Vibrational Sum Frequency Generation by the Quadrupolar

Apr 25, 2013 - In reality, however, VSFG at the nonpolar benzene/air interface has been observed with traditional homodyne-detected VSFG. Here we repo...
5 downloads 0 Views 908KB Size
Letter pubs.acs.org/JPCL

Vibrational Sum Frequency Generation by the Quadrupolar Mechanism at the Nonpolar Benzene/Air Interface Korenobu Matsuzaki,†,‡ Satoshi Nihonyanagi,† Shoichi Yamaguchi,† Takashi Nagata,‡ and Tahei Tahara*,† †

Molecular Spectroscopy Laboratory, RIKEN, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan Department of Chemistry, School of Science, The University of Tokyo, 3-8-1 Komaba, Meguro-ku, Tokyo 153-8902, Japan



S Supporting Information *

ABSTRACT: The interface selectivity of vibrational sum frequency generation (VSFG) spectroscopy is explained under the dipole approximation as resulting from the breakdown of inversion symmetry at the interface. From this viewpoint, VSFG is not expected to occur at the interface consisting of centrosymmetric molecules, because the inversion symmetry is preserved even at the interface. In reality, however, VSFG at the nonpolar benzene/air interface has been observed with traditional homodyne-detected VSFG. Here we report a heterodyne-detected VSFG study of the benzene/air interface. The result strongly indicates that VSFG at this interface cannot be explained within the framework of the dipole approximation. The selection rule and polarization dependence of the observed VSFG signal show that the quadrupole transition plays an essential role because of the field discontinuity at the interface. This finding implies the applicability of interfaceselective VSFG to the nonpolar interfaces comprising centrosymmetric molecules, which opens a new possibility of VSFG spectroscopy. SECTION: Surfaces, Interfaces, Porous Materials, and Catalysis nterfaces are important in many fields of science and technology, and it is crucial to obtain molecular-level understanding of their properties. However, the information of the interface is difficult to be obtained except for the solid surface under the vacuum, because the detection of the signal from the interface is usually hindered by the background signal from the bulk in ordinary spectroscopic measurements. Vibrational sum frequency generation (VSFG) spectroscopy1,2 has an intrinsic interface selectivity and has widely been utilized to study interfaces, especially liquid interfaces. VSFG is a second-order nonlinear process, and under the electric dipole approximation, it occurs only in the region where inversion symmetry is broken. Because molecules are randomly oriented in the bulk liquid, the bulk region practically has inversion symmetry. In contrast, inversion symmetry is broken at the interface because the interfacial molecules exhibit net alignment due to the anisotropic environment. Therefore, VSFG is observed only at the interface, and hence the molecular information of the interface can be selectively obtained. Because of this principle of the interface selectivity in VSFG, it is usually believed that VSFG spectroscopy cannot be applied to the study of nonpolar interfaces comprised of centrosymmetric molecules. From the viewpoint of the selection rule, the vibrational modes that are both IR and Raman active are observed in VSFG spectroscopy, whereas there exists the mutual exclusion rule between IR and Raman active modes for centrosymmetric molecules. Therefore, no vibrational modes should be observed at such nonpolar interfaces in VSFG spectroscopy. Contrary to this expectation, Hommel and Allen reported that VSFG is actually observable at the nonpolar

