Probing the defect chemistry in gallium arsenide using the Atomistically Informed Device Engineering (AIDE) Method

Atomic defects play a central role on semiconductor device performance. However, characterizing defect formation often depends on dynamical processes which are largely inaccessible to experimental observations and immensely challenging for state-of-the-art theoretical models to implement. In this work, we present a multiscale Atomistically Informed Device Engineering (AIDE) method which utilizes experimental (capture cross-sections, doping populations, etc.) and atomistic (defect levels, migration barriers, etc.) information to probe the dynamical behavior of defects. We demonstrate the capabilities of our AIDE method on an irradiated Si-doped GaAs diode by i) obtaining the quasi-Fermi levels of the system, ii) generating the correct defect charge states via charge carrier reactions, and iii) implementing a diffusion-driven defect-defect reaction. The AIDE method is generalizable to other semiconductors and can be used to probe defect behavior at defect concentrations and timescales which are inaccessible to experimental and theoretical approaches. With these capabilities, longstanding mysteries in semiconductors such as the missing Ga vacancy in GaAs can be studied and resolved.