Wafer processing quality prediction method

Abstract

The embodiment of the invention provides a wafer processing quality prediction method. The method comprises the following steps: obtaining a stress distribution diagram of a wafer to be processed, wherein the stress distribution diagram comprises a plurality of data points; determining M data groups in the stress profile based on the M reference lines, wherein each data group is at least divided into a first sub-data group and a second sub-data group; determining the characteristics of the corresponding reference lines based on the first stress value distribution of the first sub-data group corresponding to each data group and the second stress value distribution of the second sub-data group corresponding to each data group; and predicting a predicted processing result of the wafer to be processed based on the characteristics of each reference line.

Description

Technical Field

[0001] The present invention relates to a processing quality prediction method, and more particularly to a wafer processing quality prediction method. Background Art

[0002] During the process of epitaxy, polishing, etc. on a wafer, it is often necessary to wait until the processing is completed before knowing the processing quality of the wafer after processing.

[0003] Therefore, for those skilled in the art, if a technical solution can be developed to predict the processing quality before processing the wafer, relevant pre-processing operations can be performed on the wafer in advance, thereby achieving the effect of improving the quality of the wafer after processing. Summary of the invention

[0004] In view of this, the present invention provides a wafer processing quality prediction method, which can be used to solve the above technical problems.

[0005] An embodiment of the present invention provides a chip processing quality prediction method, which is suitable for a chip processing quality prediction device, including: obtaining a stress distribution map of a chip to be processed; obtaining a side view of the stress distribution map on a specific axis, wherein the side view includes multiple data points on the specific axis, and each data point has a corresponding stress value; in response to determining that the multiple data points include at least N specific data points, determining that a predicted processing result of the chip to be processed is an abnormal result, wherein the stress value corresponding to each specific data point is higher than a preset stress threshold, wherein N is a positive integer; and in response to determining that the multiple data points include less than N specific data points, determining that the predicted processing result of the chip to be processed is a normal result.

[0006] An embodiment of the present invention provides a chip processing quality prediction method, which is suitable for a chip processing quality prediction device, including: obtaining a grayscale image of a chip to be processed; determining at least one region of interest in the grayscale image, wherein the at least one region of interest includes multiple pixels; finding at least one specific pixel in the at least one region of interest, wherein the grayscale value of each specific pixel is higher than a grayscale threshold; determining a specific proportion of the at least one specific pixel in the multiple pixels; and predicting a predicted processing result of the chip to be processed based on the specific proportion.

[0007] An embodiment of the present invention provides a chip processing quality prediction method, which is suitable for a chip processing quality prediction device, including: obtaining a stress distribution map of a chip to be processed, wherein the stress distribution map includes multiple data points, each data point has a corresponding stress value; determining M data groups in the stress distribution map based on M reference lines, wherein each data group is at least divided into a first sub-data group and a second sub-data group, and M is a positive integer; determining a characteristic of each corresponding reference line based on a first stress value distribution of the first sub-data group and a second stress value distribution of the second sub-data group corresponding to each data group; and predicting a predicted processing result of the chip to be processed based on the characteristics of each reference line. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 FIG. 4 is a schematic diagram of a wafer processing quality prediction device according to an embodiment of the present invention.

[0009] Figure 2 is a flow chart of a wafer processing quality prediction method according to a first embodiment of the present invention.

[0010] Figure 3 1 is a diagram showing different stress distribution diagrams and corresponding side views according to the first embodiment of the present invention.

[0011] Figure 4A and Figure 4B It is a schematic diagram of determining and predicting processing results according to the first embodiment of the present invention.

[0012] Figure 5 is a flow chart of a wafer processing quality prediction method according to a second embodiment of the present invention.

[0013] Figure 6 It is a grayscale diagram corresponding to different wafers to be processed according to the second embodiment of the present invention.

[0014] Figure 7 It is a schematic diagram of determining and predicting processing results according to the second embodiment of the present invention.

[0015] Figure 8 is a flow chart of a wafer processing quality prediction method according to a third embodiment of the present invention.

[0016] Fig.9A and Fig. 9B is a schematic diagram of a determination data set according to a third embodiment of the present invention.

[0017] Fig.10 is a flow chart of determining the characteristics of a reference line according to the third embodiment of the present invention.

[0018] FIG. 11A to FIG. 11E FIG. 4 is a schematic diagram of characteristics of a determination reference line according to a third embodiment of the present invention.

