Silica Nanoarchitectures with Tailored Pores Based on the Hybrid Three

J. Phys. Chem. B 2006, 110, 15269-15274

15269

Silica Nanoarchitectures with Tailored Pores Based on the Hybrid Three- and Four-Membered Rings Dongju Zhang† and R. Q. Zhang* Center of Super-Diamond and AdVanced Films (COSDAF) and Department of Physics and Materials Science, City UniVersity of Hong Kong, Hong Kong SAR, China ReceiVed: April 4, 2006; In Final Form: June 15, 2006

Inspired by the recent developments in the controlled synthesis of porous materials, we present herein the structural prediction of silica nanoarchitectures by using the three- (3MRs) and four-membered rings (4MRs), which are more frequently found in the nanometer-sized particles than in the bulk form, as building blocks. The proposed models include the active molecular rings, thin nanowires, hollow nanotubes, discrete fullerenelike cages, and porous zeolite-like three-dimensional networks. Their geometrical and electronic structures and properties were studied by performing density functional calculations. These silica nanostructures were proved, using molecular dynamics simulations, to possess intrinsic structural stabilities with highly symmetrical geometries and regular nanochannels. These atomically well-defined clusters, (SiO)n, are chemically more reactive than those proposed earlier and are energetically more favorable for n > 20 in high-level density functional calculations over the corresponding two-membered ring (2MR) chains and rings as well as the pure 3MR networks. The nanoparticles and nanodevices based on them are expected to have potential technological applications that mainly make use of their characteristic geometrical structures (nanosized pores) and novel electronic properties.

Introduction The controlled synthesis of porous materials is an ongoing challenge in the field of materials science. The great interest in porous materials is motivated by their numerous applications in sensing,1 catalysis,2,3 gas storage and separation,4,5 and drug delivery.6,7 After intense development of zeolite-based materials in the early 1990s,8 the desire to obtain size- and shape-selective silica frameworks fueled the demand for their application in the areas of fine chemicals, pharmaceuticals, and nanotechnology.9 Recently, silica nanoclusters10-13 and nanoparticles14,15 have attracted great interest, largely due to their importance in microelectronics, photonics/optics, and catalysis. Experimentally, these nanosized materials, ranging in size from a small number of SiO2 units to a few nanometers in diameter,16,17 have been controllably synthesized via different methods such as laser ablation, plasma discharge, secondary ion mass spectrometry, and flame oxidation. Their electronic and optical properties were found to be very different from those of the surfaces of bulk materials, as shown by experiments on light absorption18,19 and photoluminescence.20-21 A remarkable property of bulk silica is that it can form a number of different polymorphs in which the local atomic coordinations are similar but the global networks formed are different. Such a polymorph phenomenon is also expected to occur in nanosized silica materials. It is generally well-known that silica nanomaterials, as well as silica surfaces and thin films, possess characteristic structures where sufficiently small silica rings (an Si-O-Si-O‚‚‚ ring with n Si atoms is referred to as an n-membered ring), such as two- (2MRs), three- (3MRs), and four-membered-rings (4MRs), frequently occur, quite unlike the situation in bulk, where a SiO4 tetrahedral network mostly * Corresponding author. E-mail: [email protected]. † On leave from School of Chemistry and Chemical Engineering, Shandong University, Jinan, Shandong 250100, China.

consists of six- (6MRs) and eight-membered-rings (8MRs). For example, 2MRs were found to exist in silica-w at high temperature22,23 and on the surfaces of amorphous and crystalline silica at high temperature,24-28 and 3MRs and 4MRs were confirmed to have a higher concentration in a synthetic nanometer-sized amorphous silica (called fumed silica) than in bulk.29 In addition, 3MRs and 4MRs are also believed to exist in vitreous silica, as shown by two sharp defect lines in the Raman spectrum,30,31 which have been assigned to the breathing vibrations of oxygen atoms in 3MRs and 4MRs.32-33 Previous research has indicated that the geometrical structures of silica clusters depend on their sizes.34-36 For the smallest silica clusters with n e 5, linear 2MR chains are the most stable structures. This high stability is related to their small ratios of nonbridging oxygen (NBO) atoms, which destabilize clusters as structural defects. For the slightly larger clusters with n ) 6-12, on the other hand, the ground state isomers mainly involve 3MRs and even 4MRs because of their smaller intrinsic strains, although the larger number of NBO atoms in these rings plays an adverse role in stabilizing clusters.36,37 With increasing cluster size, silica clusters prefer to form fully coordinated structures in various types, including rings, cages, and tubes. These defect-free structures for n g 18 were proven to be energetically very favorable and particularly resistant to collapse or rupture.30-32 They may now be specially tailored in view of the technical feasibility of controllably producing silica clusters in various sizes16,17 and are expected to have potential applications in nanotechnology. In contrast to these discrete, fully coordinated structures, three-dimensionally extendable structures are important types of silica nanostructures. In a recent study, we proposed a structural model of such silica clusters based on pure 3MR units.40,41 As indicated by the density functional theoretical calculations, the 3MR-based clusters are energetically more favorable than the corresponding 2MR chains and rings for n g 840 and, in particular, a closed and extendable cage-

