System for indirectly monitoring and controlling a process with particular application to plasma processes

• US 5,711,843 A
• Filed: 02/21/1995
• Issued: 01/27/1998
• Est. Priority Date: 02/21/1995
• Status: Expired due to Fees

• Current Assignee: Orincon Technologies Incorporated (Martin Marietta Corporation)
• Original Assignee: Orincon Technologies Incorporated (Martin Marietta Corporation)
• Inventors: Jahns, Gary L.
• Primary Examiner(s): TUNG, TA HSUNG
• Application Number:
  US08/392,631
• Time in Patent Office: 1,071 Days
• Field of Search: 156/626.1, 156/643.1, 156/646.1, 156/345, 216/58, 216/60, 216/65-67, 216/85
• US Class Current: 156/345.24
• CPC Class Codes: B24B 37/013 Devices or means for detecting lapping completion
• CPC Class Codes: H01J 37/32935 Monitoring and controlling tubes by information coming from the object and/or discharge
• CPC Class Codes: H01J 37/32972 Spectral analysis
• CPC Class Codes: H01J 37/3299 Feedback systems
• CPC Class Codes: H01L 22/20 Sequence of activities consisting of a plurality of measurements, corrections, marking or sorting steps
• CPC Class Codes: H01L 22/26 Acting in response to an ongoing measurement without interruption of processing, e.g. endpoint detection, in-situ thickness measurement endpoint detection arrangements in CMP apparatus B24B37/013, in discharge apparatus H01J37/32

Abstract

The invention enables real-time control of a process using information regarding process properties that are indirectly related to the state of the process. A set of properties that characterize the process environment (fingerprint) is measured and used by a process results estimator to infer information regarding the state of the process and by a process condition monitor for monitoring the process to ascertain whether a particular type of condition exists. In one embodiment, optical emission spectra (OES) are used as the fingerprint. The process results estimator is sufficiently powerful to enable the process state to be inferred even when the relationship between the process environmental properties and the process state is complicated and difficult to describe with traditional mathematical models. In one embodiment, the process results estimator is embodied by a neural network. The process condition monitor can also be embodied by a neural network. Because the invention does not directly measure the process state, the invention is particularly useful in situations in which it is difficult or undesirable to directly determine the state of the process during the process such as monitoring and control of a plasma process such as plasma etching.

Claims

1. A system for controlling a process implemented by a processing apparatus, the process being characterized by process state having a plurality of process properties, each process property having a process property value at each instant in time, comprising:

means for detecting a process fingerprint, the process fingerprint being indirectly representative of a plurality of process property values;

a process condition monitor for monitoring the process in real-time to ascertain whether a particular process condition exists;

a process property estimator for estimating a process property value at an instant in time based on the process fingerprint;

means for comparing the process property value estimate to a target process property value to produce a process property value difference; and

means for determining a control response as a function of the process property value difference, the control response adapted to change the process property value and the process state of the processing apparatus in real-time based on the process property value difference.

2. A system as in claim 1, wherein:

the process condition monitor is a condition monitor neural network; and

the process property results estimator is a property estimator neural network.

3. A system as in claim 2, wherein the condition monitor neural network is a fuzzy min-max neural network.

4. A system as in claim 2, wherein the property estimator neural network is a multilayer perceptron neural network.

5. A system as in claim 2, wherein:

the condition monitor neural network is a fuzzy min-max neural network; and

the property estimator neural network is a multilayer perceptron neural network.

6. A system as in claim 1, wherein the means for determining a control response further comprises:

a fixed translator that uses the process property value difference to produce a coarse estimate of the control response; and

an adaptive translator that uses the process property value difference and the coarse estimate to produce a fine-tuned control response.

7. A system as in claim 1, wherein the process is a plasma process.

8. The system of claim 1, wherein the process signature detecting means comprises means for detecting electromagnetic radiation from the process, the electromagnetic radiation having a plurality of wavelengths.

9. A system for controlling a plasma process implemented by a processing apparatus, the process being characterized by a process state having a plurality of process properties, each process property having a process property value at each instant in time, comprising:

means for detecting a process fingerprint, the process fingerprint being indirectly representative of a plurality of process property values;

a multilayer perceptron neural network for producing a process property value estimate at an instant in time for a particular process property based on said process fingerprint;

means for comparing the process property value estimate to a target process property value to produce a process property value difference; and

means for determining a control response as a function of the process property value difference, the control response adapted to change the process property value and the process state of the processing apparatus in real-time based on the process property value difference.

10. The system of claim 9, wherein the process signature detecting means comprises means for detecting electromagnetic radiation from the process, the electromagnetic radiation having a plurality of wavelengths.

11. A system for controlling a plasma process implemented by a processing apparatus, the process being characterized by a process state having a plurality of process properties, each process property having a process property value at each instant in time, comprising:

means for detecting a process fingerprint, the process fingerprint being indirectly representative of a plurality of process property values;

a fuzzy min-max neural network for monitoring the process to ascertain whether a particular process property value exists and to generate a process property value estimate;

means for comparing the process property value estimate to a target process property value to produce a process property value difference; and

means for determining a control response as a function of the process property value difference, the control response adapted to change the process property value and the process state of the processing apparatus in real-time based on the process property value difference.

12. The system of claim 11, wherein the process signature detecting means comprises means for detecting electromagnetic radiation from the process, the electromagnetic radiation having a plurality of wavelengths.