[0019] Fig.12 It is a schematic diagram of determining and predicting processing results according to the third embodiment of the present invention.

[0020]

Explanation of symbols

[0021] 100: Wafer processing quality prediction device

[0022] 102: Storage circuit

[0023] 104: Processor

[0024] 311~314,910: Stress distribution diagram

[0025] 311a~314a,311b~314b,411,412: Side view

[0026] 611~616: Grayscale image

[0027] 611a~616a,611b~616b: Area of ​​interest

[0028] 911~91M: Reference line

[0029] 911a: first line segment

[0030] 911b: Second line segment

[0031] A1: First stress threshold

[0032] B1: Second stress threshold

[0033] T1: Preset stress threshold

[0034] T2: Specific quantity threshold

[0035] Q1~Q4: Quadrant

[0036] S210~S250,S510~S550,S810~S840,S1011~S1022: Steps DETAILED DESCRIPTION

[0037] Please refer to Figure 1 , which is a schematic diagram of a wafer processing quality prediction device according to an embodiment of the present invention. In different embodiments, the wafer processing quality prediction device 100 can be implemented as various intelligent devices and / or computer devices, but is not limited thereto.

[0038] exist Figure 1 In the embodiment, the wafer processing quality prediction device 100 includes a storage circuit 102 and a processor 104 .

[0039] The storage circuit 102 is, for example, any type of fixed or removable random access memory (RAM), read-only memory (ROM), flash memory, hard disk or other similar device or a combination of these devices, and can be used to record multiple program codes or modules.

[0040] The processor 104 is coupled to the memory circuit 102 and may be a general purpose processor, a special purpose processor, a conventional processor, a digital signal processor, a plurality of microprocessors, one or more microprocessors combined with a digital signal processor core, a controller, a microcontroller, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), any other type of integrated circuit, a state machine, an Advanced RISC Machine (ARM) based processor, and the like.

[0041] In an embodiment of the present invention, the processor 104 can access the modules and program codes recorded in the storage circuit 102 to implement the wafer processing quality prediction method proposed by the present invention, the details of which are described as follows.

[0042] Please refer to Figure 2 , which is a flow chart of a wafer processing quality prediction method according to the first embodiment of the present invention. The method of this embodiment can be Figure 1 The wafer processing quality prediction device 100 is executed as follows: Figure 1 Description of components shown Figure 2 Details of each step.

[0043] First, in step S210, the processor 104 obtains a stress distribution map of a wafer to be processed (hereinafter referred to as A). In step S220, the processor 104 obtains a side view of the stress distribution map in a specific axis direction, wherein the side view includes a plurality of data points in the specific axis direction (e.g., the Y axis direction or the X axis direction), and each data point corresponds to a stress value.

[0044] In the embodiment of the present invention, the wafer A to be processed is, for example, a silicon carbide wafer cut from a crystal ingot, but is not limited thereto.

[0045] Please refer to Figure 3 , which are different stress distribution diagrams and corresponding side views drawn according to the first embodiment of the present invention.

[0046] exist Figure 3In FIG. 5 , the stress distribution graphs 311 - 314 are, for example, stress distribution graphs measured from four different wafers to be processed.

[0047] In this embodiment, the side views 311a-314a are, for example, side views of the stress distribution diagrams 311-314 along the Y axis, and the side views 311b-314b are, for example, side views of the stress distribution diagrams 311-314 along the X axis, but are not limited thereto.

[0048] exist Figure 3 In the figure, the side views 311a-314a respectively include a plurality of data points on the Y axis, and each data point corresponds to a stress value. In addition, the side views 311b-314b respectively include a plurality of data points on the X axis, and each data point has a corresponding stress value.

[0049] After obtaining the side view of the considered stress distribution diagram in a specific axial direction (which includes multiple data points), the processor 104 determines in step S230 whether the multiple data points include at least N specific data points, wherein the stress value corresponding to each specific data point is higher than the preset stress threshold, and N is a positive integer. In different embodiments, N can be set by the designer as any value (e.g., 2) according to requirements.

[0050] In one embodiment, in response to determining that the plurality of data points include at least N specific data points, the processor 104 determines in step S240 that the predicted processing result of the wafer A to be processed is an abnormal result. In another embodiment, in response to determining that the data points include less than N specific data points, the processor 104 determines in step S250 that the predicted processing result of the wafer A to be processed is a normal result.