10.1021/jp062103v CCC: $33.50 © 2006 American Chemical Society Published on Web 07/19/2006

15270 J. Phys. Chem. B, Vol. 110, No. 31, 2006

Zhang and Zhang

Figure 1. Geometries of the silica clusters based on hybrid 3MR and 4MR units optimized at the B3LYP/6-31G(d) level. Black sticks depart from Si atoms, and gray sticks from O atoms.

like structure was formed as n ) 36, which could exist in mesoporous silica materials.41 For an additional increase in cluster size, as found in fumed silica,29 relatively larger 4MRs are expected to be an important component of silica networks due to both the intrinsically smaller strains in 4MRs and the greatly reduced NBO atoms with growing cluster size. To better understand the microscopic structures and properties of nanoscaled silica materials, we present here a series of new structural models of silica nanostructures, including the active molecular rings, thin nanowires or nanorods, hollow nanotubes, discrete fullerene-like cages, and porous zeolite-like three-dimensional (3D) networks, derived using 3MR and 4MR as building blocks (“superatoms”), and study their geometrical and electronic structures as well as other related properties by performing density functional theoretical calculations. Models and Computational Details As shown in Figure 1, our new silica nanoclusters are based on uniformly hybrid 3MR and 4MR units, where adjacent 3MR and 4MR are mutually perpendicular. The structure for n ) 6 is the smallest cluster among the desired structures. With increasing cluster size, naturally curved molecular chains are formed, and closed molecular rings occur at n ) 10, 15, and 20, containing two, three, and four pairs of hybrid 3MR and 4MR units, respectively. Further, the clusters can be incorporated into semicage- and cage-like structures when their sizes amount to 28 and 36, respectively. It is particularly interesting that this cage-like structure can be extended into a 3D network, which is expected to be energetically more favorable than the corresponding pure 3MR-based network41 as n is large enough, for the smaller intrinsic strain in 4MRs. Such energetically preponderant structures may occur in some silica nanostructures or new bulk-silica polymorphys. We have performed density functional theoretical calculations (geometrical optimizations and molecular dynamics simulations) for a series of silica clusters designed here, (SiO2)n, with size varying from n ) 6-80 using both SIESTA 1.342 and Gaussian 9843 codes. The GGA/DZP level geometrical optimizations, for all clusters considered here, were first performed using the

Perdew-Burke-Ernzerhof’s generalized gradient approximation (GGA) functional44 with the double-ζ plus polarization orbital (DZP). To obtain more accurate geometries and energetics, further optimizations at the B3LYP/6-31G(d) level for the clusters with n e 36 were carried out at the B3LYP/6-31G(d) level of theory, which has been confirmed to be accurate enough for describing silica systems.35 The so-optimized structures were further subjected to the calculations for harmonic vibrational frequencies at the same level of theory to examine whether the optimized structures are the real minima on the potential energy surfaces. The energetic stability of the silica clusters are evaluated by calculating their monomer binding energies (Eb’s), defined as the difference between the total energy of a cluster and the energy of the corresponding isolated SiO2 components, as given by eq 1,

