Ingot evaluation methods

Abstract

This invention provides a method for evaluating crystal ingots. The wafer images corresponding to multiple wafers cut from multiple crystal ingots are divided into multiple regions. The statistical number of multiple defect types included in each region is calculated. Based on the statistical number of defect types included in the regions, multiple statistical parameters are obtained. Regression analysis is performed using the statistical parameters corresponding to each wafer and the bending value to obtain multiple regression coefficients. A set of bending prediction equations is established based on the regression coefficients.

Description

Technical Field

[0001] This invention relates to an evaluation and testing method, and more particularly to an ingot evaluation method. Background Technology

[0002] The semiconductor manufacturing industry has extremely stringent requirements for product quality, and the quality of incoming ingots has a significant impact on the quality of the resulting wafers. Therefore, how to filter out substandard ingots before processing is one of the key challenges in this field. Summary of the Invention

[0003] This invention relates to a method for evaluating crystal ingots, which can predict the quality of crystal ingots using evaluation slices.

[0004] According to an embodiment of the present invention, the ingot evaluation method includes: dividing the wafer image corresponding to each of the multiple wafers cut from multiple ingots into multiple regions; calculating the statistical number of multiple defect types included in each region; obtaining multiple statistical parameters based on the statistical number of defect types included in the region; performing regression analysis using the statistical parameters corresponding to each wafer and the bending value to obtain multiple regression coefficients; and establishing a set of bending prediction equations based on the regression coefficients.

[0005] In an embodiment of the present invention, the step of obtaining statistical parameters based on the statistical quantity of defect types included in the region includes: for any defect type, performing the following steps: calculating the statistical quantity corresponding to the defect type in each region; calculating the standard deviation of the statistical quantity corresponding to the defect type in the region as one of the statistical parameters; and sorting the statistical quantities corresponding to the defect type in the region, and taking the A highest-valued statistical quantities in descending order of value to calculate the average value as one of the statistical parameters, where A is a positive integer.

[0006] In embodiments of the present invention, the defect types include threading edge dislocations (TED), threading screw dislocations (TSD), and basal plane dislocations (BPD). The step of performing regression analysis using the statistical parameters and bending values ​​corresponding to each wafer, and obtaining the regression coefficients, includes inputting the statistical parameters and bending values ​​corresponding to each wafer into the following formula to perform regression analysis.

[0007] Bow(k)=P1×Std(TED)+P2×Avg(TED);

[0008] Bow(k)=P3×Std(TSD)+P4×Avg(TSD);

[0009] Bow(k)=P5×Std(BPD)+P6×Avg(BPD).

[0010] Where Bow(k) is the bending value of the k-th wafer, P1 to P6 are regression coefficients, Std(TED), Std(TSD), and Std(BPD) are the standard deviations of the statistical quantities of the three defect types corresponding to the three defect types of through-edge dislocation, through-screw dislocation, and basal plane dislocation, respectively, and Avg(TED), Avg(TSD), and Avg(BPD) are the averages of the highest A statistical quantities corresponding to the three defect types of through-edge dislocation, through-screw dislocation, and basal plane dislocation, respectively.

[0011] In an embodiment of the present invention, the step of obtaining statistical parameters based on the statistical quantity of defect types included in the region includes: for any two defect types, performing the following steps: calculating the total statistical quantity of the two defect types in each region; calculating the standard deviation of the statistical quantity of the region corresponding to the total statistical quantity of the two defect types as one of the statistical parameters; and sorting the statistical quantities of the region corresponding to the total statistical quantity of the two defect types, taking the highest A statistical quantities and calculating the average value as one of the statistical parameters.

[0012] In an embodiment of the present invention, the defect types include through-edge dislocations, through-screw dislocations, and basal plane dislocations. Regression analysis is performed using the statistical parameters and bending values ​​corresponding to the wafers, and the step of obtaining regression coefficients includes: inputting the statistical parameters and bending values ​​corresponding to each wafer into the following formula to perform regression analysis.

