Journal of the Chilean Chemical Society
On-line version ISSN 0717-9707
J. Chil. Chem. Soc. vol.50 no.4 Concepción Dec. 2005
J. Chil. Chem. Soc., 50, N° 4 (2005), págs: 745-752
THE INTERACTION OF ATOMIC OXYGEN WITH SMALL COPPER CLUSTERS. A THEORETICAL APPROACH.
Departamento de Química, Facultad de Ciencias Básicas, Universidad de Antofagasta, Casilla 170, Antofagasta, Chile. E-mail: firstname.lastname@example.org
A theoretical study of atomic oxygen adsorption on small Cun (n 8) clusters was carried out using density functional methods. The copper-oxygen system is important to understand the mechanism of the oxygen oxidation catalyzed by copper. It is found that the oxygen reactivity is strongly dependent of the size of copper clusters and Cun with n odd exhibiting the highest reactivity. The adsorption energy has values from 37 to 93 kcal/mol. In general, in clusters with even number of atoms the oxygen adsorption is carried out on a top type site, on the other hand, in clusters with odd number of atoms the adsorption is carried out on highly coordinated sites. In comparison with bare Cun clusters, the relative stability of copper clusters under oxygen adsorption is not largely affected, exhibiting only small modifications in the structure of the cluster. Electronic and symmetrical effects seen to govern the oxygen adsorption process.
Keywords: copper cluster, oxygen adsorption, quantum chemical calculations.
The study of small metal clusters is an active area of research. Their properties may show both similarities to and differences from their bulk counterparts. Clusters containing from a few to several thousands of atoms have been investigated to study the evolution of properties of bulk matter from properties of atoms . In addition, the basic properties of small metal clusters are believed to be an important link in the understanding of fundamental mechanism of catalysis and several chemical transformations [2,3]. In particular, the study of clusters on heterogeneous catalysis has an important relevance toward potential technological applications. A major goal in catalysis research is to design catalysts that can achieve perfect selectivity and desirable activity. However, between activity and selectivity, the latter is difficult to achieve and control. Therefore, in the synthesis of a designed heterogeneous catalyst, it would be desirable to have complete control over the formation of active site, the environment around the active site, the binding sites and their locations relatives to the active sites, and the path to access these functionalities. In other words, we would like to have complete control from the atomic scale to macroscopic scale and to achieve this it is that the study of clusters is relevant. Gases adsorption on small transition and novel metal clusters only recently they have begun to be studied [4-10]. However, studies in neutral metal clusters are more attractive due to the potential applications like supported catalyst because they are neutral in nature. About the oxygen adsorption on copper clusters exist scarce studies limited to molecular adsorption in the range from six to nine copper atoms [11,12]. The theoretical studies showed that the molecular adsorption is strongly size selective, finding that clusters with closed shells exhibit a low reactivity with molecular oxygen compared with clusters with open shells. However, these studies are not conclusive due to the limited number of analyzed copper clusters and to be partial studies of the adsorption process. It has been generally accepted that dissociative adsorption of oxygen is thermodynamically more favourable compared to molecular adsorption, yet, the dissociative adsorption can be kinetically hindered, which can be the driving force for the adsorption of the molecular oxygen. In all oxidative process it is important to favor the dissociative adsorption preserving the stability of the catalyst. Therefore, an important stage in the use of clusters like reagent centers is to know its behavior with regard to the dissociation, that is, to know the behavior of the cluster with regard to the atomic oxygen interaction.
This article presents a theoretical study of the interaction of atomic oxygen with neutral Cun (n£8) clusters. The interest is to characterize the energy and the electronic structure of the interaction system and to compare its stability with bare copper clusters studied previously , using density functional methods.
METHODOLOGY AND COMPUTATIONAL DETAILS
The electronic energy and properties have been calculated by solving the Kohn–Sham equations in an atomic basis set formed by gaussian functions. For the copper atoms the Stuttgart pseudopotential  with the corresponding basis set has been used. The pseudopotential replaces ten core electrons thus nineteen valence electrons are considered. For the oxygen atom the D95+(d) basis set  has been used. The calculations have been done using the B3LYP [17-21] exchange correlation functional, which is of the hybrid type. It consists of a careful mixing of Hartree-Fock exchange, calculated with Kohn-Sham orbitals, and the B88 exchange functional  plus the LYP correlation functional . The use of B3LYP functional has been used en previous calculations with good agreement with experimental data [14,22]. The initial geometries of Cun clusters were those optimized in the previous work . The interaction of oxygen atom with copper clusters was fully optimized without symmetry restrictions in different adsorption sites present on the cluster surface. Top, bridge and threefold hollow type sites are possible to identify on the cluster surface. All calculations were realized using in spin-unrestricted wavefunctions for the O-cluster system. For Cu3O, Cu5O and Cu6O a doublet spin multiplicity was used. For Cu2O, Cu4O, Cu6O and Cu8O a triplet spin multiplicity was used. All the calculations have been done using the Gaussian 98 program .
