Spatial overlapping between UV and VUV FEL

The spatial overlap of the pump and probe laser beams at the interaction region is crucial in two-color experiments. We have taken the advantage of the imaging apparatus and overlapped the two laser beams precisely by monitoring the image and TOF of photoions.

Pyrazine in the molecular beam was ionized by one-color 2-photon ionization using either pump or probe laser. The intensities of both lasers were increased for the alignment. The photoion signals were observed by changing the polarity of the voltages applied to the electrodes from that for electron detection. The crossing point of the laser beam with the flight tube axis was determined from the TOF of photoions. This was because the arrival time of the ion depends on the distance of the ionization point from the repeller. As the ionization point is closer to the MCP detector, it is farther away from the repeller and the ion gains less kinetic energy in the electric field. This leads to a smaller speed of ion in the field free region and consequently a longer flight time of the ion. Figure 5.8 (a) and (b) show the mass spectra of pyrazine observed by 161-nm FEL radiation and 260-nm UV radiation, respectively. The parent ion signal is used as an indicator of the overlap of the laser beams. Since the VUV-FEL beam cannot be moved in the laboratory, the position of the UV laser beam was adjusted to overlap with the VUV beam.

To overlap the two beams in the plane perpendicular to the flight axis, the arrival positions of ions on the MCP detector were imaged by the CCD camera. Figure 5.9 (a) and (b) show the ion image of pyrazine observed by 161-nm FEL radiation and 260-nm UV radiation, respectively. The UV laser beam direction was changed by moving a focusing lens until the ion image created by a UV laser overlaps with that by a VUV-FEL.

116

5.3.4 Results

(A) S1–S0 Internal Conversion (IC)

Figure 5.10 (a) shows a time profile of the pyrazine ion observed by (1+1’) REMPI with 260-nm pump and 161-nm probe pulses. The pump pulse excited pyrazine from S₀ to S₂. It is ideal that each pulse does not produce any ionization signal to have high contrast ratio of the two-color signal against the one-color signals. However, the pump pulse created a small amount of one-color two-photon ionization signal. 161 nm is also resonant with an electronic transition.³⁶ Therefore, one-color two-photon ionization signals were observed for both pump and probe pulses. However, the sum of these one-color signals is still 3–4 times smaller than the two-color signal, and was already subtracted in Fig. 5.10. Least-squares fitting of the time profile by assuming a single exponential decay yielded a lifetime of ~ 20  3 ps. Previously, TRPEI by Horio et al.²⁷ using 264-nm pump and 198-nm probe have shown that the S₂ state dephases to S1 in ~ 23 fs, and TRPEI by Wang et al. reported that these S1 state decays in ca.

22 ps.^25,26 Owing to a limited time resolution (< 5 ps) of this experiment, the ultrafast S₂–S1

internal conversion within 30 fs cannot be observed. The time constant of the decay of ionization signal is in good agreement with the lifetime of hot S₁ state reported by Wang et al.

The lifetime of these highly vibrationally excited S1 molecules are determined by IC to the S0

state.¹⁶

Figure 5.10 (b) shows a photoelectron image observed at the delay time of about 5 ps. In a [1+1’] pump and probe ionization experiment with laser polarization vectors parallel to each other, the photoelectron angular distribution (PAD) is characterized by:

( ) [ ( ) ( )]

where Pn is the n-th order Legendre polynomial, is the angle between the photoelectron velocity and the linear polarization of the laser, βn is anisotropy parameter. The least-squares fitting of the photoelectron angular distribution extracted from Fig. 5.10 (b) provides 2 =

117

−0.15 ± 0.05 and 4 = −0.03 ± 0.10. The low anisotropy is typical for valence excited states of molecules.

