Light-induced cis-trans isomerisation around a carbon double bond is the mechanism by which our eyes detect light.1 Since our sight is our dominant sense, it is necessary for this process to be reversible, highly reproducible over the span of a human lifetime, and fast.2 These factors imply that this chemical process is highly-directed, or coherent, which means that there is concerted molecular motion in the electronically-excited state, driving the reaction to the electronic ground state along a predictable reaction path.3 For the efficient passage of population from the electronically excited state to the electronic ground state, a strong coupling needs to be present that will allow radiationless transfer, converting the potential energy introduced by the absorbed photon into kinetic energy, allowing it to be detected.4 This coupling is a conical intersection and allows coherent, fast transfer between the two electronic states.5 The two key achievements of our paper are the development and application of techniques for an “apples-to-apples” comparison of the dynamics in a molecular system between gas and liquid, and demonstrating that the ground state recovery in an olefinic cis-trans isomerisation can be coherent, even in complex environments.
To investigate olefinic cis-trans isomerisation, we used the molecule cis-stilbene as a model system. This molecule has certain advantages: it is a liquid with reasonably high vapour pressure, enabling gas- and liquid-phase experiments, the cis-trans isomerisation can only occur around one bond, while also having large bulky moieties that need to be moved during the extent of the reaction, similar to what there would be in a biomolecule. To obtain a signal of the reaction, there are a few things that are desirable. As the reaction passes between electronic states, there is a change in the electronic structure. This is accompanied by a large change in the nuclear geometry, as the molecule twists around the ethylenic double bond. This means that we need to be able to reliably initiate the reaction at a certain time, and track the change in electronic and nuclear structure in time. Time-resolved photoelectron spectroscopy is a perfect technique for this because the photoelectron spectrum is very sensitive to both changes in the electronic structure and the nuclear structure. Using ultrashort laser pulses (~30*10-15 s in duration), we can add time-resolution. In a pump-probe scheme, we use one laser beam to start the reaction in the electronically excited state. Then we probe this as a function of time using a second laser beam that ionises the molecule, ejecting an electron. This electron will have an energy and an intensity which is related to the electronic state it came from, the electronic structure, and the nuclear geometry. To encode the time into this technique, we separate the output beam from our laser system into two paths, where one part is used to generate extreme ultraviolet light that we use to ionise, and the other portion generates the ultraviolet excitation beam. We use mirrors on a translation stage to change the path length between the two beams, which creates a time delay between them. We can then apply this to gas and liquid jets of cis-stilbene, and measure and compare the dynamics between the two.
The excitation in this experiment prepares the first electronically excited state, which promotes an electron from the highest-occupied molecular orbital to the lowest unoccupied molecular orbital, which is a π*←π transition. This introduces antibonding character into the electronic configuration, which changes the hybridisation felt by the ethylenic carbons. The nuclear motion that this then initiates is a hydrogen-out-of-plane motion, which is very fast. This pyramidalises the ethylenic carbon atoms, and takes the system to a flat region of the potential energy surface, where the system starts to rotate around the ethylenic double bond. The combination of these two motions takes the reaction to the conical intersection region, and mediates the crossing between first electronically excited state, and the ground electronic state. As we probe this reaction in time, we see modulations in the intensity of the photoelectrons that we measure. This modulation occurs because the change in nuclear coordinates changes the ionization probability in time. By taking the Fourier transform of these modulations, we are able to extract three frequencies in both liquid and gas. These frequencies are related to the torsional motion around the ethylenic double bond. When comparing these excited state frequencies in the gas and liquid phase, we see a consistent frequency red-shift in the liquid-phase, when compared to the gas phase. Physically, this means that the motion of the torsion in liquid is slower than it is in gas. Our hypothesis for the origin of this is that the friction of the environment slows down the motions, resulting in red-shifted frequencies and a longer-lived excited state. To test this, we applied state-of-the-art theoretical dynamics simulations where the liquid was simulated by applying a time-dependent friction factor, which can be tuned to increase or decrease the effect. What we saw from these results was that, even though this solvent model is a reasonably simple approach, compared to methods which model the solvent molecules individually, we get a very good reproduction of the gas and liquid photoelectron spectrum through this method.
This side-to-side comparison of the gas and liquid dynamics of cis-stilbene shows that the primary effect that slows the dynamics is friction. However, despite this friction effect, a coherent reaction path is still observed in the liquid phase measurements. This finding that olefinic cis-trans isomerisations can still exhibit coherence in liquid begins to explain why rhodopsin has been naturally selected by evolution for the vision process.
For more information, please see our recent publication in Nature Chemistry: https://www.nature.com/articles/s41557-022-01012-0
1. G. Wald, Science, 1968, 162, 230—239.
2. O. Flender et al., Phys. Chem. Chem. Phys., 2016, 18, 14941—8.
3. S. Takeuchi et al., Science, 2008, 322, 1073—7.
4. X. Hu, K. Schulten, Phys. Today, 1997, 28—34.
5. D. R. Yarkony, J. Phys. Chem. A, 2001, 105, 6277—93.