A Neural Network Model Prediction System for Optimizing Process Parameters of Semiconductor Thin Films

Abstract

The invention proposes a semiconductor thin film process parameter optimization system predicted by a neural network model. The invention includes a semiconductor thin film process parameter optimization device integrated with neural network model prediction of multiple online monitoring functions. The present invention combines semiconductor thin film quality feature vectors and manually calibrated quality grades to construct a semiconductor thin film quality training set and perform training to obtain a trained semiconductor thin film quality grade prediction neural network model; use the semiconductor thin film quality training set to predict growth process parameters. Neural network model Conduct training to obtain the neural network model for predicting the growth process parameters after training; further optimize the growth process parameter eigenvector of the semiconductor film sample through the semiconductor film quality grade prediction neural network model and the growth process parameter prediction neural network model, so that the sample grows to the semiconductor film sample. Quality predetermined level. Realize the optimization of growth conditions, the growth of low dislocation density, low defect and high quality epitaxial semiconductor film.

Description

technical field [0001] The invention belongs to the technical field of epitaxial semiconductor thin film growth, and in particular relates to a semiconductor thin film process parameter optimization system predicted by a neural network model. Background technique [0002] Whether it is LD, LED or power electronic devices based on nitride, silicon carbide, diamond and other base materials, their performance is highly dependent on the quality control of the growth process of semiconductor thin film materials. In the growth process of semiconductor thin film materials, on-line monitoring technology, as the premise of controlling the growth process, is an important part of semiconductor material growth technology. In order to achieve the uniformity, repeatability, controllability, and low defect requirements of the growth film, during the material growth process, it is necessary to control the parameters that directly affect the material properties, such as reaction temperature,...

AI Technical Summary

Problems solved by technology

However, there is currently no semiconductor thin film material growth equipment with integrated mass spectrometer on-line monitoring

[0006] Another important measuring instrument, the reflective high-energy electron diffractometer is the standard measuring equipment for molecular beam epitaxy equipment. The growth process and growth mode of the semiconductor thin film can be observed in real time through the electron diffraction pattern, but it cannot capture the instantaneous change information of the material structure.

Images