Computational Materials Science 49 (2010) S15–S20 Contents lists available at ScienceDirect Computational Materials Science journal homepage: www.elsevier.com/locate/commatsci First principles study of the physisorption of hydrogen molecule on graphene and carbon nanotube surfaces adhered by Pt atom Pham Tien Lam a, Phan Viet Dung a, Ayumu Sugiyama a, Nguyen Dinh Duc b, Tatsuya Shimoda a,c, Akihiko Fujiwara a, Dam Hieu Chi a,b,c,* a b c Japan Advanced Institute of Science and Technology, 1-1 Asahidai, Nomi, Ishikawa 923-1292, Japan Vietnam National University, 144 Xuan Thuy, Cau Giay, Hanoi, Vietnam ERATO Shimoda Nano-Liquid Process Project, Japan Science and Technology Agency, 2-5-3-Asahidai, Nomi, Ishikawa 923-1211, Japan a r t i c l e i n f o Article history: Received August 2009 Received in revised form 12 February 2010 Accepted 12 February 2010 Available online 27 April 2010 Keywords: DFT Fuel cell Carbon nanotube Catalysts a b s t r a c t Adsorptions of hydrogen, oxygen and carbon monoxide molecules on surfaces of single wall carbon nanotubes (SWNTs) and graphene adhered by a Pt atom have been investigated by density functional theory calculation (DFT) Our calculations show that the Pt adatom significantly promotes the physisorption of hydrogen in a region around it with radius of about Å The physisorption configuration in which oxygen molecule aligned parallelly to the surfaces of SWNTs and graphene are most preferred In contrast, both of the physisorption configurations in which CO molecule aligned parallelly and perpendicularly with the carbon end towards the graphene and SWNTs surfaces were preferred The obtained results suggested that the modification of the electronic structure by adhesion of Pt atom on surfaces of the support materials can modify their physisorption properties of gas molecules Ó 2010 Elsevier B.V All rights reserved Introduction Poly electrolyte membrane fuel cells (PEMFCs) have been considered to be the most promising, among different types of fuel cell, because they operate at low temperature and give high specific power and power density [7,2,1] In PEMFCs platinum catalysts, as an active component, is the most important component for electro-catalysts [7] This brings up a major barrier to commercial application of fuel cells that suffer from high cost and low stability Carbon materials are considered as the best support materials for electro-catalysts in fuel cells because of its conductivity, surface area, corrosion resistance and low cost [2,1] Among various types of carbon materials, carbon nanotubes with high surface area, good electronic conductivity, and high chemical stability, have been found to be an ideal support material for Pt clusters [4] Highly dispersed and size-controlled small Pt clusters (less than nm) made from dispersed single Pt atoms were achieved by using carbon nanotube supports [16] The motion of Pt clusters on CNT was also * Corresponding author at: Japan Advanced Institute of Science and Technology, 1-1 Asahidai, Nomi, Ishikawa 923-1292, Japan Tel.: +81 76 151 1584; fax: +81 76 151 1535 E-mail address: dam@jaist.ac.jp (D.H Chi) 0927-0256/$ - see front matter Ó 2010 Elsevier B.V All rights reserved doi:10.1016/j.commatsci.2010.02.041 observed experimentally by high-resolution transmission electron microscopy Further, the superior electro-catalytic activity and the high tolerance to carbon monoxide poisoning of nanoparticle supported on carbon nanotube have been confirmed by several studies [20,22,12] In addition, our previous theoretical studies reveal several novel properties of Pt clusters on SWNTs, including the substrate mediated interaction and the structural fluxionality [8,9] Fundamental information regarding properties of Pt nano clusters on SWNTs under gas environment is strongly required for designing catalyst for fuel