Time-Resolved Sum-Frequency Generation Spectroscopy of

Time-Resolved Sum-Frequency Generation Spectroscopy of Cyclohexane Adsorbed on Ni(111) under Ultrashort NIR Laser Pulse Irradiation. Jun Kubota* ...
0 downloads 0 Views 212KB Size
J. Phys. Chem. B 2005, 109, 20973-20978

20973

Time-Resolved Sum-Frequency Generation Spectroscopy of Cyclohexane Adsorbed on Ni(111) under Ultrashort NIR Laser Pulse Irradiation Jun Kubota,*,† Akihide Wada,† and Kazunari Domen‡ Chemical Resources Laboratory, Tokyo Institute of Technology, 4259 Nagatsuda, Midori-ku, Yokohama 226-8503, Japan, and Department of Chemical System Engineering, The UniVersity of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan ReceiVed: July 5, 2005; In Final Form: September 14, 2005

The adsorption of cyclohexane on Ni(111) was studied by infrared-visible sum-frequency generation (SFG) spectroscopy with and without near-infrared (NIR) pump pulse irradiation. Two adsorption states of cyclohexane were found in the monolayer region, a low-coverage state showing SFG peaks at 2740, 2815, and 2865 cm-1, and a high-coverage state showing peaks at 2740, 2815, and 2905 cm-1. Both states coexisted on the saturated Ni(111) surface. The broad peak at 2740 cm-1 was due to the softened CH stretching mode of the axial CH groups of cyclohexane that point toward the Ni(111) surface. The peaks at 2815 and 2865 (or 2905) cm-1 were due to the symmetric and asymmetric stretching modes of CH2 groups, respectively, that were free from the surface. Irradiation with NIR pulses caused a temporary jump in temperature at the Ni(111) surface and enhanced the intensity of the 2905 cm-1 peak, but weakened the other peaks. This indicates that the temperature jump excited the cyclohexane molecules from the low-coverage state to the high-coverage state. The dynamics of the structural change observed in the adsorbed cyclohexane on NIR irradiation is discussed.

1. Introduction Infrared-visible sum-frequency generation (SFG) spectroscopy is a modern surface vibrational spectroscopy that uses nonlinear optical processes with a pulsed laser.1-3 An advantage of SFG spectroscopy is its short time resolution, which enables the observation of vibrational spectra on the picosecond time scale. Previously, we have used SFG spectroscopy to perform time-resolved observations with pump pulse irradiation.4-8 Irradiation with near-infrared (NIR) pulses causes a temporary temperature jump at the surface, with the surface temperature jumping by a few or several hundred degrees, before recovering on the subnanosecond time scale.4-12 The surface species is thermally excited by the irradiation and sometimes changes its chemical structure. The desorption or decomposition of surface molecules are kinetically limited in the short temperature jump, so that the molecules keep staying on the surface even if the surface is heated to a higher temperature than the conventional desorption temperature.10-12 This technique is regarded as a powerful tool for the identification of unstable states in surface molecules involved in thermal reactions. However, only a few examples have been reported for the time-resolved spectroscopy at surfaces with NIR pump pulses for identification of unstable species and structural change.4-8 The surface temperature jump simultaneously causes the thermal excitation of low-frequency vibrations, which, as a result of anharmonic coupling, results in a frequency shift and band broadening in the fundamental vibrational modes.11 This phenomenon has been well studied on CO adsorbed on metal surfaces.9,11 The chemistry of cyclohexane on metal surfaces is of considerable importance in the catalysis of hydrocarbons and, * Corresponding author. E-mail: [email protected]. Fax: +8145-924-5282. † Chemical Resources Laboratory, Tokyo Institute of Technology. ‡ Department of Chemical System Engineering, The University of Tokyo.