The binding energy of the adsorption process for the optimum process was calculated as
Eb = E(CunO) – E(Cun) – E(O); n £ 1-8 0000000(1)
being E(CunO), E(Cun) and E(O) the energies of oxygen-copper cluster, bare copper cluster and atomic oxygen, respectively.
RESULTS AND DISCUSSION
Figure 1 displays the optimal geometrical structures of oxygen-copper cluster systems with the corresponding geometrical parameters. If these structures are compared with structures of bare copper clusters studied in previous work , it can see that the interaction of oxygen atom with each one copper clusters only modifies slightly the cluster geometry. Distortions in general are observed in local environment, i.e., in the region of the oxygen adsorption. Lengthening of the Cu-Cu distance from 0.01 to 0.37 Å is found. The Cu-Cu average distance in CunO and in bare Cun clusters are plotted in Figure 2. In this figure it can see clearly that the Cu-Cu distance is larger in CunO clusters in comparison with bare Cun. In clusters with an odd number of copper atoms the oxygen is adsorbed on highly coordinated adsorption sites. In Cu3, oxygen atom is adsorbed on a threefold hollow type site, in change, in Cu5 and Cu7 is adsorbed on bridge type sites, in analogy with sites present on copper surfaces. For clusters with even number of copper atoms the oxygen is adsorbed on top type sites (Cu2 and Cu6), and Cu4 on a bridge type site. In Cu8, the oxygen adsorption is on a threefold hollow site but distorted, i.e., the adsorption process has not ideal C3v symmetry (see Fig.3). The distance between oxygen and the nearest copper atom (O-Cu), values from 1.74 Å to 1.90 Å are found. Oxygen occupying top sites present the minor O-Cu distance (about 1.75 Å). Oxygen occupying highly coordinated sites present the greater O-Cu distance (1.83-1.90 Å). In particular, in Cu3 the oxygen adsorption produces a bigger symmetry of the cluster from C2v to C3v with a O-Cu distance of 1.87 Å. It is surprising that this result is in good agreement with experimental data found for the oxygen adsorption on a Cu(111) surface (1.83±0.02 Å), in which oxygen is adsorbed on a threefold hollow site.
ENERGIES AND ELECTRONIC PROPERTIES
For all the studied clusters, the total energy, binding energy per atom, HOMO and LUMO energies, gap LUMO-HOMO and difference of gaps between CunO and Cun clusters are collected in Table I. First, it can be seen that all CunO clusters are stable on the atomic oxygen adsorption with adsorption energy from 37 to 94 kcal/mol. These values show than the interaction oxygen-Cu cluster is in the order of a chemisorption. Clusters with even number of copper atoms exhibit in general the minor adsorption energy in comparison with clusters with odd number of atoms. It is interesting to note that the greater adsorption energy is found in the Cu3O cluster that it presents a major number of O-Cu interactions and in addition, it is of the order found for the atomic oxygen adsorption on a Cu(111) surface in a similar adsorption site (100 kcal/mol) . Binding energies per atoms of CunO clusters and bare Cun are displayed in Figure 4. This figure shows clearly that the adsorption of oxygen atom makes to decrease the stability of the copper cluster, with the exception for the Cu3 cluster where it is slightly bigger. However, starting from Cu6 cluster the binding energy is more regular, tending to increase. Thus, it would be expected that in the sufficiently big copper clusters the oxygen adsorption would not affect in a significant form the stability of the cluster.