[0051] Please refer to Figure 4A and Figure 4B , which is a schematic diagram of determining and predicting processing results according to the first embodiment of the present invention.

[0052] exist Figure 4A In the example, it is assumed that the processor 104 obtains the side view 411 (which includes, for example, 8 data points) after executing step S220. In this case, the processor 104 may determine Figure 4A Whether there is a specific data point among the 8 data points in the stress value higher than the preset stress threshold T1.

[0053] In this embodiment, since the stress values ​​of the eight data points shown are all lower than the preset stress threshold T1, the processor 104 may determine Figure 4AWhen the number of specific data points in the scenario is less than N (e.g., 2), step S250 is executed accordingly to determine that the predicted processing result of the wafer A to be processed is a normal result. That is, the processor 104 can predict that the processed wafer obtained after the wafer A to be processed is used for processing in the future should have good / acceptable quality, but it is not limited thereto.

[0054] exist Figure 4B In the example, it is assumed that the processor 104 obtains the side view 412 (which includes, for example, 8 data points) after executing step S220. In this case, the processor 104 may determine Figure 4B Whether there is a specific data point among the 8 data points in the stress value higher than the preset stress threshold T1.

[0055] In this embodiment, since the stress values ​​of the third data point and the eighth data point shown are higher than the preset stress threshold T1, the processor 104 may determine Figure 4B The number of specific data points in the scenario is 2. When N is assumed to be 2, the processor 104 may perform step S240 accordingly after determining that the number of specific data points is not less than N, and determine that the predicted processing result of the wafer A to be processed is an abnormal result. That is, the processor 104 may predict that the processed wafer obtained after the wafer A to be processed is used for processing in the future should have poor quality, but it is not limited thereto.

[0056] In one embodiment, after relevant personnel learn that the predicted processing result of the chip A to be processed is an abnormal result, corresponding pre-processing can be performed on the chip A to be processed in an attempt to improve the quality of the processed chip obtained after the chip A to be processed is used for processing in the future, but it is not limited to this.

[0057] Please refer to Figure 5 , which is a flow chart of a wafer processing quality prediction method according to a second embodiment of the present invention. The method of this embodiment can be Figure 1 The wafer processing quality prediction device 100 is executed as follows: Figure 1 Description of components shown Figure 5 Details of each step.

[0058] First, in step S510, the processor 104 obtains a grayscale image of the to-be-processed wafer A. Next, in step S520, the processor 104 determines a region of interest in the grayscale image.

[0059] Please refer to Figure 6 , which is a grayscale diagram corresponding to different chips to be processed according to the second embodiment of the present invention. Figure 6In the embodiment, the grayscale images 611-616 may correspond to different wafers to be processed, and the processor 104 may determine the corresponding regions of interest 611a-616a, 611b-616b in each of the grayscale images 611-616.

[0060] Taking the grayscale image 611 as an example, the processor 104 may determine the regions of interest 611a and 611b therein. In one embodiment, the processor 104 may divide the grayscale image 611 into a plurality of quadrants, and determine the desired region of interest in at least one of the quadrants. Figure 6 In the embodiment, the processor 104 may, for example, divide the grayscale image 611 into quadrants Q1 to Q4, and determine the regions of interest 611a and 611b in the quadrants Q1 and Q4, respectively, but the present invention is not limited thereto.

[0061] In one embodiment, the processor 104 may also directly determine an area at the upper right corner and the lower right corner of the grayscale image 611 to serve as the regions of interest 611 a and 611 b , but is not limited thereto.

[0062] Based on similar concepts, the processor 104 may determine corresponding regions of interest 612a-616a, 612b-616b in the grayscale images 612-616, the details of which are not further elaborated herein.

[0063] For ease of explanation, it is assumed that the grayscale image corresponding to the chip A to be processed is Figure 6 Therefore, the grayscale image 613 and the corresponding regions of interest 613a and 613b will be taken as an example for explanation below, and a person with ordinary knowledge in the art should be able to infer the operations performed by the processor 104 based on other grayscale images.

[0064] exist Figure 6 In FIG. 6 , the regions of interest 613 a and 613 b may include a plurality of pixels.

[0065] Then, in step S530, the processor 104 finds specific pixels in the regions of interest 613a and 613b, wherein the grayscale value of each specific pixel is higher than the grayscale threshold. In different embodiments, the designer may select the grayscale threshold according to the requirements, but is not limited thereto.