as a result, has been extensively studied over the last few decades.13-28 In particular, the axial CH bonds of cyclohexane that point toward the surface are weakened by the interaction with the metal surface; the softened CH vibration is of interest in surface chemistry. The nature of the softened CH bond is of relevance for the C-H bond activation of cyclohexane, which results in the formation of benzene and other cycloalkenes. SFG spectroscopy has been applied to the adsorption and reaction of cyclohexene and cyclohexadiene on Pt(111);29,30 however, these studies did not involve pump-probe (time-resolved) experiments. On a flat Ni(111) surface, the molecular adsorption of cyclohexane occurs below 200 K, and cyclohexane is desorbed without dehydrogenation on heating.15,25,28 The (x7 × x7)R19.1° structure has been reported in the coverage range between θ ) 0.04 and 0.143 (0.143 is the saturation coverage).25 The carbon skeleton of the adsorbed cyclohexane is known to be almost parallel to the surface, so that the axial CH groups that point toward the surface are significantly perturbed by chemisorption. The perturbed CH bonds are significantly weakened, and the corresponding C-H stretching vibration has been observed at 2720-2730 cm-1 by high-resolution electron energy loss spectroscopy (HREELS)15 and infrared reflection absorption spectroscopy (IRAS)28. At higher coverage, lateral intermolecular interactions are considered to cause the cyclohexane molecules to incline slightly from C3V symmetry.25 The change in the symmetry of the adsorbed cyclohexane with increasing coverage has been thoroughly investigated on a Cu(111) surface by IRAS, and at high coverage, the tilting of the carbon skeleton out of the plane parallel to the surface has been reported.27 The structural change in the cyclohexane layer that occurs with increasing coverage on Pt(111)26 and Cu(111)27 surfaces has also been reported to behave as a phase transition. In this study, we apply time-resolved SFG spectroscopy to cyclohexane adsorbed on a Ni(111) surface. The dependence

10.1021/jp0536706 CCC: $30.25 © 2005 American Chemical Society Published on Web 10/14/2005

20974 J. Phys. Chem. B, Vol. 109, No. 44, 2005

Kubota et al.

of the spectra on coverage and temperature was first examined without pump pulse irradiation. Then the dynamic behavior of cyclohexane under irradiation with near-infrared (NIR) pulses was investigated. The structural change observed in cyclohexane on irradiation is discussed. 2. Experimental Section The laser system used in this work has been described previously.4-8 The SFG spectra were obtained using frequencytunable IR and 532 nm visible pulses generated from a modelocked Nd:YAG laser (pulse width of 35 ps fwhm, repetition rate of 10 Hz). The 532 nm pulses were first converted to frequency-tunable NIR pulses by a parametric generator/ amplifier by using two β-BaB2O4 (BBO) crystals. Frequencytunable IR (1300-3000 cm-1) pulses were then obtained by differential frequency generation (DFG) by using the tunable NIR pulses and the fundamental output of the Nd:YAG (1064 nm) by using an AgGaS2 (AGS) crystal. The energies of the IR and visible pulses at the sample surface were 30 µJ/pulse (at 2000 cm-1) and 100 µJ/pulse, respectively. Both the IR and visible pulses were p-polarized at the sample surface. NIR pump pulses at 1064 nm and energy of 2 mJ/pulse were passed through a variable optical delay and irradiated onto the sample surface. The beam diameters of the IR, visible, and pump pulses were approximately 2, 3, and 4 mm, respectively. The SFG signal was detected by using a photomultiplier tube after passing through optical filters and a monochromator. The Ni(111) sample was handled in an ultrahigh vacuum chamber with a base pressure of 2 × 10-8 Pa. The sample surface was cleaned by cycles of Ar+ bombardment and annealing at 1000 K. The cleanness and long-range structure of the prepared surfaces were checked by using Auger electron spectroscopy (AES) and low-energy electron diffraction (LEED), respectively, by the use of a four-grid retarding field analyzer. A quadropole mass analyzer was used for temperatureprogrammed desorption (TPD) measurements. Reagent-grade cyclohexane was used after dehydration with dried molecular sieves and purification by freeze-pump-thaw cycles in a vacuum system. Exposure in this paper is shown in Langmuir units (L), where 1 Langmuir ) 1 × 10-6 Torr s, 1 Torr ) 133 Pa, and the apparent pressure, as measured by an ionization gauge without calibration, was used to estimate exposure. 3. Results and Discussion 3.1. TPD Spectra of Cyclohexane/Ni(111). The adsorption/ desorption behavior of cyclohexane on Ni(111) was first confirmed by using TPD. The TPD spectra of cyclohexane on Ni(111) at various exposures are shown in Figure 1. The adsorption of cyclohexane was performed at 110 K at the exposure indicated, and the surface was heated at a heating rate of 3 K s-1. Because the mass analyzer used was less sensitive to the parent peak of cyclohexane at m/e ) 84, the most intense fragment peak of cyclohexane at m/e ) 56 (C4H8+) was recorded. Cyclohexane is known to desorb molecularly without any decomposition at Ni(111).13,16 Below 2.5 L, a single TPD peak appeared at 195 K, which corresponded to desorption from the first layer of cyclohexane. Upon increasing the exposure, another peak appeared at 155 K; this peak originated from the condensed multilayer. The complete first layer is known to have a cyclohexane/Ni ratio of 0.143.25 The TPD results presented here are in good agreement with previous reports.25,28 3.2. SFG Spectra of Cyclohexane/Ni(111). The temperature dependence of the SFG spectra of cyclohexane/Ni(111) is shown in Figure 2. 10 L of cyclohexane was exposed to Ni(111) at

