- Email: [email protected]
Adsorption of carbon monoxide, carbon dioxide and methane on hexagonal boron nitride with high titanium coverage
Surface Science 637–638 (2015) 48–52
Contents lists available at ScienceDirect
Surface Science journal homepage: www.elsevier.com/locate/susc
Adsorption of carbon monoxide, carbon dioxide and methane on hexagonal boron nitride with high titanium coverage I. Carrillo a, J.M. Ramírez b, L.F. Magaña b,⁎ a b
Instituto Tecnológico de Tláhuac, Av. Estanislao Ramírez # 301, Colonia Ampliación Selene, C.P. 13420 México, D.F., Mexico Instituto de Física, Universidad Nacional Autónoma de México, Apartado Postal 20-364, C.P. 01000 México, D.F., Mexico
a r t i c l e
i n f o
Article history: Received 30 September 2014 Accepted 8 March 2015 Available online 14 March 2015 Keywords: Density functional calculations Hexagonal boron nitride Carbon monoxide Carbon dioxide Methane
a b s t r a c t We used density functional theory and molecular dynamics to explore the adsorption of CO, CO2, and CH4 on a layer of h-BN with high titanium coverage. After optimization, we found that each titanium atom is located above each of the hexagons of the surface. We considered atmospheric pressure and 300 K. It is found that the first molecule is adsorbed and dissociated on the surface. The CO2 molecule is broken into O and CO. The methane molecule is physisorbed, and not dissociated. It is desorbed at 370 K. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Carbon monoxide, carbon dioxide and methane are environmental problems. CO is a toxic gas which, at normal conditions is tasteless, odorless, and colorless. It comes primarily from incomplete combustion of carbon containing fuels. It is not-irritating, and it is an asphyxiant. At high levels, it is deadly. The investigation of the adsorption of carbon monoxide on palladium and transition metals is an attractive research topic. There are theoretical and experimental studies on the adsorption of carbon monoxide on the Ti [001] surface [1–4]. On the other hand, many industrial processes around the world produce CO2 and CH4. These gases are two of the main greenhouse gases emitted today [5]. At the present, the removal of these gases from the atmosphere is mainly done using amine baths. This process presents several problems. Among them, we have additional processing and corrosion control [6]. Another procedure is by chemically capturing CO2 using oxides [7]. However, to remove these greenhouse gases, adsorption processes have been considered too. Among the materials that have been studied for these processes, we have activated carbons, zeolites, and porous metal–organic frameworks [8,9]. In previous works, we studied the adsorption of these three gases on a titanium–graphene system with high metal coverage [10,11]. We found that the carbon monoxide molecule is adsorbed and dissociated on this surface [10]. The carbon dioxide molecule is adsorbed and dissociated into two parts: O and CO [11]. Finally, we obtained that the methane molecule adsorbs with no dissociation. It is physisorbed, and desorbs at 600 K [11]. ⁎ Corresponding author.
http://dx.doi.org/10.1016/j.susc.2015.03.003 0039-6028/© 2015 Elsevier B.V. All rights reserved.
In this work, we studied the adsorption of these gases considering a high titanium coverage over a different substrate. Here, the system we investigated is a layer of the hexagonal boron nitride with high titanium coverage (h-BNTi). The influence of the substrate may be important. Boron nitride is a synthetic material. It was discovered in the late nineteenth century. Boron and nitrogen being neighbors of carbon in the periodic table have atomic radii comparable to the latter. In its hexagonal form (h-BN), boron nitride has a crystal structure similar to that of graphite. This is the structure that we considered in this work, although boron nitride can also exhibit a cubic structure [12]. Currently, it is possible to prepare boron nitride nanotubes and sheets [13]. In particular, large-scale fabrication of h-BN nanosheets were reported [14]. These nanotubes and nanosheets can be used, among other applications, as a mechanical substrate for graphene. In addition, a BN nanosheet is a unique insulating two-dimensional system to explore [14–16]. The aim of this work is to provide new insights for the seeking of functionalization methods that could widen the application fields of boron nitride nanomaterials. In particular, we believe that the use of titanium atoms decorating the h-BN surface in the way that we propose, is preferable to working with the titanium surface alone for this type of applications, because the h-BN surface is lighter and could be potentially easier to work with, in a practical device. Previous studies have found that functionalized boron nitride surfaces could be suitable for electronic applications such as gas sensors [17]. 2. Method We performed density functional theory (DFT) calculations, with the local density approximation (LDA), and the general gradient
I. Carrillo et al. / Surface Science 637–638 (2015) 48–52
approximation (GGA), molecular dynamics (MD), within the Born– Oppenheimer approximation with the Quantum Espresso code [18]. For exchange–correlation energies, we used the Perdew–Zunger expression [19], for LDA, and the Perdew–Burke–Ernzerhof (PBE) approximation for GGA [20]. We used the norm conserving Troullier– Martins pseudopotentials [21], in the fully separable form of Kleinman– Bylander [22]. We considered as valence electronic states for boron: 2s22p1, for nitrogen: 2s22p3, for hydrogen: 1 s, for carbon: 2s22p2, for oxygen: 2s22p4, for titanium: 3s23p63d24s2. We performed non-relativistic, and non-spin polarized calculations. The cut-off energy was 1100 eV, and we took 40 k points within the Monkhorst–Pack special k point scheme [23]. The threshold energy convergence was 1.2 × 10− 5 eV. For the validation of our pseudopotentials, we obtained by energy minimization in the CO2 molecule, for the length of the C–O bond 1.193 Å and for the O–C–O angle, 179.9° (the experimental values are 1.163 Å and 180°, respectively [24]). In the same manner, with the hydrogen pseudopotential we obtained for the H–C chemical bond length in CH4 1.097 Å and for the H–C–H angle 109.36° (the experimental values are 1.087 Å and 109.4°, respectively [24]). In the same way, and using the pseudopotential for titanium, we obtained for the lattice parameter a = 2.863 Å and for c = 4.5444 Å, the experimental values are 2.950 and 4.683 Å, respectively [24]. In the case of h-BN we obtained for the B–N chemical bond length 1.447 Å, and for the lattice parameter c 6.661 Å; the corresponding experimental values are 1.446 Å, and 6.661 Å, respectively [24]. We represented the system h-BNTi using a hexagonal unit cell with a = b = 5.0 Å and c = 40 Å, with periodic conditions, containing four Ti atoms, four B atoms, and four N atoms (see Fig. 1). In the c direction, we took a large enough separation to ensure that there is no interaction between adjacent Ti and h-BN sheets. We used at MD 300 K and the definition for the adsorption energy ΔE = E(h-BN + Ti) − [E(h-BN) + E(Ti)]; where E(h-BN + Ti) is the energy of the final optimized configuration; [E(h-BN) + E(Ti)] is the energy of the initial system, which is h-BN alone plus the energy of Ti alone with no interaction between them. We found that the adsorption energy is −6.12 eV/Ti with LDA, and −6.16 with GGA. The Ti atoms are contained in two different planes. This configuration is similar, not exactly equal to that of Ti–graphene with high metal coverage (C2Ti) [11]. In the following, and in order to compare these two systems, we describe some parameters of the optimized h-BNTi system, and we write in a parenthesis the corresponding value of the parameter of the optimized C2Ti system. The Ti atoms are contained in two different planes. The nearest Ti atom to the h-BN plane is at 2.965 (1.828) Å. The other plane of titanium atoms is at a distance of 4.14 (3.239) Å from the h-BN (graphene) plane. In this way, the distance from the titanium lower plane to the titanium upper plane is 1.176 (2.815) Å. The distance Ti–Ti is 2.89 (4.22) Å and 2.418 (2.44) Å on the upper plane and on the lower plane, respectively, see Fig. 1a. From a Lowdin charge analysis, we found that the Ti atoms of the upper plane are positive, with a charge +0.413 (+0.335), and the Ti
Fig. 1. Unit cell oh h-BN with high Ti coverage (h-BNTi), after optimization.