[0066] Next, in step S540 , the processor 104 determines a specific ratio (hereinafter referred to as R) of the specific pixel in the plurality of pixels.

[0067] Please refer to Figure 7 , which is a schematic diagram of determining and predicting processing results according to the second embodiment of the present invention.

[0068] exist Figure 7In the embodiment, after obtaining the grayscale image 613 corresponding to the chip A to be processed and the corresponding regions of interest 613a and 613b, the processor 104 can find out the pixels with grayscale values ​​higher than the grayscale threshold in the regions of interest 613a and 613b as specific pixels, and determine the corresponding specific ratio R accordingly.

[0069] For example, assuming that the regions of interest 613a and 613b include 100 pixels in total, and the number of specific pixels found is 65, the processor 104 may accordingly determine that the corresponding specific ratio R is 65%, but the present invention is not limited thereto. For another example, assuming that the regions of interest 613a and 613b include 200 pixels in total, and the number of specific pixels found is 30, the processor 104 may accordingly determine that the corresponding specific ratio R is 15%, but the present invention is not limited thereto.

[0070] Thereafter, in step S550 , the processor 104 predicts a predicted processing result of the wafer A to be processed based on the specific ratio R.

[0071] In one embodiment, in response to determining that the specific ratio R is lower than a first ratio threshold, the processor 104 determines that the predicted processing result is good. In another embodiment, in response to determining that the specific ratio R is between a first ratio threshold and a second ratio threshold, the processor 104 determines that the predicted processing result is fair, wherein the second ratio threshold is higher than the first ratio threshold. In yet another embodiment, in response to determining that the specific ratio R is higher than the second ratio threshold, the processor 104 determines that the predicted processing result is bad.

[0072] exist Figure 7 In the example, it is assumed that the first ratio threshold and the second ratio threshold are 10% and 25% respectively, but it is not limited thereto. In this case, if the obtained specific ratio R is lower than 10%, the processor 104 may determine that the predicted processing result is good; if the obtained specific ratio R is between 10% and 25%, the processor 104 may determine that the predicted processing result is ordinary; if the obtained specific ratio R is higher than 25%, the processor 104 may determine that the predicted processing result is bad, but it is not limited thereto.

[0073] In one embodiment, the processor 104 may also determine the degree of failure of the corresponding predicted processing result based on the specific ratio R. Figure 7 For example, if the specific ratio R is 5%, the processor 104 can determine that the degree of badness of the predicted processing result is 2 (corresponding to good); if the specific ratio R is 15%, the processor 104 can determine that the degree of badness of the predicted processing result is 4 (corresponding to ordinary); if the specific ratio R is 35%, the processor 104 can determine that the degree of badness of the predicted processing result is 8 (corresponding to bad), but it is not limited to this.

[0074] Please refer to Figure 8, which is a flow chart of a wafer processing quality prediction method according to a third embodiment of the present invention. The method of this embodiment can be Figure 1 The wafer processing quality prediction device 100 is executed as follows: Figure 1 Description of components shown Figure 8 Details of each step.

[0075] First, in step S810, the processor 104 obtains a stress distribution map of the wafer A to be processed (eg Figure 3 ), wherein the stress distribution diagram includes a plurality of data points, and each data point has a corresponding stress value.

[0076] Thereafter, in step S820 , the processor 104 determines M data sets in the stress distribution map based on the M reference lines.

[0077] Please refer to Fig.9A and Fig. 9B , which is a schematic diagram of a determination data set according to a third embodiment of the present invention.

[0078] exist Fig.9A In the embodiment, assuming that the stress distribution diagram 910 is a stress distribution diagram corresponding to the chip A to be processed, the processor 104 can obtain corresponding M data sets based on reference lines 911-91M, where the reference lines 911-91M are, for example, M parallel virtual lines, but are not limited thereto.

[0079] For example, after determining reference line 911, processor 104 may determine that all data points located on reference line 911 belong to the first data group among the M data groups. For another example, after determining reference line 912, processor 104 may determine that all data points located on reference line 912 belong to the second data group among the M data groups. For another example, after determining reference line 91M, processor 104 may determine that all data points located on reference line 91M belong to the Mth data group among the M data groups.

[0080] In different embodiments, the reference lines 911-91M can be designed by the designer to have any direction and / or length according to the requirements, and are not limited to Fig.9A The shown state.

