Comparison of PEI with He(I) and SASE-FEL

As discussed in detail in Chapter 3, the resolution of He(I) PEI is limited by a large ionization volume. On the other hand, the VUV-FEL beam can be focused to a diameter of ~ 0.1 mm. Our trajectory calculations indicate that the electron lens system is capable of providing the energy resolution E/E of 0.04 % at 5.461 eV, if an ionization volume is 0.1 mm  3.4 mm. The resolution of 0.04 % reaches the ultimate resolution, 0.06 %, obtainable with the best commercial MCP (70 mm in diameter and 10 m in pore size). The proof of the ultimately high resolution of our PEI apparatus awaits an experiment with nanosecond UV lasers.

However, as shown in Figure 5.4, photoelectron kinetic energy distributions of Kr measured with FEL radiation (wavelength 58.4 nm) is broader than that with measured He(I) radiation under the same experimental condition; the FWHM of the peak at 7.22 eV (indicated by * in Fig. 5.4) is 246 meV with He(I) and 298 meV with FEL. In both cases, a high-resolution CCD camera (2048  2048) was used without centroiding calculation. The inferior resolution with FEL, in spite of its smaller ionization volume, is ascribed to a large effective bandwidth (0.1 eV) of SASE.¹⁴

Figure 5.5 shows a photoelectron image of pyridine with FEL at 58.4 nm and the photoelectron spectra extracted from the image. Since FEL radiation is polarized, the photoelectron distribution is cylindrically symmetric about its polarization direction (indicated by the arrow in the Fig.5.5). Similarly with Kr, the image shown in Fig. 5.5 exhibit broader structures than Fig. 4.3 measured with He(I).

One of the advantages of VUV-FEL over He(I) is much higher photon density. However, the pulse energy of the SCSS radiation is too high for single-photon ionization experiments.

We had to reduce pulse energy with metal and gas filters to avoid multiphoton ionization.

Since SCSS is operated at a low repetition rate (20–30 Hz), the reduction of pulse energy diminishes the advantage of VUV-FEL over He(I) for conventional UPS. In the next section,

112

we utilize another important advantage of VUV-FEL, namely the ultrashort pulse duration, for pump-probe experiments.

5.3 (1 + 1’) pump-probe experiment of pyrazine

5.3.1 Introduction

Pyrazine (C₄H₄N₂, 1,4-diazabenzene) is a heterocyclic aromatic molecule that belongs to the D2h point group. The S2-S0 absorption spectrum of pyrazine in the deep UV region (230–

270 nm) exhibits a broad feature,¹⁶ which implies ultrafast decay of the S₂ state. Low-lying conical intersection (CI) between the S2 and S1 potential energy surfaces was identified by semiempirical calculations.¹⁷ Thereafter, extensive ab initio calculations have been performed to characterize S2/S1 CI more precisely,^18,19 and it has been established that the S₂/S₁ CI is located at the bottom of the S₂ diabatic PES. Thus, the ultrafast decay from S₂ has been established as the internal conversion through CI, as shown in fig. 5.6 (a). For observation of this ultrafast dynamics in real time, TRPES of pyrazine has been proposed by theoretical simulations.20 , 21 , 22

However, the time constant of the internal conversion is predicted to be only less than 30 fs.^23,24 Such high time-resolution has been hardly realized in the pump-probe experiments in deep UV region. For instance, Wang et al.²⁵ performed the first femtosecond [1+2’] TRPEI of pyrazine to study this internal conversion; however, their time-resolution (450 fs) did not allow observation of ultrafast decay from S2 to S1, and they observed only the decay of S₁ after internal conversion from S₂. The S₁-S₀ decay time constant was 22 ps. Stert et al.²⁶ performed a similar experiment. Very recently, TRPEI of pyrazine was performed with sub-20 fs deep UV pulse in our laboratory, and the time constant of the S₂S1 internal conversion was experimentally determined as 22 fs.^27,28

113

The S₁  T₁ (n *) ISC in pyrazine is also well known as the intermediate coupling case in molecular radiationless transition. Frad et al.²⁹ and Lahmani et al.³⁰ observed ISC as a biexponential fluorescence decay, in which the fast component corresponds to ISC. The biexponential fluorescence decay was extensively studied in the 1980s and a debate ensued as to whether the observed fast component was due to the predicted dephasing or Rayleigh–