49
atoms of the lower plane are negative with a charge −0.350 (positive with a charge + 0.548). Clearly, there are different charge transfers between the titanium layer and the substrates. The origin of the different charge distributions on the titanium atoms within the titanium layers can be seen, in a simple fashion, on the distinct values for the electronegativity of the atoms on each substrate. The corresponding values of the Allen electronegativity in Pauling scale are 2.051 for boron, 3.066 for nitrogen, and 2.544 for carbon. It is convenient to mention that the Pauling electronegativities are very close to these values. Clearly, the two titanium surfaces should not have identical properties. In this way, these two surfaces of titanium with a different substrate should not interact in exactly the same way with a given molecule or atom. Certainly, for the h-BnTi configuration, we still have a hexagonal distribution of Ti atoms after optimization. Each Ti is above a hexagon. On the other hand, it is well known that the Ti[001] surface is hexagonal, with the atoms on one plane, and a Ti–Ti distance of 2.95 Å [24]. In this way, the surface for our system h-BN–Ti resembles the Ti[001] surface. In Figs. 1(a), 2(a), and (3) we show this surface. Notice that the two surfaces are not identical. After the h-BNTi surface was optimized, we investigated its interaction with CO, CO2, and CH4. We defined the adsorption energy of each molecule, M: ΔE = E(h-BNTi + M) − [E(h-BNTi) + E(M)]; where E(h-BNTi + M) is the energy of the final configuration; [E(h-BNTi) + E(M)] is the energy of the original system, which is the energy of the h-BNTi solely with no interaction with M, plus the energy of the molecule alone. For each molecule, we considered different initial possible orientations with respect to the h-BNTi surface. 300 K and atmospheric pressure were considered in our calculations. We controlled the temperature in the MD calculations via velocity rescaling. The pressure in the Quantum espresso code was controlled via a variable super cell size. The Parrinello–Raman barostat is used. The zero pressure corresponds to atmospheric pressure. We describe our results in the following.
3. Results and discussion 3.1. Adsorption of the CO molecule We found that the carbon monoxide molecule is always chemisorbed and dissociated on the surface at 300 K. We considered three initial orientations of this molecule to study its adsorption on the h-BNTi system. The first one was with the molecule line parallel to the adsorption surface; the second one was with the molecule line perpendicular to the surface, with the oxygen atom facing
Fig. 2. Time evolution of the interaction of a carbon monoxide molecule, for the first orientation we considered, with the h-BN surface. The unit cell of the initial configuration of the system is shown in (a). In the (b), we show the resulting surface composition. We considered 300 K and atmospheric pressure.
50
I. Carrillo et al. / Surface Science 637–638 (2015) 48–52
Fig. 3. Time evolution of the interaction of a carbon dioxide molecule, for the first orientation we considered, with the h-BN surface. The unit cell of the initial configuration of the system is shown in (a). In the (b), we show the resulting surface composition. We considered 300 K and atmospheric pressure.
the surface; and the third was with the molecule line perpendicular to the surface but with the carbon atom facing the h-BNTi plane. For each initial position of the CO molecule, we allowed the system to follow an evolutionary process using molecular dynamics, with a time step of 1 fs. In all cases, the positively charged Ti atoms pull the negatively charged carbon atom towards the surface. This force is larger than the initial repulsion on the O atom due to the Ti atoms. Strong disruptive forces appear, and the molecule is broken. The carbon atom ends bound to three Ti atoms of the upper and the lower planes. The same happens with the titanium atom, see Fig. 2. The initial charge of the carbon atom is −0.0218, and its final charge is 0.13. Clearly, there is an electronic charge transfer from the C to the Ti atoms. For the O atom we have the initial charge is 0.0219, and its final charge is −0.16. In this case, the O atom takes electronic charge from the Ti atoms. There are reports on dissociative adsorption of CO on [0001]Ti surface [1–4], and on a polycrystalline Ti surface [25–28]. These reports are consistent with our results. The surface of h-BNTi is very similar (although not identical) to the [001]Ti surface. This dissociative adsorption of CO was also predicted for the graphene–Ti system with a high metal coverage (C2Ti). The calculated adsorption energy was − 5.22 eV [10]. Thus, these three surfaces share this property. All the three initial orientations of the molecule yielded to the same result. The adsorption energy, when we used LDA was − 2.66 eV, and −2.71 eV when we used GGA. However, we could not find experimental results for the CO adsorption energy on the (0001)Ti surface, specifically. In Ref. [4] the authors report the experimental dissociation of CO on the (0001)Ti surface, but they did not measure the CO adsorption energy. However, they made a theoretical calculation of the adsorption energy, on the (0001)Ti surface with the result of −2.04 eV. In Ref. [3] the authors report a theoretical calculation of the CO adsorption energy on the (0001)Ti surface, with a result of −2.6 eV. There are experimental reports on the CO adsorption energy in references [25] and [28], but for polycrystalline Ti. The corresponding values are −0.5 eV, and −0.88 eV, respectively. 3.2. Adsorption of the CO2 molecule As it occurred in the case of the carbon monoxide molecule, the carbon dioxide molecule is always chemisorbed and dissociated on the h-BNTi surface at 300 K. The carbon dioxide molecule is linear. We considered three initial orientations of this molecule to study its adsorption on the h-BNTi system. The first was with the molecule axis parallel to the surface. The second one was taken with this axis perpendicular to the surface. The third was with the molecule line making an angle with the graphene plane.