[0081] exist Fig. 9B , for example, when M is equal to 14, stress value distribution diagrams of M data sets (ie, data sets 1 to 14 shown) obtained for a certain stress distribution diagram are shown.

[0082] In one embodiment, each data group may be divided into at least a first sub-data group and a second sub-data group. For example, for the i-th reference line among the M reference lines (which corresponds to the i-th data group among the M data groups), the processor 104 may first divide the i-th reference line into at least a first line segment and a second line segment, and determine the data points located on the first line segment as belonging to the first sub-data group in the i-th data group, and determine the data points located on the second line segment as belonging to the second sub-data group in the i-th data group. In the embodiment of the present invention, i is an index value.

[0083] For example, for reference line 911 (e.g., the first reference line), the processor 104 may divide the reference line 911 into a first line segment 911a and a second line segment 911b, and determine the data points located on the first line segment 911a as belonging to the first sub-data group of the first data group, and determine the data points located on the second line segment 911b as belonging to the second sub-data group of the first data group.

[0084] In one embodiment, assuming that the reference line 911 includes 5102 data points from left to right, the processor 104 may determine the 0th to 2552nd data points as belonging to the second sub-data group of the first data group, and determine the 3400th to 5102nd data points as belonging to the first sub-data group of the first data group, but the present invention is not limited thereto. In this example, the 2553th to 3399th data points may be regarded as data points corresponding to noise, and the processor 104 may not consider them in subsequent operations, thereby reducing the amount of calculation required for subsequent operations, but the present invention is not limited thereto.

[0085] exist Fig. 9B In the scenario, each data group can be divided into a corresponding first sub-data group and a second sub-data group based on the same principle, but is not limited thereto.

[0086] Then, in step S830 , the processor 104 determines the characteristics of each corresponding reference line based on the first stress value distribution of the first sub-data set and the second stress value distribution of the second sub-data set corresponding to each data set.

[0087] In an embodiment of the present invention, the processor 104 may be based on Fig.10 The content implements step S830.

[0088] Please refer to Fig.10 , which is a flow chart of determining the characteristics of a reference line according to the third embodiment of the present invention.

[0089] First, in step S1011, a specific number of at least one first specific data point in the first data set of the i-th reference line is obtained, wherein the stress value corresponding to each first specific data point is higher than a first stress threshold (eg, 0.3, hereinafter referred to as A1).

[0090] Next, in step S1012, the processor 104 determines whether the specific number exceeds a first number threshold (eg, 6). If so, the processor 104 may continue to execute step S1016 to determine that the characteristic of the i-th reference line is bad; if not, the processor 104 may continue to execute step S1013.

[0091] In step S1013, the processor 104 obtains a first number of at least one first reference data point in the first sub-data group corresponding to the i-th reference line and a second number of at least one second reference data point in the second sub-data group corresponding to the i-th reference line, wherein the stress value corresponding to each first reference data point and the stress value corresponding to each second reference data point are higher than a second stress threshold (e.g., 0.1, hereinafter referred to as B1).

[0092] Next, in step S1014, the processor 104 determines whether the second number is less than a second number threshold (eg, 50). If so, the processor 104 may continue to execute step S1015; if not, the processor 104 may continue to execute step S1017.

[0093] In step S1015, the processor 104 determines whether the first number is greater than the second number. If so, the processor 104 may continue to execute step S1016 to determine that the characteristic of the i-th reference line is bad; if not, the processor 104 may continue to execute step S1022 to determine that the characteristic of the i-th reference line is good.

[0094] In step S1017, the processor 104 determines whether the first number is greater than a second number threshold (eg, 50). If not, the processor 104 may continue to execute step S1022 to determine that the characteristic of the i-th reference line is good; if so, the processor 104 may continue to execute step S1018.

[0095] In step S1018, the processor 104 determines the reference number based on the smaller one of the first number and the second number. In one embodiment, the processor 104 may use one-third of the smaller one of the first number and the second number as the reference number, but is not limited to this. For example, if the first number and the second number are 100 and 30, respectively, the processor 104 may use one-third of the second number (i.e., 10) as the reference number. For another example, if the first number and the second number are 60 and 100, respectively, the processor 104 may use one-third of the first number (i.e., 20) as the reference number, but is not limited to this.