I

© 2013 American Chemical Society

benzene/air interface using traditional homodyne-detected VSFG spectroscopy.3 They argued that the anisotropic environment of the interface makes benzene become “polar” at the interface, and hence interfacial benzene becomes VSFG active. In this paper, we report a heterodyne-detected VSFG (HDVSFG) study of the benzene/air interface. HD-VSFG is advantageous over the traditional homodyne-detected VSFG in a number of respects.2,4−10 In particular, because the nonresonant background is separated in the real part, the imaginary χ(2) (Imχ(2)) spectra obtained with HD-VSFG is free from the spectral distortion due to the interference between resonant and nonresonant contributions, and hence we can precisely determine the resonant frequency of molecular vibrations. This allows us to precisely compare the vibrational frequencies observed in HD-VSFG with those observed by the conventional vibrational spectroscopy for the bulk. The obtained data indicate that the VSFG signal is actually generated not mainly because the benzene molecules become polar due to large structural distortion but because the quadrupole transition substantially occurs in the interface region. Nevertheless, the VSFG signal is still generated from the interfacial region, and hence this finding opens a new possibility of VSFG spectroscopy for the study of nonpolar interfaces. Received: April 20, 2013 Accepted: April 25, 2013 Published: April 25, 2013 1654

dx.doi.org/10.1021/jz400829k | J. Phys. Chem. Lett. 2013, 4, 1654−1658

The Journal of Physical Chemistry Letters

Letter

comparison shows that the vibrational bands observed in the Imχ(2) spectra are divided into two groups: One exhibits the same vibrational frequencies as bulk liquid benzene, and the other shows different frequencies. For instance, within the experimental error, the 3035 (3034), 3070, and 3091 (3089)cm−1 bands observed in the SSP, SPS, and PSS Imχ(2) spectra are considered identical to the three E1u bands observed in IR spectra of the bulk. On the other hand, all the bands in the PPP polarization combination exhibit substantially different frequencies from those in either IR or Raman spectrum of the liquid. These frequencies of the vibrational bands in the PPP polarization combination are located in between the frequencies of benzene molecules in liquid and gas. Actually, the peak frequency of 3066 cm−1 in the PPP spectrum (3067 cm−1 in the SSP spectrum) is higher than the liquid-phase frequency of 3062 cm−1 but lower than the gas-phase frequency of 3074 cm−1.11 Because these bands in the liquid and gas have been assigned to the Raman-active A1g mode,11,12 it is natural to assign the 3067-cm−1 band in the Imχ(2) spectra to the same Raman active vibration. Moreover, the appearance of the frequency in between the values in the gas and liquid phases suggests that these bands are attributable to the vibrations of benzene at the interface. In this manner, the four bands observed in the PPP spectrum can be assigned to the vibrations of benzene at the interface: the band at 3066 cm−1 is the Raman active A1g mode, and the bands at 3046, 3079, and 3098 cm−1 are the IR active E1u modes of the interfacial benzene molecules. These vibrational frequencies assigned to the interfacial benzene molecule are tabulated in Table 1 with the frequencies observed in the IR and Raman spectra of the liquid and gas. Figure 1. Imχ(2), IR, and Raman spectra of benzene. The Imχ(2) of the benzene/air interface were obtained by HD-VSFG with (a) SSP, (b) SPS, (c) PSS, and (d) PPP polarization combinations (see Supporting Information (SI) for the experimental details). (e) IR and (f) Raman spectra were acquired from bulk liquid-phase benzene.

Table 1. CH Stretch Frequencies [cm−1] of Interfacial Benzene Observed in HD-VSFG Measurements with SSP and PPP Polarization Combinationsa E1u (IR) gas11,13 HD-VSFG

(2)

Figure 1(a−d) shows the Imχ spectra of the benzene/air interface measured with HD-VSFG in the CH stretch region. These data confirm that VSFG is indeed observable at the nonpolar benzene/air interface as reported with conventional homodyne-detected VSFG.3 The well-resolved Imχ(2) spectra show numbers of vibrational bands for each polarization combination, and more importantly, disclose that different bands were observed for the SSP, SPS, PSS, and PPP polarization combinations (The first, second, and third letters correspond to the polarization of the sum frequency, visible, and IR light, respectively). In the CH stretch region, benzene has six CH fundamental modes that belong to the A1g, E2g, B1u, and E1u irreducible representations of the D6h point group. Among them, the A1g and E2g modes are Raman active, E1u is IR active, and the B1u mode is inactive both in Raman and IR. Consequently, in the bulk spectra, the A1g and E2g modes are found at 3062 cm−1 and 3048 cm−1 in the Raman spectrum, respectively, while the E1u modes are split into three in the IR spectrum due to the Fermi resonance as shown in Figure 1e,f. Now, Figure 1 allows us to directly compare the vibrational frequencies observed in the HD-VSFG measurements with those observed in Raman and IR spectra of liquid benzene. In this analysis, we consider that the Imχ(2) bands labeled with their frequencies in Figure 1a−d are physically meaningful, and neglect other features whose intensity is smaller than the baseline fluctuation. The careful

