Method of monitoring and/or controlling a semiconductor manufacturing apparatus and a system therefor

• US 20030199108A1
• Filed: 05/16/2003
• Published: 10/23/2003
• Est. Priority Date: 09/06/2001
• Status: Active Grant

• Current Assignee: Akira Kagoshima, Daisuke Shiraishi, Hideyuki Yamamoto, Hiroyuki Kitsunai, Junichi Tanaka, Shoji Ikuhara, Toshio Masuda
• Original Assignee: Akira Kagoshima, Daisuke Shiraishi, Hideyuki Yamamoto, Hiroyuki Kitsunai, Junichi Tanaka, Shoji Ikuhara, Toshio Masuda
• Inventors: Tanaka, Junichi, Shiraishi, Daisuke, Ikuhara, Shoji, Yamamoto, Hideyuki, Kitsunai, Hiroyuki, Masuda, Toshio, Kagoshima, Akira
• Granted Patent: US 6,828,165 B2
• Time in Patent Office: Days
• US Class Current: 438/14
• CPC Class Codes: G05B 13/048 using a predictor
• CPC Class Codes: G05B 19/41865 characterised by job scheduling, process planning, material flow
• CPC Class Codes: G05B 2219/31357 Observer based fault detection, use model
• CPC Class Codes: G05B 2219/32058 Execute program as function of deviation from predicted state, result
• 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
• CPC Class Codes: H01L 21/67276 Production flow monitoring, e.g. for increasing throughput program-control systems per se G05B19/00, e.g. total factory control G05B19/418
• CPC Class Codes: H01L 22/20 Sequence of activities consisting of a plurality of measurements, corrections, marking or sorting steps
• CPC Class Codes: H01L 2924/00 Indexing scheme for arrangements or methods for connecting or disconnecting semiconductor or solid-state bodies as covered by H01L24/00
• CPC Class Codes: H01L 2924/0002 Not covered by any one of groups H01L24/00, H01L24/00 and H01L2224/00
• CPC Class Codes: Y02P 90/02 Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Abstract

A method and system are provided for controlling and/or monitoring a semiconductor processing apparatus while predicting its processing results. The system includes a sensor for monitoring a processing state of the processing apparatus, a sensed data storage unit for preserving sensed data sent from the sensor, an input device for inputting measured values for processing results of semiconductor devices processed by the processing apparatus, a processing result measured value storage unit for preserving the inputted processing result measured values, a model equation generation unit for generating a model equation from preserved sensed data and processing result measured values, a model equation storage unit for preserving the generated model equation, a model equation based prediction unit for predicting processing results from the preserved model equation and the sensed data, and a process recipe control unit for controlling processing conditions of the processing apparatus from predicted processing results.

Claims

1. A semiconductor processing apparatus for processing a semiconductor wafer, comprising:

a sensor for monitoring a processing state of said semiconductor processing apparatus;

processing result input means for inputting measured values for processing results of a semiconductor wafer processed by said semiconductor processing apparatus;

a model equation generation unit relying on sensed data acquired by said sensor and said measured values to generate a model equation for predicting a processing result using said sensed data as an explanatory variable;

a processing result prediction unit for predicting a processing result based on said model equation and said sensed data; and

a process recipe control unit for comparing said predicted processing result with a previously set value to control a processing condition of said semiconductor processing apparatus such that a deviation between said predicted processing result and said previously set value is corrected,

wherein said model equation generation unit generates a model equation using a PLS method (Partial Least Square method).

6. A semiconductor processing apparatus according to claim 1, further comprising:

a principal component extraction unit for extracting a principal component based on sensed data acquired by said sensor; and

a fault detection unit for detecting a fault in processing based on variations in the principal component extracted by said extraction unit,

wherein said semiconductor processing apparatus stops the processing when said fault detection unit detects a fault if said model equation generation unit has not generated a model equation.

7. A semiconductor processing apparatus according to claim 1, further comprising:

a sensed data storage unit for preserving said sensed data; and

a processing result measured value storage unit for preserving a processing result inputted by said processing result input means,

wherein said model equation generation unit generates said model equation based on the sensed data and the measured value preserved in said respective storage units associated therewith, and preserves the generated model equation in a model equation storage unit.

2. A semiconductor processing apparatus for processing a semiconductor wafer, comprising:

a sensor for monitoring a processing state of said semiconductor processing apparatus;

processing result input means for inputting measured values for processing results of a semiconductor wafer processed by said semiconductor processing apparatus;

a model equation generation unit relying on sensed data acquired by said sensor and said measured values to generate a model equation for predicting a processing result using said sensed data as an explanatory variable;

.

a processing result prediction unit for predicting a processing result based on said model equation and said sensed data; and

a process recipe control unit for comparing said predicted processing result with a previously set value to control a processing condition of said semiconductor processing apparatus such that a deviation between said predicted processing result and said previously set value is corrected,

wherein said model equation generation unit generates a model equation using robust regression.