[0096] In step S1019, the processor 104 finds a reference number of first data points with higher stress values ​​in the first sub-data group corresponding to the i-th reference line, and determines the first reference value based on the stress values ​​corresponding to each first data point. For example, assuming that the reference number is 10, the processor 104 can use the data points corresponding to the top 10 high stress values ​​in the first sub-data group corresponding to the i-th reference line as the above-mentioned first data points, and determine the first reference value accordingly. In one embodiment, the processor 104 can, for example, sum up the stress values ​​of the above-mentioned first data points to determine the first reference value, but is not limited thereto.

[0097] Next, in step S1020, the processor 104 finds a reference number of second data points with higher stress values ​​in the second sub-data group corresponding to the i-th reference line, and determines the second reference value based on the stress value corresponding to each second data point. Continuing with the above example, assuming that the reference number is 10, the processor 104 can use the data points corresponding to the top 10 high stress values ​​in the second sub-data group corresponding to the i-th reference line as the above second data points, and determine the second reference value accordingly. In one embodiment, the processor 104 can, for example, sum up the stress values ​​of the above second data points to determine the second reference value, but is not limited thereto.

[0098] Then, in step S1021, the processor 104 determines whether the first reference value is greater than the second reference value. If so, the processor 104 may continue to execute step S1016 to determine that the characteristic of the i-th reference line is bad; if not, the processor 104 may continue to execute step S1022 to determine that the characteristic of the i-th reference line is good.

[0099] For ease of understanding, the following is supplemented by FIG. 11A to FIG. 11E To illustrate, FIG. 11A to FIG. 11E FIG. 4 is a schematic diagram of characteristics of a determination reference line according to a third embodiment of the present invention.

[0100] exist Fig.11A In the example, assuming that the first sub-data set and the second sub-data set of the i-th data set respectively have the first stress value distribution and the second stress value distribution shown, the processor 104 may execute Fig.10 method.

[0101] exist Fig.11A In the scenario, assuming that the first stress value distribution and the second stress value distribution shown in the figure instruct the processor 104 to execute steps S1011, S1012, and S1016 in sequence, the processor 104 can determine accordingly Fig.11A The corresponding characteristic of the i-th reference line is bad.

[0102] exist Fig. 11BIn the example, assuming that the first sub-data set and the second sub-data set of the i-th data set respectively have the first stress value distribution and the second stress value distribution shown, the processor 104 may execute Fig.10 method.

[0103] exist Fig. 11B In the scenario, assuming that the first stress value distribution and the second stress value distribution shown in the figure instruct the processor 104 to execute steps S1011 to S1016 in sequence, the processor 104 can determine accordingly Fig. 11B The corresponding characteristic of the i-th reference line is bad.

[0104] exist Fig. 11C In the example, assuming that the first sub-data set and the second sub-data set of the i-th data set respectively have the first stress value distribution and the second stress value distribution shown, the processor 104 may execute Fig.10 method.

[0105] exist Fig. 11C In the scenario, assuming that the first stress value distribution and the second stress value distribution shown in the figure instruct the processor 104 to execute steps S1011 to S1014, S1017 to S1021, and S1016 in sequence, the processor 104 can determine accordingly Fig. 11C The corresponding characteristic of the i-th reference line is bad.

[0106] exist Fig.11D In the example, assuming that the first sub-data set and the second sub-data set of the i-th data set respectively have the first stress value distribution and the second stress value distribution shown, the processor 104 may execute Fig.10 method.

[0107] exist Fig.11D In the scenario, assuming that the first stress value distribution and the second stress value distribution shown in the figure instruct the processor 104 to execute steps S1011 to S1014 and S1017 to S1022 in sequence, the processor 104 can determine accordingly Fig.11D The corresponding characteristic of the i-th reference line is good.

[0108] exist Fig.11E In the example, assuming that the first sub-data set and the second sub-data set of the i-th data set respectively have the first stress value distribution and the second stress value distribution shown, the processor 104 may execute Fig.10 method.

[0109] exist Fig.11E In the scenario, assuming that the first stress value distribution and the second stress value distribution shown in the figure instruct the processor 104 to execute steps S1011 to S1014, S1017 to S1021, and S1016 in sequence, the processor 104 can determine accordingly Fig.11D The corresponding characteristic of the i-th reference line is bad.

[0110] After determining whether the characteristics of each reference line 911 ˜ 91M are good or bad, in step S840 , the processor 104 predicts a predicted processing result of the wafer A to be processed based on the characteristics of each reference line 911 ˜ 91M.

