Debris determination method

Abstract

The present invention is a debris determination method for determining whether or not debris is generated around a hard laser mark after forming the hard laser mark on the back surface of a wafer or after polishing the back surface of the wafer after forming the hard laser mark, characterized in that: a parameter of thickness deviation of the wafer is measured by a flatness measuring machine; then, statistical data of parameters of thickness deviation in a region (hereinafter referred to as region A) including the hard laser mark is extracted, statistical data of parameters of thickness deviation in a region (hereinafter referred to as region B) adjacent to the region A is extracted, and the statistical data of the region A and the statistical data of the region B are compared and a difference is obtained. When the difference is equal to or greater than a predetermined threshold value, it is determined that debris is generated. As a result, provided is a debris determination method capable of accurately detecting local thickness changes caused by debris.

Description

Technical Field

[0001] The present invention relates to a method for determining debris, and more particularly to a method for determining debris in a hard laser marking forming area. Background Art

[0002] In order to identify the individual silicon wafers, there is a process (hard laser marking (hereinafter also referred to as HLM) process) of printing the individual number on the flat part of the wafer end using a solid-state laser. The hard laser mark is a dotted line of text formed by hitting the wafer with a high-output solid-state laser. The hard laser mark is printed at the specified positions on the front and back of the silicon wafer in accordance with the SEMI standard.

[0003] Due to the high-output laser, the physical properties of silicon change around the dot. Therefore, in the subsequent wafer grinding process, the periphery of the dot cannot be ground in the same state as the non-dot portion, and sometimes a local protrusion with a steep thickness (a protrusion with a height of tens of nm to several μm) is formed.

[0004] This is called debris. If HLM is printed on the back side, debris will be generated on the back side. When the area containing the debris is adsorbed by the wafer suction cup during the device process, the shape of the debris will be transferred to the active surface of the device (wafer surface), and it will cause defocusing during exposure, which may cause a significant reduction in the device manufacturing yield.

[0005] Therefore, before entering the device process, screening must be performed in advance to detect debris generated on the back side of the wafer.

[0006] As a method for detecting the debris, a method is known in which an optical interferometric flatness measuring machine (WaferSight manufactured by KLA) is used to determine the presence or absence of debris based on, for example, the value of ESFQR. However, the flatness parameter of ESFQR is limited to calculating the difference between the maximum thickness and the minimum thickness of any region, and it is not easy to accurately detect whether debris has occurred.

[0007] The reason is that although the optical interferometer flatness measuring machine has high measurement accuracy, the dynamic range relative to the displacement is narrow. When the displacement at the location where debris is generated is measured, it is easy to exceed the dynamic range and the measured value contains incorrect values. These incorrect values ​​will be processed as the maximum and minimum values ​​for obtaining ESFQR. Therefore, it is not easy to correctly detect whether debris is generated in the ESFQR of the optical interferometer flatness measuring machine. Therefore, in the existing debris detection method as described above, there is a problem that it is not possible to classify whether debris is generated with good accuracy.

[0008] As another debris detection method, there is also a technique that uses a laser microscope to evaluate the back side of the wafer at high resolution. However, the total throughput is not sufficient, and it will clamp the active surface of the device, which will become a destructive inspection. Therefore, it cannot be used as a debris evaluation method for products.

[0009] Furthermore, as a method for evaluating the shape of a wafer from a different point of view from ESFQR, there are technologies disclosed in Patent Documents 1 to 3. These technologies are methods for indicating the starting point of the collapse and flipping of the peripheral shape of the silicon wafer and for quantifying them, but cannot detect localized thickness changes such as debris. In addition, the shape evaluation of the wafer is preferably performed by excluding the area of ​​1 mm to 2 mm from the outermost periphery of the wafer, and the shape data is obtained by excluding the laser mark part.

[0010] Prior art literature

[0011] Patent Literature

[0012] Patent Document 1: Japanese Patent Application Publication No. 2006-5164

[0013] Patent Document 2: Japanese Patent Application Publication No. 2004-20286

[0014] Patent Document 3: Japanese Patent Application Publication No. 2003-86646 Summary of the invention

[0015] 1. Technical issues to be resolved

[0016] There may be a situation where debris with a steep shape is generated around the laser mark point in the HLM printing area. In order to measure the flatness of the wafer with good accuracy, it is preferable that the dynamic range of the flatness measuring machine is narrow. However, if the dynamic range of the flatness measuring machine is narrow, the steep protrusion of the debris will easily exceed the dynamic range of the flatness measuring machine.

