Process monitor

• US 6,894,786 B1
• Filed: 07/12/1999
• Issued: 05/17/2005
• Est. Priority Date: 07/11/1998
• Status: Expired due to Fees

• Current Assignee: Surface Technology Systems Plc (KLA Corporation), Vorgem Limited
• Original Assignee: Vorgem Limited
• Inventors: Beckmann, William George, Holbrook, Mark Burton, Grange, Jacques Andre
• Primary Examiner(s): Turner, Samuel A.
• Assistant Examiner(s): CONNOLLY, PATRICK J
• Application Number:
  US09/743,164
• Time in Patent Office: 2,136 Days
• Field of Search: 356/450, 356/454, 356/519, 356/451, 438/9, 438/7
• US Class Current: 356/454
• CPC Class Codes: G01N 21/73 using plasma burners or torches

Abstract

A substrate is etched in a vacuum enclosure in a process which generates plasma light emission. The process is monitored by passing emitted light via a window, a thin film narrow band filter and a “Fabry-Perot” etalon to a detector. The output signal from the detector is analyzed by shape recognition techniques to derive a measure of the progress of the process. The shape recognition preferably makes use of digital filtering and comparison with reference data derived from the theoretical analysis or from a calibration run.

Claims

1. A method for automatically determining the progress of a process, said process being one based upon the use of a plasma which exhibits a changing spectral characteristic as the process progresses, said spectral characteristic being selected from the group consisting of absorption and radiation;

the method comprising;

continuously monitoring at least part of said spectral characteristic to provide a spectral signal, and

electronically applying shape recognition techniques to the shape of said signal.

2. The method of claim 1, in which said shape recognition techniques are applied to the shape of said signal in the time domain.

3. The method of claim 2, further including the preliminary step of applying shape recognition techniques to said spectral characteristics in the optical frequency domain.

4. The method of claim 1, in which the shape recognition is achieved by digital filtering.

5. The method of claim 1, in which the shape recognition is achieved by a series of masks derived from different time epochs.

6. The method of claim 5, in which said series of masks is derived using a method selected from the group comprising a Gradiometer transform, a Fourier transform, a Laplace transform, a Kohonen self-organising map, a cellular neural network paradigm, polynomial interpolated measures, and fractals.

7. A process control system for automatically determining the progress of a process, said process being one based upon the use of a plasma which exhibits a changing spectral characteristic as the process progresses, said spectral characteristic being selected from the group consisting of absorption and radiation, the system comprising:

monitoring means for continuously monitoring at least part of said spectral characteristic to provide a spectral signal, and

shape recognition means for applying shape recognition techniques to the shape of said signal to compare said signal with a predicted output to provide a measure of the progress of the process.

8. The system of claim 7, in which said shape recognition means comprises digital filtering means operating upon said signal in the time domain.

9. The system of claim 8, in which said monitoring means includes spectral detection means adapted to perform shape recognition on said spectral characteristic in the optical frequency domain.