Ingot Evaluation Method

Abstract

The present invention provides a method for evaluating an ingot, comprising the following steps: obtaining multiple statistical parameters based on defect information of a first wafer at a first end and a second wafer at a second end of each of a plurality of ingots; measuring a first curvature of the first end and a second curvature of the second end of each ingot, thereby labeling the ingot into one of a plurality of corresponding categories; and establishing a classification model based on the statistical parameters of the ingots and the labeled categories.

Description

Technical Field

[0001] The present invention relates to an evaluation and detection method, and in particular to an ingot evaluation method. Background Art

[0002] The semiconductor manufacturing industry has stringent quality requirements. To improve product quality, wafers that do not meet quality specifications are screened out before packaging and are not packaged. As wafer manufacturing processes become increasingly complex, traditional visual inspection is no longer sufficient. Consequently, machine vision has been developed to reduce costs, increase speed, and minimize misjudgments. The quality of incoming ingots significantly impacts the quality of the resulting wafers. Therefore, filtering out substandard ingots before processing remains a key challenge in this field. Summary of the Invention

[0003] The present invention is directed to a crystal ingot evaluation method, which can predict the quality of the crystal ingot based on the wafers obtained through the crystal ingot processing process.

[0004] According to an embodiment of the present invention, a method for evaluating an ingot includes: obtaining a plurality of statistical parameters based on defect information of a first wafer at a first end and a second wafer at a second end of each of a plurality of ingots; measuring a first curvature of the first end and a second curvature of the second end of each ingot, thereby marking the ingot into one of a plurality of corresponding categories; and establishing a classification model based on the statistical parameters of the ingot and the categories to which it is marked.

[0005] In an embodiment according to the present invention, the steps of obtaining statistical parameters based on the defect information of the first wafer at the first end and the second wafer at the second end of each ingot include: dividing the first wafer and the second wafer into at least two regions in a specified direction; calculating the first defect ratio and the second defect ratio corresponding to the two regions divided by the first wafer based on the defect information of the first wafer; calculating the third defect ratio and the fourth defect ratio corresponding to the two regions divided by the second wafer based on the defect information of the second wafer; calculating the first difference based on the first defect ratio and the second defect ratio; calculating the second difference based on the third defect ratio and the fourth defect ratio; calculating the third difference based on the first difference and the second difference; and using the first difference, the second difference and the third difference as statistical parameters.

[0006] In an embodiment according to the present invention, the first defect ratio and the second defect ratio are respectively the ratios of at least one defect type in the two regions divided by the first wafer, and the third defect ratio and the fourth defect ratio are respectively the ratios of at least one defect type in the two regions divided by the second wafer.

[0007] In an embodiment according to the present invention, the defect information includes a plurality of defect coordinates, each defect coordinate corresponds to a defect type, and the defect types include threading screw dislocation (TSD) and basal plane dislocation (BPD).

[0008] In an embodiment according to the present invention, the step of measuring a first curvature of a first end and a second curvature of a second end of each ingot includes: measuring the curvature of a plurality of chips included in each ingot one by one; dividing each ingot into a first end section, a second end section and a middle section, wherein the middle section is located between the first end section and the second end section; calculating the average value of the curvature of the chips located in the first end section as the first curvature; and calculating the average value of the curvature of the chips located in the second end section as the second curvature.

[0009] In an embodiment according to the present invention, the categories include a first category, a second category, and a third category, and the step of marking the ingot to a corresponding one of the categories includes: determining whether the first curvature and the second curvature are within a preset range; if both the first curvature and the second curvature are within the preset range, marking the corresponding ingot as the first category; if one of the first curvature and the second curvature is within the preset range and the other of the first curvature and the second curvature is not within the preset range, marking the corresponding ingot as the second category; and if both the first curvature and the second curvature are not within the preset range, marking the corresponding ingot as the third category.

[0010] In an embodiment of the present invention, after establishing the classification model, it also includes: measuring the defect information of the first chip and the second chip at the front and rear ends of the crystal ingot to be tested to obtain statistical parameters; and inputting the statistical parameters into the classification model to predict one of the categories corresponding to the crystal ingot to be tested.

[0011] Based on the above, the defect concentration of wafers obtained through the ingot processing process is calculated to identify the distribution of defects on the wafer. Finally, a corresponding classification model is established through artificial intelligence classification algorithms. Based on this, the classification model is used to identify the quality of the ingot. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 is a block diagram of an analysis system according to an embodiment of the present invention;

[0013] Figure 2 is a schematic diagram of a crystal ingot according to an embodiment of the present invention;

[0014] Figure 3is a flow chart of a method for evaluating an ingot according to one embodiment of the present invention;

[0015] Figure 4A and Figure 4B FIG. 1 is a schematic diagram of a wafer partition according to an embodiment of the present invention.

