Optimization of mid-infrared quantum cascade detectors

In this work, the optimization of the design of quantum cascade detectors is considered. The initial design studied was a detector structure based on an AlGaAs/GaAs heteropair, consisting of four quantum wells. A genetic algorithm  was utilized to optimize the responsivity and detectivity of the design under study by varying of the widths and chemical composition of the first five layers of the cascade. The responsivity and detectivity were simplified to the  characteristics, that can be evaluated based on the solutions of the Schrödinger and Boltzmann equations. The results demonstrate a strong dependence on the optimization algorithm parameters and designate significant change  from the initial design. We have shown that to achieve optimal output characteristics and improve the convergence rate, one must use larger populations and high mutation probability in the genetic algorithm. The analysis of the obtained designs also shows that additional regularization techniques are required to achieve better output characteristics of the device. Specifically, the appropriate weighting of the set of optimized characteristics must be  determined before final optimization.