The photoelectron kinetic energy distribution (PKED) extracted from Fig. 5.10 (b) is showed in Fig. 5.11 (a). Comparing with the previous study with 264-nm UV pump and 198-nm DUV probe (red curve in Fig. 5.11a), the advantage of 161-nm VUV radiation over 198-nm DUV light is clear. With the latter, the observed distribution is truncated around its maximum at the electron binding energy (EBE) of 10.96 eV. In contrast, the entire Franck-Condon envelope is observed with the former. The high-resolution He(I) photoelectron spectrum of jet-cooled pyrazine adopted from Oku et al.³⁷ is also shown in Fig.

5.11 (a). Their spectrum indicates that the first ionization energy is 9.29 eV, while our PKED exhibits a peak at ca. 10.6 eV. The S₂  S1 IC transforms the difference of the electronic energies of S₂ and S₁, 0.86 eV, into the vibrational energy in S₁. This energy is approximately conserved upon ionization, which shifts the peak nearly 1 eV (Fig. 5.11 (b)). The S1 state has an electronic configuration of (n,  *), whereas the D0 and D1 states are of n^–1 and ^–1. Therefore, frozen core approximation predicts that ionization from S1 predominantly occurs to D₀. This experiment clearly shows that ionization from S₁ to D₁ is actually minor, if present.

(B) S1–T1 Intersystem Crossing (ISC)

Figure 5.12 (a) shows the ionization signal upon the 324-nm pump and 161-nm probe excitation. The signal exhibits an initial decay, due to S₁-T₁ ISC, and reaches a plateau at later time: the least squares fitting of the decay curve yields the lifetime of S₁ state as 114  17 ps, which is consistent with the previous reports (110 ps).^31,33,34 Figure 5.12 (b) shows the PKEDs and the photoelectron (Abel-inverted) images observed at 8, 58 and 408 ps. A 8 ps, photoelectron signal from the S₁ state appears clearly as the outer ring, whereas it diminished at longer time delays (408 ps) due to S1–T1 ISC. Previous studies concluded that the S1 state dephases exclusively to T₁. Therefore, we anticipated that ionization occurs with similar

118

efficiencies from S1 and T1 at 161 nm, and that the total ionization intensity does not diminish.

The observed decay, which contradicts with our expectation, may suggest that T₁ state has some deactivation processes and its population is lost. From the energy resolution of our PEI spectrometer and the broad bandwidth (0.1 eV) of SCSS, the energy resolution of PKED is estimated to be 0.2-0.3 eV. Accordingly, no distinct vibrational structure is discernible.

5.4 Conclusion

The new PEI spectrometer described in earlier chapters has been used for PEI with a VUV-FEL. The background electrons observed with a conventional three-plate electrostatic lens were almost completely eliminated by our new lens. Owing to the characteristics of SASE, the resolution was inferior with VUV-FEL to that with He(I). However, we utilized the short pulse duration of the VUV-FEL for the first time-resolved photoelectron imaging of polyatomic molecules by synchronization with a fs tunable UV laser. The time evolution of photoelectron intensity and photoelectron images clearly exhibited the features of ultrafast S₁– T1 intersystem crossing and S1–S0 internal conversion in pyrazine in a supersonic molecular beam. This study clearly demonstrates the feasibility of picosecond TRPEI of photo-induced dynamics of large polyatomic molecules with an FEL and a tunable UV laser.

119

Figure 5.1 Schematic illustration of the principle of a pump-probe experiment with ionization for probing. The pump pulse excites an electronic state and after some delay (eg. t₁, t₂) the probe pulse ionizes the excited neutral molecule.

120

Figure 5.2 Schematic configuration of the SCSS test accelerator extracted from ref. 13.

121

Figure 5.3 Photoelectron image observed for photoionization of a pulsed supersonic beam of Kr at a stagnation pressure of 0.1 MPa with FEL (60 nm) using a conventional three-plate electrostatic lens: (a) before and (b) after background subtraction. (c) Photoelectron image of Kr measured with newly-design electrostatic lens. No background image was subtracted.