Table II presents the calculated atomic net charge on the oxygen and copper atoms in CunO clusters using Mulliken population analysis. This result shows that the adsorption is characterized for a strong charge transfer from copper cluster to oxygen atom. The charge transfer goes from 0.33e to 0.70e, a typical behavior of a chemisorption. It can be seen that the charge transfer on the oxygen atom increases with the increase the size of copper cluster, with the exception of Cu3O and Cu8O clusters, where apparently in these two last systems an optimal interaction oxygen-copper cluster is favored to the charge transfer. Table III presents the calculated atomic net charges in bare copper clusters. The comparison of results in Tables II and III show in general that after the oxygen adsorption, adsorption sites in equivalent positions in the copper cluster increase its net charge as it is evident in the clusters Cu2O, Cu4O, Cu6O and Cu7O. In clusters Cu5O and Cu8O also present adsorption sites where there are an increase of the net charge induced by the oxygen adsorption. These results suggest that the copper cluster can adsorb more than an oxygen atom. However, from these results is impossible to conclude if the adsorption of a second oxygen atom can maintain the stability of the copper cluster.
It is interesting to analyze the nature of the oxygen adsorption on copper clusters with respect to physical or reactivity parameters to establish some correlation that it permits us to characterize and predict the behavior of the adsorption process. In the literature there are some works that it points on this sense treating to predict the formation of bare copper clusters [25,26]. Comparison between the adsorption energy and binding energy per atom for CunO clusters is shown in Figure 5. One can see that there is a good correlation between both properties, finding the biggest adsorption energy for the oxygen atom when the stability of clusters is bigger and, in general, this behavior is presented in clusters with odd number of copper atoms. With a similar purpose in Figure 6 is compared the adsorption energy and the difference of gaps between Cun and CunO clusters in function of the copper atom number (n), using data collected in Table I. As the gap is a measure of the chemical reactivity related with the molecular harness , thus oxygen adsorption is more favorable when the system trend to a major harness, i.e., when the difference of gaps has a positive value and, being presented in clusters with odd number of copper atoms. From Table III the results show in general that the oxygen adsorption is realized in a region of molecular cluster with high charge density. This result is expected due to the oxygen atom behaves as electron acceptor, i.e., as a nucleophilic reagent. It can be seen that for the Cu6 and Cu7 clusters is found a high charge density in vertexes and the plane D5h, respectively, doing the preferred region by the oxygen adsorption. In Cu5 cluster the situation is similar; oxygen prefers to adsorb in the molecular region with high charge density, far from deficient charge centers.
Thus a biggest interaction with oxygen atom is achieved in the molecular plane of copper cluster in the region of the Cu-Cu bond. In addition, it is seen that another important aspect of the adsorption process seems to be the conservation of a high symmetry of the adsorption system: C×h for Cu2O, C3v for Cu3O and C2v for Cu6 and Cu7 (see Fig. 1). Therefore the combination of the electronic and symmetrical aspects seems to govern the optimal oxygen adsorption on the copper clusters.
The Fukui function is another parameter very used to predict the molecular reactivity. In particular, a novel method has been developed recently to evaluated condensed Fukui function by mean of an alone calculation maintaining the spin multiplicity of the system and, of this way to obtain chemical reactivity of different atomic sites present in a molecule for electrophilic, nucleophilic or radical attacks, defining the fk-, fk+ , fkº and functions, respectively [28, 29]. This method evaluates the condensed Fukui function for different attacks as the electron density from frontier molecular orbital coefficients (HOMO and LUMO) and the overlap matrix (for more details to see references 28 and 29). Using this method and considering that oxygen atom is a charge acceptor and the copper cluster behaves as a charge donor, therefore the most appropriate condensed Fukui function is the fk-, expecting high values for this function describing the electrophilic behavior at the specific atomic sites in copper clusters. Table IV summarizes the condensed electrophilic Fukui functions obtained for different Cun clusters. These results show that the fk- function predicts in correct form the site for the oxygen adsorption in the cases when the adsorption is produced at simple atoms in copper cluster (Cu2 and Cu6), but it is not able to predict in those cases in than oxygen adsorption takes place in highly coordinated sites (Cu3, Cu4, Cu5, Cu7 y Cu8). From these results can be concluded that it is necessary to develop appropriated functions for surfaces that they take into account as much the geometry as the symmetry in addition to charge density of the adsorption site to predict a correct reactivity. This it is not an easy task due to the complexity of the surfaces and varied systems but it is necessary to develop. Studies in this sense are at the moment in development in our work group.
