基于SVR的脉冲激光沉积TiN/AlN多层薄膜的工艺优化  

Process parameters optimization for TiN/AlN multilayer films fabricated by pulsed laser deposition via SVR

作　　者：皇思洁[1] 蔡从中[1] 曾庆文[1] 

机构地区：[1]重庆大学应用物理系,重庆401331

出　　处：《功能材料》2013年第14期2074-2078,共5页Journal of Functional Materials

基　　金：中央高校基本科研业务费资助项目(WLYJSBJRCTD201102);教育部新世纪优秀人才支持计划资助项目(NCET-07-0903);留学回国人员科研启动基金资助项目(2008101-1);重庆市自然科学基金资助项目(CSTC2006BB5240)

摘　　要：根据脉冲激光沉积(PLD)法在单晶Si试样表面沉积制备多层TiN/AlN硬质膜实验数据,应用基于粒子群算法(PSO)寻优的支持向量回归(SVR)方法,建立不同工艺参数下沉积的TiN/AlN多层膜的AlN膜厚及TiN薄膜硬度的SVR预测模型。在相同的训练与测试样本集下,将SVR所得的AlN膜厚预测值与免疫径向基函数(IRBF)神经网络的计算结果进行比较。结果表明,SVR模型训练和预测结果的平均绝对百分误差要比IRBFNN模型的小,其预测精度更高,预测效果更好。应用SVR的TiN薄膜硬度模型对PLD法沉积TiN薄膜的工艺参数进行了优化,分析了多因素对PLD法沉积TiN薄膜硬度的交互作用和影响。该方法可为人们利用PLD法沉积TiN/AlN多层功能薄膜提供科学的理论指导,具有重要的理论意义和实用价值。Based on the experimental dataset of TiN/AlN multilayer films ablated on the monocrystalline silicon substrate via pulse laser deposition（PLD） technique,the support vector regression（SVR） combined with particle swarm optimization（PSO） is proposed to construct models for prediction the thickness of AlN thin films and hardness of TiN thin films in TiN/AlN multilayer films deposited under different process parameters.The predicted hardness of AlN films via established SVR model is compared with that obtained by immune radial basis function（IRBF） neural network using identical training and test samples.It is demonstrated that the mean absolute percentage errors achieved by SVR both for the training set and test set are smaller than those of IRBFNN,respectively.SVR possesses higher prediction accuracy and a better forecast capacity than IRBFNN.The established SVR model for TiN hardness was further employed to analysis and optimized the PLD deposition process,and was utilized to depict the interaction influences of multi-factor on the hardness of deposited TiN films.The method introduced in this study can provide scientific and theoretical guide in fabrication of TiN/AlN multilayer films via PLD,and it would be of important practical application value.

关 键 词：脉冲激光沉积 TIN AlN硬质多层膜 支持向量回归 回归分析 工艺优化 

分 类 号：O4-39[理学—物理] TP391[自动化与计算机技术—计算机应用技术]