Figure 1. TPD spectra of cyclohexane from a Ni(111) surface at various exposures. The heating rate was 3 K s-1. The adsorption of cyclohexane was performed at 110 K and the exposure stated.

Figure 2. SFG spectra of cyclohexane adsorbed on Ni(111) with increasing temperature. Cyclohexane was exposed to Ni(111) at 110 K and an exposure of 10 L. The adsorbed Ni(111) was gradually heated to the stated temperature, and the SFG acquisition was carried out at that temperature.

110 K, and the surface was heated to the temperatures stated. At 110 K, SFG signals were observed at 2920 and 2835 cm-1 with a broad shoulder at 2700 cm-1. Although some peaks or bends can be seen in the spectral region below 2800 cm-1 in the SFG spectra through this study, they were considered to originate from noise of fluctuation of laser power and beam pattern. No common behavior was observed for such noises, and cyclohexane does not have corresponding vibrational mode to these noises. The frequency of the peaks shown here is the apparent peak-top frequency, and detailed spectral curve fits are presented in the next section. The positions of the peaks at 2920 and 2835 cm-1 were different from those observed at higher temperatures, as a result of which they were assigned to cyclohexane in the multilayer. In theory, the inner part of the multilayer has centrosymmetry, and so cannot generate a SFG signal.1-3 However, the molecules near the topmost or bottommost layers in the multilayer have a lesser degree of symmetry and thus can generate SFG signals.

SFGS of Cyclohexane Adsorbed on Ni(111)

J. Phys. Chem. B, Vol. 109, No. 44, 2005 20975

Figure 3. Schematic drawing of cyclohexane adsorbed on Ni(111).

