Numerical optimization of semiconductor waveguide structure

Research on radiation sources in the IR and THz ranges operating at room temperature is still highly attractive to this day. Waveguides play a critical role in these structures and their improvement is also required. This paper studies  the optimization of waveguides based on GaAs material with different doping levels of layers to reduce absorption losses and increase the optical confinement factor. The optimization is carried out in three steps: selection of  optimization parameters, determination of initial values of parameters and Bayesian optimization. The thickness and doping level of heavily doped layers are chosen as optimization parameters. The results show the Bayesian  algorithm converges to the desired values rather quickly. It was found that the dependence of the waveguide output characteristics on concentration is weaker than on layer thickness. An increase in layer thickness leads to an increase in losses. Weak asymmetry in the structure can lead to a slight improvement in the confinement factor value.