Informed Chemical Classification of Organophosphorus Compounds via Unsupervised Machine Learning of X-ray Absorption Spectroscopy and X-ray Emission Spectroscopy

Samantha Tetef, Vikram Kashyap, William M. Holden, Alexandra Velian, Niranjan Govind, and Gerald T. Seidler
The Journal of Physical Chemistry A 2022 126 (29), 4862-4872
DOI: 10.1021/acs.jpca.2c03635

UW Scalable Quantum Computing Lab

In March 2022, I began research as part of the UW SQRLab headed by Professor Sara Mouradian. I have been working on analyzing crosstalk error in trapped ion quantum computing using simulations. “Crosstalk” error refers to when a laser pointed at one ion has non-negligible intensity at neighboring ion positions as well. I am attempting to create a model for the error associated with a Mølmer–Sørensen gate, which entangles two ions through coupling to a particular vibrational mode of the ion string, in the presence of crosstalk. My hope is that this will allow one to choose beforehand which vibrational mode to use in order to achieve the highest fidelity entangling gate.

Argonne National Laboratory Spectroscopy Group

I joined Argonne National Labs’s Spectroscopy Group at the Advanced Photon Source August 2021 as a research aide to Dr. Chengjun Sun and worked through July 2022 on the Argonne X-ray Emmission Analysis Package in Python (pyAXEAP), a python package aimed at automating the process of calibrating and calculating X-ray emmision spectra (XES) from 2D pixel-array-detectors on the beamline. During the school year I worked part-time remotely, but during June 2022 I spent 2 weeks on-site at the APS.

UW Seidler Lab

I joined the UW Seidler Lab March 2020 and began working on a project designing a refrigerated, vacuum-sealed, holder for a sample that some collaborators in the Chemistry department wanted to analyze mid-reaction. The project was postponed due to COVID and so I transitioned to a project applying unsupervised machine learning to categorizing the calculated X-ray spectra of organic molecules containing phosphorus. Our goal was to see if the computer was able to, while only looking at the spectra, sort the spectra by similarity into groups that we could then correlate with various common molecule structure classes. The computer was able to do this surprisingly well, giving us a measure of how much chemical structure information is encoded in the spectra. This work became my first scientific paper.