Mapping and Optimization of Oscillator Strength in Quantum Bragg Mirror Detectors as a Function of their Dimensions

In this work, we estimate the oscillator strength of the optical transitions in quantum Bragg mirror detectors (QBMDs) and determine which structure provides its maximum value. The thicknesses of the quantum wells (QWs) and barriers of the probed structures strongly influence the wavefunctions and, thus, the optical transitions of the active region. For this reason, the structure geometry must be carefully chosen to achieve high oscillator strength at the desired operation energy. Since the QBMD has several QWs and barriers to be optimized, we have performed a series of approximately 350,000 simulations to map the transition energy and the oscillator strength to fully understand the role played by the thickness of each layer of a base structure with 7 quantum wells in total. Due to the computational cost of conducting these simulations for every possible configuration, we shifted to genetic algorithms to find optimal solutions with less constraints in the simulation. Finally, we present a validation of our genetic algorithm, which shows that it consistently finds high-performing structures as validated by the high oscillator strength values obtained. Overall, our study highlights the promising potential of genetic algorithms as a powerful tool for optimizing the properties of QBMDs.

I. Introduction

Quantum Bragg Mirror Detector (QBMD) is a new class of mid-infrared photodetectors that has been under scrutiny since 2016 [ 1 – 3 ]. They operate based on transitions between a bound state in a quantum well and a localized or leaky electronic state in the continuum (also called, respectively, bound state in the continuum - BIC - or quasi-bound state in the continuum - quasi-BIC). When designing the conduction band profile of a QBMD, the material, the periodicity, and the thickness of the different layers must be carefully chosen because they strongly influence the device’s operation. Detailed calculations must be performed to design an optimal structure with the desired performance. To better understand how these parameters affect the device’s performance, we have conducted a series of simulations to calculate the electron wavefunctions, map the transition energies originating from the ground state, and find the oscillator strength for each structure [4] . During the simulations the desired energy transition was restricted from 280 to 320 meV. This mapping serves as a guide to pinpoint the ideal heterostructure with a desired transition energy for a particular conduction band profile. However, when the number of variable parameters increases, (such as the number of QWs or unrestricted thicknesses of QWs), finding a better structure with a high oscillator strength could be computationally unfeasible due to the computation time required. To circumvent this situation, genetic algorithms (GAs) can be used. To validate the effectiveness of the GA that finds structures with high oscillator strength, optimization tests were performed only in the search space of the simulations used in the mapping. The solution found by the GA was always among the 20 best in the entire search space. We have then used GAs to find structures with higher oscillator strength but now without restricting the number of quantum wells in the calculation. In summary, this work not only demonstrates the feasibility of using GAs to optimize a QBMD heterostructure but provides insights into the performance of the QBMDs and presents guidelines on how to design them for specific applications.