Design and Optimization of Silicon Avalanche Photodiodes With Desired Breakdown Voltage Using Bayesian Optimization

In this study, we introduce an innovative optimization methodology for the design and optimization of avalanche photodiodes (APDs) using analytical calculations, Bayesian optimization (BO), and technology computer-aided design (TCAD) simulations. The parameters under optimization include the thickness and doping concentration of the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">p</i> and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i</i> regions. Our approach aims to tailor APDs to specific breakdown voltage (BV) requirements while achieving superior performance characteristics, such as gain, responsivity, and bandwidth. Through BO, we efficiently explore the design space to identify optimal configurations. The optimized APDs exhibit enhanced performance compared to previous studies, with superior responsivity, gain, and cutoff frequency while having the desired BV. The optimization process successfully designs APDs with BVs ranging from 30 to 50 V, achieving gains of 100, responsivities of 0.29 A/W, and impressive cutoff frequencies exceeding 0.9 GHz for a photon wavelength of 650 nm and an intensity of 0.1 W/cm<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><roman>2</roman></sup> with the BV error less than 0.5 V relative to the target value. This approach demonstrates the effectiveness of BO in optimizing APDs for specific applications, addressing the challenges of balancing multiple performance metrics and meeting targeted BV requirements.