Have a UROP opening you would like to submit?
Please fill out the form.
Looking for life in martian sediments by analyzing spectra of mineral mixtures
12: Earth, Atmospheric and Planetary Sciences
Jian Gong firstname.lastname@example.org
Spectroscopic analyses of natural and experimental samples, including Absorbance, Raman, Fourier-Transform Infrared (FTIR), X-Ray Diffraction (XRD), Energy-dispersive X-ray (EDS), are used to identify minerals and amorphous solid phases in planetary and terrestrial samples. For example, multiple spectroscopic techniques are used to identify mineral mixtures in the cores acquired by the rover missions on Mars. Typically, the spectrum of a sample is compared to some known standards. This works well for relatively pure samples, but gets more complicated in the case of heterogeneous samples and unknown mixtures due to the overlaps in peaks and added noise. One approach is to collect multiple types of spectra from a sample and form a set. This set is then analyzed using multivariate statistical techniques, especially Principle Component Analysis (PCA) analysis. The grouped dataset can be transformed into clustered subsets, where each subset should correspond to one component of the mixture. This requires no initial standards, but through the later addition of standards, it becomes apparent which cluster is associated with which known standard. This technique essentially classifies and identifies multiple spectra by the groupings of clusters, rather than any individual spectrum, thereby tackling both the “heterogeneous” and the “noise” problem at once. This remote summer UROP project aims to develop flexible algorithms and pipelines that can efficiently transform spectral datasets (FT-IR, Raman, XRD, EDS) into data that can be readily processed and interpreted by comparisons to standards. This project will have immediate applications to multiple projects currently on-going in the Bosak Lab at MIT, including the experimental weathering of early Mars analog sediments. The project will be supervised by Dr. Jian Gong, a postdoctoral scholar in the Bosak lab.
programming, spectral analysis, some familiarity with minerals and materials