亚阈值状态下MOSFET二维双区和单区静电势模型的比较  

Comparison between two-dimensional dual-region electrostatic potential model and single-region electrostatic potential model of MOSFET in subthreshold state

作　　者：张满红[1] 袁至衡 ZHANG Manhong;YUAN Zhiheng(Institute of Modern Electronic Science,North China Electric Power University,Beijing 102206,China)

机构地区：[1]华北电力大学现代电子科学研究所,北京102206

出　　处：《现代电子技术》2017年第10期128-132,137,共6页Modern Electronics Technique

基　　金：国家自然科学基金资助项目(61176080)

摘　　要：定义一个平均误差,该误差可以估算求解MOSFETs二维双区和单区静电势模型电势分布所用源漏端边界条件的偏差情况。首先根据长沟道模型近似确定衬底耗尽层厚度,通过平均误差计算发现双区模型最大源漏偏差远小于0.06 V,而单区模型相应的源漏偏差大于0.1 V。当器件沟道长度为亚微米级时,利用电压掺杂转换模型的耗尽层厚度计算方法对两种模型做出校正,双区电势模型在校正后的源漏条件偏差有明显的减小,单区模型的源漏偏差却会增大,尤其在短沟道以及衬底高掺杂浓度时误差较大。结果表明,双区静电势模型更为精准。An average error is defined to estimate the deviation of the source(S)and drain(D)boundary conditions tosolve the potential distribution of the two-dimensional single-region and dual-region electrostatic models of MOSFETs.The thick-ness of the substrate depletion layer is determined approximately according to the long channel model.The average error is calcu-lated to find out that the maximum S-D deviation of the dual-region model is far lower than0.06V,and the corresponding S-Ddeviation of the single-region model is higher than0.1V.If the channel length of the device belongs to submicron order,the de-pletion layer thickness calculation method of the voltage-doping transformation model is used to correct the two models.The cor-rected S-D condition deviation of the dual-region electrostatic potential model is decreased obviously,but the corrected S-D con-dition deviation of the single-region electrostatic potential model is increased,especially for the short channel and the substratewith high doping density.The results indicate that the dual-region electrostatic potential model is more accurate than the single-region electrostatic potential model.

关 键 词：单区模型 双区模型 特征函数 边界条件 平均误差 

分 类 号：TN4[电子电信—微电子学与固体电子学] TN32