With increasing temperature, the multilayer begins to desorb, with the first layer remaining on the surface at 150 K, as can be seen in the TPD spectra of Figure 1. At 150 K, peaks at 2905, 2865, and 2815 cm-1, accompanied by a broad feature at 2700 cm-1, were observed. These peaks can be attributed to the saturated first layer. Upon further heating, the peak at 2905 cm-1 was the first to decrease in intensity, while the peak at 2865 cm-1 disappeared. The peak at 2815 cm-1 and the broad shoulder at 2700 cm-1 also gradually decreased in intensity. The difference in behavior of the peaks at 2905 and 2865 cm-1 indicates that the first layer consists of two types of adsorbed species. The broad feature at 2700 cm-1 was assigned to the softened CH stretching mode of cyclohexane in the first layer.15,28 The signal from the softened band in the SFG spectra appears to be weaker than that observed in IRAS28 and HREELS.15 SFG susceptibility is a function of the product of the transition dipole and the Raman susceptibility,1-3 such that the intensity of bands observed by SFG is sometimes different from that observed by IRAS and HREELS. Furthermore, the vibrational bands in SFG spectra interfere with each other, resulting in unique spectral shape of SFG. Details of the spectral shape and intensity of each component are discussed in the next section. The peaks at 2905 and 2865 cm-1 are attributed to the asymmetric stretching mode of the free CH2 groups.26-28 The peak at 2815 cm-1 is assigned to the symmetric stretching mode of the free CH2 groups.26-28 Figure 3 shows an illustration of adsorbed cyclohexane with its corresponding CH stretching modes. It was suspected that the 2865 cm-1 band was attributable to the equatorial CH because the frequency of this band is similar to that of the equatorial CH mode.28 However, as seen in the spectrum at 170 K, the 2905 cm-1 peak was very weak and the 2865 and 2815 cm-1 peaks were major. Thus, the 2865 cm-1 peak should be assigned to the asymmetric stretching mode of the free CH2 groups rather than the equatorial CH mode. It is unlikely that the peak of the equatorial CH mode appears without the appearance of the asymmetric stretching mode of the CH2 groups. The exposure dependence of the SFG spectra of cyclohexane/ Ni(111) is displayed in Figure 4. The peaks at 2865 and 2815 cm-1 first appeared at 0.2 L, with the peak at 2905 cm-1 appearing at higher coverage. Thus, the peaks at 2815 and 2865 cm-1 were considered to originate from low-coverage states of the adsorbed cyclohexane. The peak at 2865 cm-1 saturated at around 0.2 L, indicating that the coverage of cyclohexane in the low-coverage state was approximately 10% of that in the monolayer. Between 2.5 and 5 L, the surface was covered with the monolayer, as seen in Figure 1, and the SFG spectrum was almost the same as that at 150 K, shown in Figure 2. At higher exposures, a peak at 2905 cm-1 is observed; this originates from cyclohexane in the high-coverage state. The symmetric CH2 and softened CH modes of cyclohexane in the high-coverage state were observed at 2815 and 2700 cm-1, respectively; these are the same frequencies that occur in the low-coverage state.

Figure 4. SFG spectra of cyclohexane adsorbed on Ni(111) at 110 K and various exposures. The SFG acquisition was carried out at 110 K.

The effect of increasing coverage on cyclohexane has been extensively studied on Pt(111)26 and Cu(111)27 by means of IRAS. On these surfaces, the CH2 asymmetric stretching mode has been reported to a shift in frequency with increasing coverage. The frequencies of the other peaks are, however, little affected by the coverage. The spectra observed in this study seems to be similar to that for Cu(111) and Pt(111). While the two kinds of adsorbed cyclohexane behave as a phase transition on Cu(111) and Pt(111), the two states, which correspond to low and high coverage, coexist at saturation coverage on Ni(111). The saturation coverage of cyclohexane on Ni(111) is known to be 0.143 with respect to the surface Ni atoms, and the (x7 × x7)R19.1° structure has been reported to be between θ ) 0.04 and 0.143.25 In the (x7 × x7)R19.1° island, the cyclohexane molecules are highly (closed) packed and intermolecular interactions and steric hindrance cause the symmetry to be less than C3V.25 We consider that the low- and high-coverage states correspond to the isolated and highly packed molecules, respectively. Therefore, the high-coverage state observed in this study can be attributed to the (x7 × x7)R19.1° islands. The low-coverage state can be assigned to the relatively free molecules outside of the (x7 × x7)R19.1° domains. This suggests that, even at saturation coverage, the molecules in the low-coverage state still remain at the domain boundaries. On the Ni(111) surface, it was found that all of cyclohexane molecules cannot be compressed into the closed pack (x7 × x7)R19.1° structure, even at the saturation coverage. The coexistence of two kinds of structures might have an energetic advantage. Figure 5 illustrates the domains for the low- and high-coverage states of cyclohexane on Ni(111). 3.3. Simulation of SFG Spectra. SFG light is generated by a nonlinear optical process, as a result of which the spectral shape is theoretically different from those of IRAS or HREELS, which sometimes show a dispersive spectral shape.1-3 Thus, the apparent peak position and intensity do not directly indicate the correct resonant frequency and the magnitude of SFG susceptibility for each vibrational mode.