122

Figure 5.4 Comparison of PKE distributions of Kr in photoionization by He(I) radiation (solid) and FEL radiation (wavelength: 58.4 nm) (dashed).

123

Figure 5.5 (a) Symmetrized photoelectron image observed for photoionization of a pulsed supersonic beam of 10 % pyridine seeded in He with FEL ( = 58.4 nm). The left half is the raw image, whereas the right half is the slice image obtained by taking the inverse Abel transform. (b) Comparison of the photoelectron spectra extracted from the image obtained using FEL in Fig. 8(a) (dash) with the photoelectron spectrum obtained using He(I) (solid) measured with a 512  512 pixels CCD camera.

124

Figure 5.6 Schematic energy diagram of (a) S₂–S1 internal conversion (IC) and (b) S₁–T1

intersystem crossing (ISC) in pyrazine.

125

Figure 5.7 The setup for pump-probe experiment using a fs UV laser and a VUV free-electron laser.

126

Figure 5.8 Mass spectrum of pyrazine obtained from: (a) 2-photon ionization by 161-nm FEL;

(b) 2-photon ionization by 260-nm UV laser.

Int ens ity (a. u. )

p-C₄H₄N⁺

127

Figure 5.9 Ion image of 10% pyrazine seeded in He observed by: (a) FEL radiation; (b) UV radiation.

128

Figure 5.10 (a) (1+1’) REMPI signal with 260-nm pump and 161-nm probe light. (a) Time profile of the pyrazine ion signal and least-squares fit of a single exponential decay curve. (b) Observed photoelectron image at the delay time of about 5 ps. A one-color background signal was already subtracted from the image.

129

Figure 5.11 (a) Photoelectron kinetic energy distribution in He(I) photoelectron spectroscopy of the ground-state pyrazine (black) (ref. 37), 264-nm UV pump and 198-nm DUV probe experiment (red) (ref. 27) and 260-nm UV pump and 161-nm VUV probe (blue) (this work).

(b) Schematic energy diagram of ionization process. UV absorption spectrum of pyrazine vapor at room temperature and time-spectrum of VUV FEL are shown as insets.

130

Figure 5.12 (a) (1+1’) REMPI with 324-nm pump and 161-nm probe. (a) Time profile of pyrazine ion signal and least-squares fit of a single exponential decay curve. (b) Photoelectron kinetic energy distribution observed at various time delays. The inset shows the corresponding photoelectron (Abel-inverted) images.

