Using superlattice ordering to reduce the band gap of random (In,Ga)As/InP alloys to a target value via the inverse band structure approach

Abstract

Thermophotovoltaic (TPV) devices are intended to absorb photons from hot blackbody radiating objects, often requiring semiconductor absorbers with band gap of $\ensuremath{\simeq}0.6\text{ }\text{eV}$. The random ${\text{In}}_{x}{\text{Ga}}_{1\ensuremath{-}x}\text{As}$ alloy lattice matched $({x}_{\text{In}}=0.53)$ to a (001) InP substrate has a low-temperature band gap of 0.8 eV, about 0.2 eV too high for a TPV absorber. Bringing the band gap down by raising the In concentration induces strain with the substrate, leading to a two-dimensional $(2\text{D})\ensuremath{\rightarrow}\text{three}$-dimensional (3D) morphological transition occurring before band gaps suitable for TPV applications are achieved. We use the inverse band structure approach, based on a genetic algorithm and empirical pseudopotential calculations, to search for lattice-matched InAs/GaAs multiple-repeat unit structures with individual layer thicknesses lower than the critical thickness for a $2\text{D}\ensuremath{\rightarrow}3\text{D}$ transition. Despite the fact that quantum confinement usually increases band gaps, we find a quantum superlattice structure with the required reduced gap (and a significant optical transition) that matches all target requirements. This is explained by the predominance of (potential-energy) level anticrossing effects over (kinetic) quantum confinement effects.