In all cases, the positively charged Ti atoms pull the negatively charged oxygen atoms towards the surface. This force is larger than the initial repulsion on the C atom by the Ti atoms. When the molecule approaches the surface, the molecule cannot be linear any longer, and it bends. The O atoms are under different force fields. The molecule interacts very strongly with the Ti atoms of the upper plane. The CO2 molecule is dissociated into two parts: O and CO. The first fragment, one O atom, ends bonded to four titanium atoms. The CO fraction is adsorbed on the surface in such a way that the C atom is bonded to two Ti atoms and the O atom is bonded to another two Ti atoms. This is shown in Fig. 3(b). The adsorption energy for the CO2 molecule is −2.72 eV when we utilized LDA, and −2.77 eV when employing GGA. Notice that the CO fraction created through CO2 dissociation does not also dissociate in this case. This is because when CO dissociates on the surface, the interaction is with the pristine surface. We have this same situation when CO2 is broken into CO and O. In this case, both fractions, CO and O, chemisorbed and remain in equilibrium on the surface. In this way, the fraction CO cannot face the pristine surface as in the first case. Besides, this fraction ends up bonded to the atoms on the surface. In this circumstance, the forces on this fraction are quite different from the first case, when CO interacts with the pristine surface. It is convenient to mention that in Refs. [25] to [27], and [29], we have reports on the adsorption of a CO2 molecule on a Ti surface with dissociation in CO and O. This fact is also in agreement with our results, in which the CO2 molecule adsorbs and dissociates on the h-BNTi system. This dissociative adsorption of CO2 was as well predicted for the Ti–graphene system with a high metal coverage (C2Ti). The calculated adsorption energy was − 4.97 eV [11]. In this manner, these three surfaces share this characteristic too. On the other hand, in Ref. [29] the authors performed a first principles calculation of the adsorption energy on the (0001)Ti surface, with a result of −3.32 eV. In Ref [25] the authors report an experimental CO2 adsorption energy for polycrystalline Ti of − 0.8 eV. We could not find experimental results for the CO2 adsorption energy on the (0001)Ti surface, specifically.
3.3. Adsorption of the CH4 molecule The methane molecule is a tetrahedron with a hydrogen atom at each vertex and the carbon atom in the middle (see Fig. 4(a)). The C atom and one H atom have negative charge. Two initial positions of the methane molecule were considered for our calculations. The first was with the face of the tetrahedron containing the three positive H atoms parallel to the h-BNTi surface, and the negative H atom above the carbon atom. The second one was with the same face parallel to surface, and the negative hydrogen atom below the carbon atom. We allowed the system to evolve using molecular dynamics at 300 K and the same time step. The positive Ti atoms pull the C and negative H with a force that is larger than the repulsion on the positive H atoms. In Fig. 4 we show the evolution of the system. We obtained that the methane molecule adsorbs with no dissociation. The final state (see Fig. 4(b)) is the same for any of the two initial positions of the molecule. The adsorption energy is − 0.062 eV when we used LDA, and − 0.068 eV when we utilized GGA. There are no experimental results of the methane adsorption on the surface we are studying. We found neither theoretical nor experimental results for the adsorption energy of methane on the Ti(0 0 1) surface to compare with our results. However, for the Ti–graphene system (C2Ti) the same physisorption result with no dissociation was predicted. The calculated adsorption energy was − 0.176 eV, and the desorption temperature was 600 K [11]. Finally, increasing the temperature and using MD we found that the methane molecule desorbs at 370 K. We must mention that we could not find neither theoretical nor experimental results on the adsorption of methane on h-BN, or on the (001) titanium surface.
I. Carrillo et al. / Surface Science 637–638 (2015) 48–52
51
Fig. 4. Time evolution of the interaction of a methane molecule, for the first orientation we considered, with the h-BN surface. The unit cell of the initial configuration of the system is shown in (a). In (b), we show the resulting surface composition. There is a physorption of the methane molecule. We considered 300 K and atmospheric pressure. The molecule desorbs at 370 K.
4. Conclusions Considering atmospheric pressure and 300 K, we found that the carbon monoxide and carbon dioxide molecules are always chemisorbed and dissociated on the h-BNTi surface. We also found that the CH4 molecule adsorbs with no dissociation on the surface. The surface of h-BNTi is very similar (although not identical) to the (0001) Ti surface. The dissociative adsorption of the CO molecule, that we found on the h-BNTi surface, is found experimentally on the (0001)Ti surface [1–4]. This dissociative adsorption of CO was also predicted for the Ti–graphene system with a high metal coverage (C2Ti) [10]. Thus, these three surfaces share this property. All the three initial orientations of the molecule yielded the same result. For the adsorption energy, we obtained − 2.66 eV when we used LDA, and − 2.71 eV when we used GGA. In the dissociative adsorption of the CO2 molecule on the h-BNTi surface, which we predict in this work, the molecule is broken into two fractions. One is CO and the other is O. The same dissociative adsorption of CO2 is also found experimentally on the (0001)Ti surface [25–27], and it was as well predicted for the Ti–graphene system with a high metal coverage (C2Ti) [11]. In this manner, these three surfaces share this property too. We considered three initial orientations of the molecule too, and all of them yielded the same result. The value of the adsorption energy was −2.72 eV when employing LDA, and −2.77 eV when we used GGA. For the last case, methane, we obtained that the molecule adsorbs with no dissociation. The final state is the same for any of the two initial positions of the molecule. The adsorption energy is − 0.062 eV when we used LDA, and − 0.068 eV when we utilized GGA. We could not find experimental results of the methane adsorption on the surface we are studying. We found neither theoretical nor experimental results for the adsorption energy of methane on the Ti(0 0 1) surface to compare with our results. However, for the Ti–graphene system (C 2Ti) the same physisorption result with no dissociation was predicted. The calculated adsorption energy was − 0.176 eV, and the desorption temperature was 600 K [11]. Finally, increasing the temperature and using MD, we found that the methane molecule desorbs at 370 K. The values for adsorption energies for CO and CO2, and methane desorption temperature for the system Ti/graphene (C2Ti), are different from the corresponding ones of the h-BNTi system. This can be seen as a consequence of different charge distribution on the titanium atoms, a fact that reflects the influence of the substrate.