liquid

3048 3046 (PPP) 3036

A1g (Raman) 3074 3067 (SSP) 3066 (PPP) 3062

E1u (IR)

E1u (IR)

3079

3101

3079 (PPP) 3071

3098 (PPP) 3091

a The corresponding frequencies in the gas11,13 and liquid phases are also shown.

The observation of the vibrational bands at exactly the same frequency as the bulk indicates that the signal has a quadrupole origin. Because benzene cannot show any dipole-origin VSFG signal with D6h symmetry, it is reasonable that the signal due to a quadrupolar mechanism becomes prominent. Recently, we have reported a theoretical treatment of SFG signals arising from a mechanism involving both dipole and quadrupole transitions.14 Figure 2 shows the diagrams of the VSFG processes with the electric quadrupole contribution taken into account. The diagram in Figure 2a is the ordinary VSFG process in which only dipole transitions are involved. Figure 2b,c,d depict the quadrupolar mechanism, in which one of the three transitions is replaced with a quadrupole transition (quad1, quad2, and quad3 mechanisms). In particular, in the quad3 mechanism (Figure 2d), the nonlinear polarization emits VSFG light through a quadrupole transition in the final process. An important conclusion of our previous work is that, although this polarization in the quad3 mechanism is localized at the interface, the VSFG signal only reflects the bulk properties.14 1655

dx.doi.org/10.1021/jz400829k | J. Phys. Chem. Lett. 2013, 4, 1654−1658

The Journal of Physical Chemistry Letters

Letter

induced VSFG processes. In quad1, for instance, the IR transition occurs with a dipole transition, and the following Raman-type transition involves a dipole transition and a quadrupole transition. Because of the quadrupole transition, the latter Raman-type transition is represented by a tensor of rank 3, whereas the ordinary Raman transition is expressed by a tensor of rank 2. A third rank tensor indicates that the selection rule for this transition is the same as hyper-Raman scattering, which is a two-photon excited analogue of Raman scattering. Considering that IR active modes are always hyper-Raman active, the vibrational modes detected by quad1 should be IR active. Combining this argument with the interface selectivity of the quad1 mechanism, it can be concluded that the IR active modes of interfacial molecules provide VSFG signal with the quad1 mechanism. Similarly, we can conclude that the Raman active modes of the interfacial molecules are observed with quad2 (SI for details of the selection rule). Furthermore, we can show which quadrupolar mechanisms contribute to the signal observed in a specific polarization combination. This is because the electric field of the incident light has a large gradient only for the Z component, ∂EZ/∂Z, at the interface, and the quad1 and quad2 mechanisms provide the signal through this ∂EZ/∂Z (Z denotes the axis normal to the interface). In other words, the quadrupole transition becomes significant only when the relevant incident light is Ppolarized in these mechanisms. Therefore, in the quad1 mechanism, for example, the quadrupole transition is induced by the visible light, and hence it generates signal only when the visible light is P-polarized. Consequently, the signal due to quad1 appears in the SPS and PPP polarization combinations. Similarly, for quad2, the electric field gradient of the IR light is necessary, and it appears in the SSP and PPP polarization combinations. For quad3, because this process is irrelevant to the electric field gradient, the visible and the IR light can be either S- or P-polarized. Therefore, the quad3 signal appears in all the polarization combinations. In Table 2, the selection rule for each polarization combination in the quadrupolar mechanism is summarized

