Quantum and Conventional Machine Learning Analysis of Synthesis-Structure Relationships in Transition Metal Dichalcogenide Thin Films

Vibrational modes in Raman spectra of transition metal dichalcogenides (TMDs) provide critical sample information on defects, sample thickness, and monolayer coverage. Identifying synthesis parameters that result in optimal values for these characteristics is nontrivial; however, supervised learning techniques provide a possible path towards recognizing patterns between growth conditions and characteristic features of Raman spectra acquired of the resulting samples. Leveraging data from over 300 growth trials, we use quantum as well as classical supervised learning algorithms to study the relationships between gas chalcogen precursor MOCVD synthesis parameters of MoS₂ thin films and features in Raman spectra characteristic of the thin films. We identify MOCVD growth parameters that minimize the A_1g and E_2g mode peak distance, corresponding to improved monolayer coverage. Models can be trained on data characterizing both the center and edge of samples to identify a growth recipe that minimizes the difference between the two spectra, improving the uniformity of the sample. The methodology of this machine learning investigation of synthesis–structure relationships can be applied to additional features of interest within Raman spectra, as well as to other TMDs, such as WS₂ and WSe₂.