Finally, we are adding the atomic coordinates of our calculations, in the supplementary material for on-line publication. Acknowledgments We thank Direccion General de Asuntos del Personal Academico de la Universidad Nacional Autonoma de Mexico for the partial financial support by Grant IN-106514 and we also thank UNAMSuper-Computing Center for the technical assistance. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.susc.2015.03.003. References [1] J.C. Campuzano, The Chemical Physics of Solid Surfaces and Heterogeneous Catalysis, vol. 3A, Elsevier, Amsterdam, 1990. [2] H.D. Shih, F. Jona, D.W. Jepsen, P.M. Marcus, J. Vac. Sci. Technol. 15 (1978) 596. [3] C.R. Fischer, L.A. Burke, J.L. Whitten, Phys. Rev. Lett. 344 (1982) 001. [4] M.V. Kuznetsov, D.P. Frickel, E.V. Shalaeva, N.I. Medvedeva, J. Electron Spectrosc. Relat. Phenom 96 (1998) 29. [5] T.E. Graedel, P.J. Crutzen, Atmospheric Change. An Earth System Perspective, W.H. Freeman and Company, New York, 1993. 74. [6] A. Veawab, P. Tontiwachwuthikul, A. Chakma, Ind. Eng. Chem. 38 (1999) 3917. [7] Z. Yong, V.G. Mata, A.E. Rodrigues, J. Chem. Eng. Data 45 (2000) 1093. [8] S. Himeno, T. Komatsu, S. Fujita, J. Chem. Eng. Data 50 (2005) 369. [9] P.L. Llewelyn, S. Burrelly, C. Serre, A. Vimon, M. Daturi, L. Hamon, et al., Langmuir 24 (2008) 7245. [10] I. Carrillo, L.F. Magaña, Rev. Mex. Fis. 55 (2009) 409. [11] I. Carrillo, E. Rangel, L.F. Magaña, Carbon 47 (2009) 2752. [12] Jochen Greim, Karl A. Schwetz, Boron Carbide, Boron Nitride, and Metal Borides in Ullmann's Encyclopedia of Industrial Chemistry, Wiley-VCH, Weinheim, 2005.http://dx.doi.org/10.1002/14356007.a04_295.pub2. [13] A.B. Preobrajenski, M.A. Nesterov, M.L. Ng, A.S. Vinogradov, N. Mårtensson, Chem. Phys. Lett. 446 (2007) 119. [14] Chunyi Zhi, Yoshio Bando, Chengchun Tang, Hiroaki Kuwahara, Dimitri Golberg, Adv. Mater. 21 (2009) 2889. [15] C.R. Dean, A.F. Young, I. Meric, C. Lee, L. Wang, S. Sorgenfrei, K. Watanabe, T. Taniguchi, P. Kim, K.L. Shepard, J. Hone, Nat. Nanotechnol. 5 (2010) 722. [16] J. Xue, J. Sanchez-Yamagishi, D. Bulmash, P. Jacquod, A. Deshpande, K. Watanabe, T. Taniguchi, P. Jarillo-Herrero, B.J. LeRoy, Nat. Mater. 10 (2011) 282. [17] Y.-H. Zhang, K.-G. Zhou, X.-C. Gou, K.-F. Xie, H.-L. Zhang, Y. Peng, Chem. Phys. Lett. 484 (2010) 266. [18] P. Giannozzi, S. Baroni, N. Bonini, M. Calandra, R. Car, C. Cavazzoni, D. Ceresoli, G.L. Chiarotti, M. Cococcioni, I. Dabo, A. Dal Corso, S. de Gironcoli, S. Fabris, G. Fratesi, R. Gebauer, U. Gerstmann, C. Gougoussis, A. Kokalj, M. Lazzeri, L. Martin-Samos, N. Marzari, F. Mauri, R. Mazzarello, S. Paolini, A. Pasquarello, L. Paulatto, C. Sbraccia, S. Scandolo, G. Sclauzero, A.P. Seitsonen, A. Smogunov, P. Umari, R.M. Wentzcovitch, J. Phys. Condens. Matter 21 (2009) 395502.
52 [19] [20] [21] [22] [23] [24]
I. Carrillo et al. / Surface Science 637–638 (2015) 48–52 J.P. Perdew, A. Zunger, Phys. Rev. B 23 (1981) 5048. J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 77 (1996) 3865. N. Troullier, J.L. Martıns, Phys. Rev. B 43 (1991) 1993. L. Kleinman y, D.M. Bylander, Phys. Rev. Lett. 48 (1980) 1425. H.J. Monkhorst, J.D. Pack, Phys. Rev. B 13 (1976) 5188. D.R. Lide (Ed.), Handbook of Chemistry and Physics, CRC, Boca Raton, FL, 2000, p. 9.
[25] [26] [27] [28] [29]
G.B. Raup, J.A. Dumesic, J. Phys. Chem. 89 (1985) 5240. P.H. Dawson, Surf. Sci. 65 (1977) 41. K. Kawasaki, T. Sugita, S. Ebisawa, Surf. Sci. 7 (1967) 502. J.A. Schwarz, R.S. Polizzotti, J.J. Burton, Surf. Sci. 67 (1977) 10. S.F. Li, Z.X. Guo, J. Phys. Chem. C 114 (2010) 11456.
