Method for detecting and analyzing silicon single crystal slice sample wafer

Through the Warp, thickness, BOW value detection and database analysis of silicon single crystal slice samples, the problem of single traditional detection data is solved, and the quality monitoring of multi-wire cutting machines is realized, the generation of unqualified silicon wafers is reduced, and the Detection accuracy and efficiency.

CN119069375APending Publication Date: 2024-12-03FERROTEC (NINGXIA) SEMICON TECH CO LTD
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411210487.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-30
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The traditional silicon single crystal slice sample detection data is single and incomplete, and the product quality cannot be effectively controlled.

Method used

Using Warp, thickness, BOW value detection, obtain the detection data of each batch of silicon wafer and multi-wire cutting machine model, establish a database, and perform abnormal detection, statistical analysis and trend prediction to realize the silicon wafer cut from the cutting machine. Quality monitoring.

Benefits of technology

Through detailed inspection and analysis, the multi-wire cutting machine can be adjusted in time to reduce the generation of unqualified silicon wafers, reduce waste, and improve detection accuracy and efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure HDA0005020191750000011
    Figure HDA0005020191750000011
  • Figure HDA0005020191750000012
    Figure HDA0005020191750000012
  • Figure HDA0005020191750000021
    Figure HDA0005020191750000021
Patent Text Reader

Abstract

The invention provides a silicon single crystal slice sample wafer detection and analysis method, and relates to the technical field of single crystal furnace power control systems, the silicon single crystal slice sample wafer is subjected to Warp, thickness and BOW value detection, and detection data of different indexes of silicon wafers of each batch number and the model of a corresponding multi-wire cutting machine are obtained; storing the obtained detection data of different indexes of the silicon wafer of each batch number and the model of the corresponding multi-wire cutting machine to obtain a database; abnormality detection, statistical analysis and trend prediction are conducted on the detection data according to the obtained database so that quality monitoring can be conducted on the silicon wafers cut by the multi-wire cutting machine, the multi-wire cutting machine can be adjusted in time according to statistical analysis and trend prediction conducted on the detection data, the probability that the silicon wafers are unqualified is reduced, and waste is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of single crystal furnace power control systems, and in particular relates to a silicon single crystal slice sample detection and analysis method. Background Art

[0002] In the semiconductor material production process, the inspection of silicon single crystal slice samples is an important link to ensure product quality and process stability. However, the traditional inspection of silicon single crystal slice samples is unable to control product quality due to the single and incomplete inspection data. Summary of the invention

[0003] In view of this, the present invention provides a silicon single crystal slice sample detection and analysis method capable of controlling product quality.

[0004] The technical solution adopted by the present invention to solve its technical problem is: A method for detecting and analyzing a silicon single crystal slice sample comprises the following steps: S1: Test the Warp, thickness and BOW value of silicon single crystal slices to obtain the test data of different indicators of each batch of silicon wafers and the corresponding multi-wire saw model; S2: storing the test data of different indicators of silicon wafers of each batch number and the corresponding multi-wire sawing machine model to obtain a database; S3: Perform anomaly detection, statistical analysis and trend prediction on the detection data according to the obtained database to monitor the quality of the silicon wafers cut by the cutting machine.

[0005] Preferably, in S3, the abnormality detection is specifically: retrieving the detection data of different indicators of each batch of silicon wafers in the database, and comparing the detection data of each batch of silicon wafers with the standard data of the same indicator respectively, If the test data is less than or equal to the standard data under the same indicators, it is a normal batch number silicon wafer. If under the same indicators: the test data is greater than the standard data, it is an abnormal batch number silicon wafer; The inspection data of normal batch silicon wafers and the model of the multi-wire saw corresponding to the normal batch silicon wafers are stored in the normal silicon wafer database to perform statistical analysis and trend prediction on the normal batch silicon wafers.

[0006] Preferably, the statistical analysis specifically includes: regularly performing statistical comparison on the detection data of the same indicators of normal batches of silicon wafers to obtain the accuracy of different multi-wire sawing machine models in cutting silicon wafers of different batches.

[0007] Preferably, the trend prediction is specifically as follows: regularly statistically comparing the detection data of the same indicators of normal batch silicon wafers, and drawing trend curves, and predicting the status of the corresponding multi-wire saw through the trend curves of the same indicators.