cell In this paper, we report our first principle study on the absorptions of gas molecules on surfaces of SWNTs and graphene adhered by a Pt atom Physisorptions of hydrogen, oxygen and carbon monoxide molecules on the systems have been investigated by density functional theory calculation (DFT) Our calculations show that the Pt adatom significantly promotes the adsorption of hydrogen in a region around it with radius of about Å The adsorption configuration in which the oxygen molecule is aligned parallelly to the surfaces of SWNTs and graphene are significantly more stable than the others In contrast, both of the adsorption configurations in which the CO molecule was aligned parallelly and perpendicularly with the carbon end towards the graphene and SWNTs surfaces were preferred The obtained results suggested that the modification of the electronic structure by adhesion of Pt nano clusters on surfaces of SWNTs and graphene can modify their physisorption properties of gas molecules S16 T.L Pham et al / Computational Materials Science 49 (2010) S15–S20 tionals, were very good agreement with the result of second order Møller–Plesset perturbation (MP2) calculation [19] Calculations of the interaction between hydrogen molecule and a small graphene plate (C24H12) using Hatree–Fock method and DFT method with several functionals were carried out in comparison with that using MP2/6-311++GÃÃ method for choosing the most appropriate functional (Fig 1a) All DFT [15,17] calculations were carried out with triple numerical plus polarization basis set (TNP) by Dmol3 package [11,10], and molecular orbital methods were carried out by Gaussian 03 package [3] Our calculations show that HF method almost predicts a repulsive interaction between hydrogen and garphene, while MP2 method predicts attraction interaction (Fig 1b) This result indicates correlation energy correction plays an important role in simulating Van der Waals force, since MP2 method takes into account the correlation energy correction as the perturbation On the other hand, results obtained from calculations using DFT methods strongly depend on employed functionals (Fig 1c) While LDA functionals are good agreement with MP2 result, GGA functionals Methodology 2.1 Evaluation of Van der Waals interaction DFT methods are known as low computational cost methods, and they are the most popular methods for calculating electronic structure of many-atom systems The interaction between these gas molecules and SWNT, graphene surfaces is mainly driven by Van der Waals force Unfortunately, conventional DFT methods not describe well Van der Waals interaction, dispersion interaction [14,19] It has been confirmed that local density approximation (LDA) functionals seem to well describe the physisorption of H2 on graphene and carbon nanotube surfaces, compared to experimental data [19,6] Okamoto and Miyamoto [19] have confirmed that local density approximation (LDA) functionals predict physisorption of the hydrogen molecule on graphene plate, while some generalized gradient approximation (GGA) and hybrid-DFT functionals lead to repulsion interaction Further more they also indicated that the potential energy surfaces, given by LDA func- (a) Molecular orbital method / 6-311++G** Density functional theory 100 HF MP2 50 Potential energy / meV Potential energy / meV 100 -50 -100 LDA-VWN GGA-PW91 GGA-BLYP 50 -50 -100 -150 -150 2.5 3.5 4.5 2.5 3.5 4.5 Separation / Angstrom Separation / Angstrom (b) (c) (A) (A) (B) (C) (B) (D) (C) (D) (d) Fig (a) Interaction model between graphene and hydrogen molecule (b) Potential energy surface derived by molecular orbital methods (HF and MP2/6-311++GÃÃ) (c) Potential energy surface derived by DFT methods: potential energy surfaces were estimated by taking potential energy at distance of 10 Å as zero (d) Adsorption sites of H2 