LeRoy Apker Award (2020): Stochastic resonance enhancement of the resolution of charge coupled device based thermoreflectance imaging

CCD-based thermoreflectance imaging is a non-contact, high spatial resolution technique used in a wide range of applications, including monitoring thermal lensing in lasers, the performance of photovoltaics, and defects in optoelectronics. A 4-bucket algorithm is used to perform pixel-by-pixel lock-in averaging to measure the relative change in reflectance of a sample in response to an induced thermal modulation. Prior experiments demonstrated that the technique can measure signals below the quantization limit of the camera; this enhanced resolution is posited to be due to stochastic resonance, where measurement noise dithers the signal over many bit levels. Here, we develop an experimentally validated analytical and computational model of the stochastic resonance in this system, investigating how measurement noise, when combined with the averaging required by the imaging algorithm, can be used to maximize the thermal resolution. We show that noise is required to obtain accurate thermoreflectance measurements and that in the absence of noise, the A-D conversion of the camera can lead to measurement errors. By tuning experimental parameters, stochastic resonance can be achieved for any noise level, enabling an order of magnitude or greater enhancement in the thermal resolution.