[0008] Preferably, in S1: the silicon single crystal slice sample is subjected to Warp, thickness and BOW value detection. During the detection, the silicon single crystal slice sample needs to be placed frontally.

[0009] Preferably, the method for distinguishing the front side of a silicon single crystal slice sample is as follows: before a silicon single crystal ingot is cut into silicon wafers, the batch of the silicon single crystal ingot is marked on the OF plane of the silicon single crystal ingot, and an oblique line is drawn along the length direction of the ingot on one side of the OF plane of the ingot to obtain a marked ingot; when the marked ingot is cut into silicon wafers, silicon wafers from the same batch can be combined into oblique lines with the same inclination direction to ensure that each silicon wafer from the same batch has the same direction, which is convenient for distinguishing the front side of the silicon single crystal slice sample.

[0010] Preferably, the angle between the oblique line and the horizontal plane is 20°-32°.

[0011] Preferably, one end of the oblique line coincides with an edge of the OF surface.

[0012] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides a silicon single crystal slice sample detection and analysis method, which performs Warp, thickness and BOW value detection on the silicon single crystal slice sample to obtain detection data of different indicators of each batch of silicon wafers and the model of the corresponding multi-wire cutting machine; the detection data of different indicators of each batch of silicon wafers and the model of the corresponding multi-wire cutting machine are stored to obtain a database; according to the obtained database, abnormality detection, statistical analysis and trend prediction are performed on the detection data to monitor the quality of silicon wafers cut by the multi-wire cutting machine, and according to the statistical analysis and trend prediction of the detection data, the multi-wire multi-cutting machine can be adjusted in time to reduce the probability of unqualified silicon wafers and reduce waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 This is the front view of BOW value detection.

[0014] Figure 2 This is the reverse image of BOW value detection.

[0015] Figure 3 This is a detection structure diagram of Example 1. DETAILED DESCRIPTION

[0016] The technical solutions and technical effects of the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings of the present invention.

[0017] A method for detecting and analyzing a silicon single crystal slice sample comprises the following steps: S1: Test the Warp, thickness and BOW value of silicon single crystal slices to obtain the test data of different indicators of each batch of silicon wafers and the corresponding multi-wire saw model; Specifically, the E+H Metrology tester is used to quickly and accurately test key parameters such as Warp, thickness, and BOW value of silicon single crystal slices. The tester has functions such as automatic calibration and automatic data recording, which can greatly reduce manual intervention and improve detection accuracy and efficiency.

[0018] S2: storing the test data of different indicators of silicon wafers of each batch number and the corresponding multi-wire sawing machine model to obtain a database; S3: Perform anomaly detection, statistical analysis and trend prediction on the detection data according to the obtained database to monitor the quality of the silicon wafers cut by the cutting machine.

[0019] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides a silicon single crystal slice sample detection and analysis method, which performs Warp, thickness and BOW value detection on the silicon single crystal slice sample to obtain detection data of different indicators of each batch of silicon wafers and the model of the corresponding multi-wire cutting machine; the detection data of different indicators of each batch of silicon wafers and the model of the corresponding multi-wire cutting machine are stored to obtain a database; according to the obtained database, abnormality detection, statistical analysis and trend prediction are performed on the detection data to monitor the quality of silicon wafers cut by the multi-wire cutting machine, and according to the statistical analysis and trend prediction of the detection data, the multi-wire multi-cutting machine can be adjusted in time to reduce the probability of unqualified silicon wafers and reduce waste.

[0020] Furthermore, in S3, the abnormality detection is specifically: retrieving the detection data of different indicators of each batch of silicon wafers in the database, and comparing the detection data of each batch of silicon wafers with the standard data of the same indicators respectively: If the test data is less than or equal to the standard data under the same indicators, it is a normal batch number silicon wafer. If under the same indicators: the test data is greater than the standard data, it is an abnormal batch number silicon wafer; The detection data of normal batch silicon wafers and the model of the multi-wire saw corresponding to the normal batch silicon wafers are stored in the normal silicon wafer database to perform statistical analysis and trend prediction on the normal batch silicon wafers; Based on the comparison, the staff will find out the abnormal batch of silicon wafers and adjust the corresponding multi-wire cutting machine as soon as possible to prevent excessive production of abnormal batch of silicon wafers and cause waste.