Eb ) -{E[(SiO2)n] - nE(SiO2)}/n

(1)

where E(SiO2) and E[(SiO2)n] are the energies of an isolated SiO2 component and the cluster (SiO2)n. To test the thermal stabilities of these structures, annealing processes were simulated at several selected temperatures, 500, 1000, 1500, 2000, and up to 3000 K. The simulation time step was chosen to be 1 fs, and the relaxed step was set to 1000. Atomic forces were calculated using the Hellmann-Feymann theorem and Newton’s equations were integrated by means of the Verlet algorithm. Results and Discussion Figure 1 illustrates several minimum energy structures obtained at the B3LYP/6-31G(d) level. The molecular fragment for n ) 6, consisting of a 3MR and a 4MR by sharing a vertical Si atom, is found to have Cs symmetry. It grows into a branch with D2d symmetry as n ) 8, where two 3MRs are distributed along two vertical angles of a 4MR. With further increase of the cluster size, the neutrally curved molecular chains (n ) 11, 13, and 16, which are not shown in Figure 1) and closed molecular ring (n ) 10, 15, and 20) are formed. The latter possesses a smaller ratio of NBO groups and more compact

Silica Nanoarchitectures with Tailored Pores

J. Phys. Chem. B, Vol. 110, No. 31, 2006 15271

Figure 2. Representative nanoarchitecures of silica optimized using SIESTA code at the GGA/DZP level. (a) and (b) are an elongated cage and extended rod, assembled from the rings for n ) 10, (c) is a tube, assembled the ring for n ) 15, and (d) and (e) are a squared cage and a tube, assembled from the ring for n ) 20, respectively.

structures than the former. The 4MRs in the molecular ring with n ) 10 are severely distorted due to the large ring stains, while the 3MRs are nearly in their intrinsic planar geometries. This can be attributed to the larger flexibility of 4MRs compared to the relatively rigid 3MRs. With growing ring size, however, the molecular ring strain is greatly reduced, as seen in the structure for n ) 20, where the geometries of 4MRs are similar to that in an isolated 4MR. In contrast to the fully coordinated rings proposed by Bromley,38,39 these molecular rings have NBO atoms, which are chemically reactive, as revealed by the significant magnitudes of their frontier molecular orbitals on these sites, making the rings capable of growing to form 3D structures of various types, such as straight rods, hollow tubes, extendable and closed cages, and porous networks. Figure 2 shows several representative structures assembled from these reactive molecular rings. Panels (a) and (b) arise from the molecular ring for n ) 10, containing two and four such units, respectively. The former is an elongated fully coordinated molecular cage, while the latter is a straight rod with reactive ends, which is geometrically very similar to the nanowire proposed recently by us.36 Note that 2MRs arise in these two structures as a natural result of the coalescence between the reactive sites (SiO defect groups). Similarly, panel (c) is a hollow tube, containing 4 units of the ring for n ) 15, which can further be extended into porous structures via coalescence between residual SiO groups. Panels (d) and (e) are constructed from the ring for n ) 20. The former is a spherical cage, containing two basic rings and being capped by two additional 4MRs, and the latter is tubular. We are greatly interested in these assembled structures from 3MRs and 4MRs because they may be potential candidates of tailor-made porous materials. Our GGA/DZP calculations (energy minimization and MD simulation) confirmed that all these silica nanoarchitectures were energetically and thermodynamically stable and resistant to collapse into other minima in their configuration spaces. It is noted that silica nanotubes have been successfully synthesized recently and hollow structures of the nanotubes have been clearly shown in scanning electron microscope (SEM) and transmission electron microscope (TEM) images,45 although the sizes of the nanotubes are much larger than what we describe here. These facts make

