, Q. Chen, Phys. Chem. Chem. Phys. 2013, 15, 12982 – 12987.[11] P. K. Schmidt, K. Christmann, G. Kresse, J. Hafner, M. Lischka, A. Gross,

Phys. Rev. Lett. 2001, 87, 096103.[12] I. Lopez-Corral, E. German, A. Juan, M. A. Volpe, G. P. Brizuela, Int. J. Hy-

drogen Energy 2012, 37, 6653 – 6665.[13] W. Qian, R. Hao, J. Zhou, M. Eastman, B. A. Manhat, Q. Sun et al. , Carbon

2013, 52, 595 – 604.[14] X. Xu, Y. Li, Y. Gong, P. Zhang, H. Li, Y. Wang, J. Am. Chem. Soc. 2012,

134, 16987 – 16990.[15] I. Lopez-Corral, E. German, A. Juan, M. A. Volpe, G. P. Brizuela, J. Phys.

Chem. C 2011, 115, 4315 – 4323.[16] I. Lopez-Corral, J. de Celis, A. Juan, B. Irigoyen, Int. J. Hydrogen Energy

2012, 37, 10156 – 10164.[17] I. Cabria, M. J. Lopez, J. A. Alonso, Phys. Rev. B 2010, 81, 035403.[18] A. Sigal, M. I. Rojas, E. P. M. Leiva, Phys. Rev. Lett. 2011, 107, 158701.[19] B. Yu, X. Kong, L. Wei, Q. Chen, J. Mater. Cycles Waste Manage. 2011, 13,

298 – 304.[20] J. C. Carrero-Sanchez, A. L. Elias, R. Mancilla, G. Arrellin, H. Terrones, J. P.

Laclette, M. Terrones, Nano Lett. 2006, 6, 1609 – 1616.[21] Gaussian 09 (Revision B.01), M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.

Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Men-nucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian,A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara,K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr. , J. E. Peralta, F. Ogliaro,M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, T. Keith,R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S.Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox,J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Strat-mann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L.Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dan-nenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz,J. Cioslowski, D. J. Fox, Gaussian, Inc. , Wallingford CT, 2010.

Figure 6. Simulated electronic cloud surfaces for: a) PGPdH2, b) N4GPdH2,c) PFPdH2, and d) N4FPdH2. The isovalues for (a) and (b) are 0.08 au and for(c) and (d) 0.1 au.

� 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 2014, 15, 344 – 350 349

CHEMPHYSCHEMARTICLES www.chemphyschem.org

[22] R. A. Sidik, A. B. Anderson, N. P. Subramanian, S. P. Kumaraguru, B. N.Popov, J. Phys. Chem. B 2006, 110, 1787 – 1793.

[23] X. Kong, Q. Chen, J. Mater. Chem. 2012, 22, 15336 – 15341.[24] X. Kong, Z. Sun, M. Chen, C. Chen, Q. Chen, Energy Environ. Sci. 2013, 6,

3260 – 3266.[25] X. Kong, Q. Chen, R. Li, K. Cheng, N. Yan, J. Chen, Y. Zhou, ChemPhys-

Chem 2012, 13, 1449 – 1453.

[26] X. Kong, Q. Chen, Z. Sun, ChemPhysChem 2013, 14, 514 – 519.[27] R. Thapa, D. Sen, M. K. Mitra, K. K. Chattopadhyay, Phys. B 2011, 406,

368 – 373.[28] H. Xiao, S. H. Li, J. X. Cao, Chem. Phys. Lett. 2009, 483, 111 – 114.

[29] L. J. Goossen, D. Koley, H. Hermann, W. Thiel, Chem. Commun. 2004,2141 – 2143.

[30] P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 299 – 310.[31] D. Andrae, U. Haussermann, M. Dolg, H. Stoll, H. Preuss, Theor. Chim.

Acta 1990, 77, 123 – 141.[32] I. Efremenko, E. D. German, M. Sheintuch, J. Phys. Chem. A 2000, 104,

8089 – 8096.[33] W. S. Hummers, R. E. Offeman, J. Am. Chem. Soc. 1958, 80, 1339 – 1339.

Received: October 4, 2013Published online on January 16, 2014

� 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 2014, 15, 344 – 350 350

CHEMPHYSCHEMARTICLES www.chemphyschem.org