Raman light scattering.^16,31,32 The presence of a fast component and its finite lifetime have been unambiguously proved by picosecond laser spectroscopy. Although fluorescence studies only observed the decay of the singlet state, TR-PEI allowed detection of both S₁ and T₁ states involved in ISC in pyrazine by (1+2’) resonance enhanced multiphoton ionization (REMPI)^25,33,34 and (1+1’) REMPI.³⁵ In the latter, pyrazine was excited to 0⁰ level of S₁ with a 324 nm femtosecond pump pulse and subsequently ionized with a 197 nm femtosecond probe pulse. Ionization from the S₁ zero vibrational level occurred to low vibrational states of cation, due to Franck–Condon principle, creating rapid electrons, whereas ionization from high vibrational levels of the triplet state created slow electrons, as shown in Fig. 5.6 (b).

TR-PEI clearly demonstrated a rapid decay of the ionization component from S1 and a growth of that from T₁ with a clear isosbestic point in the time-dependent photoelectron spectra. The lifetime of S1 was estimated to be 110 ps.

We revisited internal conversion and ISC in pyrazine by TRPEI using VUV-FEL.

5.3.2 Experimental setup

Figure 5.7 shows the experiment setup for pump-probe experiments using a fs UV laser and a VUV FEL. The vacuum chamber and PEI spectrometer are the same as He(I) experiment described in chapters 2 and 3. A pulsed supersonic molecular beam of 10 % pyrazine (C₄H₄N₂) seeded in He was introduced into a photoionization chamber and crossed with a femtosecond UV laser pulse and a VUV FEL pulse (161 nm). The UV laser at 324 or 260 nm excited pyrazine to the first or second excited singlet state (S₁ of S₂), respectively, and

114

the VUV laser ionizes molecules from the excited states. Electrons generated by (1+1’) REMPI were accelerated along the molecular beam propagation axis and projected onto a position-sensitive detector comprising 10-m pore MCPs, a phosphor screen (P47), and a CCD camera (512  512 pixels). Light baffles were used at the entrance and exit ports of the laser beams to reduce scattered light.

The femtosecond laser system comprises an oscillator, a regenerative amplifier, and an optical parametric amplifier (OPA). The laser system was operated at 1 kHz, and a pulse picker was used to gate a pulse train from the OPA. Although the gate width of our pulse picker was long to allow two consecutive pulses separated by 1 ms to pass through each gate, background subtraction eliminated any interference from this additional pulse. The UV laser pulse was introduced into the vacuum chamber through an optical path length of about 10 m.

The laser beam was focused with an axisymmetric lens placed in the air and reflected with an aluminum mirror inside the vacuum chamber to the molecular beam. The timing of the laser pulse was synchronized to the 238-MHz master clock of SCSS by feedback locking the cavity length of the Ti:Sapphire oscillator. The time delay between the laser and SCSS pulses was varied electronically with an expected accuracy in sub-ps and the temporal overlapping of them is monitored by the high speed photodiode.

The SCSS is typically operated at 51–61 nm with electron beam energy of 250 MeV;¹⁴ however, 61-nm radiation can directly ionize the molecule without the pump light, which makes the observation of two-color REMPI signal highly difficult. Therefore, in this study, we used 161 nm radiation to perform the pump-probe experiment. The VUV radiation intensity was attenuated to be less than 1 J/pulse with a gas filter using air as a medium. The beam size was controlled by two slits placed before the focusing mirror (see Fig. 5.7), and the VUV radiation was finally focused onto the molecular beam using a pair of elliptical and cylindrical mirrors. The repetition rate of the whole system was 20 Hz. The crossing angle between the femtosecond laser and FEL beams was 1 degree. The polarization direction of

115

UV and VUV pulses was parallel to the face of the MCP detector. Owing to the characteristics of SASE,¹⁴ the time-averaged spectrum of 161 nm radiation had an energy width (FWHM) of ~ 0.1 eV.