An expedient semi-empirical modelling approach for optimal bandgap profiling of stoichiometric absorbers: A case study of thin film amorphous silicon germanium for use in multijunction photovoltaic devices

Abstract

Listen

Translate article

Bandgap energy profiling is applied in a variety of materials for photovoltaic technologies, such as chalcogenides, III–V materials and perovskites. Bandgap profiling of the absorber layer is used to fight the fundamental loss mechanisms imposed by the bandgap energy of the absorber for the maximum voltage and current that a photovoltaic device can generate. The bandgap profile can be affected by a number of profiling strategies, such as the difference between the maximum and minimum bandgap energy, the position of the minimum bandgap energy, the width over which this minimum bandgap energy occurs and the total absorber width. These parameters have a complex effect on output characteristics of a photovoltaic device. Varying multiple parameters at once further increases the complexity, limiting the effectiveness of rigorous physical opto-electrical modelling. In this work we therefore present an expedient semi-empirical approach for the optimal bandgap profiling of stoichiometric absorbers.Using PECVD processed amorphous silicon germanium as a model, we present a unique set of semi-empirical relations that simulate the VOC and JSC of solar cells as a function of the bandgap energy profile. For this model, the influence of deposition conditions such as the relative germane flow rate, the deposition power and substrate temperature on the opto-electrical properties of a-SiGe:H films is first characterized. Opting for the relative germane flow rate to control the bandgap profiling, the experimental results of a large number of solar cells with profiled a-SiGe:H absorber are presented, varying: 1. the absorber thickness, 2. the peak germane flow rate, so minimum bandgap energy, and 3. the introduction of a plateau at the minimum bandgap energy. Using this experimental data and optical simulations, the expedience and effectiveness of the semi-empirical approach is demonstrated.