on graphene and (10, 0)SWNT surfaces adhered by a Pt atom S17 T.L Pham et al / Computational Materials Science 49 (2010) S15–S20 Table Interaction energy (Ead) and equilibrium distance (De) between H2 and graphene surface Functionals MP2 De (Å) Ead (meV) 2.89 93.1 LDA GGA VWN PWC PW92 PBE RPE HCTC BOP 2.66 86.9 2.65 87.0 3.40 26.3 3.42 14.2 3.80 12.0 3.52 38.0 Repulsion tend to underestimate interaction energy, and some GGA functionals even predict repulsion between hydrogen molecule and graphene (Table 1) Therefore, in this research we mainly used LDA functionals to evaluate the interaction between gas molecules and graphene and SWNT surfaces 2.2 Calculation models In this study, we used a periodic supper cell to simulate the graphene sheet The super cell included 128 carbon atoms with edge length of a and b of 17.07 and 19.71 Å, respectively, correspond to the C–C bond length of 1.42 Å These lattice parameters were considered to be large enough to neglect the interaction of Pt, gas molecules with their periodic image The c lattice was of 30 Å, that was large enough to neglect the interaction between graphene sheets We also applied a periodic super cells to simulate SWNTs The c lattices (which were aligned along the axes of the SWNTs) of these super cells are of 17.04 Å and 19.68 Å for (10, 0)SWNT and (5, 5)SWNT, respectively These values were chosen to correspond to the C–C bond length of 1.42 Å and to match the periodic condition The edge lengths of both a and b lattices of these super cells were of 25 Å which were large enough to ensure that there are no interactions Pt, gas molecules and their periodic images The super cells consisted of 160 carbon atom for both (10, 0)SWNT and (5, 5)SWNT We used local density approximation Vosoko–Wilk–Nusair (VWN) [21] functional to treat exchange–correlation energy Double numerical plus polarization function (DNP) basis set was used for all calculation Brillouin-zone integrations were performed by using (1 Â Â 4) k-point mesh, and (4 Â Â 1) k-point for SWNTs and graphene respectively with Monkhorsh-Pack scheme [18] All calculations were performed by DFT using Dmol3 package Results and discussion 3.1 Physisorption of hydrogen Our calculations as well as several previous publications [14,6,5] indicated that H2 prefers to be physisorbed at the center of hexagon and aligned parallelly to the surface of graphene and (10, 0)SWNT In this research, we used parallel adsorption configuration of H2 on the surface at the center of hexagon The adsorption energies of H2 on graphene and (10, 0)SWNT surfaces were calculated by formula: Ead ẳ EH2 ỵ EGraphene or SWNT EH2 =Graphene or SWNT VWN-PB BP BLYP optimization calculation confirmed that no stable physisorption configuration of hydrogen was found in a region with radius of about Å around Pt adatom When H2 move into this region, it is pulled toward the Pt adatom We also considered several adsorption sites of hydrogen around this area (Fig 1d) Our calculations clearly show that at these adsorption sites, H2 prefers to be adsorbed parallel to the surfaces at the center of hexagon Adsorption energies and equilibrium distance from center of mass of H2 to graphene and SWNT surfaces are summarized in Table It means that outside of this region the effect of the Pt adatom is not clear 3.2 Physisorption of oxygen Oxygen reduction plays an essential role in the performance of fuel cells, as oxygen reduction reaction is four-electrons-transfer reaction in which the first step is the adsorption of oxygen molecule on catalysts In this section we mainly focus on the adsorption of oxygen on carbon support materials, including graphene, (10, 0)SWNT, and (5, 5)SWNT, as well as the effect of the Pt adatom on it We considered several adsorption configurations of oxygen on graphene, (10, 0)SWNT and (5, 5)SWNT surfaces, including