13063
2005, 109, 13063-13066 Published on Web 06/18/2005
Second Harmonic Generation-Based Coherent Vibrational Spectroscopy for a Liquid Interface under the Nonresonant Pump Condition Yasushi Hirose,*,†,‡ Hiroharu Yui,‡,§ and Tsuguo Sawada‡ Kanagawa Academy of Science and Technology, 504 KSP Bldg. East Wing, 3-2-1 Sakado, Takatsu-ku, Kawasaki, Kanagawa 213-0012, Japan, Department of AdVanced Materials Science, Graduate School of Frontier Sciences, The UniVersity of Tokyo, 5-1-5-603, Kashiwanoha, Kashiwa, Chiba 277-8561, Japan, and Department of Chemistry, Faculty of Science DiVision I, Tokyo UniVersity of Science, 1-3 Kagurazaka, Shinjyuku-ku, Tokyo 162-8601, Japan ReceiVed: May 11, 2005; In Final Form: June 11, 2005
The molecular dynamics in the low-frequency region (0-500 cm-1) sensitively reflects the intermolecular interactions in a liquid. The second harmonic generation-based coherent vibrational spectroscopy (SHGCVS) was developed to monitor the low-frequency dynamics of molecules at a liquid interface, which was difficult to access by using the present spectroscopic techniques such as sum frequency generation or attenuated total reflection (ATR)-IR. Background-free detection with the transient grating (TG) optical configuration was adopted to obtain the weak signal under the electronically nonresonant pump condition. It was demonstrated that the S/N ratio of the SHG-CVS with the TG configuration was remarkably superior to that with the conventional time-resolved SHG configuration, and the improved detection limit enabled us to detect the low-frequency dynamics of coumarin 314 molecules at the air/water interface under the electronically nonresonant pump condition.
Introduction The interfacial area of a liquid is an important field for various processes, such as catalysis, electrochemical and biological reactions, and molecular sensing.1 The liquid interface is also important for the formation of a molecular assembly (e.g., a membrane, a micelle, and a liposome) and has a great influence on its structure and a reaction therein. It is well-recognized that, in chemical reactions or the formation of molecular assemblies at liquid interfaces, the spatial inhomogeneity of interactions with solvent molecules on a mesoscopic scale plays essential roles. Many molecules participate in such intermolecular interactions of a liquid, and these interactions are sensitively reflected in the low-frequency region (0-500 cm-1).2-4 Thus, an interface-specific spectroscopy for the low-frequency dynamics of molecules at a liquid interface, especially for those of the solvent molecules and surfactant molecules that compose a molecular assembly, is strongly required. The molecular motions at a liquid interface can be monitored in a frequency domain by using vibrational sum frequency generation spectroscopy (VSFGS).5 However, it is difficult to apply VSFGS to liquid interfaces in the low-frequency region, because there is no far-IR light source with sufficient intensity for an SF process, except a free-electron laser, which requires a very large facility. Attenuated total reflection infrared (ATRIR) spectroscopy is not interface-specific in the low-frequency * To whom the correspondence should be addressed. Telephone: +81 (0)44-819-2081. Fax: +81 (0)44-819-2083. E-mail:
[email protected]. † Kanagawa Academy of Science and Technology. ‡ The University of Tokyo. § Tokyo University of Science.
