Effect of Aromatic Amine−Metal Interaction on Surface Vibrational

ARTICLE pubs.acs.org/JPCC

Effect of Aromatic Amine-Metal Interaction on Surface Vibrational Raman Spectroscopy of Adsorbed Molecules Investigated by Density Functional Theory Liu-Bin Zhao, Rong Huang, Mu-Xing Bai, De-Yin Wu,* and Zhong-Qun Tian State Key Laboratory of Physical Chemistry of Solid Surfaces and College of Chemistry and Chemical Engineering, Xiamen University, Xiamen 361005, Fujian, China

bS Supporting Information ABSTRACT: The adsorption mechanism of aromatic amines on noble metal surfaces and the amine-metal interaction effects on Raman spectroscopy are studied by density functional theory. We find that donation of nitrogen lone-paired electrons plays an important role in the amine-metal interaction. The interaction will be enhanced when aniline adsorbs on positively charged metal clusters because of the electrostatic attraction. It is observed that the Raman signal of the NH2 group is sensitive to the amine-metal interaction. A significant frequency shift and abnormal giant Raman enhancement of the NH2 wagging (ωNH2) mode are predicted theoretically when aniline and its derivatives attach to noble metal surfaces with the amino group. The origin of the frequency shift and the enhanced Raman intensity as well as the line shape broadening has been discussed in detail. Using the characteristic Raman signals of ωNH2, we can infer the adsorption configuration and estimate the strength of the amine-metal interaction. Our calculated results propose that the phenomenon can be extended to other surface probe molecules containing the amino group for investigating amine-metal interactions on SERS active metal substrates.

’ INTRODUCTION Amino groups widely exist in organic molecules and biomolecules such as nucleic acid bases and amino acids. Recent studies show that an amino group can bind to metal surfaces and metal nanoparticles with strong adsorption interaction.1-4 The identification and characterization of this amine-metal interaction play an important role in studies of the electrochemical polymerization,5-7 surface catalytic oxidation of aromatic amines,8-10 and molecular electronic devices.11-13 Thereby, the adsorption of small aromatic amine molecules on metal surfaces is studied to understand complex surface processes mentioned above. Surface vibrational spectroscopy techniques, including surfaceenhanced Raman spectroscopy (SERS)14 and surface-enhanced infrared spectroscopy,15 provide information at the molecular level for studying interfacial processes due to their high detection sensitivity. Especially, the SERS technique can detect the low-frequency peaks related to the molecule-metal interaction without the interference from the solvent molecules, such as water. Since the SERS of aniline (AN) was initially reported by Jeanmaire and van Duyne,16 the SERS of aniline on silver and gold electrodes had been investigated continuously by many groups.17-23 Holze investigated the SERS of aniline on silver18 and gold19 electrodes. However, there existed a very strong fluorescent background in his Raman spectra. Tian et al. recorded high-quality SERS of aniline on rough silver and gold electrodes.21 They assigned the low-frequency bands at 306 and 326 cm-1 to the N-Ag and N-Au stretching vibrations, respectively. This indicated that aniline adsorbed on metal electrodes through its r 2011 American Chemical Society

amino nitrogen atom. It is worth mentioning that the Raman peak at around 1000 cm-1 was significantly broaden and enhanced in comparison with normal Raman spectrum of liquid aniline.21,24 In the SERS obtained by Shindo et al.17 and Ishioka et al.,22 the broad peaks were also observed at ∼850 and 1080 cm-1 when aniline adsorbed on silver and gold surfaces, respectively. Although assignment of the observed bands has been reported in the literature,24-26 no satisfactory explanations were given for these abnormal Raman bands. In summary, all the reported SERS of aniline are extremely sensitive to the experimental conditions.17,21 Similar phenomena have been observed in SERS measurements of amino-group-containing organic molecules adsorbed on silver and gold surfaces. Park et al. studied the SERS of p-aminobenzoic acid (PABA)27 and p-aminobenzonitrile (PABN)28 on silver surfaces. They considered that these two molecules were adsorbed on silver surfaces via both the amino and carboxyl/nitrile groups. They also observed new broad bands, appearing at 980 and 920 cm-1 for PABA and PABN adsorbed on silver surfaces. This band was assigned to the NH2 rocking mode based on the D2O deuterium substituted measurement, indicating that the amino group approached the silver surface. In this paper, aniline and its derivatives are chosen at first as model molecules to investigate the amine-metal interaction. Then we further investigate theoretically the influence of the amine-metal Received: December 9, 2010 Revised: January 27, 2011 Published: February 23, 2011 4174

dx.doi.org/10.1021/jp1117135 | J. Phys. Chem. C 2011, 115, 4174–4183

The Journal of Physical Chemistry C

ARTICLE

Figure 1. Schematic view of aniline adsorbed on metal surfaces for an amine-metal bond formed through charge transfer from the molecule to a metal cluster.

interaction on the Raman feature of adsorbed aromatic amines on coinage metal surfaces. The most obvious change of the surface Raman spectra pattern arises from the amino group. The interesting fact is that a significant frequency shift and an abnormal giant Raman enhancement can be observed for the NH2 wagging (ωNH2) vibration. It is found that the Raman signal of wagging mode is very sensitive to the adsorption interaction. Therefore this mode can be used to deduce the adsorption configurations of aromatic amines and the intermediates of surface reactions. Finally, this is extended to other molecules containing the amino group. Our results propose that the amino wagging vibration can be used as a probe to investigate the amine-metal interaction.

’ COMPUTATIONAL DETAILS To mimic the surface adsorption of the amino group, we adopt a molecule-metal cluster complex model. This model is proper not only to describe molecular adsorption at active sites on rough metal surfaces but also to well predict vibrational spectra. The amine-metal interaction mainly arises from donation of the amino lone pair orbital to metal unoccupied s band (see Figure 1). Here, metal clusters are used to simulate the active sites on roughed gold, silver, and copper substrates. Since the amine-metal interaction depends on metal clusters selected, we choose a series of clusters with different sizes at n = 2, 3, 4, 6, and 13. Density functional calculations were carried out with the hybrid exchange-correlation functionals approach B3LYP.29,30 The basis sets for C, N, O, and H atoms of the investigated molecules were 6-311þG(d, p), which included the polarization function to all four kinds of atoms and diffuse functions to C, N, and O atoms.31,32 For all metal atoms, the valence electrons and electrons in the inner shells were described by the basis set, LANL2DZ, and the corresponding relativistic effective core potentials, respectively.33,34 The theoretical methods were used well to calculate the binding interaction, vibrational frequency, and Raman intensity of pyridine interacting with coinage metal clusters.35-37 Natural bond orbital (NBO) analysis38,39 has been used to investigate the bonding mechanism of aniline on the metal clusters. Full