top, center, and parallel (Fig 2a) The interaction energies between O2 and the surfaces were estimated by the depth of the potential wells on potential energy surfaces The potential energy surfaces were calculated by changing the distance between O2 and the surface and calculating total energy at each point The depth of potential wells was estimated by taking the potential energy as zero when O2 was placed at the center of super cells Fig 2b shows potential energy surfaces of the singlet and triplet states of O2 molecule when it approaches to graphene surface It is apparent that in a physisorption on a graphene surface, the distance between O2 and the surface is in a range from 2.5 to 3.0 Å, the triplet sate is more stable than the singlet state Similar results were obtained for the physisorptions of O2 on (10, 0) and (5, 5)SWNT surfaces Therefore, in this research we used triplet potential energy surfaces to evaluate interaction between O2 and the surfaces Typical triplet potential energy surfaces of O2 on the surfaces are showed in Fig 2c It is apparent that the adsorption energies of O2 on the surfaces strongly depend on adsorption configurations, and the configuration in which oxygen molecule aligned parallelly to the surfaces of SWNTs and graphene are most preferred (Table 3) This result can be explained by the interaction between p electrons of O2 and p electrons of the surfaces The distance from O2 molecule to the surface seems not to depend significantly on ð1Þ The adsorption energy of hydrogen on graphene, 112.29 meV, is close to the value found by Arelano et al., 86 meV with planar periodic graphene layer [5] and by Henwood et al., 93.10 meV with the hexagonal plate consisting of 96 carbon atoms [6] The adsorption energy of H2 on (10, 0)SWNT, 107.36 meV, is close to the value found by Henwood and David Carey [14], 89.98 meV, and larger than the result found by Han and Lee [13], 34 meV, by GGAPW91 functional For the physisorption of hydrogen on SWNTs and graphene adhered by a Pt atom, our careful examination by full geometry Table Adsorption energy (Ead) of hydrogen on graphene and distance between hydrogen and graphene surface (De) Adsorption sites Pristine A B C D Ead (meV) De (Å) Graphene (10, 0)SWNT Graphene (10, 0)SWNT 112.29 112.54 113.31 112.74 112.69 107.36 108.36 112.16 114.38 106.19 2.66 2.66 2.67 2.67 2.67 2.55 2.55 2.54 2.57 2.56 S18 T.L Pham et al / Computational Materials Science 49 (2010) S15–S20 Total energy + 4987 / Ha 0.005 Top Center Parallel Parallel Singlet Parallel Triplet -0.005 -0.01 -0.015 -0.02 -0.025 -0.03 2.2 2.4 2.6 2.8 3.2 3.4 Distance / Angstrom Potential energy surface / meV (a) (b) Parallel Center -50 -100 -150 -200 2.2 2.4 2.6 2.8 3.2 3.4 Distance / Angstrom (c) (d) Fig (a) Adsorption configuration of O2 on graphene, (10, 0)SWNT surface, and (5, 5)SWNT surfaces (b) Singlet and triplet potential energy surface (c) Triplet potential energy surface with different adsorption configurations of O2 on graphene surface (d) Adsorption site of O2 and CO on graphene surface Table Adsorption energy (Ead) of O2 on graphene and SWNT surfaces and distance between the center of mass of O2 and the surfaces (De) Adsorption configuration Center Top Parallel Ead (meV) De (Å) Graphene (10, 0)SWNT (5, 5)SWNT Graphene (10, 0)SWNT (5, 5)SWNT 132.55 114.14 177.93 125.69 87.82 158.26 124.18 90.11 155.81 3.35 3.48 2.88 3.24 3.51 2.80 3.29 3.47 2.90 adsorption configurations and curvature of the surfaces Our obtained result also indicates that the adsorption energies of oxygen on SWNTs are slightly lower than that on graphene To investigate the effect of the Pt adatom, we calculated the adsorption energies of O2 on the surfaces adhered by a Pt atom Our calculations showed that on the surfaces adhered by a Pt atom, triplet still is the most stable state The physisorption