us confident of the rationality of these designed structures, which are likely to be realized with proper control by various experimental techniques. These new extended silica polymorphys, like zeolites, have regular nanosized channels of different sizes and offer many intriguing properties such as allowing the residence of guest molecules and performing chemical reactions with selected-size molecules; thus, their characteristic pore sizes are expected to offer a number of applications, including acting as catalysts, absorbents, energy storage materials, and membranes. Here, we pay special attention to the hybrid molecular ring for n ) 20, which has a regular square structure, where four 3MRs are located in the vertexes and four 4MRs at the centers of the edges. Extending along the NBO atoms in the 4MRs, the semicage- (n ) 28) and cage-like (n ) 36) structures can be formed. The latter, a spherical molecule, is a highly symmetrical structure (Oh symmetry). Compared with the pure 3MR cage-like structure,33 this 3MR-4MR hybrid one is slightly more compact, as indicated by the longest O-O distances in the two structures in the insets of Figure 6, 17.566 Å for the former and 18.499 Å for the latter. Unlike the 2MR3MR hybrid molecular cages with fully coordinated structures,40 both the pure 3MR cage-like structure and the 3MR-4MR hybrid one possess active NBO defects and so can be further extended to 3D networks. As the simplest example, Figure 3 shows two views of a double cage structure for n ) 60 extended from n ) 36. It can be noted that 6MRs are naturally formed as the cluster grows and forms the closed structures. The unique geometrical structure of this cage-like cluster is particularly interesting: six 4MRs and eight 6MRs are alternately distributed and twelve 3MRs symmetrically extend in the 3D space. Its external surface consists of two types of channels: 4MR channels with a near square cross section about 3.0 Å × 3.2 Å, and 6MR channels similar to those in R-quartz and microporous silicate materials such as zeolites. The cavity diameter of this cage is about 9.5 Å. The unique external channels and internal cavity of this cage-like structure make it a potential microreactor, which allows the adsorption of some molecules of a certain special size and shape but prevents other molecules of a larger size and shape from being adsorbed into the internal framework.

15272 J. Phys. Chem. B, Vol. 110, No. 31, 2006

Zhang and Zhang

Figure 3. Extended double cake structure for n ) 60, optimized using SIESTA code at the GGA/DZP level.

Figure 4. Monomer binding energies (Eb’s) of silica clusters calculated at the B3LYP/6-31G(d) level as a function of the cluster size n for (a) 3MR and 4MR hybrid structures, (b) 2MR chains, (c) 2MR rings, (d) pure 3MR structures, and (e) 2MR and 3MR hybrid structures. Several representative structures (one from each type) of the silica clusters are included as the insets of the figure.

To reveal the thermal stability of this new cage-like structure under high temperature, such as 1000 K, and the potential synthesis conditions of nanometer-sized silica particles, we simulated its annealing process using SIESTA code. We found that this cage-like structure did not collapse throughout the whole run, indicating its high thermal stability. To evaluate the relatively energetic stabilities of these 3MR4MR hybrid silica clusters, we calculated their monomer binding energies (Eb’s), which were defined as the energy differences between the total energy of a cluster and the energies of the corresponding isolated SiO2 monomers. The calculated results at the B3LYP/6-31G(d) level are shown in Figure 4 [curve (a)], where those for other silica models established earlier, including 2MR chains [curve (b)] and rings [curve (c)], 3MR structures [curve (d)], and 2MR-3MR hybrid structures [curve (e)], are also given for comparison. Several representative structures (one from each type) of the silica clusters are included as the insets of Figure 4. The SIESTA calculations at the GGA/DZP level produce nearly the same trend of the relatively energetic stability of these silica clusters with the B3LYP/6-31G(d) calculations. The relevant results are not shown for the reason of simplification. This fact indicates that the GGA/DZP calculations may produce just as good results as the B3LYP/6-31G(d) calculations, with a much smaller cost of CPU time. This comparison also establishes the advantage of the present GGA calculations for describing silica systems, producing reliable results with an acceptable calculation expense. This method will be applied to