geometry optimizations and analytic frequency calculations were carried out by using the Gaussian 03 package.40 The scaled quantum mechanics force field (SQMF) procedure41 was used to analyze all the fundamental vibrational bands. The vibrational frequencies are assigned according to the potential energy distribution (PED). During running normalmode analysis, we chose the scaling factors of 0.915 and 0.935 for N-H and C-H bonds, as well as 0.963 for the other internal coordinators related to the force constant matrix calculated at the B3LYP/6-311þG(d,p) level. These scaling factors were also used to correct the incomplete property of theoretical approaches and basis sets, as well as the neglect of anharmonicity for aniline-metal complexes. When certain vibrational modes involve metal atoms, its scaling factor was adopted as 1.0. Thus, they are also proper in scaling the corresponding vibrational frequencies of adsorbed PABA, PABN, and PATP systems calculated at the same theoretical level. Binding interaction energies are obtained by subtracting the energies of the substrate and molecule (in fully optimized geometries) from the total energy BE ¼ - ðEAN-Mn - EAN - EMn Þ where n denotes the number of the metal atoms in the metallic cluster and EAN-Mn and EAN (EMn) denote the energies of the complex and free aniline (metallic cluster), respectively. For retaining the full basis set associated with the entire system, the counterpoise correction of the basis set superposition error (BSSE) was considered in the binding energies in the present calculations.43,44 The correction decreases the overestimation of binding energies. Absolute Raman intensities are calculated on top of the differential Raman scattering cross section (DRSC), as published in our previous work.9 In order to make direct comparison with the SERS experiments, the simulated Raman spectra were presented in terms of the Lorentzian expansion of the DRSC magnitudes from the Raman scattering factors (RSF) under the double-harmonic approximation. 4175

dx.doi.org/10.1021/jp1117135 |J. Phys. Chem. C 2011, 115, 4174–4183

The Journal of Physical Chemistry C

’ RESULTS AND DISCUSSION Geometry and Binding Interaction. It is worth noting that in free aniline the lone-paired orbital significantly conjugates with the π orbital of the benzene ring. The theoretical25,44-46 and experimental47 studies indicate that aniline in its ground electronic state is nonplanar. The optimized parameters of aniline are consistent with the gas-phase microwave studies.47 The predicted C-N distance is about 1.398 Å (expt, 1.402 ( 0.002 Å45). The inversion angle (dihedral angle between the amino plane and the C6H5N plane) and the scissoring angle of the NH2 group are 39.09° (expt, 37.50° ( 2°47) and 112.17° (expt, 113.01° ( 2°47) calculated at the B3LYP/6-311þG(d,p) level, respectively. For aniline adsorbed on a metal surface, the nitrogen atom of its amino group binds to the metal surface. The chemisorbed aniline with the amino group bound to the surfaces shows a tilted adsorption configuration.1-3 To understand the bond mechanism of the amino group in aniline with coinage metal surfaces, we first check the frontier molecular orbital of aniline and metal clusters. The highest occupied molecular orbital (HOMO) of aniline is the p-type nitrogen lonepaired orbital which is strongly conjugated with the benzene ring π orbital while the lowest unoccupied molecular orbital (LUMO) of M4 belongs to an s-type orbital with the electron cloud mainly populated at the obtuse angle side of rhombus as shown in Figure 1. The HOMO of aniline matches well in symmetry and energy with the LUMO of M4. This leads to the formation of the N-M bond due to the partial charge transfer from aniline to metal clusters. The energy gap between the electron donor (HOMO of AN) and electron acceptor (LUMO of M4) is 2.69 and 1.55 eV for Ag4 and Au4, respectively. The binding interaction causes the nitrogen LP level lowering from -5.77 to -6.80 and -7.18 eV, while the LUMO energy of M4 rises from -3.07 to -2.18 eV in AN-Ag4 and from 4.22 to -2.78 eV in AN-Au4. NBO analysis shows that the natural charge redistribution happens, i.e., 0.0589 and 0.1384 au charge transferring from aniline moiety to metal moiety in AN-Ag4 and AN-Au4 complexes, respectively. The charge transfer direction of the adsorbate-substrate system was supported by a fact that a large dipole moment of 4.3 D was observed for aniline adsorbed on silver surfaces by Dai and co-workers.1 They concluded that the dipole moment arises from a weak charge transfer from aniline to silver in forming the N-Ag adsorption bond. Further analysis shows that a rehybridization of the nitrogen atom4 can be observed from sp2 to sp3 after adsorption. The geometry structure change in aniline after adsorption can be explained from the change of the p-π conjugation between the amino group and the benzene ring. The charge transfer decreases the charge distribution at the LP orbital of the NH2 group and weakens the p-π interaction. NBO analysis shows that the charge distribution at the LP orbital of the NH2 decrease from 1.854 au in free aniline to 1.851, 1.845, and 1.781 au in AN-Cu4, AN-Ag4, and AN-Au4, respectively. Meanwhile, the second-order perturbation energy of LP (N) f π*(C-C) decreases form 25.88 kcal/mol in free aniline to 14.12, 15.68, and 13.42 kcal/mol in AN-Cu4, AN-Ag4, and AN-Au4, respectively. Finally, it was noted that the rehybridization of amino group (from sp2.57 in free aniline to sp2.78, sp2.73, and sp2.83 in AN-Cu4, AN-Ag4, and AN-Au4, respectively) leads to a significant increase in the C-N bond distance. The change of structural parameters in AN depends on the metal clusters used. The variation trend can be summarized as Au > Cu > Ag. As presented in Table 1, the extent of structural changes in aniline is closely associated with the binding interaction between aniline and metal clusters. This interaction is quite similar to that of pyridine interacting with these metal clusters.36,37,48,49 Comparison of binding

ARTICLE

Table 1. Selected Parameters of Optimized Geometries of Aniline (AN) and Its Metallic Complexes (AN-Mn) and Binding Energies (BE) of AN-Mn H-N-H

BEa

(kcal/mol)

(kcal/mol)

C-N (Å)

(deg)

R (deg)