energies of O2 at triplet state on the surfaces of graphene and SWNTs adhered by a Pt atom with parallel configuration are summarized in Table It is apparent that adhesion of Pt promotes the interaction between oxygen and the surfaces 3.3 Physisorption of carbon monoxide In direct methanol fuel cell, CO is one of the most important intermediate substance, therefore information about the adsorption of CO on catalysts is important for understanding properties of catalysts In this section, we focus on the absorptions of CO on graphene, (10, 0)SWNT, and (5, 5)SWNT surfaces as well as the effect of the Pt atom on these adsorptions Table Adsorption energy (Ead) of O2 on graphene and SWNT surfaces Adsorption configuration Center Top Parallel Graphene (10, 0)SWNT (5, 5)SWNT Without Pt With Pt Without Pt With Pt Without Pt With Pt 132.55 114.14 177.93 132.71 117.56 179.26 125.69 87.82 158.26 129.89 89.18 159.06 124.18 90.11 155.81 121.81 91.90 160.63 We considered several adsorption configurations of CO on graphene, (10, 0)SWNT, and (5, 5)SWNT surfaces (Fig 3a) We also evaluated the interaction between CO and the surfaces by the depth of potential wells on potential energy surfaces The potential energy surfaces were calculated by changing the distance between CO and the surfaces, and calculating energy at each point (Fig 3b) These potential energy surfaces indicate that the adsorptions of CO molecules on graphene and SWNT surfaces strongly depend on the adsorption configurations We found that CO molecules prefer to S19 T.L Pham et al / Computational Materials Science 49 (2010) S15–S20 Parallel (a) Potential energy / meV Carbon center Oxy center Parallel -25 -50 -75 -100 -125 -150 2.4 2.6 2.8 3.2 3.4 Distance /Angstrom (b) Fig (a) Adsorption configurations of CO on graphene and SWNTs (b) Potential energy surface Table Adsorption energy (Ead) of CO on graphene and SWNT surfaces and distance between the center of mass of CO and the surfaces (De) Adsorption configuration Ead (meV) Carbon top Carbon center Oxygen top Oxygen center Parallel De (Å) Graphene (10, 0)SWNT (5, 5)SWNT Graphene (10, 0)SWNT (5, 5)SWNT 106.84 143.49 87.60 108.19 148.52 89.91 137.92 68.49 101.04 131.18 80.88 136.12 68.16 98.70 127.92 3.76 3.52 3.57 3.42 2.96 3.82 3.41 3.60 3.32 2.97 3.79 3.48 3.58 3.39 3.02 adsorb parallelly, or perpendicularly to the surfaces with the carbon end toward the surfaces (Table 5) The difference in adsorption energy between configurations can be attributed to the polarization of CO molecule For configuration, in which oxygen atom orients toward the surfaces, the negative charge of oxygen atom weakens the interaction between CO and graphene or SWNT surfaces, while the positive charge of carbon atom promotes the interaction between CO and the surfaces For graphene surface, the adsorption energies of two most preferring adsorption configurations, carbon center and parallel, are almost the same For (10, 0)SWNT and (5, 5)SWNT surfaces carbon center configuration seems to be more stable than parallel configuration Table Adsorption energy (Ead) of CO on graphene and SWNT surfaces Adsorption configuration Carbon center Oxygen center Parallel Graphene (10, 0)SWNT (5, 5)SWNT Without Pt With Pt Without Pt With Pt Without Pt With Pt 143.49 108.19 148.52 142.92 108.27 145.78 137.92 101.04 131.18 137.47 100.82 129.58 136.12 98.70 127.92 130.41 96.14 125.30 The distance between CO molecule and the surfaces does not seem to significantly depend on the adsorption configuration To investigate the effect of the Pt adatom, we also evaluated the adsorption of CO on graphene and SWNT surface adhered by a Pt atom with carbon center, oxygen center, and parallel configuration at the adsorption site as in Fig 2d The adsorption energies are summarized in Table In contrast with the case of O2, adhesion of Pt atom does not show