larger silica systems in our forthcoming work in view of its good performance for small- and medium-sized silica clusters. From the results shown in Figure 4, we found that the 3MR4MR hybrid clusters for n < 20 are energetically less favorable than the entire cluster models. This can clearly be attributed to their larger number of NBO atoms resulting from the attendance of 4MRs, which act as defect sites of silica clusters and play an unfavorable role in stabilizing the clusters, although the intrinsic strain in 4MRs is expected to be smaller than those in relatively restrained 2MRs and 3MRs. In accordance with early and recent reports,34,36 the 2MR molecular chains for n < 6 were again identified as the ground-state structures of the clusters [curve (b)]. However, the intrinsic internal strain on 2MRs for larger chains plays a crucial role in destabilizing (SiO2)n chains. As shown in Figure 4, the stabilities of some 3MR clusters and 2MR-3MR hybrid clusters for n > 8 exceed those of the corresponding 2MR clusters. With a further increase of the cluster size, we found that the 3MR-4MR hybrid clusters are stabilized much faster than that for the pure 3MR clusters, and eventually for n ) 28, the energetic stability of the former exceeds that of the latter. In particular, we compared the stability of the cage-like structure for n ) 36 with that of the pure 3MR structure with a similar cage-like structure and found the former to be more stable in monomer Eb by 0.032 eV per SiO2 unit than the latter. Obviously, the extra stabilization energy arises from the larger flexibilities and smaller intrinsic strains in 4MRs. These facts indicate that 4MRs may be appropriate building blocks for medium-sized silica clusters, although it is not possible for them to occur in small silica clusters. In the present calculations, the clusters for n ) 10, 15, and 20 have relatively higher stabilities than their neighbors and are energetically competitive structures with silica clusters of other types. Their vastly better stabilities relate to their unique structures: closed molecular rings with more compact geometries and a smaller number of NBO atoms, as seen in Figure 1. The higher stabilities of these ringed structures mean they have increased concentrations in production of nanosized silica and qualify for acting as potential precursors to form now silica polymorphs. According to our earlier calculations,41 the intrinsic strain in 3MR-based clusters mainly arises from NBO atoms rather than from 3MRs themselves. A similar conclusion is expected to apply to the present 3MR-4MR hybrid clusters in view of the larger flexibility of 4MR compared to 3MR. We compared the ratio of NBO atoms to total O atoms in the 3MR-4MR hybrid clusters with that in the pure 3MR clusters. It was noted that the variations of this ratio with increasing n for both types of clusters have almost the same trends as the monomer BE curves shown in Figure 4, verifying our conjecture about the crucial role of NBO atoms in destabilizing clusters. Reducing the ratio of BNO atoms by incorporating them in closed (or fully

Silica Nanoarchitectures with Tailored Pores

J. Phys. Chem. B, Vol. 110, No. 31, 2006 15273

Figure 5. HOMO-LUMO gap vs the cluster size for (a) 3MR-4MR hybrid structures, (b) 2MR chains, (c) 2MR rings, (d) pure 3MR structures, and (e) 2MR-3MR hybrid structures.

coordinated) structures is expected to be an efficient way to further stabilize these 3MR- and/or 4MR-containing clusters. It can be noted that the cluster for n ) 20, which was assembled from the active molecular ring for n ) 10 and labeled by the solid pentacle in Figure 4, shows relatively higher stability, indicating that the appropriate coalescences between the active rings could be an energetically favorable trend for the cluster growth. This result supports our design of silica nanoartwork, as shown in Figures 2 and 3. To reveal the electronic properties of these new silica clusters, we computed their energy gaps between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) at the B3LYP/6-31G(d) level of theory. Although DFT calculations cannot be expected to give accurate band gap energies, it is expected to provide a meaningful trend. The curve (a) in Figure 5 exhibits the variation of the HOMOLUMO gap as a function of the cluster size. Those for other types of silica clusters are given for comparison. We found that the gap characteristics vary with the cluster type. The gaps of 2MR molecular chains rapidly tend toward a constant, about 6.45 eV for n > 16 [curve (b)], while those of 2MR molecular rings [curve (c)] present a monotonic increase with increasing cluster size. For the 2MR-3MR hybrid clusters [curve (e)], the gap fluctuates with cluster size for the clusters with n < 9; however, it rapidly increases as n g 12. This is in good agreement with the trend of Eb shown in Figure 3, resulting from the structural transition from the NBO-containing defected structures for n < 9 to the fully coordinated caged structures for n g 12. Similarly, the overall upward trend of the energy gap with cluster size is also apparent across the whole region for the 3MR clusters [curve (d)] and 3MR-4MR hybrid clusters [curve (a)], although minor fluctuations are observed for smaller clusters. However, their upward ratios are much slower than those of the 2MR molecular rings and also than those of the 2MR-3MR hybrid clusters. The difference in the gap trends among these types of silica clusters may have a significant influence on their optical properties and chemical reactivities. On the other hand, by comparing the relative gap values of these types of silica clusters, we found that 3MR-4MR hybrid clusters have the smallest gap, which may be an indication of their higher chemical reactivities, facilitating their coalescences into the larger-sized clusters. We are particularly interested in the properties of the cagelike cluster for n ) 36 because of its potential capability as a