Figure 2. VSFG process induced by the conventional (a) and the quadrupolar (b,c,d) mechanisms. Vibrational modes detected by these diagrams are shown below each of the diagrams. Descriptive figures showing molecules detected by the corresponding mechanism are also given.16,17

(Contribution of the bulk properties to the polarization at the interface through the quad3 mechanism was first suggested by Guyot-Sionnest and Shen.15) Thus, the Imχ(2) bands observed with the bulk frequencies are assignable to the VSFG signal arising from this quad3 mechanism. The quadrupolar mechanism contains two other processes, in which IR or visible transition occurs with a quadrupolar transition (quad1 in Figure 2b and quad2 in Figure 2c). In these VSFG processes, the quadrupole transition is induced by the gradient of the relevant electric field. Because the refractive index drastically changes from the air to liquid benzene phase, the electric field gradient of the incident light becomes very large in the interface region. Therefore, the quad1 and quad2 mechanisms can be significant and give rise to the signal reflecting the properties of the interface. Because all the three quadrupolar mechanisms contain one quadrupole transition and two dipole transitions, they are expected to have similar magnitude. It implies that, as the quad3 mechanism provides prominent VSFG signals at the bulk frequencies, the signal due to the quad1 and quad2 mechanisms should also be detectable at the benzene interface. We conclude that the VSFG signals exhibiting frequencies of interfacial benzene arise from the quad1 and quad2 mechanisms. In fact, the selection rule and polarization dependence of the observed Imχ(2) signal is beautifully explained by the quadrupolar mechanism as described below. In the conventional dipolar mechanism, only those vibrational modes that are both IR and Raman active are detected because the VSFG process consists of an IR transition followed by a Raman transition (Figure 2a). As readily understood from Figure 2, a clear selection rule is also held for the quadrupole-

Table 2. Assignments of the Experimentally Observed Bands and the Theoretical Predictions experiment SSP SPS PSS PPP

interface Ramana, bulk IRb bulk IR bulk IR interface IRc, interface Raman

theory interface Raman (quad2), bulk IR (quad3) interface IR (quad1), bulk IR (quad3) bulk IR (quad3) interface IR (quad1), interface Raman (quad2), bulk IR (quad3)

a Interface Raman: Raman active modes of interfacial molecules. bBulk IR: IR active modes of bulk molecules. cInterface IR: IR active modes of interfacial molecules.

and is compared with the experimental observation. As readily seen, the experimental observation is perfectly explained with the quadrupolar mechanism. Actually, for the PSS polarization combination, the theory predicts that only IR active modes showing bulk frequencies are observed. The Raman active modes of interfacial benzene should be observed solely with the SSP and PPP polarization combinations. These predictions excellently accord with the experimental observation and rationalize the assignment of the 3067-cm−1 band to the Raman-active interfacial vibration. In the PPP polarization combination, all the mechanisms can generate signal, so that 1656

