Combinatorial materials discovery aims to find new promising materials more efficiently by fabricating and analyzing materials in parallel or using automated serial methods. The characterization of these materials involves a variety of different properties, ranging from electrical and mechanical to optical and magnetic ones. For photovoltaic and photoelectrochemical applications such as solar cells and solar water splitting, the band gap of a material is an important property. The fastest method for determining the optical band gap of thin-films is high-throughput ultraviolet-visible (UV/VIS ) spectroscopy, ideally complemented by an automated Tauc plot analysis.

Recently, several algorithms for the automated Tauc plot analysis have been developed and proposed [Ref.1-3], but the code is either not made available, not executable, or not open source. This makes it difficult for the scientific community to evaluate the efficiency and performance of existing methods, which are often validated only on a single material or a small set of measured Tauc plots. In addition, these published automated analysis algorithms are sensitive to measurement noise, which is unavoidable in optical spectroscopy, and depend on a set of user-defined parameters that must be set for each measurement. Therefore, flexibility of the analysis method is crucial when used in a high-throughput measurement system that characterizes hundreds of materials per day.

To overcome these difficulties, a new automated Tauc plot analysis algorithm has been developed that allows fast, reliable and robust on-device one direct band gap determination . The proposed ML method relies on only a few user-defined parameters for which the most appropriate default values have been identified, allowing the method to be run without any changes. Recently, the proposed fully automated ML method for the determination of band gap values has been successfully applied to 40 material libraries, including several high entropy oxides, and additionally validated with the reference band gap value for the Co-O system at room temperature .
 
References
[1] S.K. Suram; ACS Combinatorial Science, 2016, 18, 673-681.
[2] A. Escobedo-Morales; Helyion, 2019, 5, e01505.
[3] M. Schwarting; Materials Discovery, 2017, 10, 43-52.