Figure 6. Calculated IR spectra of the two cage-like clusters of (SiO2)36 for (a) the 3MR-4MR hybrid structure, and (b) the pure 3MR-based structure.

unique microreactor. Special attention has been paid to its electronic structures. We found that the HOMO state has significant magnitude on the NBO atoms, and the LUMO state mainly distributes the Si atoms bonding to NBO atoms. In other words, both the HOMO and LUMO, which are responsible for the chemical stability of a system according to frontier orbital theory, highly localize at the external region of the cage-like structure, with a total absence in its interior. These characteristically electronic properties provide support for two important properties of the cage-like structure: its 3D extendibility and its acting as a nanoscale chemical reactor container. The former is relevant to its high active SiO groups at the surface and the latter is due to the chemical inertion in its interior. The new silica polymorph based on this cage-like structure is expected to be realizable via typical techniques in view of its structural and energetic stability. To provide distinctive spectroscopic fingerprints of the structure, we computed its IR vibrational spectrum, as shown in Figure 6a. Clearly, the spectroscopic signature appears at 1072, 1132, and 1198 cm-1. The former two modes correspond to the symmetrical and asymmetrical stretch vibrations of the Si-O bonds in 3MRs, which increase by 5 and 28 cm-1 compared with the corresponding ones in the pure 3MR-based cage-like structure (Figure 6b), respectively. The latter mode was assigned to the symmetrical stretch vibrations of Si-O bonds in 4MRs. These respective spectroscopic fingerprints would guide their experimental detection by infrared matrix isolation spectroscopy.

15274 J. Phys. Chem. B, Vol. 110, No. 31, 2006 Conclusions By using the 3MRs and 4MRs as building blocks, we have presented a series of new structural models of silica nanoclusters and studied their geometrical and electronic structures and properties. High-level DFT calculations show the desired structures to be reactively more active than those proposed earlier and energetically more stable than the corresponding 2MR chains and rings and the pure 3MR-based networks for n > 20. Molecular dynamics simulations prove that these atomically well-defined clusters possess high structural stabilities with beautiful geometries and regular nanochannels. They should be considered as potential candidates for new nanodevices with desired structures and properties in view of their high structural and energetic stabilities and relatively high chemical activities. Acknowledgment. The work described in this paper is supported by a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (project no. CityU 103305). References and Notes (1) Radu, D. R.; Lai, C. Y.; Wiench, J. W.; Pruski, M.; Lin, V. S. Y. J. Am. Chem. Soc. 2004, 126, 1640. (2) Ferey, G.; Cheetham, A. K. Science 1999, 283, 1125. (3) Brunel, D.; Blanc, A. C.; Galarneau, A.; Fajula, F. Catal. Today 2002, 73, 139. (4) Ferey, G. Science 2000, 289, 994. (5) Ferey, G.; Serre, C.; Mellot-Draznieks, C.; Millange, F.; Surble, S.; Dutour, J.; Margiolaki, I. Angew. Chem., Int. Ed. 2004, 43, 6296. (6) Mal, N. K.; Fujiwara, M.; Tanaka, Y.; Taguchi, T.; Matsukata, M. Chem. Mater. 2003, 15, 3385. (7) Trewyn, B. G.; Whitman, C. M.; Lin, V. S. Y. Nano Lett. 2004, 4, 2139. (8) Cheetham, A. K.; Ferey, G.; Loiseau, T. Angew. Chem., Int. Ed. 1999, 111, 3466. (9) Zwijnenburg, M. A.; Bromley, S. T.; Jansen, J. C.; Maschmeyer, T. Chem. Mater. 2004, 16, 12. (10) Lafargue, P. E.; Gaumet, J. J.; Muller, J. F.; Labrosse, A. J. Mass Spectrom. 1996, 31, 623. (11) Wang, L. S.; Nicholas, J. B.; Dupuis, M.; Wu, H.; Colson, S. D. Phys. ReV. Lett. 1997, 78, 4450. (12) Xu, C.; Long, Y.; Zhang, R.; Zhao, L.; Qian, S.; Li, Y. Appl. Phys. A 1998, 66, 99. (13) Schenkel, T.; Schlatholter, T.; Newman, M. W.; Machicoane, G. A.; McDonald, J. W.; Hamza, A. V. J. Chem. Phys. 2000, 113, 2419. (14) Balabanova, E. Vacuum 2000, 58, 174. (15) Whyman, D. Phys. Chem. Chem. Phys. 2001, 3, 1348. (16) Wang, L. S.; Desai, S. R.; Nicholas, J. B. Z. Physica D 1997, 40, 36. (17) Beaucage, G.; Kammler, H. K.; Mueller, R.; Strobel, R.; Agashe, N.; Pratsinis, S. E.; Narayanan, T. Nature Mater. 2004, 3, 370.