dx.doi.org/10.1021/jz400829k | J. Phys. Chem. Lett. 2013, 4, 1654−1658

The Journal of Physical Chemistry Letters

Letter

Notes

they are overlapped to show the spectrum observed. Nevertheless, the observed vibrational frequencies are obviously different from the frequencies either in the liquid or gas phase. Thus, they are assignable to the Raman and IR active vibrations of interfacial benzene, which arise from quad1 (Raman active) and quad2 (IR active) mechanisms. This agreement between theoretical consideration and experimental data indicate that VSFG at the benzene/air nonpolar interface originates from the quadrupolar mechanism. In the previous homodyne-detected VSFG study, the observation of VSFG signal at the benzene/air interface was attributed to the dipole that is induced by the distortion of molecules at the interface.3 Also in a recent MD simulation study, Kawaguchi et al. claimed that the VSFG signal observed in the SSP polarization combination, which corresponds to the 3067-cm−1 band in Figure 1a, originates from symmetry-broken benzene molecules at the interface.17 We do not think that such symmetry-breaking due to a drastic structural distortion is the main origin of the observed VSFG bands of benzene. As we have already described, the vibrational frequencies of interfacial benzene in the Imχ(2) spectra are in between the frequencies in the gas and liquid phase. It is more natural to attribute this gradual frequency shift to the environmental effect on benzene that practically preserves the D6h symmetry.18,19 Furthermore, the selection rule for each polarization combination of the observed Imχ(2) spectra is perfectly explained with a theoretical consideration assuming the D6h symmetry. These facts indicate that the structure of interfacial benzene is not drastically deviated from the D6h symmetry with which the selection rule holds. We mention that Kawaguchi et al. also theoretically showed that the VSFG signal can also arise from the quadrupolar mechanism, being stimulated by our HD-VSFG data reported in this paper.17 In conclusion, we measured Imχ(2) spectra at the nonpolar benzene/air interface with HD-VSFG. The high sensitivity of HD-VSFG and well-resolved Imχ(2) spectra allowed us to observe numbers of vibrational bands of benzene. The selection rule and polarization dependence of the observed signal are explained with the mechanism involving an electric quadrupole transition. Although the signal has the quadrupole origin in this mechanism, the VSFG signal is generated by the nonlinear polarization at the interface region, and hence we are able to observe vibrational spectra of the interfacial molecules. This principle for interface selectivity of VSFG is completely different from the conventional dipole-induced VSFG and opens a new way to study interfaces with VSFG spectroscopy. We note that we have already observed VSFG at another nonpolar liquid/air interface, i.e., cyclohexane/air interface, and the observation is well accounted for by the quadrupolar mechanism.



The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by a Grant-in-Aid for Scientific Research (A) (No. 24245006) and a Grant-in-Aid for Scientific Research (B) (No. 22350014) from Japan Society for the Promotion of Science (JSPS).



ASSOCIATED CONTENT

S Supporting Information *

The experimental details, the derivation of the selection rule for quadrupole-induced VSFG, and the |χ(2)|2 spectra of benzene. This material is available free of charge via the Internet at http://pubs.acs.org.