10.1021/jp0524476 CCC: $30.25
region, because the penetration depth becomes several tens of micrometers. In the present paper, we developed the second harmonic generation-based coherent vibrational spectroscopy (SHG-CVS) as a method to investigate the low-frequency dynamics at interfacial areas. Since the SHG-CVS is a time-domain Raman spectroscopy and uses only an ultrashort pulsed visible laser as a light source,6 the above limitation for VSFGS does not exist in the SHG-CVS. In addition, we adopted the TG optical configuration to detect the weak signal under the electronically nonresonant pump condition. The interface between air and aqueous solution of coumarin 314 (c314) was investigated by this method, and it was demonstrated that the S/N ratio of the SHG-CVS was remarkably improved by the TG optical configuration. Because of the improved detection limit, the lowfrequency dynamics of c314 molecules at the air/water interface was detected without the electronic resonance enhancement by the pump beam. Theory of the SHG-CVS The SHG-CVS is a fourth-order time-domain nonlinear Raman spectroscopy.6 It is an analogy of χ(3)-based time-domain nonlinear Raman spectroscopy for bulk phases such as Ramaninduced optical Kerr effect spectroscopy (RIKES) and impulsive stimulated Raman spectroscopy (ISRS).7 In an isotropic medium such as a liquid, even order nonlinear susceptibility is nonzero only at the interface or surface, where inversion symmetry breaks. Therefore, the SHG-CVS can selectively observe the liquid interface. Figure 1 shows the energy diagram of the SHGCVS. The ultrashort pump pulse with duration of a few tens of © 2005 American Chemical Society
13064 J. Phys. Chem. B, Vol. 109, No. 27, 2005
Letters
ISHG(2Ω, ∆t) - I0SHG(2Ω)
xI0
SHG
∝ χ(4)(∆t)
(3)
(2Ω)
where I0SHG(2Ω) ) |E(2)(2Ω)|2 is the intensity of the unmodulated second harmonic light of the probe beam and is determined by the intensity of the signal without the pump beam. SHG-CVS with the TG Optical Configuration Figure 1. Energy diagram of the SHG-CVS. ν and ν′ are the vibrational ground state and excited state, respectively. ω is the frequency of ν′. Ω is the frequency of the pump pulse and probe pulse. ∆t is the delay time between the pump and probe pulses.
Figure 2. (a) Schematic illustration of the SHG-CVS with a TG optical configuration. (b) Phase-matching condition of the TRSHG optical configuration and that of the TG optical configuration.
femtoseconds excites the coherence of the low-frequency motions of a liquid through a stimulated Raman process. After a delay time of ∆t, a probe beam (frequency Ω) interacts with the coherently excited molecules and generates a signal light with second harmonic frequency (2Ω). The coherent motions of the excited molecules oscillatorily modulate the fourth-order nonlinear susceptibility χ(4), which generates the second harmonic frequency field E(4)(2Ω, ∆t)
E(4)(2Ω, ∆t) ∝ χ(4)(∆t)Epr2 ∝
∑j Aj cos(ωj∆t + φj) exp(-γj∆t)
(1)
where Epr is the field of the probe beam. Aj, ωj, φj, and γj are the amplitude of the coherent motion, eigenfrequency, initial phase, and relaxation constant of the jth mode. In the previous report on the SHG-CVS, a conventional timeresolved second harmonic generation (TRSHG) optical configuration has been used.6,8 In the TRSHG setup, as shown in Figure 2b, the 2Ω frequency lights generated from the fourthorder process (signal) and that from the SHG process of the probe beam (not modulated by the pump beam) were generated in the same direction. Therefore, the detected intensity of the 2Ω lights, ISHG(2Ω, ∆t) was presented as follows
ISHG(2Ω, ∆t) ) |E(2)(2Ω) + E(4)(2Ω, ∆t)|2 = |E(2)(2Ω)|2 + 2E(2)(2Ω)E(4)(2Ω, ∆t)
(2)
Equation 3 shows that, in the TRSHG configuration, the signal/base ratio was substantially small, and high stability of the background SH intensity I0SHG(2Ω) was required. However, the stability of the I0SHG(2Ω) was relatively low for a liquid interface because of a mechanical instability and a small nonlinear susceptibility. Therefore, electronic resonance enhancement by the pump beam was required for the measurement of a liquid interface.8 As a result, the SHG-CVS was difficult to apply for investigation of the dynamics of small solvent molecules such as water and alcohol or surfactant molecules, which play essential roles in many processes at a liquid interface, because the electronic transitions of these molecules exist only in the UV region. Then, we adopted the TG optical configuration to the SHGCVS to detect the weak signal of a liquid interface without electronic resonance enhancement. In the TG optical configuration (Figure 2a), the 2Ω signal light was spatially separated from 2Ω lights from second-order processes (SHG and SFG) because of the difference in phase-matching conditions (Figure 2b). Therefore, the signal was detected under a backgroundfree condition, and a high S/N ratio was achieved even under the nonresonant pump condition. The intensity of the detected signal ITG(2Ω, ∆t) was presented as follows