[0013] Bow(k)=P7×Std(TED&TSD)+P8×Avg(TED&TSD);

[0014] Bow(k)=P9×Std(TSD&BPD)+P10×Avg(TSD&BPD);

[0015] Bow(k)=P11×Std(TED&BPD)+P12×Avg(TED&BPD).

[0016] Where Bow(k) is the bending value of the k-th wafer, and P7 to P12 are regression coefficients. Std(TED&TSD) is the standard deviation of the total number of statistical quantities corresponding to the two defect types of through-edge dislocations and through-screw dislocations in the region, and Avg(TED&TSD) is the average of the A highest statistical quantities corresponding to the total number of statistical quantities corresponding to the two defect types of through-edge dislocations and through-screw dislocations. Std(TSD&BPD) is the standard deviation of the total number of statistical quantities corresponding to the two defect types of through-screw dislocations and basal plane dislocations in the region, and Avg(TSD&BPD) is the average of the A highest statistical quantities corresponding to the total number of statistical quantities corresponding to the two defect types of through-screw dislocations and basal plane dislocations. Std(TED&BPD) is the standard deviation of the total number of statistical quantities corresponding to the two defect types of through-edge dislocations and basal plane dislocations in the region, and Avg(TED&BPD) is the average of the A highest statistical quantities corresponding to the total number of statistical quantities corresponding to the two defect types of through-edge dislocations and basal plane dislocations.

[0017] In an embodiment of the present invention, the step of obtaining statistical parameters based on the statistical number of defect types included in the region includes: calculating the total statistical number of defect types in each region; calculating the standard deviation of the statistical number of defect types corresponding to the region as one of the statistical parameters; and sorting the statistical numbers corresponding to the total defect types in the region, taking out the highest A statistical numbers to calculate the average as one of the statistical parameters, where A is a positive integer.

[0018] In an embodiment of the present invention, regression analysis is performed using the statistical parameters and bending values ​​corresponding to each wafer, and the step of obtaining regression coefficients includes: inputting the statistical parameters and bending values ​​corresponding to each wafer into the following formula to perform regression analysis.

[0019] Bow(k)=P13×Std(TED&TSD&BPD)+P14×Avg(TED&TSD&BPD).

[0020] Where Bow(k) is the bending value of the k-th wafer, and P13 to P14 are regression coefficients. Std(TED&TSD&BPD) is the standard deviation of the total number of statistical quantities of the three defect types (through edge dislocation, through screw dislocation, and base plane dislocation) in the region, and Avg(TED&TSD&BPD) is the average of the A highest statistical quantities among the total number of statistical quantities of the three defect types (through edge dislocation, through screw dislocation, and base plane dislocation).

[0021] In an embodiment of the present invention, after establishing the set of curvature prediction equations based on the regression coefficients, the method further includes: using the set of curvature prediction equations to calculate the curvature value of the wafer to be tested processed from the wafer to be tested; determining whether the post-processing curvature value is within a specified range; if the post-processing curvature value is within the specified range, determining that the quality of the wafer to be tested is good; and if the post-processing curvature value is not within the specified range, determining that the quality of the wafer to be tested is poor.

[0022] In an embodiment of the invention, after the step of dividing the wafer image corresponding to each wafer cut from the ingot into the region, the method further includes: discarding the regions located at the four corners of the wafer image, and calculating the statistical number of defect types included in the remaining regions.

[0023] Based on the above, a set of curvature prediction equations is established using known ingots and processed wafers. This set of curvature prediction equations is used to predict the quality of the ingot to be tested, thereby filtering out ingots that will cause poor processing geometry quality. This can significantly improve the overall processing quality and reduce production costs. Attached Figure Description

[0024] Figure 1 This is a block diagram of an analysis system according to an embodiment of the present invention;

[0025] Figure 2 This is a schematic diagram of a crystal ingot according to an embodiment of the present invention;

[0026] Figure 3 This is a flowchart of an ingot evaluation method according to an embodiment of the present invention;

[0027] Figure 4 This is a schematic diagram of a wafer image according to an embodiment of the present invention.

