一种反求杂质浓度的数值方法——基于CV数据的逐点反求多次循环法

A Point by Point Multiple Sweeps Numerical Algorithm for Dopant Profiling Based on CV Data

根据CV数据用数值模拟反求杂质浓度时,遇到了不确定性问题。本文认为这实际上是用CV数据反求杂质浓度时的非单值性问题。针对数值方法的非单值性问题,本文提出了逐点反求多次循环的算法。本算法已成功地应用于分布陡峻的高斯离子注入分布及矩形阶跃分布,反求的杂质浓度分布与理论杂质浓度分布符合得很好。

Uncertainty of numerical simulation in CV profiling is in fact a problem of uniqueness. The problem has been discussed in this work. A point by point multiple sweeps numerical algorithm consists of three aspects: (1) The dopant profile is obtained by several sweeps, and smaller convergent criterion is used for a following sweep; (2) Each sweep is composed of a point by point extraction process starting from the lowest bias point to the highest bias point; (3) The gradient sign at any point of the extracting dopant profile in the overall simulation process is defined by the sign of the same point of the profile obtained by the depletion approximation. The algorithm can not only deal with very steep dopant profiles, but also can cope with the extreme case, a step change of dopant distribution. Good agreement between the extracted profiles and the theoretical ones has been obtained.