3. A semiconductor processing apparatus for processing a semiconductor wafer, comprising:

a sensor for monitoring a processing state of said semiconductor processing apparatus;

processing result input means for inputting measured values for processing results of a semiconductor wafer processed by said semiconductor processing apparatus;

a model equation generation unit relying on sensed data acquired by said sensor and said measured values to generate a model equation for predicting a processing result using. said sensed data as an explanatory variable;

a processing result prediction unit for predicting a processing result based on said model equation and said sensed data; and

a process recipe control unit for comparing said predicted processing result with a previously set value to control a processing condition of said semiconductor processing apparatus such that a deviation between said predicted processing result and said previously set value is corrected,

wherein said model equation generation unit generates a model equation using principle component robust regression.

4. A semiconductor processing apparatus for processing a semiconductor wafer, comprising:

a sensor for monitoring a processing state of said semiconductor processing apparatus;

processing result input means for inputting measured values for processing results of a semiconductor wafer processed by said semiconductor processing apparatus;

a model equation generation unit relying on sensed data acquired by said sensor and said measured values to generate a model equation for predicting a processing result using said sensed data as an explanatory variable;

a processing result prediction unit for predicting a processing result based on said model equation and said sensed data; and

a display unit for displaying said predicted value or a deviation of said predicted value from a previously set value.

5. A semiconductor processing apparatus for processing a semiconductor wafer, comprising:

a multiplicity of sensors for monitoring processing states of said semiconductor processing apparatus;

a principal component extraction unit for extracting a principal component based on a multiplicity of sensed data acquired by said multiplicity of sensors; and

a fault detection unit for detecting a fault in processing based on variations in the principal component extracted by said extraction unit.

8. A method of controlling processing of an semiconductor object to be processed in a semiconductor processing apparatus using a sensor for monitoring a processing state of said semiconductor processing apparatus, comprising the steps of:

inputting measured values for processing results of a semiconductor object processed by said semiconductor processing apparatus to hold the input measured values;

based on sensed data acquired by said sensor and said held measured values, generating a model equation for predicting a processing result using said sensed data as an explanatory variable;

predicting a processing result of the semiconductor object based on said model equation and said sensed data; and

comparing said predicted processing result with a pre-set value and controlling a processing condition of said semiconductor processing apparatus based on the comparison result so as to connect an error between said predicted processing result and said preset value,

wherein said model equation generation step generate a model equation using a PLS method (Partial Least Square method).

13. A semiconductor processing controlling method according to claim 8, further comprising the steps of:

extracting a principal component based on sensed data acquired by said sensor;

detecting a fault in processing based on variations in the principal component extracted by said extraction step, and

stopping the processing operation of said semiconductor processing apparatus when said fault detection unit detects a fault if said model equation generation step has not generated a model equation.

14. A semiconductor processing controlling method according to claim 8, further comprising:

preserving said sensed data in a sensed data storage unit; and

preserving a processing result inputted by said processing result input step in a processing result measured value storage unit,

wherein said model equation generation step generates said model equation based on the sensed data and the measured value preserved in the respective storage units associated therewith, and preserves the generated model equation in a model equation storage unit.

9. A method of controlling processing of an semiconductor object to be processed in a semiconductor processing apparatus using a sensor for monitoring a processing state of said semiconductor processing apparatus, comprising the steps of:

inputting measured values for processing results of a semiconductor object processed by said semiconductor processing apparatus to hold the input measured values;

based on sensed data acquired by said sensor and said held measured values, generating a model equation for predicting a processing result using said sensed data as an explanatory variable;

predicting a processing result of the semiconductor object based on said model equation and said sensed data; and

comparing said predicted processing result with a pre-set value and controlling a processing condition of said semiconductor processing apparatus based on the comparison result so as to connect an error between said predicted processing result and said preset value,

wherein said model equation generation step generates a model equation using robust regression.

10. A method of controlling processing of an semiconductor object to be processed in a semiconductor processing apparatus using a sensor for monitoring a processing state of said semiconductor processing apparatus, comprising the steps of:

inputting measured values for processing results of a semiconductor object processed by said semiconductor processing apparatus to hold the input measured values;

based on sensed data acquired by said sensor and said held measured values, generating a model equation for predicting a processing result using said sensed data as an explanatory variable;

predicting a processing result of the semiconductor object based on said model equation and said sensed data; and

comparing said predicted processing result with a pre-set value and controlling a processing condition of said semiconductor processing apparatus based on the comparison result so as to connect an error between said predicted processing result and said preset value,

wherein said model equation generation step generates a model equation using principle component robust regression.

11. A semiconductor processing monitoring method for monitoring processing of a semiconductor object to be processed in a semiconductor processing apparatus using a sensor for monitoring a processing state of said semiconductor processing apparatus, comprising the steps of:

inputting measured values for processing results of a semiconductor object processed by said semiconductor processing apparatus;

based on sensed data acquired by said sensor and said measured values, generating a model equation for predicting a processing result using said sensed data as an explanatory variable;

predicting a processing result of the semiconductor object based on said model equation and said sensed data; and

displaying said predicted value or a deviation of said predicted value from a pre-set value.

12. A semiconductor processing monitoring method for monitoring processing of a semiconductor wafer in a semiconductor processing apparatus using a multiplicity of sensors for monitoring processing states of said semiconductor processing apparatus, comprising the steps of:

extracting a principal component based on a multiplicity of sensed data acquired by said multiplicity of sensors; and

detecting a fault in processing based on variations in the principal component extracted by said extraction step.