AN

1.398

112.17

39.09

AN-Cu2

1.437

107.72

51.67

17.02

12.49

AN-Cu4

1.440

108.10

51.85

22.39

16.78

AN-Cu6 AN-Ag2

1.436 1.428

107.86 108.52

51.21 45.59

14.41 8.71

9.80 7.29

AN-Ag4

1.433

108.47

50.53

12.84

11.28

AN-Ag6

1.427

108.62

49.19

7.06

5.69

AN-Ag13

1.429

108.42

49.85

8.09

6.47

AN-Au2

1.442

107.73

52.11

20.88

17.99

AN-Au4

1.445

107.83

52.44

22.09

19.18

AN-Au6

1.437

108.13

50.99

13.54

10.85

AN-Au13

1.440

108.27

51.30

14.00

11.09

species

a

BE

Considering BSSE correction.

energies in AN-Mn complexes indicates that the bonding strength is stronger in AN-Cun and AN-Aun than AN-Agn. This can be explained that the energy gaps between s and d bands are smaller in gold and copper (about 2.0 eV) than that in silver (about 3.8 eV).48 As a consequence, the d orbitals in gold and copper positively participate in chemical bonding with the NH2 group. The adsorption energies are predicted to be 19.18, 11.28, and 16.78 kcal/mol after considering BSSE correction for AN-Au4, AN-Ag4, and AN-Cu4, respectively. Thus, the ability of an unoccupied orbital on metal clusters to accept the LP electrons weakens as the following trend of Au4 > Cu4 > Ag4.37 This is in agreement with the geometric structure changes of C-N bond length and H-N-H angle. This also leads to the vibrational frequency shift of the ωNH2 vibration and C-N stretching mode of aniline adsorbed on the three metals. For the same metal clusters in different sizes, the binding energy changes in a descending trend: M4 > M2 > M13 > M6. These metal clusters may reflect diverse active sites at metal surfaces. The variation of binding energies can be interpreted by the matching extent of energy levels and symmetry between the interaction orbitals of aniline and metal clusters. Compared with the LUMO orbital of Ag4, the Ag6 cluster is slightly higher, which means that the Ag4 matches well with the HOMO of aniline in energy. Indeed, this is supported by the calculated binding energies. For example, the binding energy in AN-Ag4 (12.84 kcal/mol) is also the most close to the experimental value of 17.70 kcal/mol derived from TPD measurement.1 In the present calculation, the binding energy of AN-Ag4 is the largest when aniline binds to silver clusters, in accordance with previous studies on the pyridine system.37 So the tetramer metallic cluster can describe well the adsorption on metal surfaces. Simulated Raman Spectra of AN-Mn. To understand the change of vibrational Raman spectral changes of aniline due to adsorption interactions, we first calculate the normal Raman spectrum of free aniline. As seen in Figure 2, the calculated Raman spectrum of free aniline shows excellent accordance with the liquid Raman spectrum.21,24 Strong Raman bands predicted at 815, 990, 1027, 1272, 1613, and 1632 cm-1 can be attributed to the ring breathing vibration (v1), the ring triangle deformation (v12), the mixed vibration from C-C stretching and C-H in-plane bending (v18a), the C-N stretching, the C-C stretching vibration (v8a), and the NH2 scissoring vibration, respectively, based on the PED from normal-mode 4176

dx.doi.org/10.1021/jp1117135 |J. Phys. Chem. C 2011, 115, 4174–4183

The Journal of Physical Chemistry C

ARTICLE

Table 2. Vibrational Frequencies (ω, cm-1) and Raman Activities (SR, Å4/amu) of NH2 Related Modes in Aniline and Its Metallic Complexes AN

AN-Cu4

AN-Ag4

AN-Au4

ω

283

543

494

600

SR

0.26

6.53

2.51

0.01

ω

564

982

934

1029

SR

6.76

45.34

101.91

233.40

rocking

ω SR

1043 0.04

1133 2.43

1131 1.29

1137 0.29

scissoring

ω

1632

1624

1627

1620

SR

25.64

8.95

8.89

6.06

twisting wagging

Figure 2. Calculated Raman spectra of aniline interacting with coinage metal clusters. An incident wavelength of 632.8 nm and a line width of 10 cm-1 were used in the simulated Raman spectra. Peaks labeled with an asterisk belong to NH2 wagging mode.

Figure 4. Dependence of amino wagging frequencies on N-M force constants in AN-Ag4 and AN-Au4 complexes on the basis of normalmode analysis.

Figure 3. Characterization of the NH2 vibrations in free aniline.

analysis. The assignment is in accordance with previous assignment by Wojciechowski et al.25 Figure 3 presents four vibrational modes (twisting, wagging, rocking, and scissoring) involved in the NH2 group. It is noted that the Raman intensity of the wagging vibration at 564 cm-1 is relatively weak in free aniline. This agrees our predicted RSF value about 6.76 Å4/amu, corresponding to the DRSC value about 0.95  10-30 cm2 sr-1 mol-1 at the excitation wavelength 632.8 nm. The NH2 twisting vibration at 283 cm-1 and the rocking one at 1043 cm-1 have even smaller RSFs about 0.26 and 0.04 Å4/amu, corresponding to the DRSC values about 0.98  10-31 and 0.26  10-32 cm2 sr-1 mol-1, respectively. Their Raman signals can hardly be detected in the normal Raman spectrum. The adsorption interaction results in a significant change in Raman bands of the NH2 group. After adsorption, the most distinct change is that a very strong band appears at about 934 cm-1 on silver and 1029 cm-1 on gold in our simulated Raman spectra (Figure 2). On the basis of our normal-mode analysis, this band can be attributed to the ωNH2 mode from the PED value (see Supporting Information).

Surprisingly, the band remarkably changes in its vibrational frequency, which blue shifts from 564 cm-1 to 900-1000 cm-1, simultaneously accompanying with a significant increase in the DRSC values from 0.40  10-30 to ∼6.36  10-30 cm2 sr-1 mol-1 (RSFs change from 6.76 to ∼233.40 Å4/amu) for free aniline and AN-M4, respectively. As listed in Table 2, the NH2 twisting mode blue shifts to 543, 494, and 600 cm-1 by binding to Cu, Ag, and Au clusters. However, the Raman intensities of the twisting mode in both free aniline and its adsorption states are quite weak. Although the vibrational frequency of the NH2 rocking vibration blue shifts to higher wavenumber by about 90 cm-1 at adsorption states, its Raman intensity is still weak. Furthermore, it is very interesting that a set of double peaks (v8a and NH2 scissoring) around 1600 cm-1 in normal Raman spectrum of free aniline change to a single band in AN-M4. Our theoretical calculation shows that the v8a mode increases in the Raman activity while the NH2 scissoring decreases obviously. The frequency shift of the ωNH2 mode strongly depends on the binding interaction between the amino group and metals. Here the binding interaction is stronger for gold and copper than for silver. The ωNH2 frequency in AN-M4 declines as the following trend: Au > Cu > Ag (see Table 2). As shown in Figure 4, the fundamental frequency of the amino wagging mode has a nearly linear relationship on the changes of the force constants of the N-M bond. Accordingly, the frequency shift of the ωNH2 vibration can be used to correlate directly to the strength of the interaction of aniline with metal surfaces. Experimentally, the abnormal enhancement of the ωNH2 mode in SERS of aniline was not realized. For aniline adsorbed on silver17 and gold,22 the strongest SERS band was observed at ∼1000 cm-1, which 4177

dx.doi.org/10.1021/jp1117135 |J. Phys. Chem. C 2011, 115, 4174–4183

The Journal of Physical Chemistry C

Figure 5. Simulated Raman spectra of aniline adsorbed on silver electrode surfaces above the potential of zero charge (PZC) with models of aniline interacted with cationic Ag clusters (Ag4þ, Ag7þ, and Ag13þ). An incident wavelength of 632.8 nm and a line width of 10 cm-1 were used in the simulated Raman spectra.

