The goal of our research is to elucidate microscopic dynamics of prompt chemical and photophysical processes in condensed phases. We put particular emphasis on the initial stages of the processes and the coupling between the electron and nuclear degrees of freedom, where the nuclear coordinates include both inter- and intra-molecular, as they play critical roles in most of the spectroscopy and determine the fate of a photoexcited system. Our primary research tool is the femtosecond time-domain spectroscopy that can determine the structure and energy levels of a molecule directly while a molecule is undergoing chemical reactions and physical processes. Depending on the system under scrutiny, a variety of femtosecond techniques including those developed in our own laboratory are employed; time-resolved fluorescence (TRF), 3 pulse photon echo peak shifts (3PEPS), spectrally resolved four wave mixing (SRFWM), transient grating (TG), tau-resolved transient grating (TRTG), and transient absorption (TA). Summarized below are some of our recent results made possible by the state-of-the-art instruments developed in our laboratory. 

Creation and Detection of Vibrational Wave Packets in Electronic Excited States  

When a molecule is excited impulsively by a short pulse of light, vibrational wave packets are created, which are manifested in the oscillations of the time trace in various time-resolved spectroscopies, TA being a typical example. The nature of the vibrational modes excited and their dynamics provide a wealth of information on the molecular dynamics, structures of the states involved, and the coupling between them. In general, the coherent nuclear wave packet can be launched by any nuclear rearrangements occurring faster than roughly half of the vibrational period. Excited state intramolecular proton transfer (ESIPT) is one such example. 
Observation of the dynamics of the vibrational wave packet in the excited state can be done unambiguously by TRF as far as the time resolution is high enough. A TA signal may also give excited state dynamics through the stimulated emission component, although a TA signal may be complicated by the ground state bleach, excited state absorption, and product absorption. We have developed a noncollinear fluorescence upconversion technique that now gives near perfect time resolution with 20 fs pump pulses. 

Development of New (and Old) Time domain Techniques  

Time-resolved fluorescence (TRF): We have developed the TRF technique to unprecedented time resolution of around 30 fs (FWHM) employing noncollinear sum frequency generation and a Cassegrainian pair as an imaging optic (Opt. Lett. 30, 96-98 (2005); Opt. Express 16, 20742 (2008)). In addition, full emission spectra can be recorded without the conventional spectral reconstruction method. The time delay, phase matching angle of the BBO crystal, monochromator are moved in unison to record the fluorescence spectra for a given time while compensating group velocity dispersion. 

Tau (t) Resolved Transient Grating (TRTG) : Third order nonlinear spectroscopic techniques such as 3PEPS, TG, and TA give information on the dynamics of the system through the transition frequency correlation function of the probe, M(t). Although they all more or less follow M(t), only the 3PEPS is background-free. That is, 3PEPS is free from the population relaxation component to give M(t) directly including inhomogeneity. However, it requires a long acquisition time due to the scanning of two time delays in 3PEPS. We devised a nonlinear time domain scheme, electronic coherence period (denoted as t conventionally) resolved transient grating (TRTG), which provides M(t) with only a population delay scan. TRTG is similar to TG except the electronic coherence period t is set to a finite value, usually a few femtosecond, and the third order signals at two phase matching directions are compared and subtracted. TRTG was implemented based on a diffractive optics to set the t delay accurately within a few tens of attoseconds. The significance of TRTG was verified by theoretical calculation as well as experiments for a dye in various solvents. (Initial idea was published in J. Chem. Phys. 120, 5269-5274 (2004), and the full account is to be published). 

Spectrally resolved four wave mixing (SRFWM) : Spectral dispersion of a four wave mixing signal has been reported before. We have further developed the technique to a few cm-1 precision in center frequency measurement, which enables us to resolve both the ground state and the excited state dynamics through the ground state bleach (GB) and excited state stimulated emission (ESE) components in a TG experiment, respectively. As an example of SRFWM, detailed frequency-time resolved spectra of a dye molecule in methanol following impulsive excitation by 20 fs pulses is presented. The transient spectra are successfully separated into the GB and ESE, and the center frequencies of each component are measured as a function of time delay, which allows the observation of the nuclear dynamics in each state separately. Both GB and ESE show oscillations due to the vibrational wave packet motions in each state. The ambiguity in the assignment of an oscillation to either the ground or the excited electronic states, which exists in all third order nonlinear time domain spectroscopies, can be resolved unambiguously in this SRFWM experiment. More importantly, the solvation dynamics can be obtained from the time dependent blue shift of the ground state hole spectrum, which is purely due to the fluctuation of the solvent molecules around the probe in the ground state, and from the usual time dependent Stokes shift of the excited state particle spectrum, which is due to the dissipation of the probe in the excited state. 

Development of the Light Sources 

To perform those experiments for various chemical systems, it is essential to make our own femtosecond laser to meet the stringent requirements. An ideal laser for femtosecond spectroscopies requires the wavelength tunability especially in the visible, high yet controllable pulse repetition rate, and moderate pulse energy for wavelength conversion by nonlinear processes. A cavity-dumped Ti:sapphire laser meets those requirements except the limited wavelength tunability. One way to get a wavelength tunability is to build a Ti:sapphire amplifier and an optical parametric amplifier, although that is not economic and the repetition rate is limited to a few kHz. We have developed a cavity-dumped optical parametric oscillator based on a periodically poled lithium niobate (Opt. Lett. 30, 1855-1857 (2005)). A notable feature is to implement positive group velocity dispersion for much higher pulse energy, and many feed back mechanism including the cavity length and pump beam pointing to stabilize the OPO. The output is enough for various nonlinear processes to efficiently convert the wavelength well into UV.