[0017] For example, in the measurement of displacement using an optical interferometer, when a height difference exceeding about 1 / 4 of the wavelength (λ) of the light source used occurs within a 1mm moving range, the interferometer cannot correctly transmit back the displacement corresponding to the actual displacement according to the principle of the optical interferometer. Therefore, the HLM print area where a steep shape change occurs is excluded from the target area for flatness measurement.

[0018] At this time, if the HLM printing area is not excluded and the flatness including the HLM printing area is measured using a flatness measuring machine, the displacement to be measured cannot be correctly measured for shapes caused by debris that exceed the dynamic range of the flatness measuring machine, resulting in excessive or insufficient output.

[0019] Currently, the presence of debris is confirmed by using flatness parameters such as ESFQR of the area including the HLM printing area without excluding it. Parameters such as ESFQR, which are called local flatness (Japanese: サイトフラットネス), are expressed as the difference (range) between the maximum and minimum values ​​of the specified area, even if the reference surface is obtained in a different way.

[0020] When measuring shape changes such as chips that exceed the dynamic range of the flatness measuring machine, as mentioned above, an excessively large or small displacement is output, and this value is often used as the maximum or minimum value when calculating the local flatness. In other words, a value that has not been correctly measured is treated as a representative value for calculating the local flatness, resulting in the local flatness value being output as an excessively large or small value.

[0021] As mentioned above, conventional flatness parameters such as ESFQR are used to detect debris near HLM (hard laser marking). Such flatness parameters are calculated by the difference between representative values ​​such as the maximum and minimum values ​​in an arbitrary range. Therefore, the local shape change caused by the debris may be hidden in the shape change of the wafer itself within the local area, and it is not easy to make the shape change specific to the HLM printing area obvious.

[0022] Furthermore, the steep shape change caused by the debris will also exceed the dynamic range of the flatness measuring machine. Therefore, when calculating the local flatness such as ESFQR, there is a high possibility that incorrect information exceeding the dynamic range will be processed as the local maximum or minimum value.

[0023] This has the problem that it is difficult to accurately detect a local thickness change caused by the debris.

[0024] The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to provide a debris determination method capable of accurately detecting a local thickness change caused by debris.

[0025] (II) Technical solution

[0026] The present invention is completed in order to achieve the above-mentioned purpose, and provides a debris determination method. After a hard laser mark is formed on the back side of a wafer, or after the back side of the wafer is ground after the hard laser mark is formed, whether debris is generated around the hard laser mark is determined. The method is characterized in that the parameters of the thickness deviation of the wafer are measured by a flatness measuring machine, and then the statistical data of the parameters of the thickness deviation of the area containing the hard laser mark (hereinafter referred to as area A) are extracted, and the statistical data of the parameters of the thickness deviation of the area adjacent to the area A (hereinafter referred to as area B) are extracted, the statistical data of the area A and the area B are compared to obtain the difference, and when the difference is above a specified threshold, it is determined that debris has been generated.

[0027] In addition, the thickness deviation mentioned here refers to the unevenness of thickness measured at a plurality of points, for example.

[0028] According to such a debris determination method, the HLM printing area is separated from other areas, and the statistical values ​​of the thickness data of each area (for example, mean value, standard deviation) are obtained. By comparing the two, debris can be detected, and local thickness changes caused by debris can be accurately detected.

[0029] Specifically, considering the spot size of HLM (diameter 50μm), the location where the flatness measuring machine outputs too large or too small values ​​is limited to the vicinity of the spot center of HLM. In addition, as the debris area increases, the area that exceeds the dynamic range of the flatness measuring machine also increases. With this, the area where the displacement is accurately measured within the dynamic range and the displacement is large also increases. Therefore, by using an evaluation method that reflects the measurement value of this area to capture its change, it is possible to determine whether debris has occurred. Specifically, instead of using only restrictive representative values ​​such as "maximum value" and "minimum value" to evaluate debris like ESFQR, the idea is to compare statistical values ​​such as the average value, center value, standard deviation, etc. derived from all measured values ​​in the HLM printing area with the same statistical values ​​of the displacement of the area where HLM is not printed. This makes it possible to accurately measure the displacement within the dynamic range and reflect changes in areas with large displacements. It is then possible to determine whether debris has occurred based on the change in the statistical value, separate the HLM printing area from other areas, and obtain the statistical values ​​of their respective thickness data (for example, the average value and standard deviation, center value). By comparing the two, it is possible to determine whether debris has occurred. As a result, it is confirmed that the determination can be made with good accuracy.

[0030] At this time, the statistical data may be set to any one of a mean value, a standard deviation, a mode value, a central value, and a variance.

[0031] Thus, the statistical data of the region A and the region B are compared to obtain a difference, and when the difference is equal to or greater than a predetermined threshold, it can be determined that debris has occurred.