can be attributed to the v12 vibration.21 Although some preternatural bumps appeared at around ∼950 and 1080 cm-1 in the SERS of aniline on silver17 and gold,22 respectively, the abnormally enhanced SERS bands were overlooked. Now we assign it to the wagging vibration of the NH2 group on the basis of our present theoretical calculations. Figure 5 presents the simulated Raman spectra of aniline linking to positively charged silver clusters, in analogy to the case that aniline adsorbs on a silver electrode with applied potentials positive to the potential of zero charge (PZC). Compared with aniline interacting with neutral Ag clusters, the ωNH2 vibrational frequency further blue shifts to about 1000 cm-1 in the cases of cationic Ag clusters. The band is responsible for a wide band near 993 cm-1 at the potential of -0.3 V (vs SCE), positive to the PZC value about -0.9 V on a silver electrode.21 It is noted that as the size of the metallic cluster increases, the simulated Raman spectra are closer to SERS spectra of aniline on silver electrodes. This can be understood by the overestimating the electrostatic interaction between aniline and positively charged metal clusters with small size. At the same time the N-Ag and the C-N stretching frequencies exhibit a blue shift and a red shift, respectively compared with that in neutral Aganiline complexes. This indicates that these vibrational frequencies are strongly sensitive to the amine-metal interaction. The electrostatic interaction will strengthen the N-metal bond for aniline adsorbed on a positively charged metal surface.46 The binding energies of the AN-Ag4þ and AN-Ag13þ clusters are 24.34 and 13.11 kcal/mol, respectively, which are significantly larger than that of the corresponding neutral clusters (see Table 1). A recent study on the intermediates of benzyl chloride reduction at silver electrode by SERS provided further proof for the line shape broadening of the amino wagging vibrational band in SERS.50 Both in situ SERS experiments and DFT simulation showed that surface benzyl species adsorbed on silver exhibited strong and broad Raman peaks around ∼800 cm-1. The SERS band was attributed to some vibrations related to the CH2 wagging vibration. There the wagging vibration formed strong coupling with the benzene ring vibrations. The binding interaction also influences the vibrational modes related to the C-N bond and the benzene ring moiety. The C-N stretching frequency red shifts significantly from 1272 cm-1 in free aniline to 1220, 1227, and 1212 cm-1 for Cu, Ag, and Au clusters, respectively. This is in conformity with the change of binding energies

ARTICLE

as Ag < Cu < Au. In a previous study, such a red shift from 1277 to 1245 cm-1 was invoked by Holtz to infer the aniline adsorption orientation on Ag electrodes via its NH2 group.18 Gao et al. also observed the red shift of the C-N stretching frequency from 1280 cm-1 in bulk aniline to 1230 cm-1 in the adsorbed state on Au electrodes.20 Although the 32 and 50 cm-1 frequency shifts on Ag and Au electrodes are slightly smaller than our calculated values, the red-shift extent of the C-N stretching further approves that the NAu bond is stronger than the N-Ag bond. The C-N stretching frequency red shifts to 1236 cm-1 detected by a high-resolution electron energy loss spectroscopy study of aniline on Cu(110) by Plank et al.3 For the vibrational modes in the benzene ring moiety, the vibrational frequencies do not affect much due to the adsorption interaction, in agreement with slight frequency shifts in the present calculations. However, the relative Raman intensities of some vibrational modes change markedly after adsorptions. For example, the v1 mode, which has a medium Raman intensity in free aniline, evidently decreases its intensity in AN-Ag4 and AN-Au4. The theoretical calculation result agrees well with the SERS experiments.17,21,23 Finally, we pay our attention to the N-M bond. The vibrational frequency depends on the strength of the N-M bond. Our calculations predicted that the N-M stretching frequencies for the three substrates are 361 (Cu), 296 (Ag), and 356 cm-1 (Au), respectively. It is worth noting that the N-Ag stretching frequency is very close to the observed value of 296 cm-1 by Tian et al.21 when aniline adsorbed on a rough silver electrode at the potential of -0.6 V. From above results, we note that the amino group of aniline binding to metal surfaces gives a characteristic Raman signal that could be used to estimate the adsorption configuration and bonding strength of the amine-metal interaction. In the next subsection, we will use the unique Raman feature to analyze the adsorption behavior of aniline derivatives, such as three para-substituted aniline molecules, PABA, PABN, and PATP. They have bifunctional groups and can form the metal-molecule-metal sandwich structure on rough metal surfaces. Para-Substituted Anilines. In this subsection, the effect of different absorbate-substrate interactions on the SERS is studied through three typical double-functional aromatic amines, PABA, PABN, and PATP (p-aminothiophenol). It is well-known that carboxyl,27,51-55 nitrile,28,56,57 and thiol10,58-63 groups can form strong chemical interactions on metal surfaces. Recent studies especially on metal-molecule-metal junction systems11-13 suggested that the amino group also forms chemisorption interaction with gold surfaces. Thus, the identification and characterization of the amino-lead interaction and adsorption structures appear to be of great importance in molecular electronics. Park and co-workers studied experimentally the SERS of PABA27 and PABN28 on silver surfaces. It was found that these two molecules were adsorbed on the silver surfaces via both carboxyl (nitrile) group and the amino group. Generally speaking, the former ones (-COOand -CN groups) can form strong adsorption interaction with silver surfaces. But the latter one can be evidenced directly by the broad bands observed at 980 and 920 cm-1, which were assigned to the amino rocking vibration in previous papers. We assign the bands to the amino wagging vibration based on the PED. Recently, we also calculated the Raman spectrum of PATP adsorbed on silver surface with both single-end and double-end configurations by density functional theory.9 It was found that a new peak that appeared at about 900 cm-1 should be assigned to the amino wagging vibration when the amino end and thiol end coadsorbed on silver cluster. However, this abnormal Raman feature did not attract great attention 4178

dx.doi.org/10.1021/jp1117135 |J. Phys. Chem. C 2011, 115, 4174–4183

The Journal of Physical Chemistry C

ARTICLE

Figure 7. Simulated surface Raman spectra of N,N-dideuterated ANAg4 and AN-Au4. The excitation wavelength and the line width of 10 cm-1 were the same as those given in Figure 2. Peaks labeled with an asterisk belong to NH2 wagging mode.