[0016] Description of Reference Numerals

[0017] 110: Measuring instruments

[0018] 120: Analytical device

[0019] 200: Ingot

[0020] 210: First end section

[0021] 220: Middle section

[0022] 230: Second end section

[0023] A, B, C, D: Area

[0024] H1~H5、T1~T5、W:Whip

[0025] S305-S320: Steps of the Ingot Evaluation Method DETAILED DESCRIPTION

[0026] Reference will now be made in detail to exemplary embodiments of the present invention, examples of which are illustrated in the accompanying drawings. Whenever possible, the same reference numerals are used in the drawings and the description to refer to the same or like parts.

[0027] Figure 1 is a block diagram of an analysis system according to one embodiment of the present invention. Figure 1 The analysis system includes a measuring instrument 110 and an analysis device 120. The measuring instrument 110 and the analysis device 120 can perform data transmission between them, for example, by wired or wireless communication.

[0028] Measuring instrument 110 is, for example, an automated optical inspection (AOI) instrument, but the present invention is not limited thereto; measuring instrument 110 can be any instrument. AOI instruments are high-speed, high-precision optical image inspection systems that incorporate measurement lens technology, optical lighting technology, positioning measurement technology, electronic circuit testing technology, image processing technology, and automation applications. They utilize machine vision as a standard inspection technology. Measuring instrument 110 uses optical instruments to obtain the surface condition of the finished product and then employs computer image processing technology to detect defects such as foreign matter and pattern anomalies.

[0029] The analysis device 120 is an electronic device with computing capabilities, which can be implemented using a personal computer, laptop, tablet computer, smart phone, or any other device with computing capabilities, but the present invention is not limited thereto. The analysis device 120 receives defect information of multiple known wafers from the measuring instrument 110 (i.e., one or more defect coordinates of a defect and the defect type corresponding to each defect coordinate) and uses this information for training to obtain a classification model. This model is then used to predict the quality of the wafers after the wafers are processed using measurement data of the wafers obtained from the ingot to be tested, thereby determining whether the ingot to be tested meets expectations and identifying the quality of the ingot. The processing process may include cutting, grinding, polishing, etc., and may include etching the wafer before measuring the defect information, but the present invention is not limited thereto.

[0030] Figure 2 is a schematic diagram of a crystal ingot according to an embodiment of the present invention. Figure 2 After processing, the ingot 200 can be processed to obtain multiple chips W. In this embodiment, the ingot 200 includes a first end and a second end. For example, the first end of the ingot 200 is the head end, and the second end of the ingot 200 is the tail end. Alternatively, the first end of the ingot 200 can be considered the tail end, and the second end can be considered the head end. Alternatively, the left end of the ingot 200 can be considered the first end, and the right end can be considered the second end. Alternatively, the right end can be considered the first end, and the left end can be considered the second end. The present invention is not limited to this. In addition, the ingot 200 is divided into a first end section 210, a middle section 220, and a second end section 230. The middle section 220 is located between the first end section 210 and the second end section 230. Furthermore, the first end section 210 includes five chips H1-H5, and the second end section 230 includes five chips T1-T5. The number of chips included in the first end section 210 and the second end section 230 is for illustration only and is not intended to be limiting. In other embodiments, the first end section 210 , the middle section 220 , and the second end section 230 may be set by dividing the crystal ingot 200 into three equal parts, or the crystal ingot 200 may be divided into three parts as needed.

[0031] Figure 3 This is a flow chart of a method for evaluating an ingot according to an embodiment of the present invention. Figure 3 In step S305 , a plurality of statistical parameters are obtained based on defect information of a first wafer at a first end and a second wafer at a second end of each of the plurality of ingots.

[0032] by Figure 2Taking the ingot 200 as an example, the measuring instrument 110 is used to perform optical inspection on the wafer H1 (first wafer) and the wafer T1 (second wafer) at the head and tail ends (first end and second end) to detect whether there are defects in multiple coordinate positions included in each of the wafers H1 and T1, and record the coordinate positions with defects and their defect types. Generally speaking, the defect types output by the AOI instrument include threading edge dislocation (TED), threading screw dislocation (TSD), and basal plane dislocation (BPD). After analysis, it can be seen that TSD and BPD have a significant impact on quality. Therefore, in this embodiment, TSD and BPD are used for statistical analysis.

[0033] Specifically, in a specified direction (such as vertical or horizontal), wafer H1 and wafer T1 are each divided into at least two regions, and the proportion of various defect types in the two regions is calculated to obtain multiple statistical parameters. For example, Figure 4A and Figure 4B FIG. 1 is a schematic diagram of a chip partition according to an embodiment of the present invention. Figure 4A As shown in FIG, the wafer W is divided into two regions A and B in the vertical direction by passing through the center of the circle, and the number of TSD plus BPD in region A and region B is calculated respectively. Figure 4B As shown, the wafer W is divided into two upper and lower regions C and D in the horizontal direction by passing through the center of the circle, so as to calculate the number of TSDs and BPDs in the regions C and D respectively.

