The goal of my project was to analyze the origin of selectivity in a particular Suzuki-Miyaura cross-coupling reaction using computational chemistry methods. The Suzuki-Miyaura reaction (SMC) is a particularly synthetically useful reaction because it can create carbon-carbon bonds, and is effective at coupling heterocycles, which is notoriously difficult in cross-coupling reactions. The catalytic cycle of the SMC consists of multiple steps, but the initial oxidative addition step requires the most activation energy and is generally considered the rate determining step. Therefore, by analyzing how different ligands interact in this transition state, we can understand why they produce the results they do. This can lead to rational selection of reaction conditions and possibly even rational catalyst design.
To conduct all my calculations, I most importantly learned how to use the computational chemistry software, Gaussian. I would use Gaussian to create my inputs to send to the supercomputer, which can handle calculations including optimizations, energy in solvent, and even tracing the energy over the course of an entire reaction. I also used Gaussian to analyze the transition states I found using multiple methods. I could simply look over the file and analyze bond lengths and such to locate favorable coordinations or sterically unfavorable interactions that would change selectivity. I could also analyze the individual energies of distorted molecules to see if that imparts selectivity, a method called distortion-interaction analysis. Finally, using Gaussian I could look at the HOMO and LUMO of each molecule to explain selectivity.
However, beyond the concrete skills I learned in Gaussian, I also enhanced my ‘chemical sense’ through my research. I had to learn how to determine what was a physically reasonable transition state, what constitutes a bond or interaction, and what structures various ligands were likely to form. I also became more familiar with the theory and literature associated with my project. Gaussian uses a model called Density Functional Theory to conduct its calculations, which is a simplification of Schrodinger’s equation for large systems that uses functionals. By understanding this, I can use tricks to reduce the computational cost of my calculations (like creating symmetrical molecules that simplify the wave function) and understand the limitations of my method. I also had to read many articles in order to troubleshoot my data and learn the mechanism behind this reaction. At first, my theoretical data did not match my experimental data, so I had to take it into my own hands to determine what was wrong and make progress at fixing it.
Finally, I had to learn how to use other software in order to effectively communicate my results, like Chemdraw. Chemdraw is an industry standard for drawing molecules, so it was essential for depicting my data. Moreover, I had to speak at group meetings every other week, so through that I honed my public speaking and effective diagraming skills.
Overall, I am very grateful for the impact this project has had on my education at Penn. Just being in a group where I am the only undergraduate was a bit intimidating at times, but I think I improved my confidence in my own work through it. Plus, just learning about what projects my peers were working on helped to further my knowledge of organic chemistry. This project also really helped me rethink my perspective on research; before, I used to dread working in a field where I was destined to be stuck in the lab all day. But having my own project where I felt like I could control the pacing and direction helped me feel excited about entering academia. I really hope to continue working in the Kozlowski Group throughout the rest of my time at Penn!