[0021] Furthermore, the statistical analysis is specifically as follows: regularly compare the test data of the same indicators of normal batches of silicon wafers to obtain the accuracy of different multi-wire sawing machine models in cutting different batches of silicon wafers. When customers need it, the appropriate multi-wire sawing machine can be selected according to the accuracy of the test data and the type of silicon wafer.

[0022] Furthermore, the trend prediction is specifically as follows: regularly statistically comparing the detection data of the same indicators of normal batch silicon wafers, and drawing trend curves, and predicting the status of the corresponding multi-wire saw through the trend curves of the same indicators.

[0023] Specifically, statistics can be conducted weekly or monthly based on actual conditions. The operating conditions of each multi-wire cutting machine can be predicted based on the trend curve of the same indicator. If the detection data shows an upward trend, it needs to be fed back to the staff of the multi-wire cutting machine to adjust the multi-wire cutting machine to prevent the occurrence of unqualified silicon wafers.

[0024] like Figure 1 , 2 As shown, further, in S1: the silicon single crystal slice sample is subjected to Warp, thickness, and BOW value detection. When detecting the BOW value, the silicon single crystal slice sample needs to be placed on the front side, because the BOW values ​​of the curvature after measurement on the front and back sides of the silicon wafer are different. If it is not guaranteed that all measurements are made on the front side, the silicon wafer detection value is incorrect and needs to be re-detected. If the staff fails to discover it in time, it will lead to overall errors in subsequent data anomaly detection, statistical analysis, and trend prediction, and incorrect detection results, affecting customer use.

[0025] Furthermore, the method for distinguishing the front side of the silicon single crystal slice sample is as follows: before the silicon single crystal ingot is cut into silicon wafers, the batch of the silicon single crystal ingot is marked on the OF surface of the silicon single crystal ingot to prevent the wrong goods; and an oblique line is drawn along the length direction of the ingot on one side of the OF surface of the ingot to obtain a marked ingot. When the marked ingot is cut into silicon wafers, the silicon wafers of the same batch can be combined into oblique lines with the same inclination direction to ensure that each silicon wafer of the same batch has the same direction, which is convenient for distinguishing the front side of the silicon single crystal slice sample and preventing the silicon wafers of the same batch from being placed forward or backward, resulting in inaccurate detection results.

[0026] Furthermore, the angle between the oblique line and the horizontal plane is 20°-32°, which reduces the possibility that the oblique line on the crystal ingot is worn and washed away by mortar, cutting fluid and steel wire during the cutting process, making it difficult for workers to identify it.

[0027] Furthermore, one end of the oblique line coincides with an edge of the OF surface.

[0028] The use of the present invention is described by the following examples. Example

[0029] A method for detecting and analyzing a silicon single crystal slice sample comprises the following steps: S1: Before the silicon single crystal ingot is cut into silicon wafers, the batch of the silicon single crystal ingot is marked on the OF surface of the silicon single crystal ingot; and an oblique line is drawn along the length direction of the ingot on one side of the OF surface of the ingot to obtain a marked ingot. When the marked ingot is cut into silicon wafers, the silicon wafers of the same batch can be combined into oblique lines with the same inclination direction. The front of the silicon single crystal sliced ​​sample is placed in the E+H Metrology tester for Warp, thickness, and BOW value detection, and the detection data of different indicators of each batch of silicon wafers and the corresponding multi-wire sawing machine model are obtained; S2: storing the test data of different indicators of silicon wafers of each batch number and the corresponding multi-wire sawing machine model to obtain a database; S3: performing abnormality detection, statistical analysis and trend prediction on the detection data according to the obtained database to monitor the quality of the silicon wafers cut by the cutting machine; The specific abnormality detection is as follows: retrieve the test data of different indicators of each batch of silicon wafers in the database, and compare the test data of each batch of silicon wafers with the standard data of the same indicators: If the test data is less than or equal to the standard data under the same indicators, it is a normal batch number silicon wafer. If under the same indicators: the test data is greater than the standard data, it is an abnormal batch number silicon wafer; The detection data of normal batch silicon wafers and the model of the multi-wire saw corresponding to the normal batch silicon wafers are stored in the normal silicon wafer database to perform statistical analysis and trend prediction on the normal batch silicon wafers; The specific statistical analysis is as follows: the test data of the same indicators of normal batches of silicon wafers are statistically compared every month to obtain the accuracy of different multi-wire sawing machine models in cutting different batches of silicon wafers. When customers need it, the appropriate multi-wire sawing machine can be selected based on the accuracy of the test data and the type of silicon wafer.