REFERENCES

(1) Zhu, X. D.; Suhr, H.; Shen, Y. R. Surface Vibrational Spectroscopy by Infrared-Visible Sum Frequency Generation. Phys. Rev. B 1987, 35, 3047−3050. (2) Nihonyanagi, S.; Yamaguchi, S.; Tahara, T. Direct Evidence for Orientational Flip-Flop of Water Molecules at Charged Interfaces: A Heterodyne-Detected Vibrational Sum Frequency Generation Study. J. Chem. Phys. 2009, 130, 204704. (3) Hommel, E. L.; Allen, H. C. The Air−Liquid Interface of Benzene, Toluene, m-Xylene, and Mesitylene: A Sum Frequency, Raman, and Infrared Spectroscopic Study. Analyst 2003, 128, 750− 755. (4) Nihonyanagi, S.; Yamaguchi, S.; Tahara, T. Water Hydrogen Bond Structure near Highly Charged Interfaces Is Not Like Ice. J. Am. Chem. Soc. 2010, 132, 6867−6869. (5) Mondal, J. A.; Nihonyanagi, S.; Yamaguchi, S.; Tahara, T. Structure and Orientation of Water at Charged Lipid Monolayer/ Water Interfaces Probed by Heterodyne-Detected Vibrational Sum Frequency Generation Spectroscopy. J. Am. Chem. Soc. 2010, 132, 10656−10657. (6) Nihonyanagi, S.; Ishiyama, T.; Lee, T.-k.; Yamaguchi, S.; Bonn, M.; Morita, A.; Tahara, T. Unified Molecular View of the Air/Water Interface Based on Experimental and Theoretical χ(2) Spectra of an Isotopically Diluted Water Surface. J. Am. Chem. Soc. 2011, 133, 16875−16880. (7) Mondal, J. A.; Nihonyanagi, S.; Yamaguchi, S.; Tahara, T. Three Distinct Water Structures at a Zwitterionic Lipid/Water Interface Revealed by Heterodyne-Detected Vibrational Sum Frequency Generation. J. Am. Chem. Soc. 2012, 134, 7842−7850. (8) Nihonyanagi, S.; Singh, P. C.; Yamaguchi, S.; Tahara, T. Ultrafast Vibrational Dynamics of a Charged Aqueous Interface by Femtosecond Time-Resolved Heterodyne-Detected Vibrational Sum Frequency Generation. Bull. Chem. Soc. Jpn. 2012, 85, 758−760. (9) Stiopkin, I. V.; Jayathilake, H. D.; Bordenyuk, A. N.; Benderskii, A. V. Heterodyne-Detected Vibrational Sum Frequency Generation Spectroscopy. J. Am. Chem. Soc. 2008, 130, 2271−2275. (10) Stiopkin, I. V.; Weeraman, C.; Pieniazek, P. A.; Shalhout, F. Y.; Skinner, J. L.; Benderskii, A. V. Hydrogen Bonding at the Water Surface Revealed by Isotopic Dilution Spectroscopy. Nature 2011, 474, 192−195. (11) Thakur, S. N.; Goodman, L.; Ozkabak, A. G. The Benzene Ground-State Potential Surface. 1. Fundamental Frequencies for the Planar Vibrations. J. Chem. Phys. 1986, 84, 6642−6656. (12) Shimanouchi, T. Tables of Molecular Vibrational Frequencies Consolidated Volume I. Nat. Stand. Ref. Data Ser., Nat. Bur. Stand. 1972, 39, 160 pages. (13) Page, R. H.; Shen, Y. R.; Lee, Y. T. Infrared-Ultraviolet DoubleResonance Studies of Benzene Molecules in a Supersonic Beam. J. Chem. Phys. 1988, 88, 5362−5376. (14) Yamaguchi, S.; Shiratori, K.; Morita, A.; Tahara, T. Electric Quadrupole Contribution to the Nonresonant Background of Sum Frequency Generation at Air/Liquid Interfaces. J. Chem. Phys. 2011, 134, 184705. (15) Guyot-Sionnest, P.; Shen, Y. R. Local and Nonlocal Surface Nonlinearities for Surface Optical Second-Harmonic Generation. Phys. Rev. B 1987, 35, 4420−4426. (16) Shiratori, K.; Morita, A. Molecular Theory on Dielectric Constant at Interfaces: A Molecular Dynamics Study of the Water/ Vapor Interface. J. Chem. Phys. 2011, 134, 234705.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. 1657

dx.doi.org/10.1021/jz400829k | J. Phys. Chem. Lett. 2013, 4, 1654−1658

The Journal of Physical Chemistry Letters

Letter

(17) Kawaguchi, T.; Shiratori, K.; Henmi, Y.; Ishiyama, T.; Morita, A. Mechanisms of Sum Frequency Generation from Liquid Benzene: Symmetry Breaking at Interface and Bulk Contribution. J. Phys. Chem. C 2012, 116, 13169−13182. (18) Angus, W. R.; Bailey, C. R.; Hale, J. B.; Ingold, C. K.; Leckie, A. H.; Raisin, C. G.; Thompson, J. W.; Wilson, C. L. Structure of Benzene. Part VII. Discussion of Coincidental Frequencies in the Infra-Red and Raman Spectra of Benzene and Hexadeuterobenzene. J. Chem. Soc. 1936, 966−971. (19) Neddersen, J. P.; Mounter, S. A.; Bostick, J. M.; Johnson, C. K. Nonresonant Hyper-Raman and Hyper-Rayleigh Scattering in Benzene and Pyridine. J. Chem. Phys. 1989, 90, 4719−4726.

1658

dx.doi.org/10.1021/jz400829k | J. Phys. Chem. Lett. 2013, 4, 1654−1658