Method and apparatus for monitoring a process by employing principal component analysis

• US 6,368,975 B1
• Filed: 07/07/1999
• Issued: 04/09/2002
• Est. Priority Date: 07/07/1999
• Status: Expired due to Term

• Current Assignee: Applied Materials Incorporated
• Original Assignee: Applied Materials Incorporated
• Inventors: Sarfaty, Moshe, Lymberopoulos, Dimitris, Balasubramhanya, Lalitha, Davidow, Jed
• Primary Examiner(s): Utech, Benjamin L.
• Assistant Examiner(s): DEO, DUY VU NGUYEN
• Application Number:
  US09/348,972
• Time in Patent Office: 1,007 Days
• Field of Search: 438/5, 438/6, 438/7, 438/8, 438/9, 438/10, 438/11, 438/12, 438/13, 438/690, 438/706, 438/712, 438/710, 216/58, 216/59, 216/60
• US Class Current: 438/706
• CPC Class Codes: G01N 21/274 Calibration, base line adjustment, drift correction
• CPC Class Codes: G01N 21/3563 for analysing solids; Preparation of samples therefor
• CPC Class Codes: G01N 21/359 using near infrared light
• CPC Class Codes: G01N 2201/1293 resolving multicomponent spectra
• CPC Class Codes: G05B 23/024 Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks

Abstract

A method and apparatus for monitoring a process by employing principal component analysis are provided. Correlated attributes are measured for the process to be monitored (the production process). Principal component analysis then is performed on the measured correlated attributes so as to generate at least one production principal component; and the at least one production principal component is compared to a principal component associated with a calibration process (a calibration principal component). The calibration principal component is obtained by measuring correlated attributes of a calibration process, and by performing principal component analysis on the measured correlated attributes so as to generate at least one principal component. A principal component having a feature indicative of at least one of a desired process state, process event and chamber state then is identified and is designated as the calibration principal component. Preferably the at least one production principal component is compared to the calibration principal component by computing the inner product of the calibration and production principal components.

Claims

1. A method of detecting an endpoint of a semiconductor fabrication process, comprising:

(a) performing a calibration process in which a first workpiece is exposed to a plasma;

(b) collecting first optical emission spectroscopy (OES) data for electromagnetic radiation emitted by the plasma during the calibration process;

(c) determining a timing of an endpoint that occurred during the calibration process;

(d) performing principal component analysis with respect to a window of the first OES data that corresponds to the determined endpoint timing to compute a principal component of the first OES data in the window.

(e) after steps (a)-(d), performing a production process in which a second workpiece is exposed to a plasma;

(f) collecting second OES data for electromagnetic radiation emitted by the plasma during the production process;

(g) for a series of windows of the second OES data, performing principal component analysis to compute a respective principal component for each window of the second OES data and comparing the principal component computed for each window with the principal component computed at step (d), wherein the comparing includes calculating inner products of the principal component computed at step (d) with the principal components computed at this step (g); and

(h) detecting an endpoint of the production process based on a result of step (g).

2. The method of claim 1, wherein step (h) includes detecting a transition in the calculated inner products.

3. The method of claim 1, wherein the workpieces are silicon wafers each having a multilayer semiconductor structure.

4. The method of claim 3, wherein the calibration process and the production process each include etching a layer of a multilayer semiconductor structure.

5. The method of claim 4, wherein the etched layer includes silicon dioxide.

6. The method of claim 4, wherein the etched layer includes a metal.

7. The method of claim 4, wherein the etched layer includes polysilicon.

8. The method of claim 4, wherein the etched layer includes a bottom-anti-reflective-coating.

9. The method of claim 1, wherein the first and second OES data are collected for electromagnetic emissions having wavelengths from about 180 to 850 nanometers.

10. The method of claim 1, wherein the first and second OES data is mean-centered before performing the principal component analyses.

11. The method of claim 1, wherein step (c) includes microscopic examination of the workpiece.

12. The method of claim 1, wherein steps (f)-(h) are performed concurrently with step (e).

13. The method of claim 1, wherein the series of windows of the second OES data are generated at intervals of about one second.

14. The method of claim 1, wherein the endpoint detected at step (h) is completion of an etching process.