[0111] In one embodiment, the processor 104 may find at least one specific reference line among the reference lines 911 - 91M, wherein the characteristic of each specific reference line is bad. That is, the processor 104 may find one or more reference lines with bad characteristics as the specific reference line.

[0112] Then, in response to determining that the number of the specific reference lines is higher than the specific number threshold, the processor 104 may determine that the predicted processing result of the wafer A to be processed is an abnormal result. In another embodiment, in response to determining that the number of the specific reference lines is not higher than the specific number threshold, the processor 104 may determine that the predicted processing result of the wafer A to be processed is a normal result.

[0113] Please refer to Fig.12 , which is a schematic diagram of determining and predicting processing results according to the third embodiment of the present invention.

[0114] exist Fig.12 In the example, assuming that the specific number threshold T2 is determined to be 2.5, the processor 104 may obtain the number of specific reference lines corresponding to the wafer A to be processed, based on Fig.12 The predicted processing result of the wafer A to be processed is determined.

[0115] For example, if the number of reference lines corresponding to the characteristic of the wafer A to be processed is 2, the processor 104 may determine that the predicted processing result of the wafer A to be processed is a normal result. For another example, if the number of reference lines corresponding to the characteristic of the wafer A to be processed is 3, the processor 104 may determine that the predicted processing result of the wafer A to be processed is an abnormal result.

[0116] In one embodiment, the processor 104 may first determine the number of corresponding specific reference lines based on the stress distribution diagrams corresponding to other wafers to be processed, and obtain the corresponding bows of these other wafers to be processed after being processed. Then, the processor 104 may generate a corresponding specific reference line number and bow according to the number of corresponding specific reference lines and bows of these other wafers to be processed. Fig.12 .

[0117] Afterwards, after a certain wafer to be processed (eg wafer A to be processed) is obtained, the predicted processing result of wafer A to be processed can be determined based on the number of specific reference lines corresponding to the wafer to be processed, but it is not limited thereto.

[0118] In summary, the method proposed in the embodiment of the present invention can perform correlation analysis on the stress distribution map and / or grayscale map of the wafer to be processed to obtain the corresponding predicted processing result. In this way, the relevant personnel can decide whether to perform pre-processing on the wafer to be processed based on the obtained predicted processing result, in an attempt to avoid the wafer from having poor quality (such as excessive bowing) after processing.

[0119] Although the present invention has been disclosed as above by way of embodiments, it is not intended to limit the present invention. Any person having ordinary knowledge in the technical field may make some changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be determined by the scope of the attached patent application.

Claims

1. A wafer processing quality prediction method, suitable for a wafer processing quality prediction device, comprising: Obtaining a stress distribution map of a wafer to be processed; Obtaining a side view of the stress distribution diagram along a specific axial direction, wherein the side view includes a plurality of data points along the specific axial direction, and each of the data points has a corresponding stress value; In response to determining that the data points include at least N specific data points, determining that a predicted processing result of the wafer to be processed is an abnormal result, wherein the stress value corresponding to each of the specific data points is higher than a preset stress threshold, wherein N is a positive integer; as well as In response to determining that the data points include less than N specific data points, it is determined that the predicted processing result of the wafer to be processed is a normal result. The method of claim 1 , wherein the specific axis is an X axis.

3. The method of claim 1, wherein the wafer to be processed is a silicon carbide wafer cut from a crystal ingot.

4. A wafer processing quality prediction method, suitable for a wafer processing quality prediction device, comprising: Obtaining a grayscale image of a wafer to be processed; Determining at least one region of interest in the grayscale image, wherein the at least one region of interest includes a plurality of pixels; Finding at least one specific pixel in the at least one region of interest, wherein the grayscale value of each specific pixel is higher than a grayscale threshold; Determining a specific ratio of the at least one specific pixel among the pixels; and A predicted processing result of the wafer to be processed is predicted based on the specific ratio. 5 . The method of claim 4 , wherein the grayscale image is divided into a plurality of quadrants, and the at least one region of interest is located in at least one of the quadrants.

6. The method of claim 5, wherein the quadrants include a first quadrant, a second quadrant, a third quadrant and a fourth quadrant, the at least one region of interest includes a first region of interest and a second region of interest, and the first region of interest and the second region of interest are located in the first quadrant and the fourth quadrant, respectively.