Contents lists available at ScienceDirect
Surface Science journal homepage: www.elsevier.com/locate/susc
Adsorption of carbon monoxide, carbon dioxide and methane on hexagonal boron nitride with high titanium coverage I. Carrillo a, J.M. Ramírez b, L.F. Magaña b,⁎ a b
Instituto Tecnológico de Tláhuac, Av. Estanislao Ramírez # 301, Colonia Ampliación Selene, C.P. 13420 México, D.F., Mexico Instituto de Física, Universidad Nacional Autónoma de México, Apartado Postal 20-364, C.P. 01000 México, D.F., Mexico
a r t i c l e
i n f o
Article history: Received 30 September 2014 Accepted 8 March 2015 Available online 14 March 2015 Keywords: Density functional calculations Hexagonal boron nitride Carbon monoxide Carbon dioxide Methane
a b s t r a c t We used density functional theory and molecular dynamics to explore the adsorption of CO, CO2, and CH4 on a layer of h-BN with high titanium coverage. After optimization, we found that each titanium atom is located above each of the hexagons of the surface. We considered atmospheric pressure and 300 K. It is found that the first molecule is adsorbed and dissociated on the surface. The CO2 molecule is broken into O and CO. The methane molecule is physisorbed, and not dissociated. It is desorbed at 370 K. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Carbon monoxide, carbon dioxide and methane are environmental problems. CO is a toxic gas which, at normal conditions is tasteless, odorless, and colorless. It comes primarily from incomplete combustion of carbon containing fuels. It is not-irritating, and it is an asphyxiant. At high levels, it is deadly. The investigation of the adsorption of carbon monoxide on palladium and transition metals is an attractive research topic. There are theoretical and experimental studies on the adsorption of carbon monoxide on the Ti [001] surface [1–4]. On the other hand, many industrial processes around the world produce CO2 and CH4. These gases are two of the main greenhouse gases emitted today [5]. At the present, the removal of these gases from the atmosphere is mainly done using amine baths. This process presents several problems. Among them, we have additional processing and corrosion control [6]. Another procedure is by chemically capturing CO2 using oxides [7]. However, to remove these greenhouse gases, adsorption processes have been considered too. Among the materials that have been studied for these processes, we have activated carbons, zeolites, and porous metal–organic frameworks [8,9]. In previous works, we studied the adsorption of these three gases on a titanium–graphene system with high metal coverage [10,11]. We found that the carbon monoxide molecule is adsorbed and dissociated on this surface [10]. The carbon dioxide molecule is adsorbed and dissociated into two parts: O and CO [11]. Finally, we obtained that the methane molecule adsorbs with no dissociation. It is physisorbed, and desorbs at 600 K [11]. ⁎ Corresponding author.
http://dx.doi.org/10.1016/j.susc.2015.03.003 0039-6028/© 2015 Elsevier B.V. All rights reserved.
In this work, we studied the adsorption of these gases considering a high titanium coverage over a different substrate. Here, the system we investigated is a layer of the hexagonal boron nitride with high titanium coverage (h-BNTi). The influence of the substrate may be important. Boron nitride is a synthetic material. It was discovered in the late nineteenth century. Boron and nitrogen being neighbors of carbon in the periodic table have atomic radii comparable to the latter. In its hexagonal form (h-BN), boron nitride has a crystal structure similar to that of graphite. This is the structure that we considered in this work, although boron nitride can also exhibit a cubic structure [12]. Currently, it is possible to prepare boron nitride nanotubes and sheets [13]. In particular, large-scale fabrication of h-BN nanosheets were reported [14]. These nanotubes and nanosheets can be used, among other applications, as a mechanical substrate for graphene. In addition, a BN nanosheet is a unique insulating two-dimensional system to explore [14–16]. The aim of this work is to provide new insights for the seeking of functionalization methods that could widen the application fields of boron nitride nanomaterials. In particular, we believe that the use of titanium atoms decorating the h-BN surface in the way that we propose, is preferable to working with the titanium surface alone for this type of applications, because the h-BN surface is lighter and could be potentially easier to work with, in a practical device. Previous studies have found that functionalized boron nitride surfaces could be suitable for electronic applications such as gas sensors [17]. 2. Method We performed density functional theory (DFT) calculations, with the local density approximation (LDA), and the general gradient
I. Carrillo et al. / Surface Science 637–638 (2015) 48–52
approximation (GGA), molecular dynamics (MD), within the Born– Oppenheimer approximation with the Quantum Espresso code [18]. For exchange–correlation energies, we used the Perdew–Zunger expression [19], for LDA, and the Perdew–Burke–Ernzerhof (PBE) approximation for GGA [20]. We used the norm conserving Troullier– Martins pseudopotentials [21], in the fully separable form of Kleinman– Bylander [22]. We considered as valence electronic states for boron: 2s22p1, for nitrogen: 2s22p3, for hydrogen: 1 s, for carbon: 2s22p2, for oxygen: 2s22p4, for titanium: 3s23p63d24s2. We performed non-relativistic, and non-spin polarized calculations. The cut-off energy was 1100 eV, and we took 40 k points within the Monkhorst–Pack special k point scheme [23]. The threshold energy convergence was 1.2 × 10− 5 eV. For the validation of our pseudopotentials, we obtained by energy minimization in the CO2 molecule, for the length of the C–O bond 1.193 Å and for the O–C–O angle, 179.9° (the experimental values are 1.163 Å and 180°, respectively [24]). In the same manner, with the hydrogen pseudopotential we obtained for the H–C chemical bond length in CH4 1.097 Å and for the H–C–H angle 109.36° (the experimental values are 1.087 Å and 109.4°, respectively [24]). In the same way, and using the pseudopotential for titanium, we obtained for the lattice parameter a = 2.863 Å and for c = 4.5444 Å, the experimental values are 2.950 and 4.683 Å, respectively [24]. In the case of h-BN we obtained for the B–N chemical bond length 1.447 Å, and for the lattice parameter c 6.661 Å; the corresponding experimental values are 1.446 Å, and 6.661 Å, respectively [24]. We represented the system h-BNTi using a hexagonal unit cell with a = b = 5.0 Å and c = 40 Å, with periodic conditions, containing four Ti atoms, four B atoms, and four N atoms (see Fig. 1). In the c direction, we took a large enough separation to ensure that there is no interaction between adjacent Ti and h-BN sheets. We used at MD 300 K and the definition for the adsorption energy ΔE = E(h-BN + Ti) − [E(h-BN) + E(Ti)]; where E(h-BN + Ti) is the energy of the final optimized configuration; [E(h-BN) + E(Ti)] is the energy of the initial system, which is h-BN alone plus the energy of Ti alone with no interaction between them. We found that the adsorption energy is −6.12 eV/Ti with LDA, and −6.16 with GGA. The Ti atoms are contained in two different planes. This configuration is similar, not exactly equal to that of Ti–graphene with high metal coverage (C2Ti) [11]. In the following, and in order to compare these two systems, we describe some parameters of the optimized h-BNTi system, and we write in a parenthesis the corresponding value of the parameter of the optimized C2Ti system. The Ti atoms are contained in two different planes. The nearest Ti atom to the h-BN plane is at 2.965 (1.828) Å. The other plane of titanium atoms is at a distance of 4.14 (3.239) Å from the h-BN (graphene) plane. In this way, the distance from the titanium lower plane to the titanium upper plane is 1.176 (2.815) Å. The distance Ti–Ti is 2.89 (4.22) Å and 2.418 (2.44) Å on the upper plane and on the lower plane, respectively, see Fig. 1a. From a Lowdin charge analysis, we found that the Ti atoms of the upper plane are positive, with a charge +0.413 (+0.335), and the Ti
Fig. 1. Unit cell oh h-BN with high Ti coverage (h-BNTi), after optimization.
