Semiconductor film process parameter optimization system for neural network model prediction

Abstract

The invention provides a semiconductor film process parameter optimization system based on neural network model prediction. The invention comprises a semiconductor film process parameter optimizationdevice for neural network model prediction integrated with multiple online monitoring functions. According to the method, a semiconductor film quality training set is constructed and trained in combination with semiconductor film quality feature vectors and manually calibrated quality levels, and a trained semiconductor film quality level prediction neural network model is obtained; training the growth process parameter prediction neural network model through the semiconductor film quality training set to obtain a trained growth process parameter prediction neural network model; furthermore, the growth process parameter feature vector of the semiconductor film sample is optimized through the semiconductor film quality grade prediction neural network model and the growth process parameter prediction neural network model, so that the growth of the sample reaches a predetermined grade of the quality of the semiconductor film sample. And the growth of the epitaxial semiconductor film withoptimized growth conditions, low dislocation density, low defects and high quality is realized.

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