Solar energy conversion processes start from photon-matter interactions and a photon triggered chemical reaction almost always generates changes in nuclear geometry of those involved. The availability of high repetition rate, intense and femtosecond hard X-ray pulses could enable the visualization of nuclear geometry evolution in many light-induced reactions in solar energy conversion processes, including ‘‘the transition state’’ (TS)152 at the top of the potential barrier crossing trajectory between reactants and products. It is believed that a TS has ‘‘no lifetime’’ or a barrier crossing time of a fraction of the vibrational period. Almost all XTA experiments carried out so far have investigated transient structures that equilibrated in an excited state or transient state potential well rather than followed atomic or electron motions coherently. Hence, the TS has eluded experimental observation especially in the solution phase. Visualization of coherent vibrational motions leading to the TS in chemical reactions will transform our understanding of chemistry by providing information about the displacement pathways of all of the atoms involved, thereby replacing the simplistic descriptions now employed. Many important chemical reactions in solar energy conversion processes involve atomic displacements, such as natural and artificial photosynthesis and catalysis. Despite decades of study using spectroscopic probes, mechanisms of isomerization about a carbon-carbon double bond, and/or multiple double bonds, remain uncertain. Theoretical theoretical studies on barrier crossing and conical intersections for isolated small molecules have been carried out153 which, combined with femtosecond structural studies, could provide new knowledge in chemical sciences. Some examples shown in the previous section are excellent candidates for such studies. To visualize the TS, the X-ray pulse duration must be shorter than the period of vibrational motion that forms the products.
Most XTA measurements conducted so far employed continuous energy tunability in about 1-keV range from synchrotron sources and collected X-ray absorption spectra step-by-step at each individual X-ray photon energy defined by the monochromator. The current femtosecond X-ray free electron laser (XFEL) sources provide either monochromatic or narrowly distributed energy spectra that are not continuously tunable for the step-by-step approach. For example, the Linear Coherent Light Source (LCLS) at the Stanford Linear Accelerator Center (SLAC) currently has about a 50-eV band width centered around 7.1 keV due to intrinsic bandwidth of self-amplified-spontaneous – emission (SASE).154
A preliminary XTA measurement at the X-ray pump-probe station on [Fe(II)(bpy)3]2+ in water demonstrated the feasibility of XANES measurements with ~50eV spectral range, which is sufficient to cover the pre-edge, edge and above edge region (M. Cammarata, D. Fritz et al, unpublished results). Combining the electronic configuration information obtained from the pre-edge/edge region with the structural information obtained by MIA42 44 from the XANES region and the full XAFS spectra from synchrotron sources, it is possible to extract the dynamics of electronic transition and nuclear geometry changes simultaneously and to verify directly the Franck-Condon principle using the fs X-ray pulses from XFELs.
The first examples of such studies that are relevant to solar energy conversion are those of transition metal complexes functioning as light harvesters and catalysts. For example, the ground state metal to nitrogen stretching frequency is in a range of 500-700 cm 155 corresponding to vibrational periods of about 50-70 fs. In more recent examples of platinum dimer complexes, coherent vibrational motions of Pt-Pt stretching have been observed,156,157 which are in the range 100-250 cm 1. Therefore, it is possible to follow structural reorganizations resulting from electron transfer directly and hence the reorganization energy if the metal-to-ligand bonds are directly involved. The internal reorganization energy appearing in the Marcus Equation of electron transfer158 161 applied to organic and organometallic materials has been approximated by a C = C bond stretching frequency of around 1500 cm 1,161 as a generic value, but the actual reorganization is vague owing to lack of the information on transient structures of excited state and charge separated state. Using XTA with fs time resolution, correlations in metal electronic configurations associated with photoinduced electron transfer and the reorganization of nuclear structures included implicitly in the Marcus Equation of electron transfer can be visualized for the first time. This information will help the design of artificial photosynthetic systems for solar fuel or electricity generation.