[0028] Explanation of reference numerals in the attached figures

[0029] 110: Measuring Instruments

[0030] 120: Analytical device

[0031] 20-1~20-N: Crystal ingots

[0032] 21-1~21-N, W: wafer

[0033] 400: Chip Image

[0034] C1~C36: Area

[0035] E1: First end

[0036] E2: Second end

[0037] S305~S325: Steps in Ingot Evaluation Methods Detailed Implementation

[0038] Reference will now be made in detail to exemplary embodiments of the invention, examples of which are illustrated in the accompanying drawings. Wherever possible, the same component reference numerals are used in the drawings and description to denote the same or similar parts.

[0039] Figure 1 This is a block diagram of an analysis system according to an embodiment of the present invention. Please refer to... Figure 1 The analysis system includes a measuring instrument 110 and an analysis device 120. Data transmission between the measuring instrument 110 and the analysis device 120 can be performed, for example, via wired or wireless communication.

[0040] Measuring instrument 110 is, for example, an Automated Optical Inspection (AOI) instrument. AOI instruments are high-speed, high-precision optical image inspection systems that incorporate measuring lens technology, optical illumination technology, positioning measurement technology, electronic circuit testing technology, image processing technology, and automation technology applications. They utilize machine vision as the standard inspection technology. Measuring instrument 110 uses optical instruments to obtain the surface condition of the finished product and then uses computer image processing technology to detect defects such as foreign objects or pattern anomalies.

[0041] The analysis device 120 is an electronic device with computing capabilities, which can be implemented using a personal computer, laptop calculator, tablet computer, smartphone, or any device with computing capabilities; this invention is not limited thereto. The analysis device 120 receives measurement data (i.e., the coordinate positions of defects and the types of defects) from multiple known wafers from the measuring instrument 110, and uses this data to train a predictive model (a set of curvature prediction equations) for subsequent use of the measurement data of the wafer under test to obtain the quality of the wafer after the ingot under test has been processed into a wafer.

[0042] Figure 2 This is a schematic diagram of a crystal ingot according to an embodiment of the present invention. Please refer to... Figure 2After the crystal ingots 20-1 to 20-N undergo processing, one wafer is taken from each ingot to obtain multiple wafers 21-1 to 21-N. The processing can include cutting, grinding, and polishing, but this invention is not limited to these. An optical inspection is performed on each wafer 21-1 to 21-N using a measuring instrument 110 to detect defects at various coordinate positions and record the coordinate positions and types of defects. The defect types include threading edge dislocations (TED), threading screw dislocations (TSD), and basal plane dislocations (BPD). The plurality of wafers 21-1 to 21-N can be wafers obtained by processing at any position on the ingot. In some preferred embodiments, the plurality of wafers 21-1 to 21-N are wafers obtained by processing at a position close to the first end E1 and the second end E2 (head and tail ends) of the ingot 20-1 to 20-N. In other embodiments, they can be wafers obtained by processing at a position outside the head and tail end regions of the ingot 20-1 to 20-N. The present invention is not limited thereto.

[0043] Figure 3 This is a flowchart of an ingot evaluation method according to an embodiment of the present invention. Please refer to... Figure 3 In step S305, the wafer images corresponding to the wafers 21-1 to 21-N cut from the multiple ingots 20-1 to 20-N are divided into multiple regions. Figure 4 This is a schematic diagram of a wafer image according to an embodiment of the present invention. Please refer to... Figure 4 The wafer image 400 is divided into multiple regions C1 to C36. This number of divisions is merely one example; in other embodiments, the wafer image 400 can be divided into any number of regions as needed, and the invention is not limited thereto. Each of wafers 21-1 to 21-N has a corresponding wafer image similar to wafer image 400.

[0044] Next, in step S310, the statistical counts of multiple defect types included in each region are calculated. For example, taking wafer 21-1 as an example, and using wafer image 400 as a wafer image of wafer 21-1 for explanation. The analysis device 120 calculates the statistical counts of TED, TSD, and BPD included in each region C1 to C36 of the wafer image 400 corresponding to wafer 21-1 based on the measurement data corresponding to wafer 21-1 (i.e., the coordinate positions of defects and defect types). The same applies to other wafers 21-2 to 21-N.