Figure 8. Potential energy curves along amino wagging vibration in free aniline and AN-Au4.

Figure 6. Simulated surface Raman spectra of PABA (A), PABN (B), and PATP (C) on silver and gold surfaces with both single end and double ends at the B3LYP/6-311þG(d,p)/LAN2DZ level. The excitation wavelength and the line width of 10 cm-1 were the same as Figure 2. The peaks highlighted in blue and red represent the amino wagging mode.

in other theoretical studies on PATP in metal-molecule-metal junction structures.64,65 Figure 6 presents the simulated surface Raman spectra of PABA, PABN, and PATP on silver and gold surfaces with both single-end and double-end cases. It can be seen that the Raman features of three para-substituted anilines in the single-end adsorption are different from their double-end cases. Note that the single-end adsorption takes place mainly through -COO-, -CN, or -SH. By inspecting the simulated Raman spectra of the two adsorption configurations, we

can find that the major changes are from the vibrational bands related to the amino group. These variations can be summarized as the following three points: First, the amino group approaching a metal surface will exhibit unique Raman signals. The Raman signal can be used to ascertain the adsorption configuration and the strength of the binding interaction. For example, the new Raman peaks appearing at around 900 and 1000 cm-1 for PABA, PABN, and PATP may be considered as an indicator for the formation of a double-end configuration on silver and gold surfaces, respectively. These strong Raman bands can be attributed to the ωNH2 vibration on metal surfaces. The assignment is different from the previous suggestion from Park et al., who assigned the broad and strong SERS band to the NH2 rocking vibration.27,28 Our calculated result shows that the NH2 rocking vibration possesses extremely low Raman signals. It was noted that such a new peak has not been observed for the SERS of PATP,10,56 indicating that PATP may adsorb on a noble metal surface mainly via a single-end adsorption. Zhou et al. studied SERS of PATP trapped in a gap site of a gold substrate and silver nanoparticles.58 They observed three strong SERS bands appeared at 1142, 1391, and 1436 cm-1 but no band is associated with the amino group. Our recent studies suggested that the SERS bands arise from the azobenzene-like surface species due to a surface 4179

dx.doi.org/10.1021/jp1117135 |J. Phys. Chem. C 2011, 115, 4174–4183

The Journal of Physical Chemistry C

ARTICLE

catalytic coupling reaction of PATP molecules adsorbed on silver and gold nanoscale structural surfaces.9,10,58 In the earlier study, these new bands were assigned to the nontotally symmetric b2 vibrational modes of PATP.59 The so-called b2 modes should be actually ag modes of azobenzene.9,66 Second, the frequencies of the C-NH2 stretching vibration of PABA, PABN, and PATP at the single-end adsorption on Ag (Au) are predicted at 1280 (1289), 1310 (1314), and 1272 (1281) cm-1. These frequencies are very close to 1273, 1297, and 1279 cm-1 in their free molecules. However, the frequencies significantly red shift to 1229 (1218), 1252 (1234), and 1228 (1218) cm-1 when the amino

Figure 9. Optimized structures of adenine, guanine, cytosine, o-aminopyridine, m-aminopyridine, and p-aminopyridine interacting with metal along the amino group approaching the surfaces.

group binds to silver (gold). Experimentally, the peak at 1279 cm-1 in the ordinary Raman spectrum of PABA red shifts to 1253 cm-1 in the SERS spectrum on silver.27 Such a red shift is invoked to conclude that the amino group of PABA also interacted with a silver surface. Nevertheless, the Raman intensities of the C-NH2 stretching vibration are quite weak in general; the frequency shift of this mode is hardly to be identified in SERS measurements of other aromatic amines such as PABN and PATP. Thus, the C-NH2 stretching vibration does not seem to be a good probe to study the amine-metal interaction. Third, from our simulated spectra we can find that at a singleend adsorption state a set of double peaks appear at around 1600 cm-1 corresponding to the v8a and the NH2 scissoring vibrations. But for the double-end adsorption configuration only one band could be observed. This phenomenon is similar to that of aniline. The above spectral feature can be explained as the amine-metal interaction decreases its Raman intensity of the scissoring mode. The frequency shift of the ωNH2 mode depends on the electronic properties of para-substituted functional groups. In electron-withdrawing (EW) substituted aniline the frequency of the vibration is lower than that of aniline. On the contrary, the frequency increases in electron donating (ED) substituted aniline. This is due to the EW group enhancing the p-π conjugation effect between amino group and benzene ring while the ED group weakens it. The p-π conjugation interaction would further change the hybridization of the amino groups. In the EW(ED) substituted aniline the amino group is inclined to sp2(sp3). The change of hybridization property will lead to the shape change of a potential energy surface along this vibration and finally results in the frequency shift of this mode. The

Table 3. Inversion Angle (degree), C-NH2 Bond Length (Å), Calculated Vibrational Frequency (cm-1), and Raman Activity (Å4/amu) of NH2 Wagging Mode in Adenine (ADE), Guanine (GUA), Cytosine (CYT), ortho-Aminopyridine (OAP), metaAminopyridine (MAP), para-Aminopyridine (PAP) Molecules and Their Metallic Complexes species ADE

