Genetic Algorithm Application for the Optimization of Solid State Devices using a Novel Modeling Technique

Abstract

A new method for developing a realistic physical model of any type of solid state device is presented. Application to model advanced multi-junction solar cells; Thermo-photovoltaics; sensors; as well as other novel solid state devices are introduced in this presentation. The primary goal of multijunction solar cell design is to maximize the output power for a given solar spectrum. The construction of multijunction cells places the individual junction layers in series, thereby limiting the overall output current to that of the junction layer producing the lowest current. The solution to optimizing a multijunction design involves both the design of individual junction layers which produce an optimum output power and the design of a series-stacked configuration of these junction layers which yields the highest possible overall output current. This paper demonstrates the use of Genetic Algorithm in a two-part process to refine a given multijunction solar cell design for near-optimal output power for a desired light spectrum. This approach can similarly be utilized to optimize the parameters of any Solid state device to yield any desired performance.