Accurate Determination of Semiconductor Diffusion Coefficient Using Confocal Microscopy

Confocal microscopy is a versatile tool capable of directly monitoring photoexcited carrier transport and recombination in thin film and single crystal samples. The diffusion coefficient, an important material property for designing efficient optoelectronic devices, is often determined by fitting the evolution of the photoexcited carrier population to a simplified Gaussian function. Although this model has gained widespread adoption, its application to different material systems and its sensitivity to various experimental conditions has not been explored. Here, we simulate the diffusive processes in metal halide perovskites and find that the diffusion coefficient can be inaccurately fit when higher-order processes, such as bimolecular and Auger recombination, dominate. Significant fitting error (> 5%) is introduced if the initial photoexcited carrier density exceeds 1x10$^1$$^7$ cm$^-$$^3$ and if the material diffusion coefficient is less than ~ 1 cm$^2$ s$^-$$^1$, both conditions commonly encountered in confocal microscopy measurements of perovskites. In addition, we find that grain size and grain boundaries present in polycrystalline thin films impact the carrier density temporal profiles, introducing more error in the diffusion coefficient fits. This analysis highlights important considerations in the interpretation of confocal microscopy data and provides critical steps towards the development of more robust diffusion models.