inversion angle R/deg

C-N bond length/Å

frequencies/cm-1

Raman activity/Å4/amu

0.10

1.353

47

0.01

ADE-Cu4 ADE-Ag4

47.92 43.59

1.409 1.399

939 841

13.11 48.18

ADE-Au4

49.07

1.416

981

122.35

GUA

38.44

1.379

572

1.62

GUA-Cu4

55.33

1.428

1013

22.50

GUA-Ag4

52.86

1.418

946

46.93

GUA-Au4

56.18

1.435

1046

178.75

CYT

14.85

1.360

199

1.30

CYT-Cu4 CYT-Ag4

49.64 46.15

1.418 1.406

952 863

49.92 23.15

CYT-Au4

51.24

1.427

998

54.94

OAP

34.10

1.383

472

1.00

OAP-Cu4

52.96

1.433

1003

22.24

OAP-Ag4

50.94

1.424

930

69.08

OAP-Au4

54.11

1.441

1040

84.17

MAP

37.90

1.394

558

9.18

MAP-Cu4 MAP-Ag4

51.39 50.15

1.436 1.429

979 915

12.57 40.70

MAP-Au4

52.40

1.441

1027

152.59

PAP

32.17

1.383

440

1.89

PAP-Cu4

50.11

1.429

952

49.92

PAP-Ag4

48.29

1.421

898

65.32

PAP-Au4

51.06

1.435

1004

96.64

4180

dx.doi.org/10.1021/jp1117135 |J. Phys. Chem. C 2011, 115, 4174–4183

The Journal of Physical Chemistry C

ARTICLE

ND2 wagging vibration. In contrast, the Raman intensity of the ND2 rocking vibration in the adsorbed molecules is very weak though its vibrational frequency also has a quite large red shift from 1131 to 860 cm-1 on silver or 1137 to 861 cm-1 on gold. Another vibration strongly affected by deuteration is the amino scissoring (δND2) mode. The frequency of the δND2 vibration shifts to 1147 and 1150 cm-1 in C6H5ND2-Au4 and C6H5ND2-Ag4 as a result of the decrease of reduced masses Abnormal Raman Feature of NH2 Wagging Vibration. From the above observation, the wagging mode displays a large blue-shifted frequency and very intense Raman intensity, indicating that its Raman signal can be used as a “surface probe vibration” to characterize the adsorption behavior of aromatic amine molecules. Nevertheless, the significant frequency shift and giant Raman intensity enhancement as well as line shape broadening of ωNH2 vibration have not been realized in the literature. The origins of these abnormal spectral features are discussed as follows: (1) Frequency shift. Natural bond orbital analysis suggests the hybridization changes of the N atom from sp2.57 in free aniline to sp2.78, sp2.73, and sp2.83 in AN-Cu4, AN-Ag4, and ANAu4, respectively. In contrast to a symmetric double-well potential energy curve (PEC) in free aniline,42,43,67 the PEC along the wagging motion changes to a single potential well as illuminated in Figure 8 when it interacts with Au4. Niu et al.68 and Rauhut and Pulay26 derived the frequency of 541 cm-1 (from the average of the two transition, 1 f 3 and 0 f 2) under the harmonic approximation of the frequency of the wagging mode in free aniline. Our calculated frequency (561 cm-1) is close to that value. After adsorptions on copper, silver, and gold, the scaled frequencies of the mode are predicted at 982, 934, and 1029 cm-1, respectively. The large blue shift in vibrational frequency is due to the increase of curvature of the inversion barriers as shown in Figure 8. (2) Raman intensity enhancement. The large enhancement in the Raman intensity of the wagging vibration can be attributed to the change of the hybridization effect from sp3 to sp2 during periodic vibrations, which will further influence the charge distribution in the molecular moiety and metal moiety in adsorbed systems. When the -NH2 group vibrates along the wagging mode, its hybridization changes periodically between sp3 and sp2. The LP orbital mainly distributes in the amino group in the sp3 structure and strongly conjugates with the benzene ring in the sp2 configuration. As a sequence, the charge transfer from the LP orbital to an unoccupied metallic orbital in AN-Mn influences the charge distribution in aniline and metal by the hybridization effect. (3) Line-shape broadening. In previous experimental studies, a broad band appears at around 850 cm-1 on a silver electrode17 and 1080 cm-1 on a gold electrode22 in SERS of aniline. Similar observations were also reported for PABA27 and PABN.28 In these studies, the broad bands were not fully interpreted. From our calculations, the broad bands are suggested to be attributed to the ωNH2 vibration. We think the broadening of the line width on the wagging Raman band can be illuminated by the vibrational coupling of the mode with other vibrations and the intermolecular dipole-dipole coupling from a large local dipole moment.69 For aniline adsorbed on silver, the dipole moment can increase from 1.59 D70 in free aniline to about 4.3 D1 due to the formation of the adsorbate-substrate complex. The dipole moments in the

Figure 10. Variation of the wagging frequencies along the NH2 inversion angle and C-N bond length for various amino-containing compounds: adenine (black), guanine (red), cytosine (green), oaminopyridine (blue), m-aminopyridine (cyan), and p-aminopyridine (magenta) linking to Cu (square), Ag (circle), and Au (triangle) with the amino group.

ωNH2 frequencies of the investigated three molecules in their double-end configuration are PATP > PABA > PABN. This trend confirms the SERS experimental result: 980 and 920 cm-1 for PABA and PABN, respectively.27,28 Deuterium-Substituted Isotopic Effect. The isotopic effect plays an important role in vibrational spectroscopy for inferring molecular structures and understanding the intermolecular interaction. The isotopic effect on surface Raman spectra of aniline is also studied here. Figure 7 presents the simulated Raman spectra of aniline deuterium substituted species, C6H5ND2-Au4 and C6H5ND2-Ag4. It is noted that the wagging mode displays an obvious frequency shift by isotopic effect. The frequency moves downward to 812 and 697 cm-1 for C6H5ND2-Au4 and C6H5ND2-Ag4, respectively. In C6H5ND2-Ag4, the ωND2 coordinator is coupled with C-H out-of-plane bending (γC-H) and it contributes to several normal modes at 679, 697, and 770 cm-1. Theses modes are marked with an asterisk in Figure 7. The PED of the ωND2 mode has the largest value in the 697 cm-1 mode so that this mode should be assigned to be the eigen vibration of ωND2 in C6H5ND2-Ag4. This is supported by the SERS of PABA and PABN in D2O solution by Park et al.27,28 They observed that a broad and intense band at 980 and 920 cm-1 shifts to 740 and 676 cm-1 in the SERS spectra of PABA and PABN in D2O medium, respectively. They assigned the band to a ND2 rocking mode (see Figure 3). On the basis of our calculations, the band should be attributed to the 4181

