efficient excited states in large systems

We recently published a paper titled “Pushing the Limits of EOM-CCSD with Projector-Based Embedding for Excitation Energies” where we calculated the interaction of light with some small molecules that are in solution using state of the art techniques. In this post, I’m planning on giving a general introduction to why we did this research and the ways impact it may have.

The interaction between light and molecules is central to all branches of physical sciences, with our understanding of the physical process involved going back to the quantum revolution 100 years ago. Being able to work out the amount of energy needed for light to affect a molecule and the strength of that interaction is valuable in many areas that affect modern day life, such a photosynthesis, designing better solar cells or even how to build better phone screens.

The use of computers in chemistry makes it possible to predict how chemical reactions occur with little cost or damage to the environment and can be a helpful guide to experiment. Computational chemists aim to find ways to calculate many chemical properties as efficiently as possible without losing accuracy in our predictions.

In the case of light-chemical interactions, two methods that are commonly used are a quick and somewhat rough method called TD-DFT and a more accurate and expensive method called EOM-CCSD. Our work combines the methods in such a way that makes it possible to work our how light interacts with chemicals accurately and quickly.

The approach that we took was to treat different chemicals with different accuracy using a method called “Embedding” (placing one method inside another). By doing this we were able to accurately calculate the effect of light hitting a molecule in solvent between 100x and 1000x faster than possible before.

In the long term, our research may help enable highly accurate predictions of light-chemical interactions in academia and industry. For example, calculating how chlorophyll in plants absorbs light or how to test new solar cell designs without having to actually build them.