20976 J. Phys. Chem. B, Vol. 109, No. 44, 2005

Kubota et al.

Figure 5. Sketch of ordered (x7 × x7)R19.1° and disordered domains of cyclohexane on Ni(111). The relatively free molecules in the disordered domains were moved by the pumping as discussed in section 3.4.

Figure 6. Simulation of the SFG spectral shape for a monolayer of cyclohexane adsorbed on Ni(111) at 110 K. Each component of the vibrational mode is also shown in this figure.

The intensity of SFG light, ISF, can be expressed as the function of IR frequency, ω, by

ISF ∝ |χNR

(2)

+

∑n ω

n0

χn(2)



e

- ω + iΓn

n

|

2

(1)

where χNR(2) is the nonresonant term of second-order nonlinear susceptibility. In the resonant term, χn(2), ωn, Γn, and θn represent the resonant term of second-order nonlinear susceptibility, resonant frequency, dumping factor, and the relative phase of the nth mode, respectively. The resonant SFG susceptibility corresponds to the product of the transition dipole and the Raman tensor. The simulation was performed for a full monolayer of cyclohexane on Ni(111) at 150 K, as shown in Figure 6. Each vibrational component, |(χn(2))/(ωn0 - ω + iΓn)|2, is also depicted in Figure 6. The spectral shape observed fits well with the four resonant vibrational modes. It is noted that, essentially, the broad band at 2700 cm-1, attributed to the softened CH bond, has a strong SFG susceptibility, but that this appears weak in the real spectrum because of interference (relative phase shift) with the other bands. The difference in the shape of the spectra obtained by using SFG and those obtained with IRAS28 and HREELS15 are caused by interference among the modes, which occurs in the nonlinear process, as shown in eq 1.