dx.doi.org/10.1021/jp1117135 |J. Phys. Chem. C 2011, 115, 4174–4183

The Journal of Physical Chemistry C present calculation are 1.74 and 4.50 D for AN and AN-Ag4, respectively. Another possible reason for the broadening of the ωNH2 mode is that its frequency is greatly sensitive to the amine-metal interaction. According to our calculations, the ωNH2 frequencies are 909, 934, 900, and 914 cm-1 in ANAg2, AN-Ag4, AN-Ag6, and AN-Ag13, respectively. For positively charged surfaces, the frequencies distribute to 1023, 1024, and 1003 cm-1 in AN-Ag4þ, AN-Ag7þ, and AN-Ag13þ, respectively. The predicted result is supported by SERS measurements published by previous studies.17,21,23,27,28 Therefore, the broad and intense Raman band in these previous SERS studies should be attributed to the ωNH2 vibration. Other Amino-Containing Compounds. We finally extend our calculations to other surface probe molecules. Figure 9 presents the optimized structures of some amino group containing compounds such as three DNA bases (adenine, guanine, and cytosine) as well as o-, m-, and p-aminopyridine (OAP, MAP, and PAP) binding to M4 (M = Cu, Ag, and Au) clusters via their amino group. Table 3 lists the Raman shift and Raman activity of ωNH2 modes combining with the C-N bond length and the inversion angle. The vibrational frequency and Raman intensity of ωNH2 mode strongly depend on the binding interaction. The amino group in adenine changes from a nearly planar structure to pyramidalization structure.4 The C-N bond length increases by 0.046-0.061 Å and inversion angle increases by 4050°; meanwhile, its vibrational frequency blue shifts remarkably from 47 cm-1 in free adenine to the region of 800-1000 cm-1 with different metal clusters. Similarly, for other amino-containing compounds, a large blue shift of the ωNH2 frequency can also be predicted. The frequency shift exhibits a near linear relationship to the C-N bond length and the inversion angle (see Figure 10). For different metals, the frequencies display in descending order in Au > Cu > Ag. In addition, the DRSC values show an abnormal enhancement as these probe molecules interact with metal clusters along the amino group (see Table 3). Therefore, our result suggests that this would be a universal propensity when an aromatic amino group approaches a metal surface.

’ CONCLUSIONS We have investigated the adsorption interaction of aniline and related compounds with metal surfaces along the amino group. The theoretical results clearly illustrate that the adsorption results in crucial changes in Raman spectral feature of the amino group. The most interesting point is that there is a significant blue shift in the fundamental frequency and an abnormal Raman intensity enhancement for the wagging vibration of the amino group due to the amine-metal interaction. The frequency shifts by about 500 cm-1 and the Raman intensity is enhanced by about 35-fold with respect to free aniline. The significant frequency shift of the amino wagging vibration can be interpreted by the changing of the wagging mode potential energy curve from a double-well potential to a single-well potential. Simultaneously, the curvature of the inversion barriers increases obviously, while the giant Raman intensity enhancement of this mode is on account of the electron density redistribution between molecule and metal when the hybridization of amino group changes between sp3 and sp2 along the wagging vibration. Finally, the line shape broadening may arise from the dipole-dipole interaction and the vibrational coupling as well as complex surface adsorption sites. The spectroscopic features can be extended to identify adsorption orientations along the specific group of surface

ARTICLE

species and to infer chemical reaction intermediates and mechanisms of molecules adsorbed on metal surfaces.

’ ASSOCIATED CONTENT

bS

Supporting Information. A table of comparison of theoretical harmonic frequencies and Raman scattering activities of amine-metal interactions. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT We are grateful for the financial support of this work by the NSF of China (Nos. 20973143 and 91027009) and the Ministry of Science and Technology of China (973 Program Nos. 2007CB815303 and 2009CB930703). D.Y.W. is grateful for support from the funding project (2010121020) and HPC of Xiamen University. ’ REFERENCES (1) Rockey, T. J.; Yang, M.; Dai, H.-L. Surf. Sci. 2005, 589, 42. (2) Hoft, R. C.; Ford, M. J.; McDonagh, A. M.; Cortie, M. B. J. Phys. Chem. C 2007, 111, 13886. (3) Plank, R. V.; DiNardo, N. J.; Vohs, J. M. Surf. Sci. 1995, 340, L971. (4) Preuss, M.; Schmidt, W. G.; Bechstedt, F. Phys. Rev. Lett. 2005, 94, No. 236102. (5) MacDiarmid, A. G. Angew. Chem., Int. Ed. 2001, 40, 2581. (6) Lee, Y.; Chang, C.; Yau, S.; Fan, L.; Yang, Y.; Yang, L. O.; Itaya, K. J. Am. Chem. Soc. 2009, 131, 6468. (7) Yang, L. Y. O.; Chang, C.; Liu, S.; Wu, C.; Yau, S. L. J. Am. Chem. Soc. 2007, 129, 8076. (8) Grirrane, A.; Corma, A.; Garcia, H. Science 2008, 322, 1661. (9) Wu, D. Y.; Liu, X. M.; Huang, Y. F.; Ren, B.; Xu, X.; Tian, Z. Q. J. Phys. Chem. C 2009, 113, 18212. (10) Huang, Y. F.; Zhu, H. P.; Liu, G. K.; Wu, D. Y.; Ren, B.; Tian, Z. Q. J. Am. Chem. Soc. 2010, 132, 9244. (11) Venkataraman, L.; Klare, J. E.; Nuckolls, C.; Hybertsen, M. S.; Steigerwald, M. L. Nature 2006, 442, 904. (12) Quek, S. Y.; Venkataraman, L.; Choi, H. J.; Louie, S. G.; Hybertsen, M. S.; Neaton, J. B. Nano Lett. 2007, 7, 3477. (13) Tsutsui, M.; Taniguchi, M.; Kawai, T. J. Am. Chem. Soc. 2009, 131, 10552. (14) Wu, D. Y.; Li, J. F.; Ren, B.; Tian, Z. Q. Chem. Soc. Rev. 2008, 37, 1025. (15) Osawa, M. Bull. Chem. Soc. Jpn. 1997, 70, 2861. (16) Jeanmaire, D. L.; Van Duyne, R. P. J. Electroanal. Chem. 1977, 84, 1. (17) Shindo, H. J. Chem. Soc., Faraday Trans. 1 1986, 82, 45. (18) Holze, R. Electrochim. Acta 1987, 32, 1527. (19) Holze, R. J. Electroanal. Chem. 1988, 250, 143. (20) Gao, P.; Gosztola, D.; Weaver, M. J. J. Phys. Chem. 1988, 92, 7122. (21) Tian, Z. Q.; Lei, L. C.; Jing, X. B. Acta Physico-Chim. Sin. 1988, 4, 458. (22) Ishioka, T.; Uchida, T.; Teramae, N. Chem. Lett. 1998, 27, 765. (23) Roth, E.; Hope, G. A.; Schweinsberg, D. P.; Kiefer, W.; Fredericks, P. M. Appl. Spectrosc. 1993, 47, 1794. (24) Evans, J. C. Spectrochim. Acta 1960, 16, 428. 4182