7. The method of claim 4, wherein the step of predicting the predicted processing result of the wafer to be processed based on the specific ratio comprises: In response to determining that the specific ratio is lower than a first ratio threshold, determining that the predicted processing result is good; In response to determining that the specific ratio is between the first ratio threshold and a second ratio threshold, determining that the predicted processing result is normal, wherein the second ratio threshold is higher than the first ratio threshold; and In response to determining that the specific ratio is higher than the second ratio threshold, the predicted processing result is determined to be bad.

8. A wafer processing quality prediction method, suitable for a wafer processing quality prediction device, comprising: Obtaining a stress distribution map of a wafer to be processed, wherein the stress distribution map includes a plurality of data points, each of which has a corresponding stress value; Determine M data groups in the stress distribution map based on M reference lines, wherein each of the data groups is at least divided into a first sub-data group and a second sub-data group, and M is a positive integer; Determine a characteristic of each of the reference lines based on a first stress value distribution of the first sub-data set and a second stress value distribution of the second sub-data set corresponding to each of the data sets; and A predicted processing result of the wafer to be processed is predicted based on the characteristics of each of the reference lines. 9 . The method of claim 8 , wherein the M reference lines include an i-th reference line, and the M data sets include an i-th data set corresponding to the i-th reference line, wherein i is an index value.

10. The method of claim 9, wherein the first sub-data group of the i-th data group includes the data points located on a first line segment of the i-th reference line, and the second sub-data group of the i-th data group includes the data points located on a second line segment of the i-th reference line.

11. The method of claim 9, wherein the step of determining the characteristic of each of the reference lines based on the first stress value distribution of the first sub-data set and the second stress value distribution of the second sub-data set corresponding to each of the data sets comprises: Obtaining a specific number of at least one first specific data point in the first data set corresponding to the i-th reference line, wherein the stress value corresponding to each of the first specific data points is higher than a first stress threshold; In response to determining that the specific quantity exceeds a first quantity threshold, the characteristic of the i-th reference line is determined to be defective.

12. The method of claim 9, wherein the step of determining the characteristic of each of the reference lines based on the first stress value distribution of the first sub-data set and the second stress value distribution of the second sub-data set corresponding to each of the data sets comprises: Obtaining a specific number of at least one specific data point in the first sub-data group corresponding to the i-th reference line, wherein the stress value corresponding to each of the specific data points is higher than a first stress threshold; In response to determining that the specific number of the at least one specific data point does not exceed a first number threshold, obtaining a first number of at least one first reference data point in the first sub-data set corresponding to the i-th reference line and a second number of at least one second reference data point in the second sub-data set corresponding to the i-th reference line, wherein the stress value corresponding to each of the first reference data points and the stress value corresponding to each of the second reference data points are higher than a second stress threshold; Determining whether the second quantity is less than a second quantity threshold; In response to determining that the second quantity is less than a second quantity threshold, determining whether the first quantity is greater than the second quantity; In response to determining that the second quantity is less than the second quantity threshold, determining that the characteristic of the i-th reference line is defective.

13. The method of claim 12, wherein in response to determining that the second amount is not less than the second amount threshold, the method further comprises: Determining whether the first quantity is greater than the second quantity threshold; In response to determining that the first quantity is not greater than the second quantity threshold, determining that the characteristic of the i-th reference line is good.

14. The method of claim 13, wherein in response to determining that the first quantity is greater than the second quantity threshold, the method further comprises: determining a reference quantity based on the smaller one of the first quantity and the second quantity; Finding the reference number of first data points with higher stress values ​​in the first sub-data group corresponding to the i-th reference line, and determining a first reference value based on the stress value corresponding to each of the first data points; Finding the reference number of second data points with higher stress values ​​in the second sub-data group corresponding to the i-th reference line, and determining a second reference value based on the stress value corresponding to each of the second data points; In response to determining that the first reference value is greater than the second reference value, determining that the characteristic of the i-th reference line is defective; In response to determining that the first reference value is not greater than the second reference value, the characteristic of the i-th reference line is determined to be good.

15. The method of claim 8, wherein the step of predicting the predicted processing result of the wafer to be processed based on the characteristics of each of the reference lines comprises: Finding at least one specific reference line among the reference lines, wherein the characteristic of each specific reference line is bad; In response to determining that the number of the at least one specific reference line is higher than a specific number threshold, determining that the predicted processing result of the wafer to be processed is an abnormal result; In response to determining that the number of the at least one specific reference line is not higher than the specific number threshold, it is determined that the predicted processing result of the wafer to be processed is a normal result.

Images