The resonance frequencies of the vibrational modes obtained by the simulation are listed in Table 1, alongside the results of previous studies on cyclohexane adsorbed on various metal surfaces carried out by using IRAS and HREELS. The obtained resonance frequencies are slightly different from the apparent frequencies at the peak tops. 3.4. Dynamic Behavior under Pump Irradiation. Transient spectra of cyclohexane/Ni(111) under NIR pump pulse irradiation are shown in Figure 7. A Ni(111) surface covered with a cyclohexane multilayer was annealed at 150 K prior to the measurements; this formed monolayer of cyclohexane on the surface. According to a simulation of the surface temperature profile by using a heat-diffusion equation,5 the temperature at the surface reached a maximum at ∼50 ps upon pump pulse irradiation. Irradiation increased the intensity of the peak at 2905 cm-1, while it appeared to weaken the other peaks. In particular, the peak at 2865 cm-1 showed a significant decrease in intensity. To clarify the changes caused by irradiation, the timedependence of the signal intensity at each resonant frequency was recorded, as shown in Figure 8. Irradiation increased the intensity of the signal at 2905 cm-1, as can be seen in Figure 8a. The increase in the SFG signal can be assigned to an increase in molecular density, a change in molecular orientation, or an ordering of the long-range structure with spectral sharpening. We consider that the change observed is caused by an increase in the number of molecules that have a peak at 2905 cm-1. Cyclohexane is known to adsorb almost parallel to the surface, so that NIR pumping cannot enhance the CH2 symmetric signal by changing the molecular orientation. Further ordering in the (x7 × x7)R19.1° islands is not expected to occur, and no sharpening of the 2905 cm-1 peak was observed. A weakening of the SFG peaks under NIR pumping has been generally reported for most surface molecules because of anharmonic coupling with the thermally excited low-frequency modes.8,9,11 Because SFG intensity is proportional to the square of molecular density,1-3 a change of 10% in the SFG peak corresponds to an increase in the amount of molecules of ∼5%. The intensity of the 2905 cm-1 peak, however, might be suppressed by thermal broadening, which originates in anharmonic coupling with low-frequency modes. This means that the increase in the amount of molecules may be considerably higher than 5%. This conclusion would mean that the number of cyclohexane molecules in the high-coverage state was increased by the temperature jump. As can be seen in Figure 2, the cyclohexane molecules in the low-coverage state are more stable than those in the high-coverage state. Thermal excitation, initiated by pump pulse irradiation, converted cyclohexane from the low coveragestate into the high-coverage state. The other peaks, at 2865, 2815, and 2740 cm-1, were weakened by the irradiation, as can be seen in Figure 8b-c. For the population of the high-coverage state to increase, the number of cyclohexane molecules in the low-coverage state must have been correspondingly decreased by the irradiation. Thus, the change of the signal at 2865 cm-1 was attributed to the transformation between the two states. At the same time, these peaks would also have been suppressed by the thermal effect resulting from anharmonic coupling with the lowfrequency modes. The weakening of the signals at 2815 and 2740 cm-1 was thus considered to be due to the thermal weakening but not due to the transformation between the two states. Because the signal change is caused by both thermal effect and transformation, it is not possible to quantitatively estimate the magnitude of either of them.

SFGS of Cyclohexane Adsorbed on Ni(111)

J. Phys. Chem. B, Vol. 109, No. 44, 2005 20977

TABLE 1: Frequencies (cm-1) of CH Stretching Modes of Cyclohexane Adsorbed on Various Metal Surfaces Ni(111) modes

multilayer

νas(CH2) νs(CH2) ν(equatorial CH) ν(softened CH)

2920 2835

method reference

SFG this work

a

Ni(111)

low θa

high θa

2860 2822

2903 2822

∼2740

∼2740

Ni(111)

Cu(111)

Pt(111)

monolayer

phase A

phase B

1L

2L

2904 2845 2892 2730

2925 2851 2851 2820 2776

2903 2843 2887 2600

2922 2853

2720

2905 2848 2895 2770

HREELS 15

IRAS 27

IRAS 28

} 2900

IRAS 26

Resonance frequencies obtained from the fitting in the eq 1, but not the apparent peak top frequencies.

Figure 7. Time-resolved SFG spectra of cyclohexane adsorbed on Ni(111) at 110 K. The multilayer of cyclohexane on Ni(111) was annealed at 150 K to form a complete monolayer prior to the measurements.

The time profiles of the SFG intensities, shown in Figure 8, were similar to those of the surface temperature.5,9-12 In our previous work, the temperature jump of this apparatus was estimated to be 300 K by the direct observation of thermal shift of the O-D band of OD/NiO/Ni(111) under the pump irradiation and by the theoretical calculation on the one-dimensional heat diffusion equation.5 Although the inhomogeneity and fluctuation of the beam pattern and energy made an accurate estimate of the surface temperature jump difficult to determine experimentally, we considered the temperature jump to be 300 K in the present experiments. From the jump of 300 K, we can consider that the energy difference between the low- and highcoverage states of the cyclohexane molecule is approximately 10-30 kJ mol-1, which is the same as that published previously.4,5,8 In conclusion, cyclohexane molecules in the low-coverage state were changed by the irradiation to the high-coverage state. The low-and high-coverage states corresponded to the relatively free molecules and the molecules in the (x7 × x7)R19.1° structure, respectively. Therefore, NIR pumping had the effect of transformation of cyclohexane molecules between the two states, as illustrated in Figure 9. This transformation was not observed when using conventional heating (Figure 2); this indicates that, in order to stimulate the molecules into undergoing this transformation, the surface must have been instantaneously heated above the desorption temperature of cyclohexane. From the entropy point of view of thermodynamics, an ordered structure would become disordered at higher temperatures. This suggests that the transition from the low-coverage state of cyclohexane to the high-coverage state has a disadvantage in the entropy because the (x7 × x7)R19.1° structure seems to be more ordered than the relatively free molecules in the low-coverage state. However, the experimental results show

