A fundamental understanding of material and device physics is key to the design of solar cells with high efficiency and long operational life. however the physical processes of light absorption (charge carrier generation) and subsequent electrical power generation or loss (transport and recombination) are complex and difficult to directly monitor.

Transient charge extraction measurements can provide a snapshot of these processes, and are used to assess material and device characteristics (electrical performance, material quality, presence of defects). however current methods for processing these results are limited in scope and require significant manual analysis and interpretation before meaningful insights are obtained.

This project will use machine learning systems (generative representation learning) to model the transient charge extraction measurement, thereby enabling an improved understanding of the dynamics of charge carriers (transport, recombination) in operational solar cells, and provide a predictive model to classify behaviour and investigate underlying causal relationships.

School

Photovoltaic and Renewable Energy Engineering

Research Area

Photovoltaics | Charge extraction | Time-series analysis | Machine learningArtificial intelligence

Through this research project you will gain an understanding of the physics of charge carrier solar cells, from light absorption to current extraction, and how these processes are influenced by material type and device architecture, as well as develop valuable skills regarding time-series data analysis and the development and training of machine learning models.

Applicant suitability: familiarity (basic skills, experience desirable) with coding in python (data processing, analysis, visualisation); basic understanding of machine learning systems (statistics, optimisation).

This will enable the characterisation and analysis of, as well as development of new design rules for, novel photovoltaic material systems and device architectures.

Deliverables: prepared training datasets (charge extraction transients); performance validation (statistical assessment) and visual examples of model output (generated charge carrier transient prediction) for each configuration of learning model; analysis of the influence of varying training dataset, model architecture, and learning methodology.