dx.doi.org/10.1021/jp1117135 |J. Phys. Chem. C 2011, 115, 4174–4183

The Journal of Physical Chemistry C (25) Wojciechowski, P. M.; Zierkiewicz, W.; Michalska, D.; Hobza, P. J. Chem. Phys. 2003, 118, 10900. (26) Rauhut, G.; Pulay, P. J. Phys. Chem. 1995, 99, 3093. (27) Park, H.; Lee, S. B.; Kim, K.; Kim, M. S. J. Phys. Chem. 1990, 94, 7576. (28) Park, S. H.; Kim, K.; Kim, M. S. J. Mol. Struct. 1993, 301, 57. (29) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (30) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (31) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980, 72, 650. (32) McLean, A. D.; Chandler, G. S. J. Chem. Phys. 1980, 72, 5639. (33) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270. (34) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284. (35) Wu, D. Y.; Liu, X. M.; Duan, S.; Xu, X.; Ren, B.; Lin, S. H.; Tian, Z. Q. J. Phys. Chem. C 2008, 112, 4195. (36) Wu, D. Y.; Hayashi, M.; Shiu, Y. J.; Liang, K. K.; Chang, C. H.; Yeh, Y. L.; Lin, S. H. J. Phys. Chem. A 2003, 107, 9658. (37) Wu, D. Y.; Hayashi, M.; Chang, C. H.; Liang, K. K.; Lin, S. H. J. Chem. Phys. 2003, 118, 4073. (38) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. Rev. 1988, 88, 899. (39) Carpenter, J. E.; Weinhold, F. J. Mol. Struct.: THEOCHEM 1988, 169, 41. (40) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. A.; Montgomery, J., Jr; Vreven, K.; Kudin, T. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, E.; Yazyev, R. O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; L. Martin, R.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, Revision C.02; Gaussian, Inc.: Wallingford, CT, 2004. (41) Kozlowski, P. M.; Rush, T. S.; Jarzecki, A. A.; Zgierski, M. Z.; Chase, B.; Piffat, C.; Ye, B. H.; Li, X. Y.; Pulay, P.; Spiro, T. G. J. Phys. Chem. A 1999, 103, 1357. (42) Simon, S.; Duran, M.; Dannenberg, J. J. J. Chem. Phys. 1996, 105, 11024. (43) Boys, S. F.; Bernardi, F. Mol. Phys. 1970, 19, 553. (44) Schultz, G.; Portalone, G.; Ramondo, F.; Domenicano, A.; Hargittai, I. Struct. Chem. 1996, 7, 59. (45) Bludsky, O.; Sponer, J.; Leszczynski, J.; Spirko, V.; Hobza, P. J. Chem. Phys. 1996, 105, 11042. (46) Alcolea Palafox, M.; Gill, M.; Nuzez, N. J.; Rastogi, V. K.; Mittal, L.; Sharma, R. Int. J. Quantum Chem. 2005, 103, 394. (47) Lister, D. G.; Tyler, J. K.; Høg, J. H.; Larsen, N. W. J. Mol. Struct. 1974, 23, 253. (48) Wu, D. Y.; Ren, B.; Jiang, Y. X.; Xu, X.; Tian, Z. Q. J. Phys. Chem. A 2002, 106, 9042. (49) Wu, D. Y.; Ren, B.; Xu, X.; Liu, G. K.; Yang, Z. L.; Tian, Z. Q. J. Chem. Phys. 2003, 119, 1701. (50) Wang, A.; Huang, Y. F.; Sur, U. K.; Wu, D. Y.; Ren, B.; Rondinini, S.; Amatore, C.; Tian, Z. Q. J. Am. Chem. Soc. 2010, 132, 9534. (51) Suh, J. S.; DiLella, D. P.; Moskovits, M. J. Phys. Chem. 1983, 87, 1540. (52) Venkatachalam, R. S.; Boerio, F. J.; Roth, P. G. J. Raman Spectrosc. 1988, 19, 281. (53) Ibrahim, A.; Oldham, P. B.; Stokes, D. L.; Vo-Dinh, T. J. Raman Spectrosc. 1996, 27, 887.

ARTICLE

(54) Sun, S.; Birke, R. L.; Lombardi, J. R.; Leung, K. P.; Genack, A. Z. J. Phys. Chem. 1988, 92, 5965. (55) Yang, X. M.; Tryk, D. A.; Hashimoto, K.; Fujishima, A. J. Phys. Chem. B 1998, 102, 4933. (56) Gao, P.; Weaver, M. J. J. Phys. Chem. 1985, 89, 5040. (57) Muniz-Miranda, M.; Pergolese, B.; Bigotto, A. J. Phys. Chem. C 2008, 112, 6988. (58) Li, J. F.; Huang, Y. F.; Ding, Y.; Yang, Z. L.; Li, S. B.; Zhou, X. S.; Fan, F. R.; Zhang, W.; Zhou, Z. Y.; Wu, D. Y.; Ren, B.; Wang, Z. L.; Tian, Z. Q. Nature 2010, 464, 392. (59) Osawa, M.; Matsuda, N.; Yoshii, K.; Uchida, I. J. Phys. Chem. 1994, 98, 12702. (60) Zhou, Q.; Li, X.; Fan, Q.; Zhang, X.; Zheng, J. Angew. Chem., Int. Ed. 2006, 45, 3970. (61) Wang, Y.; Zou, X.; Ren, W.; Wang, W.; Wang, E. J. Phys. Chem. C 2007, 111, 3259. (62) Wang, Y.; Chen, H.; Dong, S.; Wang, E. J. Chem. Phys. 2006, 125, No. 044710. (63) Yang, X. M.; Tryk, D. A.; Hashimoto, K.; Fujishima, A. J. Raman Spectrosc. 1998, 29, 725. (64) Sun, M.; Xu, H. ChemPhysChem 2009, 10, 392. (65) Liu, S.; Zhao, X.; Li, Y.; Zhao, X.; Chen, M. J. Chem. Phys. 2009, 130, No. 234509. (66) Fang, Y.; Li, Y.; Xu, H.; Sun, M. Langmuir 2010, 26, 7737. (67) Kydd, R. A.; Krueger, P. J. Chem. Phys. Lett. 1977, 49, 539. (68) Niu, Z.; Dunn, K. M.; Boggs, J. E. Mol. Phys. 1985, 55, 421. (69) Sechkarev, A. V. Izv. Vyssh. Uchebn. Zaved. Fiz. 1965, 1, 5. (70) Fischer, I. Nature 1950, 165, 239.

4183

dx.doi.org/10.1021/jp1117135 |J. Phys. Chem. C 2011, 115, 4174–4183