49
atoms of the lower plane are negative with a charge −0.350 (positive with a charge + 0.548). Clearly, there are different charge transfers between the titanium layer and the substrates. The origin of the different charge distributions on the titanium atoms within the titanium layers can be seen, in a simple fashion, on the distinct values for the electronegativity of the atoms on each substrate. The corresponding values of the Allen electronegativity in Pauling scale are 2.051 for boron, 3.066 for nitrogen, and 2.544 for carbon. It is convenient to mention that the Pauling electronegativities are very close to these values. Clearly, the two titanium surfaces should not have identical properties. In this way, these two surfaces of titanium with a different substrate should not interact in exactly the same way with a given molecule or atom. Certainly, for the h-BnTi configuration, we still have a hexagonal distribution of Ti atoms after optimization. Each Ti is above a hexagon. On the other hand, it is well known that the Ti[001] surface is hexagonal, with the atoms on one plane, and a Ti–Ti distance of 2.95 Å [24]. In this way, the surface for our system h-BN–Ti resembles the Ti[001] surface. In Figs. 1(a), 2(a), and (3) we show this surface. Notice that the two surfaces are not identical. After the h-BNTi surface was optimized, we investigated its interaction with CO, CO2, and CH4. We defined the adsorption energy of each molecule, M: ΔE = E(h-BNTi + M) − [E(h-BNTi) + E(M)]; where E(h-BNTi + M) is the energy of the final configuration; [E(h-BNTi) + E(M)] is the energy of the original system, which is the energy of the h-BNTi solely with no interaction with M, plus the energy of the molecule alone. For each molecule, we considered different initial possible orientations with respect to the h-BNTi surface. 300 K and atmospheric pressure were considered in our calculations. We controlled the temperature in the MD calculations via velocity rescaling. The pressure in the Quantum espresso code was controlled via a variable super cell size. The Parrinello–Raman barostat is used. The zero pressure corresponds to atmospheric pressure. We describe our results in the following.
3. Results and discussion 3.1. Adsorption of the CO molecule We found that the carbon monoxide molecule is always chemisorbed and dissociated on the surface at 300 K. We considered three initial orientations of this molecule to study its adsorption on the h-BNTi system. The first one was with the molecule line parallel to the adsorption surface; the second one was with the molecule line perpendicular to the surface, with the oxygen atom facing
Fig. 2. Time evolution of the interaction of a carbon monoxide molecule, for the first orientation we considered, with the h-BN surface. The unit cell of the initial configuration of the system is shown in (a). In the (b), we show the resulting surface composition. We considered 300 K and atmospheric pressure.
50
I. Carrillo et al. / Surface Science 637–638 (2015) 48–52
Fig. 3. Time evolution of the interaction of a carbon dioxide molecule, for the first orientation we considered, with the h-BN surface. The unit cell of the initial configuration of the system is shown in (a). In the (b), we show the resulting surface composition. We considered 300 K and atmospheric pressure.
the surface; and the third was with the molecule line perpendicular to the surface but with the carbon atom facing the h-BNTi plane. For each initial position of the CO molecule, we allowed the system to follow an evolutionary process using molecular dynamics, with a time step of 1 fs. In all cases, the positively charged Ti atoms pull the negatively charged carbon atom towards the surface. This force is larger than the initial repulsion on the O atom due to the Ti atoms. Strong disruptive forces appear, and the molecule is broken. The carbon atom ends bound to three Ti atoms of the upper and the lower planes. The same happens with the titanium atom, see Fig. 2. The initial charge of the carbon atom is −0.0218, and its final charge is 0.13. Clearly, there is an electronic charge transfer from the C to the Ti atoms. For the O atom we have the initial charge is 0.0219, and its final charge is −0.16. In this case, the O atom takes electronic charge from the Ti atoms. There are reports on dissociative adsorption of CO on [0001]Ti surface [1–4], and on a polycrystalline Ti surface [25–28]. These reports are consistent with our results. The surface of h-BNTi is very similar (although not identical) to the [001]Ti surface. This dissociative adsorption of CO was also predicted for the graphene–Ti system with a high metal coverage (C2Ti). The calculated adsorption energy was − 5.22 eV [10]. Thus, these three surfaces share this property. All the three initial orientations of the molecule yielded to the same result. The adsorption energy, when we used LDA was − 2.66 eV, and −2.71 eV when we used GGA. However, we could not find experimental results for the CO adsorption energy on the (0001)Ti surface, specifically. In Ref. [4] the authors report the experimental dissociation of CO on the (0001)Ti surface, but they did not measure the CO adsorption energy. However, they made a theoretical calculation of the adsorption energy, on the (0001)Ti surface with the result of −2.04 eV. In Ref. [3] the authors report a theoretical calculation of the CO adsorption energy on the (0001)Ti surface, with a result of −2.6 eV. There are experimental reports on the CO adsorption energy in references [25] and [28], but for polycrystalline Ti. The corresponding values are −0.5 eV, and −0.88 eV, respectively. 3.2. Adsorption of the CO2 molecule As it occurred in the case of the carbon monoxide molecule, the carbon dioxide molecule is always chemisorbed and dissociated on the h-BNTi surface at 300 K. The carbon dioxide molecule is linear. We considered three initial orientations of this molecule to study its adsorption on the h-BNTi system. The first was with the molecule axis parallel to the surface. The second one was taken with this axis perpendicular to the surface. The third was with the molecule line making an angle with the graphene plane.
