Bayesian statistical model for quantifying defects in monolayer MoS2

The optoelectronic properties of molybdenum disulfide (MoS2) are influenced significantly by sulfur vacancy defects. While tools like electron microscopy can yield precise measurements of vacancies for small samples, they are not suitable for industrial-scale production. Estimates obtained from more scalable approaches like Raman spectroscopy are also subject to large uncertainties. This work introduces a Bayesian model that combines the information from observable Raman shifts with prior information to provide a posterior probability that comprehensively defines what is known about the sulfur vacancy in a MoS2 sample, in the context of measurement uncertainty and model error. This methodology serves as a robust tool to infer the sulfur vacancy concentration variation based on the observed Raman feature in a scalable way.