Apparatus and method of segmenting sensor data output from a semiconductor manufacturing facility

Patent metadata

Field	Value
Publication number	US-9696717-B2
Application number	US-201414279629-A
Country	US
Kind code	B2
Filing date	May 16, 2014
Priority date	Aug 5, 2013
Publication date	Jul 4, 2017
Grant date	Jul 4, 2017

Abstract

An apparatus and method of segmenting sensor data are provided. The apparatus includes a sensor, a first segmentation unit, a continuity evaluation unit, a second segmentation unit, and a segmentation determination unit. The sensor collects sensor data for a process of the semiconductor manufacturing facility. The first segmentation unit extracts a variation point of the sensor data to perform an abnormal difference (AD) segmentation on the sensor data based on the at least one variation point. The continuity evaluation unit evaluates a continuity ratio of the sensor data. The second segmentation unit performs a free-knot spline (FS) segmentation on the sensor data when the continuity ratio exceeds a reference ratio. The segmentation determination unit compares the AD segmentation result with the FS segmentation result and to select one of the results on the comparison result.

First claim

What is claimed is:

1. An apparatus of segmenting sensor data output from a semiconductor manufacturing facility, comprising:
a sensor operative to collect sensor data corresponding to at least one process of the semiconductor manufacturing facility;
a first segmentation unit operative to extract at least one variation point of the sensor data to perform an abnormal difference (AD) segmentation on the sensor data based on the at least one variation point, and to generate first segments;
a continuity evaluation unit operative to evaluate a continuity ratio of the sensor data;
a second segmentation unit operative to perform a free-knot spline (FS) segmentation on the sensor data when the continuity ratio exceeds a reference ratio and to generate second segments; and
a segmentation determination unit operative to compare a result of performing the AD segmentation with a result of performing the FS segmentation and to select either the first segments or the second segments based on the comparison result.

2. The apparatus of claim 1 , further comprising a pre-processing unit operative to make lengths of the sensor data equal to a first length, wherein the first segmentation unit is operative to determine a first point in the sensor data having the first length as the at least one variation point when an amount of variation at the first point is larger than a predetermined value.

3. The apparatus of claim 2 , wherein the pre-processing unit is operative to filter out abnormal data among the sensor data.

4. The apparatus of claim 2 , wherein the first segmentation unit is operative to recognize the at least one variation point as a separator to perform the AD segmentation.

5. The apparatus of claim 2 , wherein the continuity evaluation unit is operative to find a first time period corresponding to third segments among the first segments, and to evaluate the continuity ratio of the sensor data used on the first time period, and the third segments appear consecutively and each of the third segments has a non-zero value of variation.

6. The apparatus of claim 5 , wherein the continuity evaluation unit is operative to determine the sensor data as continuous-type data when the continuity ratio exceeds the reference ratio.

7. The apparatus of claim 5 , wherein the second segmentation unit is operative to perform the FS segmentation by repeatedly estimating and removing a position of a knot.

8. The apparatus of claim 1 , further comprising a post-processing unit operative to perform a post-processing process, wherein the post-processing process comprises incorporating a one-point segment into a segment nearer the one-point segment among the left and right segments with respect to the one-point segment when the one-point segment is among the first segments and the second segments.

9. The apparatus of claim 1 , wherein the reference ratio is about 50%.

10. The apparatus of claim 1 , wherein the selecting of either the first segments or the third segments is performed based on a determination of which segments of the first and second segments has less noise and a smaller number of segments.

11. A computer system comprising:
a processor; and
a non-transitory, tangible, program storage medium, readable by the computer system, embodying a program of instructions executable by the processor to perform method steps for segmenting sensor data output from a semiconductor manufacturing facility, the method comprising:
collecting sensor data corresponding to at least one process of the semiconductor manufacturing facility;
extracting at least one variation point of the sensor data;
performing an abnormal difference (AD) segmentation on the sensor data based on at least one variation point to generate first segments;
finding second segments among the first segments, wherein the second segments appear consecutively and each of the second segments has a non-zero of variation;
evaluating a continuity ratio of the sensor data based on a time period corresponding to the second segments;
performing a free-knot spline (FS) segmentation on the sensor data when the continuity ratio exceeds a reference ratio to generate third segments;
comparing a result of performing the AD segmentation with a result of performing the FS segmentation; and
selecting either the first segments or the third segments based on the comparison result.

12. The computer system of claim 11 , the method further comprising performing a pre-processing process of making lengths of the sensor data equal to a first length, wherein the extracting of the at least one variation point comprises determining a first point in the sensor data as the at least one variation point when an amount of variation at the first point is larger than a predetermined value.

13. The computer system of claim 12 , wherein the performing of the AD segmentation comprises recognizing the at least one variation point as a separator.

14. The computer system of claim 12 , wherein the evaluating of a continuity evaluation unit comprises determining the sensor data as continuous-type data when the continuity ratio exceeds the reference ratio.

15. The computer system of claim 14 , wherein the performing of the FS segmentation comprises repeatedly estimating and removing a position of a knot.

16. The computer system of claim 11 , wherein the reference ratio is about 50%.

17. The computer system of claim 11 , wherein the selecting of either the first segments or the third segments is performed based on a determination of which segments of the first and third segments has less noise and a smaller number of segments.

18. A method, implemented by an apparatus, of segmenting sensor data output from a semiconductor manufacturing facility, comprising:
collecting, by a sensor of the apparatus, sensor data corresponding to at least one process of the semiconductor manufacturing facility;
performing, by a pre-processing unit of the apparatus, a pre-processing process to make lengths of the sensor data for a plurality of wafers equal to a first length;
extracting, by a first segmentation unit of the apparatus, at least one variation point of the sensor data;
performing, by the first segmentation unit, a segmentation using a first algorithm on the sensor data based on at least one variation point to generate first segments;
finding, by a continuity evaluation unit of the apparatus, second segments among the first segments;
evaluating, by the continuity evaluation unit, a continuity ratio of the sensor data based on a time period corresponding to the second segments, wherein the second segments appear consecutively and each of the second segments has a non-zero of variation;
performing, by a second segmentation unit of the apparatus, a segmentation using a second algorithm on the sensor data when the continuity ratio exceeds a reference ratio to generate third segments;
comparing, by a segmentation determination unit of the apparatus, a result of performing the segmentation using the first algorithm with a result of performing the segmentation using the second algorithm; and
selecting, by the segmentation determination unit, either the first segments or the third segments based on the comparison result; and
identifying one or more defective wafers fabricated within the semiconductor manufacturing facility using the selected segments.

19. The method of claim 18 , wherein the first algorithm is an abnormal difference algorithm and the second algorithm is an free-knot spline algorithm.

20. The method of claim 1

Assignees

• Samsung Electronics Co Ltd
• Seoul Nat Univ R & Db Found

Inventors

• Park Hae-Sang
• Lee Young-Hak
• Kim Jung-Hee
• Jeon Heui-Sik
• Park Eunjeong Lucy
• Cho Sungzoon
• Park Jooseoung
• Yang Jiwon

Classifications

• H10P95/00 Primary

  Generic processes or apparatus for manufacture or treatments not covered by the other groups of this subclass · CPC title

• G05B2219/37224

  Inspect wafer · CPC title

• G05B23/0221 Primary

  Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods · CPC title

• Y02P90/14

  Cross-Sectional Technologies · mapped topic

• Y02P90/02

  Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS] · CPC title