ITG(2Ω, ∆t) ∝ |E(4)(2Ω, ∆t)|2 ) |χ(4)(∆t)Epr2|2 ∝ |
∑j Aj cos(ωj∆t + φj) × exp(-γj∆t)|2 (4)
Experimental Section Figure 3 shows the experimental setup of the SHG-CVS with a TG configuration. The light source was a multipass-amplified Ti:sapphire laser (Quntronix, Odin-C; 800 nm, 30 fs pulse duration, 0.6 mJ/pulse, 1 kHz repetition) with a prism-pair group velocity dispersion compensator. A part of the output was used as a pump beam, and the residual beam was used as a probe beam. The pump beam was split into two beams with a halfmirror, and the two beams were focused and crossed on the sample interface at the same time. The crossing angle of the pump beams was ∼1°. The probe beam was introduced under the phase-matching condition after passing through an optical delay line. The incident angle of all beams was ∼70°. A slit was used for avoiding the contamination by the 2Ω frequency lights from second-order processes (SHG and SFG). The temporal and spatial overlap of the pump and probe pulses was optimized by maximizing the sum frequency generated signal of 2Ω frequency from each pair of beams independently. The signal light of 2Ω frequency was detected with a photomultiplier tube (Hamamatsu Photonics, R4220P) and a Boxcar integrator (Stanford Research Systems, SR250). Scattered fundamental light was eliminated using an interference filter. The response function of the apparatus was evaluated by using a vacuumdeposited gold film (∼100 nm thickness) on a glass substrate as a sample: The response of the gold film mainly originated
Letters
J. Phys. Chem. B, Vol. 109, No. 27, 2005 13065 was about 0.4 molecule/nm2.9 Under the experimental condition (λpump ) λprobe ) 800 nm), while the S1 r S0 electronic transition of c314 showed two-photon resonance with the probe beam, there was no resonant transition with the pump beam (Figure 1). Results and Discussion
Figure 3. Experimental setup of the SHG-CVS with a TG optical configuration. PS, periscope; NDF, neutral density filter; BS, beam splitter; HM, half mirror; WP, λ/2 waveplate; CF, color filter; IF, interference filter; P, polarizer; PMT, photomultiplier tube; A/D, A/D converter.
Figure 4. (a) The SHG-CVS signal of c314 at the air/water interface with a TRSHG optical configuration and that with a TG configuration. The concentration of c314 was 20 µM. (b) The Fourier power spectrum of the SHG-CVS signal with the TG configuration (filled circle). The spectrum was normalized by the Fourier transform of the response function of the apparatus, which is called the sensitivity curve (dashed line).
from the instantaneous electronic polarizability, which was sufficiently fast to estimate the response function of the apparatus. The evaluated value was 50 fs full width at halfmaximum (fwhm) (Figure 4a). In our experiment, the low-frequency dynamics of coumarin 314 (c314) molecules at the air/water interface was investigated. C314 is a surface-active ester, and we used a 20 µM C314 aqueous solution as a sample: The surface population of c314
Figure 4a shows the SHG-CVS signal of c314 at the air/ water interface with a TRSHG optical configuration and that with a TG optical configuration. By adopting the TG configuration, the noise level was remarkably reduced, and a signal was observed at the delay time ≈ 0 for the air/c314 aqueous solution interface (no signal was observed for the air/pure water interface). This signal was not observed when either the pump beam or the probe beam was cut off, or when the temporal oVerlap of the pump beams was not accomplished, which ruled out the possibility of the contamination by 2Ω lights of secondorder processes from the interface (SHG and SFG). In addition, we did the measurements with various positions of the spatial slit in front of the detector and confirmed the absence of the background of the hyper-Rayleigh scattering at all positions. The contributions of the above processes were also denied from the dependence of the signal intensity on the laser power. Therefore, we concluded that the detected signal originated from the fourth-order process, which reflected the response of c314 molecules