[0045] Furthermore, since the proportion of wafer W in the four corners of wafer image 400 is relatively low, in the step of counting defect types, regions C33 to C36 located in the four corners of wafer image 400 can be discarded, and the statistical number of defect types included in the remaining regions C1 to C32 can be calculated. For example, Table 1 records the statistical number of various defect types in regions C1 to C32 of each of the N wafers 21-2 to 21-N.

[0046] Table 1

[0047]

[0048] Next, in step S315, multiple statistical parameters are obtained based on the statistical number of defect types included in the region. Here, the standard deviation and the mean can be used as statistical parameters.

[0049] For example, the standard deviation and mean of the statistical counts across multiple regions divided within the same wafer are calculated for each defect type. If there are three defect types, six statistical parameters can be obtained. Taking wafer 21-1 in Table 1 as an example, the statistical counts C in regions C1 to C32 are... 1-1-TED ~C 1-32-TED To calculate the standard deviation (Std(TED)) of the TED corresponding to wafer 21-1, we use the statistical data C of region C1 to C32. 1-1-TSD ~C 1-32-TSD To calculate the standard deviation (Std) of the TSD corresponding to wafer 21-1. Using the statistical data C of regions C1 to C32... 1-1-BPD ~C 1-32-BPD To calculate the standard deviation (Std(BPD)) of the BPD corresponding to wafer 21-1, and to sort C... 1-1-TED ~C 1-32-TED Take the highest A values ​​(e.g., 5) and calculate the average value Avg(TED). Sort C 1-1-TSD ~C 1-32-TSD Take the top 5 and calculate the average value Avg(TSD). Sort by C. 1-1-BPD ~C 1-32-BPD Take the top 5 and calculate the average value Avg(BPD).

[0050] Then, in step S320, regression analysis is performed using the statistical parameters corresponding to each wafer (21-1 to 21-N) and the bending (BOW) value of each wafer (21-1 to 21-N) after processing, to obtain multiple regression coefficients. For example, the statistical parameters and bending values ​​are input into the following equations (1) to (3) to perform regression analysis, to obtain multiple regression coefficients P1 to P6. Wherein, Bow(k) is the bending value of the k-th wafer after processing.

[0051] (1)Bow(k)=P1×Std(TED)+P2×Avg(TED);

[0052] (2) Bow(k)=P3×Std(TSD)+P4×Avg(TSD);

[0053] (3) Bow(k)=P5×Std(BPD)+P6×Avg(BPD).

[0054] Finally, in step S325, a set of curvature prediction equations is established based on the regression coefficients. The following set of curvature prediction equations (A) is established based on the obtained regression coefficients P1 to P6:

[0055] Bow1=P1×TStd(TED)+P2×TAvg(TED);

[0056] Bow2=P3×TStd(TSD)+P4×TAvg(TSD);

[0057] Bow3=P5×TStd(BPD)+P6×TAvg(BPD);

[0058] Pre_Bow=(Bow1+Bow2+Bow3) / 3.

[0059] In the prediction equation set (A), the curvature Bow1 is obtained using the standard deviation TStd(TED) and average value TAvg(TED) of TED in the wafer under test; the curvature Bow2 is obtained using the standard deviation TStd(TSD) and average value TAvg(TSD) of TSD in the wafer under test; and the curvature Bow3 is obtained using the standard deviation TStd(BPD) and average value TAvg(BPD) of BPD in the wafer under test. Then, the average of the curvatures Bow1, Bow2, and Bow3 is taken as the post-processing curvature value Pre_Bow.

[0060] Then, it is determined whether the post-processing bending value Pre_Bow is within the specified range. If the post-processing bending value Pre_Bow is within the specified range, the quality of the test ingot corresponding to this test wafer is determined to be good. If the post-processing bending value Pre_Bow is not within the specified range, the quality of the test ingot corresponding to this test wafer is determined to be poor. Here, the test ingot corresponding to the test wafer is the test wafer obtained from the test ingot through a processing procedure.

