Research on Diameter Prediction of Silicon Single Crystal Based on Data Driven

Abstract

Czochralski silicon single crystal growth is a dynamic time-varying process with multi-field and multi-phase coupling, complex physical changes, nonlinearity and large hysteresis, but the mechanism model based on a large number of assumptions is difficult to apply in practice. Therefore, this article is based on the long-term and massive crystal growth data of the existing CL120-97 single crystal furnace crystal pulling workshop, ignoring the complex crystal growth environment in the furnace, and analyzing the correlation of the crystal pulling parameters the affect of crystal diameter. Mining the data Contains regular information, and further builds a crystal diameter prediction model based on BP neural network. The model prediction results are verified by actual crystal pulling data. The results show that the average relative percentage error is 0.08355% for 6 groups of randomly selected crystal pulling data, which proves that the model is feasible for predicting crystal diameters at the equal diameter stage.