at the air/water interface. The origin of the SHG-CVS signal is briefly considered. Because of the use of a nonresonant pump beam, the SHGCVS signal was induced by a stimulated Raman process and originated from neither a population change of c314 nor a solvation of water molecules, as was reported in conventional TRSHG studies of an electronically excited dye molecule.9 RIKES and ISRS studies have reported that the response of molecules in a bulk solution due to a stimulated Raman process by an ultrashort laser pulse consists of two contributions: an instantaneous component of an electron system (1 ps), as well as inter- and intramolecular vibrations. In our experiment, while an oscillatory component and a long-lived one were not observed, the decay time of the SHG-CVS signal (130 fs fwhm) was apparently larger than that of the response function of the apparatus (50fs fwhm). The increase of the decay time indicated the contribution of the nucleus system of c314, and it was also confirmed in a frequency domain as a broad and structureless component below ∼100 cm-1 (Figure 4b). Because the decay time of the SHG-CVS signal (130 fs fwhm) corresponded well, we considered that the signal was assigned to an overdamped libration of c314 as well as the electronic polarizability. We could not determine that the absence of a diffusive component reflected the restriction of the diffusive motions of c314 as in the case of the molecules adsorbed on an air/solid interface10 because of the long-term instability and the sensitivity of the present apparatus. These results indicated that the SHG-CVS with the TG configuration has the potential to be applied to monitoring the dynamics of molecules at a liquid interface under the nonresonant pump condition. Further improvement of the S/N ratio was the task to be solved to apply the method to investigating the molecular dynamics of small solvent or surfactant molecules, whose signals are weaker than those of a dye molecule like c314. In the present SHG-CVS with the TG configuration apparatus, the main sources of noise were (1) the fluctuation of
13066 J. Phys. Chem. B, Vol. 109, No. 27, 2005
Letters
the liquid interface due to mechanical vibration and air convection and (2) the instability of the output of the photomultiplier tube. These problems could be improved by (1) vibration control and a more rigid sample cell and (2) the photon-counting method, respectively.
Acknowledgment. This research was supported by Grantsin-Aid for Scientific Research on Priority Areas ((no. 13129203), (B)(1)(no. 14340189), and (B)(2)(no. 14350442)) from the Japanese Ministry of Education, Science, Sports, and Culture and the Japan Science and Technology Agency.
Summary
References and Notes
The SHG-CVS with the TG optical configuration was developed as a novel interface-specific spectroscopy for a liquid interface under the electronically nonresonant pump condition. It was demonstrated that the TG configuration remarkably improved the S/N ratio of the SHG-CVS. The improved S/N ratio enabled us to observe the low-frequency dynamics of c314 at the air/water interface without the enhancement by the electronically resonant pump beam. The SHG-CVS measurement under the electronically nonresonant pump condition overcomes the limitation of the wavelength of the pump beam, which enables direct investigation of a greater variety of molecules, such as small solvent molecules or surfactant molecules whose electronic transitions exist only in the UV region. We believe that the SHG-CVS with the TG configuration opened a new window for a spectroscopic approach to the lowfrequency dynamics of molecules at a liquid interface.
(1) See, for example, Volkov, A. G.; Deamer, D. W.; Tanelian, D. L.; Martin, V. S. Liquid Interface in Chemistry and Biology; John Wiley and Sons: New York, 1998. (2) Chang, Y. J.; Castner, E. W., Jr. J. Chem. Phys. 1993, 99, 113. (3) Chang, Y. J.; Castner, E. W., Jr. J. Chem. Phys. 1993, 99, 7299. (4) Palese, S.; Schilling, L.; Miller, R. J. D.; Staver, R.; Lotshaw, W. T. J. Phys. Chem. 1994, 99, 6308. (5) Watry, M. R.; Brown, M. G.; Richmond, G. L. Appl. Spectrosc. 2001, 55, 321A. (6) Tom, H. W. K.; Chang, Y. M.; Kwak, H. Appl. Phys. B 1999, 68, 305. (7) See, for example, Vo¨hringer, P.; Scherer, N. F. J. Phys. Chem. 1995, 99, 2684 and references therein. (8) Fujiyoshi, S.; Ishibashi, T.; Onishi, H. J. Phys. Chem. B 2004, 108, 10636. (9) Zimdars, D.; Eisenthal, K. B. J. Phys. Chem. B 2001, 105, 3993. (10) Kikteva, T.; Star, D.; Lee, A. M. D.; Leach, G. W. Phys. ReV. Lett. 2000, 85, 1906.