131

16 Yamazaki, I. et al. Faraday Discuss. Chem. Soc. 1983, 75, 395.

17 Schneider, R.; Domcke, W. Chem. Phys. Lett. 1988, 150, 235.

132

23 Stock, G.; Schneider, R.; Domcke, W.; J. Chem. Phys. 1989, 90, 7184.

24 Seel, M.; Domcke, W. J. Chem. Phys. 1991, 95, 7806.

25 Wang, L.; Kohguchi, H.; Suzuki, T. Faraday Discuss. 1999, 113, 37.

26 Stert, V.; Farmanara, P.; Radloff, W. J. Chem. Phys. 2000, 112, 4460.

27 Horio, T. et al. J. Am. Chem. Soc. 2009, 131, 10392.

28 Suzuki, Y. et al. J. Chem. Phys. 2010, 132, 174302.

29 Frad, A. et al. J. Chem. Phys. 1974, 60, 4419.

30 Lahmani, F. et al. J. Chem. Phys. 1974, 60, 4431.

31 McDonald, D. B.; Fleming, G. R.; Rice, S. A. Chem. Phys. 1981, 60, 335.

32 Knee, J. L.; Doany, F. E.; Zewail, A. H. J. Chem. Phys. 1985, 82, 1042.

33 Tsubouchi, M.; Whitaker, B. J.; Suzuki, T. J. Phys. Chem. A 2004, 108, 6823.

34 Tsubouchi, M. et al. Phys. Rev. Lett. 2001, 86, 4500.

35 Suzuki, T. Whitaker, B. J. Int. Rev. Phys. Chem. 2001, 20, 313.

36 Walker, I. C.; Palmer, M. H. Chem. Phys. 1991, 153, 1669.

37 Oku, M. et al. J. Phys. Chem. A 2008, 112, 2293.

133

Chapter 6 Summary

I have employed a photoelectron imaging (PEI) spectrometer with a He(I) light source to observe speed and angular distributions of photoelectrons. PEI experiment using He(I) radiation presented in this thesis is the first attempt reported so far. The challenge in coupling an incoherent VUV light source with PEI was to overcome the interference of the numerous background photoelectrons emitted from the instrument by scattered He(I) radiation. I have examined various designs of the acceleration electrodes experimentally and computationally and identified the major source of background photoemission to be from a repeller plate.

Based on this finding, I designed a new electrostatic lens that places the repeller away from the ionization region and adds a retardation field to intercept background photoelectrons from the repeller toward the imaging detector. The resolution of PEI is enhanced by good spatial focusing of the electron trajectories starting from a finite ionization volume and also by high imaging resolution. As for the former, I used a number of electrodes behind the extractor plate to make the strength of the acceleration field varies more gradual than the simplest Eppink-Parker design using three electrodes. The energy resolution (E/E) better than 1 % was achieved even for a large ionization volume (i.e. of the order of millimeters). As for the latter, I used a super-resolution (4096  4096 pixels) imaging system developed in our laboratory (Chemical Dynamics Laboratory at RIKEN). Combining these two instruments, I obtained an energy resolution of 0.735 % at 5.461 eV (FWHM: 40 meV) with He(I) radiation.

This does not represent the ultimate resolution achievable with the PEI spectrometer. It is possible to increase the resolution by one order of magnitude, if the photon beam can be more tightly focused. With the He(I) PEI spectrometer, I have measured the photoelectron image

134

of benzene and pyridine in supersonic molecular beams. In these experiments, time-gated measurement with a pulsed beam provided a high signal contrast with respect to the background photoionization signal from residual water vapor in the ionization chamber. On the other hand, the ultimate imaging resolution with centroiding calculation is not obtainable at the high signal count rate of He(I) PEI owing to spatial overlap between the light spots of two electron impacts on the detector. A camera with a higher frame-rate and a more rapid digital image processing would overcome this problem. The photoelectron anisotropy parameters determined for benzene and pyridine are in good agreement with the literature values. The anisotropy parameters of photoelectrons from pyridine are first reported for the electron binding energies over 14 eV.

The PEI spectrometer developed for He(I) radiation is also applicable to photoelectron spectroscopy with other vacuum ultraviolet (VUV) sources such as synchrotron radiation, free electron laser (FEL) and high harmonics of the output from ultrashort pulsed lasers. I demonstrated that a VUV or EUV laser is useful for pump–probe photoelectron spectroscopy to observe ultrafast electronic dynamics. Time-resolved photoelectron imaging using a tunable femtosecond ultraviolet laser and a vacuum UV free-electron laser is performed.

Ultrafast intersystem crossing on S1 to T1 and internal conversion from S1 to S0 in pyrazine in a supersonic molecular beam were clearly observed in the temporal profiles of photoelectron intensity and photoelectron images. The VUV radiation allowed us to observe the entire Franck-Condon envelope in photoionization from a transient electronic state; this was not possible with UV-UV two-color experiment in the laboratory. This study clearly demonstrated exciting opportunities made possible with a SASE FEL and a tunable UV femtosecond laser for picosecond TR-PEI experiment on photoinduced chemical dynamics of large polyatomic molecules. On the other hand, the energy resolution of PEI was higher with He(I) radiation than VUV-FEL radiation at 58.4 nm, which is ascribed to large effective bandwidth (0.1 eV) of the laser radiation from SASE FEL.