[0034] For example, Figure 2 Taking ingot 200 as an example, analysis device 120 calculates a first defect ratio and a second defect ratio for the two regions divided by wafer H1 based on the defect information corresponding to wafer H1 (including the coordinate locations of the defects and the defect types). The first defect ratio and the second defect ratio are the proportions of the two defect types, TSD and BPD, in the two regions divided by wafer H1, respectively. A first difference is then calculated based on the first defect ratio and the second defect ratio.

[0035] Assume that the number of coordinate positions determined to have TSD in area A of wafer H1 is HA TSD , determine the number of coordinate positions with BPD as HA BPD Furthermore, it is assumed that the number of coordinate positions determined to have TSD in the region B of the wafer H1 is HB. TSD , determine the number of coordinate positions with BPD as HB BPDBased on this, the first defect ratio R1 and the second defect ratio R2 are respectively:

[0036] R1=(HA BPD +HA TSD ) / (HA BPD +HA TSD +HB BPD +HB TSD ),

[0037] R2=(HB BPD +HB TSD ) / (HA BPD +HA TSD +HB BPD +HB TSD ),

[0038] Among them, the first difference H AB It is R1-R2.

[0039] Furthermore, the analysis device 120 calculates a third defect ratio and a fourth defect ratio corresponding to the two regions divided by wafer T1 based on the defect information corresponding to wafer T1 (including the coordinate locations of the defects and the defect types). The third defect ratio and the fourth defect ratio are the respective proportions of the two defect types, TSD and BPD, in the two regions divided by wafer T1. Furthermore, a second difference is calculated based on the third defect ratio and the fourth defect ratio.

[0040] For example, assuming that the number of coordinate positions determined to have TSD in area A of wafer T1 is TA TSD , determine the number of coordinate positions with BPD as TA BPD Furthermore, it is assumed that the number of coordinate positions determined to have TSD in the region B of the wafer T1 is TB. TSD , determine the number of coordinate positions with BPD as TB BPD The third defect ratio R3 and the fourth defect ratio R4 are

[0041] R3=(TA BPD +TA TSD ) / (TA BPD +TA TSD +TB BPD +TB TSD ),

[0042] R4=(TB BPD +TB TSD ) / (TA BPD +TA TSD +TB BPD +TB TSD ),

[0043] Among them, the second difference T AB It is R3-R4.

[0044] Then, based on the first difference H AB and the second difference T AB , calculate the third difference. For example, the third difference D AB For |H AB -T AB |. The first difference H AB , the second difference T AB and the third difference D AB as statistical parameters.

[0045] If you choose Figure 4B The calculation process of its statistical parameters can refer to the above first difference H AB , the second difference T AB and the third difference D AB The calculation process of is not described here in detail.

[0046] Additionally, in other embodiments, statistical parameters may be calculated for one, three, or more defect types.

[0047] Next, in step S310, the first curvature of the first end and the second curvature of the second end of each ingot are measured, thereby labeling the ingot into one of the corresponding multiple types. Specifically, the curvature of one or more chips W included in one or more ingots 200 is measured one by one using the measuring instrument 110. For example, the curvature of the H1 chip among the chips H1 to H5 in the first end section 210 is used as the first curvature, and the curvature of the T1 chip among the chips T1 to T5 in the second end section 230 is used as the second curvature. In other embodiments, the average curvature of the chips H1 to H5 in the first end section 210 is calculated as the first curvature, and the average curvature of the chips T1 to T5 in the second end section 230 is calculated as the second curvature.

[0048] Next, a determination is made as to whether the first and second curvatures are within a predetermined range. If both the first and second curvatures are within the predetermined range, the corresponding ingot is labeled as a first category. If one of the first and second curvatures is within the predetermined range, and the other is not within the predetermined range, the corresponding ingot is labeled as a second category. If both the first and second curvatures are not within the predetermined range, the corresponding ingot is labeled as a third category.

[0049] For example, the preset range is set to -35 μm to +10 μm. Ingots whose curvature at both ends (first and second ends) falls within the preset range are classified as a first category. Ingots whose curvature at both ends (first and second ends) is outside the preset range are classified as a third category. Ingots whose curvature at only one of the ends (first and second ends) falls within the preset range are classified as a second category.

[0050] After obtaining the statistical parameters of each ingot and its corresponding category, in step S315, a classification model is established based on the statistical parameters of the ingot and its labeled category. For example, the statistical parameters of the ingot are used as the input of the classification model, and the labeled category is used as the output. The classification model is trained to adjust the parameters of the classification model.