[0061] In addition to calculating the standard deviation and mean for each type of defect as described above, we can also calculate the total statistical quantity for every two types of defects, every three types of defects, etc., thereby obtaining the corresponding standard deviation and mean.

[0062] For example, for any two defect types, perform the following steps: Calculate the total number of the two defect types in each region; calculate the standard deviation of the total number of the two defect types in each region as one of the statistical parameters; and sort the total number of the two defect types in each region, taking the highest A statistics and calculating their average as one of the statistical parameters.

[0063] Table 2 records the statistical counts of any two types of defects in regions C1 to C32 of each of the N wafers 21-2 to 21-N.

[0064] Table 2

[0065]

[0066] Taking wafer 21-1 in Table 2 as an example, the standard deviation Std(TED&TSD) of TED and TSD corresponding to wafer 21-1 is calculated using the statistical quantities C1-1-TED&TSD to C1-32-TED&TSD in regions C1 to C32. Furthermore, C1-1-TED&TSD to C1-32-TED&TSD are sorted to extract the highest A values ​​(e.g., 5 values), and the average value Avg(TED&TSD) is calculated. This process is repeated to obtain the standard deviation Std(TSD&BPD) and average value Avg(TSD&BPD) corresponding to TSD and BPD, and the standard deviation Std(TED&BPD) and average value Avg(TED&BPD) corresponding to TED and BPD. The above statistical parameters and bending values ​​are input into the following equations (4) to (6) to perform regression analysis, resulting in multiple regression coefficients P7 to P12. Here, Bow(k) is the bending value of the k-th wafer.

[0067] (4) Bow(k)=P7×Std(TED&TSD)+P8×Avg(TED&TSD);

[0068] (5) Bow(k)=P9×Std(TSD&BPD)+P10×Avg(TSD&BPD);

[0069] (6) Bow(k)=P11×Std(TED&BPD)+P12×Avg(TED&BPD).

[0070] Table 3 records the statistical number of three types of defects in regions C1 to C32 of each of the N wafers 21-2 to 21-N.

[0071] Table 3

[0072]

[0073] Taking wafer 21-1 in Table 3 as an example, the standard deviation Std(TED&TSD&BPD) of the sum of TED, TSD and BPD for wafer 21-1 is calculated using the statistical values ​​C1-1-TED&TSD&BPD to C1-32-TED&TSD&BPD in regions C1 to C32. Furthermore, C1-1-TED&TSD&BPD to C1-32-TED&TSD&BPD are sorted from largest to smallest to extract the highest A values ​​(e.g., the top 5 values), and the average value Avg(TED&TSD&BPD) is calculated. The above statistical parameters and bending values ​​are input into the following equation (7) to perform regression analysis, obtaining regression coefficients P13 to P14. Here, Bow(k) is the bending value of the k-th wafer.

[0074] (7)Bow(k)=P13×Std(TED&TSD&BPD)+

[0075] P14×Avg(TED&TSD&BPD).

[0076] Taking the three defect types as TED, TSD and BPD, 14 statistical parameters can be used, including: 6 statistical parameters of standard deviation and mean of each of TED, TSD and BPD, 6 statistical parameters of standard deviation and mean of any two defect types, and 2 statistical parameters of standard deviation and mean of the three defect types.

[0077] Table 4 shows the 14 statistical parameters corresponding to each wafer. Regression analysis was performed using the above equations (1) to (7) to obtain multiple regression coefficients P1 to P14.

[0078] Table 4

[0079]

[0080] Based on the obtained regression coefficients P1 to P14, the following set of curvature prediction equations (B) is established:

[0081] Bow1=P1×TStd(TED)+P2×TAvg(TED);

[0082] Bow2=P3×TStd(TSD)+P4×TAvg(TSD);

[0083] Bow3=P5×TStd(BPD)+P6×TAvg(BPD);

[0084] Bow4=P7×TStd(TED&TSD)+P8×TAvg(TED&TSD);

[0085] Bow5=P9×TStd(TSD&BPD)+P10×TAvg(TSD&BPD);

[0086] Bow6=P11×TStd(TED&BPD)+P12×TAvg(TED&BPD);

[0087] Bow7=P13×TStd(TED&TSD&BPD)+P14×TAvg(TED&TSD&BPD);

[0088] Pre_Bow=(Bow1+Bow2+Bow3+Bow4+Bow5+Bow6+Bow7) / 7.

