MOCVD外延生长Ga_(1-x)Al_xAs_(1-y)Sb_y半导体薄膜气固平衡和人工神经网络预报

Solid vapour Equilibrium of Preparation of Ga 1 x Al xAs 1 y Sb y Semiconductor Epilayer by MOCVD and Artificial Neural Network Prediction

本文研究了ＭＯＣＶＤ外延生长Ｇａ１－ｘＡｌｘＡｓ１－ｙＳｂｙ半导体薄膜的生长条件与外延层组成的关系，并用人工神经网络法总结有关气固平衡规律。结果表明，用气相组成，载气流量和生长温度等影响外延层组成的主要参数作为人工神经网络的输入，以固相Ｇａ１－ｘＡｌｘＡｓ１－ｙＳｂｙ中的Ａｌ和Ｓｂ的含量ｘ、ｙ作为输出，训练的人工神经网络可以预报固相组成ｘ、ｙ，得到满意结果。

The solid vapour equilibrium of metalorganic chemical vapour deposition of Ga 1 x Al xAs 1 y Sb y epilayer on GaSb substrates are studid both by experimental measurement and artificial neural network prediction. It was shown that an artificial neural network,which is trained by using the composition of source vapour,flow rate of carrying gas, growth temperature as neural metwork input vector and composition of epilayer as neural network output vector,could be satisfactorily applied to predict the composition of Ga 1 x Al xAs 1 y Sb y epilayer.