Novel Absorber Engineering Techniques to Enhance the Efficiency of Perovskite Solar Cells

Abstract

In recent years, perovskite solar cells have emerged as a promising technology in the field of photovoltaics. However, their efficiency is hindered by various factors that necessitate a comprehensive investigation. This research focuses on absorber engineering, bandgap grading optimization, and the application of machine learning algorithm. By doing so, this study aims to significantly contribute to the advancement of perovskite solar cell technology. This comprehensive thesis undertakes a multifaceted exploration of perovskite solar cells (PSCs), to address crucial challenges and enhance efficiency. The initial investigation reveals performance-limiting parameters in contemporary PSCs, emphasizing deficits in fill factor (FF) and open-circuit voltage (VOC). Absorber characteristics and device optimizations, such as carrier concentration control and shunt resistance considerations, result in a significantly improved device with enhanced VOC, FF, and power conversion efficiency (PCE). A subsequent theoretical study focuses on configurational optimization of heterojunction PSCs to minimize internal recombination, employing various design alterations. The optimized device exhibits noteworthy enhancements in PCE, FF, and VOC compared to benchmark configurations. Moving forward, a proposed bandgap grading profile seeks to maximize spectrum absorption in perovskite absorber material, targeting the Shockley Queisser (SQ) limit. Analyzing linear bandgap grading profiles, the research identifies an optimal range for efficiency, emphasizing a well-optimized small bandgap grading range to achieve 31% power conversion efficiency. Continuing the exploration, a novel double absorber layer structure is introduced, incorporating 12 different absorber layer combinations. This approach expands spectral absorption, mitigates thermalization losses, and achieves an impressive efficiency exceeding 35%. To deepen insights, a dataset of 3490 samples characterizing perovskite structures is generated. Leveraging machine learning with the Random Forest algorithm, a predictive model classifies structures and offers valuable insights into optimized PSC design. Overall, this thesis contributes significantly to the advancement of PSC technology, offering novel solutions, theoretical insights, and practical design guidelines. The findings promise higher PCE and improved overall performance, propelling PSCs toward their potential as a leading photovoltaic technology in the renewable energy landscape.