[0089] In the prediction equation set (B), the curvature Bow1 is obtained using the standard deviation TStd(TED) and average value TAvg(TED) corresponding to TED in the wafer under test; the curvature Bow2 is obtained using the standard deviation TStd(TSD) and average value TAvg(TSD) corresponding to TSD in the wafer under test; and the curvature Bow3 is obtained using the standard deviation TStd(BPD) and average value TAvg(BPD) corresponding to BPD in the wafer under test. Furthermore, the curvature Bow4, Bow5, and Bow6 are obtained using the standard deviations TStd(TED&TSD), TStd(TSD&BPD), TStd(TED&BPD), and average values ​​TAvg(TED&TSD), TAvg(TED&TSD), and TAvg(TSD&BPD) corresponding to any two defect types TED+TSD, TSD+BPD, and TED+BPD in the wafer under test. The bending degree Bow7 is obtained by using the standard deviation TStd(TED&TSD&BPD) and average value TAvg(TED&TSD&BPD) corresponding to the three defect types TED, TSD and BPD in the wafer under test. Then, the average value of the bending degrees Bow1 to Bow7 is taken as the bending value Pre_Bow after processing.

[0090] In summary, this invention utilizes a set of curvature prediction equations to establish a set of equations for the curvature of a known ingot after processing. This set of curvature prediction equations can be used to predict the quality of the ingot to be tested, thereby filtering out ingots that will cause poor processing geometry quality. This can significantly improve the overall processing quality and reduce production costs.

[0091] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for evaluating crystal ingots, characterized in that, include: The wafer images corresponding to the multiple wafers cut from multiple ingots are divided into multiple regions; Calculate the statistical number of multiple defect types included in each of the multiple regions; Based on the statistical number of the defect types included in the region, multiple statistical parameters are obtained; Regression analysis is performed using the statistical parameters and bending values ​​corresponding to each wafer to obtain multiple regression coefficients; A set of curvature prediction equations is established based on the regression coefficients. The bending value of the wafer after processing, obtained from the wafer ingot under test, is calculated using the aforementioned set of bending prediction equations. Determine whether the post-processing bending value is within the specified range; as well as If the bending value after processing is not within the specified range, the quality of the ingot to be tested is deemed poor.

2. The ingot evaluation method according to claim 1, characterized in that, The step of obtaining the statistical parameters based on the statistical number of the defect types included in the region includes: For any one of the aforementioned defect types, Calculate the statistical number of any one type of defect in each of the aforementioned regions; The standard deviation of the statistical number of the region corresponding to any one of the defect types is calculated as one of the statistical parameters; and The statistical quantities corresponding to any defect type in the region are sorted, and the A highest statistical quantities are selected in descending order of value to calculate the average value as one of the statistical parameters, where A is a positive integer.

3. The ingot evaluation method according to claim 2, characterized in that, The defect types include through-edge dislocations, through-screw dislocations, and basal plane dislocations. The regression analysis is performed using the statistical parameters and bending values ​​corresponding to each wafer, and the steps to obtain the regression coefficients include: The regression analysis is performed by inputting the statistical parameters and the bending value corresponding to each wafer into the following formula: Bow(k)=P 1×Std(TED)+P2 xAvg(TED); Bow(k)=P3×Std(TSD)+P4 xAvg(TSD); Bow(k)=P5×Std(BPD)+P6xAvg(BPD); Where Bow(k) is the bending value of the k-th wafer, P1 to P6 are the regression coefficients, Std(TED), Std(TSD), and Std(BPD) are the standard deviations of the statistical quantities of the region corresponding to the three defect types of through-edge dislocation, through-screw dislocation, and basal plane dislocation, respectively, and Avg(TED), Avg(TSD), and Avg(BPD) are the averages of the highest A statistical quantities corresponding to the three defect types of through-edge dislocation, through-screw dislocation, and basal plane dislocation, respectively.