Figure 8. Temporal changes in the SFG signals at 2905 (a), 2865 (b), 2815 (c), and 2740 cm-1 (d). The measurements were performed at 110 K. The multilayer of cyclohexane on Ni(111) was annealed at 150 K to form a complete monolayer prior to the experiment.

the opposite change. For this transition between the two states, the change of Gibbs energy, ∆G, is expressed as ∆G ) ∆H T∆S, where ∆H and ∆S are enthalpy and entropy changes, respectively, and T is temperature. Assuming that the species in the low-coverage state are in equilibrium with those in high-

20978 J. Phys. Chem. B, Vol. 109, No. 44, 2005

Kubota et al. states corresponded to relatively free molecules (stable) and (x7 × x7)R19.1° (unstable) domains on the surface, respectively. NIR pump pulse irradiation enhanced the signal intensity of the CH2 asymmetric mode of cyclohexane in the high-coverage state. The NIR pumping thermally excited the stable cyclohexane in the low-coverage state, converting it to the unstable highcoverage state. References and Notes

Figure 9. Motion of cyclohexane between ordered domains and the corresponding potential profile.

coverage states, the ratio of two kinds of species depends on temperature as

θB/θA ) exp(-∆G/RT) ) exp(∆S/R) exp(-∆H/RT) (2) where θA and θB represent the amount of adsorption for the species in low- (A) and high- (B) coverage states, respectively, and R is the gas constant. This indicates that the equilibrium is changed by the enthalpy change, ∆H, with changing temperature. As seen in Figure 2, the desorption temperature from the low-coverage state is higher than that from the high-coverage state by 20 K. This suggests that the molecules in the lowcoverage state are more stable than those in the high-coverage state, indicating ∆H is positive (∆H > 0). If the transition from the low-coverage state to the high-coverage state has ∆S < 0 because of the ordering, the ratio of θB/θA is smaller than 1 even for the infinite temperature. The fact that most of the molecules were populated in the high-coverage state even before the pumping suggested that the entropy of the high-coverage state is not smaller than that of the low-coverage state. Therefore, the temperature jump by NIR irradiation causes more population of cyclohexane molecules into the high-coverage state. In an attempt to convert stable molecules to unstable species, we have studied the thermal excitation of surface molecules by NIR pulse irradiation in the formate/NiO(111)4,5 and CO/Ni(111)8 systems. In the present study, irradiation clearly shifted the equilibrium between the two states. The present study is also an example of NIR irradiation causing molecular surface migration without desorption and the use of surface-sensitive spectroscopy to study this phenomenon. 4. Conclusion Cyclohexane adsorbed on Ni(111) was studied by SFG spectroscopy with and without NIR pump pulse irradiation. Two types of adsorbed cyclohexane were observed at low- and highcoverage regions; these showed different frequencies in their CH2 asymmetric stretching modes. The low- and high-coverage