[0030] The specific trend prediction is as follows: statistically compare the test data of the same indicators of normal batch number silicon wafers every month, and draw trend curves. Through the trend curves of the same indicators, the status of the corresponding multi-wire cutting machine is predicted.

[0031] Through the above method, the overall detection accuracy is improved, and the detection efficiency is greatly improved. And through half a year of records, such as Figure 3 As shown, the probability of placing the silicon wafer upside down during inspection is greatly reduced, which greatly improves the inspection efficiency and accuracy.

[0032] What is disclosed above is only a preferred embodiment of the present invention, which certainly cannot be used to limit the scope of rights of the present invention. A person skilled in the art can understand that all or part of the processes of the above embodiments and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A method for detecting and analyzing silicon single crystal slice samples, characterized in that: The following steps are involved: S1: Test the Warp, thickness and BOW value of silicon single crystal slices to obtain the test data of different indicators of each batch of silicon wafers and the corresponding multi-wire saw model; S2: storing the test data of different indicators of silicon wafers of each batch number and the corresponding multi-wire sawing machine model to obtain a database; S3: Perform anomaly detection, statistical analysis and trend prediction on the detection data according to the obtained database to monitor the quality of the silicon wafers cut by the cutting machine.

2. The method for detecting and analyzing silicon single crystal slice samples according to claim 1, characterized in that: In S3, the abnormality detection is specifically: retrieving the test data of different indicators of each batch of silicon wafers in the database, and comparing the test data of each batch of silicon wafers with the standard data of the same indicator respectively, If the test data is less than or equal to the standard data under the same indicators, it is a normal batch number silicon wafer. If under the same indicators: the test data is greater than the standard data, it is an abnormal batch number silicon wafer; The inspection data of normal batch silicon wafers and the model of the multi-wire saw corresponding to the normal batch silicon wafers are stored in the normal silicon wafer database to perform statistical analysis and trend prediction on the normal batch silicon wafers.

3. The method for detecting and analyzing silicon single crystal slice samples according to claim 2, characterized in that: The statistical analysis specifically includes: regularly performing statistical comparison on the test data of the same index of normal batches of silicon wafers to obtain the precision of different multi-wire sawing machine models in cutting silicon wafers of different batches.

4. The method for detecting and analyzing silicon single crystal slice samples according to claim 3, characterized in that: The trend prediction is specifically as follows: regularly statistically comparing the detection data of the same index of normal batch silicon wafers, and drawing trend curves, and predicting the status of the corresponding multi-wire saw through the trend curves of the same indexes.

5. The method for detecting and analyzing silicon single crystal slice samples according to claim 1, characterized in that: In S1: the silicon single crystal slice sample is tested for Warp, thickness and BOW value. During the testing, the silicon single crystal slice sample needs to be placed on the front side.

6. The method for detecting and analyzing silicon single crystal slice samples according to claim 5, characterized in that: The method for distinguishing the front side of a silicon single crystal slice sample is as follows: before a silicon single crystal ingot is cut into silicon wafers, the batch of the silicon single crystal ingot is marked on the OF plane of the silicon single crystal ingot, and an oblique line is drawn along the length direction of the ingot on one side of the OF plane of the ingot to obtain a marked ingot; when the marked ingot is cut into silicon wafers, silicon wafers from the same batch can be combined into oblique lines with the same inclination direction to ensure that each silicon wafer from the same batch has the same direction, which is convenient for distinguishing the front side of the silicon single crystal slice sample.

7. The method for detecting and analyzing silicon single crystal slice samples according to claim 6, characterized in that: The angle between the oblique line and the horizontal plane is 20°-32°.

8. The method for detecting and analyzing silicon single crystal slice samples according to claim 7, characterized in that: One end of the oblique line coincides with the edge of the OF surface.

Citation Information

Cited By

  • Method for detecting semiconductor monocrystalline silicon ingot before slicing

    CN120741782A

  • Method for testing a semiconductor single crystal silicon ingot before slicing

    CN120741782B