Imaging Chemical Reactions in Space and Time

Modelling and understanding reactions that are initiated by light is a difficult feat for chemists and physicists. This is due to the complex nature in which electrons move around nuclei in a molecules stable state, with many states of similar energy that are strongly interacting and dependent on each other. These strongly interacting states lead to phenomena that are inherently different from the norm. For example, in the photodissociation of CS2, where one sulphur atom is pulled from the carbon, the resulting sulphur atom is in a different electronic configuration to what it started in. Building an understanding of how such processes occur through the development of theory and experimental techniques is crucial for the technological development of society. This will be achieved by allowing for improved design of optoelectronic and photonic materials such as LEDs, solar cells, and other semi-conducting devices that rely on light driven processes.

The photodissociation of CS2 is of particular interest to theorists such as myself, as it's a relatively "small" system, with only three atoms and 38 electrons. Despite being "small" it has a lot of interesting properties due to its dense series of states with different configurations that are strongly interacting. We aim to accurately calculate the electronic structure which describes the arrangement of electrons around the nuclei of CS2 through state-of-the-art quantum chemical methods. We can think of the photodissociation process as a reaction in which the nuclei move apart in a manor that depends on the forces acting on them from the electrons configuration. Previous works involve studying the dynamics of the reaction by on-the-fly simulations in which the electronic structure is calculated in a quantum manor at each point in space for a series of trajectories and then the nuclei are moved according to Newtons classical equations of motion. These simulations have provided tremendous insight into the process. However, this work aims to pre-fit the electronic surfaces by use of artificial intelligence. Once the electronic surfaces are obtained, the nuclear motion can then be simulated on them at a lot lower cost at with a fully quantum description, giving greater insight into the reaction.

Through the development of this theory, we also aim to provide our experimental collaborators with additional tools for interpreting their experiments on the dynamics of CS2. In recent years there has been tremendous development in X-ray Free Electron Lasers (XFELs). These ultra-fast particle accelerators provide us with high energy light of great intensity. They also have the added benefit of producing light in incredibly fast pulses. The greater intensity of the light, along with the short pulses, allow us to image ultra-fast reactions on the femtosecond time scale in the gas phase - which has been a challenge for scientists in the ultra-fast community. Although there has been real developments in this experimental technique, there is still a lot of work that needs to be done in order to fully understand the science. By collaborating with experimentalists we aim to run ultra-fast scattering experiments with XFELs on CS2 and thus allow for combined greater understanding of the dynamics of CS2 with help from our simulations, and for further development in designing new XFEL experiments in order to push the boundaries of ultra-fast science.