(1) Laser Spectroscopy and Photochemistry on Metal Surfaces, Part I and II; Dai, H.-L., Ho W., Eds.; World Scientific: Singapore, 1995. (2) The Principles of Nonlinear Optics; Shen, Y. R., Ed.; John Wiley and Sons: New York, 1987. (3) Laser Applications in Surface Science and Technology; Rubahn, H.-G., Ed.; John Wiley & Sons: Chichester, U.K., 1999. (4) Bandara, A.; Kubota, J.; Onda, K.; Wada, A.; Kano, S. S.; Domen, K.; Hirose, C. J. Phys. Chem. B 1998, 102, 5951. (5) Hirose, C.; Bandara, A.; Katano, S.; Kubota, J.; Wada, A.; Domen, K. Appl. Phys. B 1999, 68, 559. (6) Kubota, J.; Wada, A.; Domen, K.; Kano, S. S. Chem. Phys. Lett. 2002, 362, 476. (7) Kubota, J.; Wada, A.; Domen, K.; Kano, S. S. Chem. Phys. Lett. 2003, 377, 217. (8) Kubota, J.; Yoda, E.; Ishizawa, N.; Wada, A.; Domen, K.; Kano, S. S. J. Phys. Chem. B 2003, 107, 10329. (9) Germer, T. A.; Stephenson, J. C.; Heilweil, E. J.; Cavanagh, R. R. J. Chem. Phys. 1994, 101, 1704. (10) Schro¨der, U.; Guyot-Sionnest, P. Surf. Sci. 1999, 421, 53. (11) Bonn, M.; Hess, Ch.; Funk, S.; Miners, J. H.; Wolf, M.; Ertl, G. Phys. ReV. Lett. 2000, 84, 4653. (12) Funk, S.; Bonn, M.; Denzler, D. N.; Hess, Ch.; Wolf, M.; Ertl, G. J. Chem. Phys. 2000, 112, 9888. (13) Nieuwenhuys, B. E.; Hagen, D. I.; Rovida, G.; Somorjai, G. A. Surf. Sci. 1976, 59, 155. (14) Madey, T. E.; Yates, J. T., Jr. Surf. Sci. 1978, 76, 397. (15) Demuth, J. E.; Ibach, H.; Lehwald, S. Phys. ReV. Lett. 1978, 40, 1044. (16) Rubloff, G. W.; Luth, H.; Demuth, J. E.; Grobman, W. D. J. Catal. 1978, 53, 423. (17) Lehwald, S.; Ibach, H. Surf. Sci. 1979, 89, 425. (18) Hoffmann, F. M.; Felter, T. E.; Thiel, P. A.; Weinberg, W. H. J. Vac. Sci. Technol. 1981, 18, 651. (19) Herz, R. K.; Gillespie, W. D.; Peterson, E. E.; Somorjai, G. A. J. Catal. 1981, 67, 371. (20) Hoffmann, F. M.; Felter, T. E.; Thiel, P. A.; Weinberg, W. H. Surf. Sci. 1983, 130, 173. (21) Waddill, G. D.; Kesmodel, L. L. Chem. Phys. Lett. 1986, 128, 208. (22) Chesters, M. A.; Parker, S. F.; Raval, R. J. Electron Spectrosc. Relat. Phenom. 1986, 39, 155. (23) Raval, R.; Pemble, M. E.; Chesters, M. A. Surf. Sci. 1989, 210, 187. (24) Raval, R.; Chesters, M. A. Surf. Sci. 1989, 219, L505. (25) Huber, W.; Zebisch, P.; Bornemann, T.; Steinru¨ck, H.-P. Surf. Sci. 1990, 239, 353. (26) Chesters, M. A.; Gardner, P. Spectrochim. Acta, Part A 1990, 46, 1011. (27) Raval, R.; Parker, S. F.; Chesters, M. A. Surf. Sci. 1993, 289, 227. (28) Yonekura, H.; Nozoye, H. Surf. Sci. 2001, 471, L134. (29) Su, X.; Kung, K. Y.; Lahtinen, J.; Shen, Y. R.; Somorjai, G. A. J. Mol. Catal., A 1999, 141, 9. (30) McCrea, K. R.; Somorjai, G. A. J. Mol. Catal., A 2000, 163, 43.