[0051] After establishing the classification model, the classification model can be further used to identify the quality of the ingot in step S320. Specifically, the statistical parameters obtained by measuring the defect information of the first wafer and the second wafer at the front and rear ends of the ingot to be tested are input into the classification model to predict the category of the ingot to be tested.

[0052] For example, Table 1 shows the training data. The curvature of the head and tail (the first end and the second end) is used to mark the category of the ingot, and the statistical parameter H is used to AB 、T AB and D AB .

[0053] Table 1

[0054]

[0055] Here, a multinomial logistic regression algorithm can be used to establish a classification model. This embodiment uses three types, so three extended logistic regression classifiers are used to reconstruct the classifiers. The first classifier selects the first category as the positive class (positive), and makes the second and third categories the negative classes (negative). The second classifier selects the second category as the positive class, and the first and third categories as the negative classes. The third classifier selects the third category as the positive class, and the first and second categories as the negative classes. Using these three classifiers, in the prediction stage, each classifier can obtain the probability of the current positive class based on the test sample. Afterwards, the classifier with the highest calculation result is selected, and its positive class can be used as the prediction result.

[0056] In addition, the collected data can be divided into a training data set and a validation data set. The validation data set is used to verify the accuracy of the classification model.

[0057] In summary, the defect concentration of wafers obtained through the ingot processing process is calculated to identify the distribution of defects on the wafer. Finally, an artificial intelligence classification algorithm is used to establish a corresponding classification model. Based on this classification model, the ingot quality can be identified.

[0058] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements 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 an ingot, characterized in that: include: Obtaining a plurality of statistical parameters based on defect information of a first wafer at a first end and a second wafer at a second end of each of the plurality of ingots; measuring a first curvature of the first end and a second curvature of the second end of each of the ingots, thereby marking the ingots into one of a plurality of corresponding categories; as well as Building a classification model based on the statistical parameters of the ingot and the categories marked therewith, The step of obtaining the statistical parameters based on the defect information of the first wafer at the first end and the second wafer at the second end of the respective ingots comprises: In a specified direction, dividing the first wafer and the second wafer into at least two areas; Calculating a first defect ratio and a second defect ratio corresponding to two areas divided by the first wafer based on the defect information of the first wafer; calculating, based on the defect information of the second wafer, a third defect ratio and a fourth defect ratio corresponding to the two regions divided by the second wafer; calculating a first difference based on the first defect ratio and the second defect ratio; calculating a second difference based on the third defect ratio and the fourth defect ratio; Calculating a third difference based on the first difference and the second difference; and The first difference, the second difference, and the third difference are used as the statistical parameters.

2. The ingot evaluation method according to claim 1, wherein: The first defect ratio and the second defect ratio are respectively the ratios of at least one defect type in the two regions divided by the first wafer, wherein the at least one defect type includes at least one of a threading screw dislocation and a basal plane dislocation. The third defect ratio and the fourth defect ratio are respectively the ratios of the at least one defect type in the two regions divided by the second wafer. 3 . The ingot evaluation method according to claim 1 , wherein the defect information comprises a plurality of defect coordinates, each of the defect coordinates corresponding to one of a plurality of defect types, the defect types comprising a threading screw dislocation and a basal plane dislocation.

4. The ingot evaluation method according to claim 1, wherein: The step of measuring the first curvature of the first end and the second curvature of the second end of each of the ingots comprises: measuring the curvature of each of the plurality of wafers included in the ingot one by one; dividing each of the ingots into a first end segment, a second end segment, and a middle segment, wherein the middle segment is located between the first end segment and the second end segment; calculating an average value of the curvature of the wafer located at the first end section as the first curvature; and An average value of the curvature of the wafer located at the second end section is calculated as the second curvature.

5. The ingot evaluation method according to claim 1, wherein: The categories include the first category, the second category and the third category, The step of marking the ingot into one of the corresponding categories includes: determining whether the first curvature and the second curvature are within a preset range; If both the first curvature and the second curvature are within the predetermined range, marking the corresponding ingot as the first category; If one of the first curvature and the second curvature is within the predetermined range and the other of the first curvature and the second curvature is not within the predetermined range, marking the corresponding ingot as the second category; and If both the first curvature and the second curvature are outside the predetermined range, the corresponding ingot is marked as the third category.

6. The ingot evaluation method according to claim 1, wherein: After establishing the classification model, the method further includes: measuring defect information of the first wafer and the second wafer at both front and rear ends of the ingot to be measured to obtain the statistical parameter; and The statistical parameters are input into the classification model to predict one of the categories corresponding to the ingot to be tested.

Images