In all cases, the positively charged Ti atoms pull the negatively charged oxygen atoms towards the surface. This force is larger than the initial repulsion on the C atom by the Ti atoms. When the molecule approaches the surface, the molecule cannot be linear any longer, and it bends. The O atoms are under different force fields. The molecule interacts very strongly with the Ti atoms of the upper plane. The CO2 molecule is dissociated into two parts: O and CO. The first fragment, one O atom, ends bonded to four titanium atoms. The CO fraction is adsorbed on the surface in such a way that the C atom is bonded to two Ti atoms and the O atom is bonded to another two Ti atoms. This is shown in Fig. 3(b). The adsorption energy for the CO2 molecule is −2.72 eV when we utilized LDA, and −2.77 eV when employing GGA. Notice that the CO fraction created through CO2 dissociation does not also dissociate in this case. This is because when CO dissociates on the surface, the interaction is with the pristine surface. We have this same situation when CO2 is broken into CO and O. In this case, both fractions, CO and O, chemisorbed and remain in equilibrium on the surface. In this way, the fraction CO cannot face the pristine surface as in the first case. Besides, this fraction ends up bonded to the atoms on the surface. In this circumstance, the forces on this fraction are quite different from the first case, when CO interacts with the pristine surface. It is convenient to mention that in Refs. [25] to [27], and [29], we have reports on the adsorption of a CO2 molecule on a Ti surface with dissociation in CO and O. This fact is also in agreement with our results, in which the CO2 molecule adsorbs and dissociates on the h-BNTi system. This dissociative adsorption of CO2 was as well predicted for the Ti–graphene system with a high metal coverage (C2Ti). The calculated adsorption energy was − 4.97 eV [11]. In this manner, these three surfaces share this characteristic too. On the other hand, in Ref. [29] the authors performed a first principles calculation of the adsorption energy on the (0001)Ti surface, with a result of −3.32 eV. In Ref [25] the authors report an experimental CO2 adsorption energy for polycrystalline Ti of − 0.8 eV. We could not find experimental results for the CO2 adsorption energy on the (0001)Ti surface, specifically.
3.3. Adsorption of the CH4 molecule The methane molecule is a tetrahedron with a hydrogen atom at each vertex and the carbon atom in the middle (see Fig. 4(a)). The C atom and one H atom have negative charge. Two initial positions of the methane molecule were considered for our calculations. The first was with the face of the tetrahedron containing the three positive H atoms parallel to the h-BNTi surface, and the negative H atom above the carbon atom. The second one was with the same face parallel to surface, and the negative hydrogen atom below the carbon atom. We allowed the system to evolve using molecular dynamics at 300 K and the same time step. The positive Ti atoms pull the C and negative H with a force that is larger than the repulsion on the positive H atoms. In Fig. 4 we show the evolution of the system. We obtained that the methane molecule adsorbs with no dissociation. The final state (see Fig. 4(b)) is the same for any of the two initial positions of the molecule. The adsorption energy is − 0.062 eV when we used LDA, and − 0.068 eV when we utilized GGA. There are no experimental results of the methane adsorption on the surface we are studying. We found neither theoretical nor experimental results for the adsorption energy of methane on the Ti(0 0 1) surface to compare with our results. However, for the Ti–graphene system (C2Ti) the same physisorption result with no dissociation was predicted. The calculated adsorption energy was − 0.176 eV, and the desorption temperature was 600 K [11]. Finally, increasing the temperature and using MD we found that the methane molecule desorbs at 370 K. We must mention that we could not find neither theoretical nor experimental results on the adsorption of methane on h-BN, or on the (001) titanium surface.
I. Carrillo et al. / Surface Science 637–638 (2015) 48–52
51
Fig. 4. Time evolution of the interaction of a methane molecule, for the first orientation we considered, with the h-BN surface. The unit cell of the initial configuration of the system is shown in (a). In (b), we show the resulting surface composition. There is a physorption of the methane molecule. We considered 300 K and atmospheric pressure. The molecule desorbs at 370 K.
4. Conclusions Considering atmospheric pressure and 300 K, we found that the carbon monoxide and carbon dioxide molecules are always chemisorbed and dissociated on the h-BNTi surface. We also found that the CH4 molecule adsorbs with no dissociation on the surface. The surface of h-BNTi is very similar (although not identical) to the (0001) Ti surface. The dissociative adsorption of the CO molecule, that we found on the h-BNTi surface, is found experimentally on the (0001)Ti surface [1–4]. This dissociative adsorption of CO was also predicted for the Ti–graphene system with a high metal coverage (C2Ti) [10]. Thus, these three surfaces share this property. All the three initial orientations of the molecule yielded the same result. For the adsorption energy, we obtained − 2.66 eV when we used LDA, and − 2.71 eV when we used GGA. In the dissociative adsorption of the CO2 molecule on the h-BNTi surface, which we predict in this work, the molecule is broken into two fractions. One is CO and the other is O. The same dissociative adsorption of CO2 is also found experimentally on the (0001)Ti surface [25–27], and it was as well predicted for the Ti–graphene system with a high metal coverage (C2Ti) [11]. In this manner, these three surfaces share this property too. We considered three initial orientations of the molecule too, and all of them yielded the same result. The value of the adsorption energy was −2.72 eV when employing LDA, and −2.77 eV when we used GGA. For the last case, methane, we obtained that the molecule adsorbs with no dissociation. The final state is the same for any of the two initial positions of the molecule. The adsorption energy is − 0.062 eV when we used LDA, and − 0.068 eV when we utilized GGA. We could not find experimental results of the methane adsorption on the surface we are studying. We found neither theoretical nor experimental results for the adsorption energy of methane on the Ti(0 0 1) surface to compare with our results. However, for the Ti–graphene system (C 2Ti) the same physisorption result with no dissociation was predicted. The calculated adsorption energy was − 0.176 eV, and the desorption temperature was 600 K [11]. Finally, increasing the temperature and using MD, we found that the methane molecule desorbs at 370 K. The values for adsorption energies for CO and CO2, and methane desorption temperature for the system Ti/graphene (C2Ti), are different from the corresponding ones of the h-BNTi system. This can be seen as a consequence of different charge distribution on the titanium atoms, a fact that reflects the influence of the substrate.