The adsorption of oxygen atom on copper clusters was theoretically studied and their properties were analyzed. It was found than clusters with odd number of copper atoms presented the highest adsorption energy and it coincides with bigger stability of the CunO system. The calculated adsorption energy and the values found for the charge transfer from the copper cluster to oxygen atom confirm that the adsorption process is a chemisorption. In general, the adsorption in highly coordinated sites is preferred to a single atomic site. The high symmetry found for CunO system and electronic effects favoring the oxygen-copper cluster interaction seem to govern the adsorption process. Indexes of molecular reactivity are inappropriate to predict the chemical reactivity in highly coordinated sites of the clusters and therefore it is necessary to develop specific indexes that it can be used in surface processes.
This work has been supported by Proyectos Específicos de Investigación (U. de Antofagasta), grant PEI-1302-04.
 Ekardt, W. Metal Clusters; Wiley: U.S.A., 1999. [ Links ]
 W. Eberhardt, Surf.Sci., 500(2002)242 [ Links ]
 H.H. Kung, M.C. Kung, Catal.Today, 97(2004)219 [ Links ]
 X.L. Ding, Z.Y. Z.Y. Li, J.L Yang, J.G. Hou , Q.S Zhu, J.Chem.Phys.,121(2004) 2558. [ Links ]
 M. Schmidt, A. Masson, C. Brechignac, Phys. Rev. Lett.,91(2004)243401 [ Links ]
 H.J. Zhai, L.S. Wang J.Chem.Phys.,122(2005) 051101. [ Links ]
 G.H. Guverlioglu, P.P. Ma, X.Y. He, R.C. Forrey, H.S, Phys.Rev.Lett.,94(2005) 026103 [ Links ]
 M. Boyukata, Z.B. Guvenc, S. Ozcelik, P. Durmus, J. Jellinek, Int.J.Mod.Phys.C.,16(2005)295. [ Links ]
 F.Y. Liu, P.B. Armentrout, J.Chem.Phys.,122(2005)194320. [ Links ]
 S. Zhao, Z.H. Li, W.N. Wang, K.N. Fan, J. Chem.Phys., 122(2005)144701. [ Links ]
 H. Gronbeck, A. Rosen, Chem.Phys.Lett.,227(1994)149. [ Links ]
 H. Gronbeck., A. Rosen, Surf.Rev.and Lett., 3(1996)687. [ Links ]
 Y.D. Kim, G. Gantefor, Q. Sun, P. Jena, Chem.Phys.Lett., 396(2004)69. [ Links ]
 P. Fuentealba, L. Padilla-Campos, Int.J.Quant.Chem., 102(2005)498. [ Links ]
 M. Dolg, U. Wedig, H. Stoll, H. Preuss, J.Chem.Phys 86 (1986) 866. [ Links ]
 T.H. Dunning, P. Hay, in Modern Theoretical Chemistry; H.H. Schaefer, Ed.; Plenum: New York, 1976; vol. 3; p.1. [ Links ]
 R. Krishnan, J.S. Binkley, R. Seeger, J.A. People, J.Chem.Phys.,72 (1980)650 [ Links ]
 A.D. Becke, J.Chem.Phys.,98 (1993)5648. [ Links ]
 A.D. Becke, Phys.Rev.A,38(1988)3098. [ Links ]
 B. Miehlich, A. Savin, H. Stoll, H. Preuss, Chem.Phys.lett.157(1989)200. [ Links ]
 C. Lee, W. Yang, R.G. Parr, Phys.Rev.B, 37(1988)785. [ Links ]
 L. Padilla-Campos, P. Fuentealba, Theor.Chem.Acc., 110(2003)414 [ Links ]
 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Replogle, J. A. Pople, Gaussian 98, Revision A.9, Gaussian, Pittsburgh, PA, 1998. [ Links ]
 Y. Xu, M. Mavrikakis, Sur.Sci.,494(2001)131. [ Links ]
 P. Jaque and A. Toro-Labbé, J.Chem.Phys.,117(2002)3208. [ Links ]
 P. Jaque and A. Toro-Labbé, J.Chem.Phys. B.,108 (2004)2568. [ Links ]
 R.G. Parr, W. Yang, Density Functional Theory of Atoms and Molecules; Oxford Press, UK, 1989. [ Links ]
 R. Contreras, P. Fuentealba, M. Galvan, P. Perez, Chem.Phys.Lett.,304(1999)4005. [ Links ]
 P. Fuentealba, P. Perez, R. Contreras,113(2000)2544