Zhang and Zhang (18) Altman, I. S.; Lee, D.; Chung, J. D.; Song, J.; Choi, M. Phys. ReV. B 2001, 63, 161402(R). (19) Suzuki, T.; Skuja, L.; Kajihara, K.; Hirano, M.; Kamiya, T.; Hosono, H. Phys. ReV. Lett. 2003, 90, 186404. (20) Glinka, Y. D.; Lin, S. H.; Chen, Y. T. Appl. Phys. Lett. 1999, 75, 778. (21) Glinka, Y. D.; Lin, S. H.; Chen, Y. T. Phys. ReV. B 2000, 62, 4733. (22) Weiss, A. Z. Anorg. Allg. Chem. 1954, 276, 95. (23) Hamann, D. R. Phys. ReV. B 1997, 55, 14784. (24) Chaing, C. M.; Zegarksi, B. R.; Dubois, L. H. J. Phys. Chem. 1993, 97, 6948. (25) Ferrari, A. M.; Garrone, E.; Spoto, G.; Ugliengo, P.; Zecchina, A. Surf. Sci. 1995, 323, 151. (26) Bakaev, V. A.; Steele, W. A. J. Chem. Phys. 1999, 111, 9803. (27) Ceresoli, D.; Bernasconi, M.; Iarlori, S.; Parrinello, M.; Tosatti, E. Phys. ReV. Lett. 2000, 84, 3887. (28) Roder, A.; Kob, W.; Binder, K. J. Chem. Phys. 2001, 114, 7602. (29) Uchino, T.; Aboshi, A.; Kohara, S.; Ohishi, Y.; Sakashita, M.; Aoki, K. Phys. ReV. B 2004, 69, 155409. (30) Stolen, R. H.; Krause, J. T.; Kurkjian, C. R. Discuss. Faraday Soc. 1970, 50, 103. (31) Galeener, F. L.; Lucovsky, G. Phys. ReV. Lett. 1976, 37, 1474. (32) Galeener, F. L. Solid State Commun. 1982, 44, 1037. (33) Pasquarello, A.; Car, R. Phys. ReV. Lett. 1998, 80, 5145. (34) Nayak, S. K.; Rao, B. K.; Khanna, S. N.; Jena, P. J. Chem. Phys. 1998, 109, 1245. (35) Chu, T. S.; Zhang, R. Q.; Cheung, H. F. J. Phys. Chem. B 2001, 105, 1705. (36) Zhang, D. J.; Zhang, R. Q. To be published. (37) Flikkema, E.; Bromley, S. T. J. Phys. Chem. B 2004, 108, 9638. (38) Bromley, S. T.; Zwijnenburg, M. A.; Maschmeyer, Th. Phys. ReV. Lett. 2003, 90, 035502. (39) Bromley, S. T. Nano Lett. 2004, 4, 1427. (40) Zhang, D. J.; Zhao, M. W.; Zhang, R. Q. J. Phys. Chem. B 2004, 108, 18451. (41) Zhao, M. W.; Zhang, R. Q.; Lee, S. T. Phys. ReV. B 2004, 69, 153403. (42) Soler, J. M.; Artacho, E.; Gale, J. D.; Garcı´a, A.; Junquera, J.; Ordejo´n, P.; Sa´nchez-Portal, D. J. Phys.: Condens. Matter 2002, 14, 2745, and references therein. (43) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.7; Gaussian, Inc.: Pittsburgh, PA, 1998. (44) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. ReV. Lett. 1996, 77, 3865. (45) Zhang, Y. H.; Lu, S. G.; Li, Y. Q.; Dang, Z. M.; Xin, J. H.; Fu, S. Y.; Li, G. T.; Guo, R. R.; Li, L. F. AdV. Mater. 2005, 17, 1056.