The idea of this paper is to develop a method for determination thermal diffusivity, linear expansion coefficient and thickness of a semiconductor sample from photoacoustic phase measurement by using neural network. The neural network has been trained on photoacoustic phases obtained from a theoretical model of measured signal for Si n-type in the range of 20Hz to 20kHz. For the analysis of parameter determination from phases, we trained phase neural networks on a large database obtained from numerical experiments in the expected range of parameter changes. An analysis of a theoretical photoacoustic model with a phase neural network is demonstrated.
The advantages of using a phase neural network with high accuracy and precision in prediction depending on the number of epochs are presented, as well as analyzes of the application of random Gaussian noise to the network in order to better predict the experimental photoacoustic signal.