4. The ingot evaluation method according to claim 1, characterized in that, The step of obtaining the statistical parameters based on the statistical number of the defect types included in the region includes: For any two of the aforementioned defect types, Calculate the total number of any two defect types in each region; The standard deviation of the statistical number of the region corresponding to the sum of the two defect types is calculated as one of the statistical parameters; and Sort the statistical quantities corresponding to the sum of any two defect types in the region, and take out the A highest statistical quantities to calculate the average value as one of the statistical parameters.

5. The ingot evaluation method according to claim 4, characterized in that, The defect types include through-edge dislocations, through-screw dislocations, and basal plane dislocations. The regression analysis is performed using the statistical parameters and bending values ​​corresponding to each wafer, and the steps to obtain the regression coefficients include: The regression analysis is performed by inputting the statistical parameters and the bending value corresponding to each wafer into the following formula: Bow(k)=P7×Std(TED&TSD)+P 8xAvg(TED&TSD); Bow(k)=P9×Std(TSD&BPD)+P 10xAvg(TSD&BPD); Bow(k)=P 11×Std(TED&BPD)+P 12xAvg(TED&BPD); Where Bow(k) is the bending value of the k-th wafer, and P7 to P12 are the regression coefficients. Std(TED&TSD) is the standard deviation of the total number of defects in the region corresponding to the combined number of through-edge dislocations and through-screw dislocations. Avg(TED&TSD) is the average of the A highest statistical values ​​among the total number of defects in the region corresponding to the combined number of through-edge dislocations and through-screw dislocations. Std(TSD&BPD) is the standard deviation of the total number of defects in the region corresponding to the combined number of through-screw dislocations and basal plane dislocations. Avg(TSD&BPD) is the average of the A highest statistical values ​​among the total number of defects corresponding to the combined number of through-screw dislocations and basal plane dislocations. Std(TED&BPD) is the standard deviation of the total number of defects in the region corresponding to the combined number of through-edge dislocations and base surface dislocations, and Avg(TED&BPD) is the average of the A highest statistical values ​​among the total number of defects in the region corresponding to the combined number of through-edge dislocations and base surface dislocations.

6. The ingot evaluation method according to claim 1, characterized in that, The step of obtaining the statistical parameters based on the statistical number of the defect types included in the region includes: Calculate the total number of defect types in each of the aforementioned regions; The standard deviation of the statistical number of the total number of the defect types corresponding to the region is calculated as one of the statistical parameters; and Sort the statistical quantities corresponding to the total number of defect types in the region, and take out the A highest statistical quantities to calculate an average value as one of the statistical parameters, where A is a positive integer.

7. The ingot evaluation method according to claim 6, characterized in that, The defect types include through-edge dislocations, through-screw dislocations, and basal plane dislocations. The regression analysis is performed using the statistical parameters and bending values ​​corresponding to each wafer, and the steps to obtain the regression coefficients include: The regression analysis is performed by inputting the statistical parameters and the bending value corresponding to each wafer into the following formula: Bow(k)=P 13×Std(TED&TSD&BPD)+P 14xAvg(TED&TSD&BPD); Where Bow(k) is the bending value of the k-th wafer, and P13 to P14 are the regression coefficients. Std(TED&TSD&BPD) is the standard deviation of the total number of statistical quantities of the three defect types (through edge dislocation, through screw dislocation, and base plane dislocation) in the region, and Avg(TED&TSD&BPD) is the average of the highest A statistical quantities among the total number of statistical quantities of the three defect types (through edge dislocation, through screw dislocation, and base plane dislocation).

8. The ingot evaluation method according to claim 1, characterized in that, After determining whether the post-processing bending value is within the specified range, the method further includes: If the bending value after processing is within the specified range, the quality of the ingot to be tested is determined to be good.

9. The ingot evaluation method according to claim 1, characterized in that, After the step of segmenting the wafer image corresponding to each wafer cut from the ingot into the region, the method further includes: Discard the regions located at the four corners of the wafer image, and calculate the statistical number of the defect types included in the remaining regions.

Images