This review discusses how ultrafast organic photochemical reactions are controlled by conical intersections, highlighting that decay to the ground-state at multiple points of the intersection space results in their multi-mode character.

Dye‐sensitized photoelectrochemical cells are promising devices in solar energy conversion. However, several limitations still have to be addressed, such as the major loss pathway through charge recombination at the dye‐semiconductor interface. Charge separating dyes constructed as push‐pull systems can increase the spatial separation of electron and hole, decreasing the recombination rate. Here, a family of dyes, consisting of polyphenylamine donors, fluorene bridges, and perylene monoimide acceptors, was investigated in silico using a combination of semi‐empirical nuclear dynamics and a quantum propagation of photoexcited electron and hole. To optimize the charge separation, several molecular design strategies were investigated, including modifying the donor molecule, increasing the π‐bridge length, and decoupling the molecular components through steric effects. The combination of a triphenylamine donor, using an extended 2‐fluorene π‐bridge, and decoupling the different components by steric hindrance from side groups resulted in a dye with significantly improved charge separation properties in comparison to the original supramolecular complex.

The spin-orbit coupling constant between the 1 A 1 and 5 T 2 states is roughly an order of magnitude larger than what has been determined experimentally from variable-temperature ultrafast time-resolved absorption measurements. 1 If you were to replace your calculated value with the experimental one, how do you think this would change your mechanistic picture?

Please clarify the case of an electronic excitation that is able to change the statistical distribution of the phonon population.Steven Johnson answered: I assume this question refers to the phenomenon of "phonon squeezing", which is experimentally seen as an oscillation of the variance of the generalized coordinate of one or more vibrational modes. This is in contrast to coherent phonons, which are typically identied with oscillations of the expectation value of the generalized coordinate.When a harmonic crystalline material is in thermal equilibrium, each of the vibrational modes can (in a quantum mechanical picture) be represented as consisting of an admixture of number states, with a Bose-Einstein distribution. The expectation value of each mode coordinate is zero and independent of mode frequency U and temperature T; the variance of the coordinate, however, depends on both U and T. In the low temperature limit, the variance is dominated by the ground-state variance, which is inversely proportional to U 2 . One way to see the "squeezing" phenomenon is to drive a sudden change in the value of U by using an ultrashort light pulse to change the electronic state of the system. If the electronic state change couples to the vibrational modes via terms in the Hamiltonian that are quadratic in the mode coordinates, the electronic state change can trigger a nearly instantaneous change in the frequency U. Aer this is done, the admixture of vibrational states that characterized the initial conditions are no longer eigenstates of the modied vibrational Hamiltonian, and will, therefore, evolve in time. This results in oscillations of the variance of the mode coordinate at a frequency of 2U. More mathematical detail and descriptions of experiments that have observed this can be found in several references, e.g., ref. 1 or 2. At high temperatures, one can also use a classical framework to describe the 80

Have you been able to carry out simultaneous time-resolved X-ray scattering and emission measurements in order to more rmly establish the mechanistic sequence and spin-structure correlations being proposed?Serhane Zerdane responded: Indeed, complementary experiments would provide valuable input, especially regarding correlated changes of spin-state and structure. When we performed the XANES experiment at LCLS, it was not possible to perform emission spectroscopy and scattering experiments simultaneously. We expect to perform these measurements during the next beamtime.Yorrick Boeije asked: Considering the 80 fs temporal resolution of the optical pump-probe measurements, how do you distinguish between local lattice torsion activation induced by (1) charge-transfer (CT) or (2) a process prior to chargetransfer, either optical excitation or the spin transition?Serhane Zerdane replied: For monitoring changes of electronic state and structure we used two complementary techniques: optical spectroscopy and X-ray spectroscopy (XANES). The time resolution was limited to 80 fs in time-resolved optical spectroscopy and this technique is indirectly sensitive to CT and spin transition through the associated electronic reorganizations. The time resolution of the XANES experiment was about 25 fs (ref. 1) and this technique is sensitive to both inter-metallic charge transfer (opposite shis for Co and Fe K edges) and spin transition (shis the rst extended X-ray absorption ne structure above the Co edge). In this way we could show that the spin transition occurs on the Co rst, and is followed by an intermetallic charge-transfer. 1