a clear influence to the interaction between CO and the surfaces Conclusions Our calculations indicated that the adsorption of the Pt atom on graphene and (10, 0)SWNTs leads to the formation of an active region with radius of about Å for the adsorption of hydrogen atom In this region the Pt adatom significantly promotes the adsorption of hydrogen by creating a deep and wide potential well on the potential energy surface for hydrogen molecules For the adsorption of oxygen, we found that oxygen molecule not change it its spin state, the most stable state is triplet state Oxygen molecules prefer to be adsorbed parallel at the center of hexagons on graphene and SWNT surfaces For the adsorption of CO, we found that CO molecules prefer to be aligned parallelly to the graphene and SWNT S20 T.L Pham et al / Computational Materials Science 49 (2010) S15–S20 surfaces, or perpendicularly to the surfaces with the carbon end towards the surfaces References [1] Jiujan Zhang, PEM Fuel Cell Electro-catalysis and Catalysts Layers, Springer, London, 2003 [2] Andrej Wieckowski, Elena R Savinova, Cnstaninos G Vayenas, Catalysis and Electrocatalysis at Nanoparticle Surfaces, MarcelDekker, Inc., New York, 2003 [3] M.J Frisch, G.W Trucks, H.B Schlegel, G.E Scuseria, M.A Robb, J.R Cheeseman, J.A Montgomery Jr., T Vreven, K.N Kudin, J.C Burant, J.M Millam, S.S Iyengar, J Tomasi, V Barone, B Mennucci, M Cossi, G Scalmani, N Rega, G.A Petersson, H Nakatsuji, M Hada, M Ehara, K Toyota, R Fukuda, J Hasegawa, M Ishida, T Nakajima, Y Honda, O Kitao, H Nakai, M Klene, X Li, J.E Knox, H.P Hratchian, J.B Cross, C Adamo, J Jaramillo, R Gomperts, R.E Stratmann, O Yazyev, A.J Austin, R Cammi, C Pomelli, J.W Ochterski, P.Y Ayala, K Morokuma, G.A Voth, P Salvador, J.J Dannenberg, V.G Zakrzewski, S Dapprich, A.D Daniels, M.C Strain, O Farkas, D.K Malick, A.D Rabuck, K Raghavachari, J.B Foresman, J.V Ortiz, Q Cui, A.G Baboul, S Cliord, J Cioslowski, B.B Stefanov, G Liu, A Liashenko, P Piskorz, I Komaromi, R.L Martin, D.J Fox, T Keith, M.A AlLaham, C.Y Peng, A Nanayakkara, M Challacombe, P.M.W Gill, B Johnson, [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] W Chen, M.W Wong, C Gonzalez, J.A Pople, Gaussian 03, Revision B.01, Gaussian, Inc., Pittsburgh PA, 2003 A.L Dicks, Power Sour 156 (2006) 128 J.S Arellano et al., Chem Phys 112 (2000) 8114–8119 J.S Arellano, L.M Molina, A Rubio, J.A Alonso, Chem Phys 117 (2002) 2281–2288 S.S.P Costamagna, S Srinivasan, Power Sour 202 (2001) 242–252 Nguyen Thanh Cuong et al., Phys Rev B 79 (2009) 235417 Hieu Chi Dam et al., Phys Rev B 79 (2009) 115426 B Delley, Chem Phys 113 (2000) 7756 B Delley, Chem Phys 92 (1990) 508 M Gangeri, S Perathoner, G Centi, Inorg Chim Acta 359 (2006) 4828–4832 Sang Soo Han, Hyuck Mo Lee, Carbon 42 (2004) 2169–2177 D Henwood, J David Carey, Phys Rev B 75 (2007) 245413 P Hohenberg, W Kohn, Phys Rev 136 (1964) 5188 Y.T Kim et al., Chem Int Ed 45 (2006) 407 W Kohn, L.J Sham, Phys Rev 140 (1965) A1133 H.J Monkhorst, J Pack, Phys Rev B 113 (1976) 5188 Y Okamoto, Y Miyanomto, Phys Chem B 105 (2001) 3470–3474 T Onoe, S Iwamoto, M Inoue, Catal Commun (2007) 701–706 S.J Vosko, L Wilk, M Nusair, Can J Phys 58 (1980) 1200 G Wu, Y.S Chen, B.Q Xu, Electrochem Commun (2005) 1237–1243 ... effect of the Pt adatom, we calculated the adsorption energies of O2 on the surfaces adhered by a Pt atom Our calculations showed that on the surfaces adhered by a Pt atom, triplet still is the. .. Pt atom on graphene and (10, 0)SWNTs leads to the formation of an active region with radius of about Å for the adsorption of hydrogen atom In this region the Pt adatom significantly promotes the. .. 5)SWNT surfaces as well as the effect of the Pt atom on these adsorptions Table Adsorption energy (Ead) of O2 on graphene and SWNT surfaces Adsorption configuration Center Top Parallel Graphene