公开(公告)号：JP2010282636A

公开(公告)日：2010-12-16

申请号：JP2010156714

申请日：2010-07-09

Applicant: Lam Res Corp ,  ラム リサーチ コーポレイション

Inventor： BARONE MARIA E ,  GOTTSCHO RICHARD A ,  VAHEDI VAHID

IPC: C23C14/54 ,  G06Q10/00 ,  C23C16/52 ,  H01J37/32 ,  H01L21/205 ,  H01L21/3065

CPC classification number: H01J37/32935

Abstract: PROBLEM TO BE SOLVED: To provide a method for predicting a processing surface profile predicting a process surface profile that a given plasma process will create on a process substrate. SOLUTION: In this method, test values of input variables are selected (200), plasma is modeled (210), results of a test process of the substrate surface profile are substantially predicted by using the result in 210 and substrate parameters given in 200 (220), initial values of the surface profile model related to the input variables and unknown coefficients are obtained (230), instructions for difference between the test surface profile and the substantial profile predicted value are generated, and optimum values of unknown coefficients for minimizing the instructions for difference are generated. COPYRIGHT: (C)2011,JPO&INPIT

Abstract translation: 要解决的问题：提供一种预测处理表面轮廓的方法，该处理表面轮廓预测给定等离子体处理将在处理衬底上产生的工艺表面轮廓。

解决方案：在此方法中，选择输入变量的测试值（200），对等离子体进行建模（210），通过使用210中的结果和给定的衬底参数，基本上预测了衬底表面轮廓的测试过程的结果 在200（220）中，获得与输入变量和未知系数相关的表面轮廓模型的初始值（230），产生测试表面轮廓和实际轮廓预测值之间的差异指令，并且产生未知系数的最佳值 生成差异指令的最小化。 版权所有（C）2011，JPO＆INPIT

公开(公告)号：AT293840T

公开(公告)日：2005-05-15

申请号：AT99907143

申请日：1999-02-18

Applicant: LAM RES CORP

Inventor： BARONE MARIA E ,  GOTTSCHO RICHARD A ,  VAHEDI VAHID

IPC: C23C14/54 ,  C23C16/52 ,  G06F19/00 ,  H01J37/32

Abstract: A method and apparatus for calibrating a semi-empirical process simulator used to determine process values in a plasma process for creating a desired surface profile on a process substrate includes providing a test model which captures all mechanisms responsible for profile evolution in terms of a set of unknown surface parameters. A set Sets of test conditions processes is are derived for which the profile evolution is governed by only a limited number of parameters. For each set of test conditions process, model test values are selected and a test substrate is substrates are actually subjected to a the test process processes defined by the test values, thereby creating a test surface profile profiles. The test values are used to generate an approximate profile prediction predictions and are adjusted to minimize the discrepancy between the test surface profile profiles and the approximate profile prediction predictions, thereby providing a final model of the profile evolution in terms of the process values.