Finally, we are adding the atomic coordinates of our calculations, in the supplementary material for on-line publication. Acknowledgments We thank Direccion General de Asuntos del Personal Academico de la Universidad Nacional Autonoma de Mexico for the partial financial support by Grant IN-106514 and we also thank UNAMSuper-Computing Center for the technical assistance. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.susc.2015.03.003. References [1] J.C. Campuzano, The Chemical Physics of Solid Surfaces and Heterogeneous Catalysis, vol. 3A, Elsevier, Amsterdam, 1990. [2] H.D. Shih, F. Jona, D.W. Jepsen, P.M. Marcus, J. Vac. Sci. Technol. 15 (1978) 596. [3] C.R. Fischer, L.A. Burke, J.L. Whitten, Phys. Rev. Lett. 344 (1982) 001. [4] M.V. Kuznetsov, D.P. Frickel, E.V. Shalaeva, N.I. Medvedeva, J. Electron Spectrosc. Relat. Phenom 96 (1998) 29. [5] T.E. Graedel, P.J. Crutzen, Atmospheric Change. An Earth System Perspective, W.H. Freeman and Company, New York, 1993. 74. [6] A. Veawab, P. Tontiwachwuthikul, A. Chakma, Ind. Eng. Chem. 38 (1999) 3917. [7] Z. Yong, V.G. Mata, A.E. Rodrigues, J. Chem. Eng. Data 45 (2000) 1093. [8] S. Himeno, T. Komatsu, S. Fujita, J. Chem. Eng. Data 50 (2005) 369. [9] P.L. Llewelyn, S. Burrelly, C. Serre, A. Vimon, M. Daturi, L. Hamon, et al., Langmuir 24 (2008) 7245. [10] I. Carrillo, L.F. Magaña, Rev. Mex. Fis. 55 (2009) 409. [11] I. Carrillo, E. Rangel, L.F. Magaña, Carbon 47 (2009) 2752. [12] Jochen Greim, Karl A. Schwetz, Boron Carbide, Boron Nitride, and Metal Borides in Ullmann's Encyclopedia of Industrial Chemistry, Wiley-VCH, Weinheim, 2005.http://dx.doi.org/10.1002/14356007.a04_295.pub2. [13] A.B. Preobrajenski, M.A. Nesterov, M.L. Ng, A.S. Vinogradov, N. Mårtensson, Chem. Phys. Lett. 446 (2007) 119. [14] Chunyi Zhi, Yoshio Bando, Chengchun Tang, Hiroaki Kuwahara, Dimitri Golberg, Adv. Mater. 21 (2009) 2889. [15] C.R. Dean, A.F. Young, I. Meric, C. Lee, L. Wang, S. Sorgenfrei, K. Watanabe, T. Taniguchi, P. Kim, K.L. Shepard, J. Hone, Nat. Nanotechnol. 5 (2010) 722. [16] J. Xue, J. Sanchez-Yamagishi, D. Bulmash, P. Jacquod, A. Deshpande, K. Watanabe, T. Taniguchi, P. Jarillo-Herrero, B.J. LeRoy, Nat. Mater. 10 (2011) 282. [17] Y.-H. Zhang, K.-G. Zhou, X.-C. Gou, K.-F. Xie, H.-L. Zhang, Y. Peng, Chem. Phys. Lett. 484 (2010) 266. [18] P. Giannozzi, S. Baroni, N. Bonini, M. Calandra, R. Car, C. Cavazzoni, D. Ceresoli, G.L. Chiarotti, M. Cococcioni, I. Dabo, A. Dal Corso, S. de Gironcoli, S. Fabris, G. Fratesi, R. Gebauer, U. Gerstmann, C. Gougoussis, A. Kokalj, M. Lazzeri, L. Martin-Samos, N. Marzari, F. Mauri, R. Mazzarello, S. Paolini, A. Pasquarello, L. Paulatto, C. Sbraccia, S. Scandolo, G. Sclauzero, A.P. Seitsonen, A. Smogunov, P. Umari, R.M. Wentzcovitch, J. Phys. Condens. Matter 21 (2009) 395502.
52 [19] [20] [21] [22] [23] [24]
I. Carrillo et al. / Surface Science 637–638 (2015) 48–52 J.P. Perdew, A. Zunger, Phys. Rev. B 23 (1981) 5048. J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 77 (1996) 3865. N. Troullier, J.L. Martıns, Phys. Rev. B 43 (1991) 1993. L. Kleinman y, D.M. Bylander, Phys. Rev. Lett. 48 (1980) 1425. H.J. Monkhorst, J.D. Pack, Phys. Rev. B 13 (1976) 5188. D.R. Lide (Ed.), Handbook of Chemistry and Physics, CRC, Boca Raton, FL, 2000, p. 9.
[25] [26] [27] [28] [29]
G.B. Raup, J.A. Dumesic, J. Phys. Chem. 89 (1985) 5240. P.H. Dawson, Surf. Sci. 65 (1977) 41. K. Kawasaki, T. Sugita, S. Ebisawa, Surf. Sci. 7 (1967) 502. J.A. Schwarz, R.S. Polizzotti, J.J. Burton, Surf. Sci. 67 (1977) 10. S.F. Li, Z.X. Guo, J. Phys. Chem. C 114 (2010) 11456.