[0032] At this time, the threshold value may be obtained according to the correlation between the difference in statistical data between the area A and the area B and whether debris is generated.

[0033] Thus, the statistical data of the region A and the region B are compared to obtain a difference, and when the difference is equal to or greater than a predetermined threshold, it can be determined that debris has occurred.

[0034] At this time, the minimum square can be obtained using thickness data from at least three points in an area closer to the wafer than area A (hereinafter referred to as area C), and the difference in statistical data between area A and area B can be obtained after standardizing the parameters of the thickness deviation.

[0035] Thus, the statistical data of the region A and the region B are compared to obtain a difference, and when the difference is equal to or greater than a predetermined threshold, it can be determined with higher accuracy that debris has been generated.

[0036] At this time, the portion where the hard laser mark is formed may be a portion along the outer side of the back surface of the wafer, and the region B may be a region adjacent to the region A in the circumferential direction.

[0037] This makes it possible to accurately determine whether debris has occurred in the HLM for identifying the individual numbers of individual silicon wafers.

[0038] (III) Beneficial effects

[0039] As described above, according to the debris determination method of the present invention, it is possible to accurately detect a local thickness change caused by debris.

[0040] Specifically, in the evaluation method based on ESFQR, no correlation is obtained between ESFQR and the displacement amount of debris. However, in the present invention, the statistical value of the thickness of the hard laser marked portion is compared with the statistical value of the thickness of other portions, and whether debris is generated can be accurately determined based on the change in the value.

[0041] This can prevent wafers with scraps from being delivered to customers. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This shows an example of the HLM printing position.

[0043] Figure 2 This is a flowchart showing an example of the HLM peripheral thickness deviation (chipping) calculation process.

[0044] Figure 3 This is an explanatory diagram showing an example of a part of the HLM peripheral thickness deviation (chipping) calculation process.

[0045] Figure 4 This is a diagram showing an example of an embodiment of the HLM peripheral thickness deviation (chipping) calculation process, which is a diagram of steps 1 to 4 and a diagram of "evaluation area and comparison area coordinates when the notch is the origin = (0, 0)".

[0046] Figure 5 It is a diagram of steps 1 to 4 when normalization is performed and when normalization is not performed.

[0047] Figure 6 is a table showing the measurement results.

[0048] Figure 7 is a graph showing the measurement results. DETAILED DESCRIPTION

[0049] Hereinafter, the present invention will be described in detail, but the present invention is not limited to these descriptions.

[0050] As described above, a debris determination method capable of accurately detecting a local thickness change caused by debris is obtained.

[0051] The inventors have repeatedly conducted intensive studies on the above-mentioned problems and have discovered a debris determination method, which determines whether debris is generated around the hard laser mark after a hard laser mark is formed on the back side of a wafer or after the back side of the wafer is ground after the hard laser mark is formed. The method is characterized in that the parameters of the thickness deviation of the wafer are measured by a flatness measuring machine, and then the statistical data of the parameters of the thickness deviation of the area containing the hard laser mark (hereinafter referred to as area A) are extracted, and the statistical data of the parameters of the thickness deviation of the area adjacent to the area A (hereinafter referred to as area B) are extracted, and the statistical data of the area A and the area B are compared to obtain the difference. When the difference is above a specified threshold value, it is determined that debris has been generated, thereby correctly detecting the local thickness change caused by the debris, thereby completing the present invention.

[0052] Hereinafter, the chip determination method of the present invention will be described.

[0053] (Debris determination method)

[0054] (Formation of HLM)

[0055] The HLM for identifying the individual wafers cut from the single crystal silicon is printed. At this time, the OCR (Optical Character Reader: Optical Character Reader) mark and T7 printing conditions are based on the SEMI standards.

[0056] Figure 1 An example of the HLM printing position when the hard laser mark 2 is printed on the wafer 1 is shown.

[0057] (Wafer grinding)

[0058] For the wafer printed with HLM, at least the surface (usually the back surface) with HLM is polished. Here, a double-sided polishing process can be performed.

[0059] (Obtain wafer thickness data)

[0060] The wafer thickness data (front side - back side) is calculated from the displacement data of both sides of the wafer using a flatness measuring machine.

[0061] The flatness measuring machine used in the present invention is not particularly limited as long as it has the target surface resolution. For example, WaferSight2 manufactured by KLA and an optical interferometer measuring machine can be used for measurement.

[0062] (Extract thickness data of HLM area and HLM surrounding area)

[0063] Figure 2 This is a flowchart showing an example of the HLM peripheral thickness deviation (chipping) calculation process.

[0064] like Figure 2 As shown, the peripheral thickness deviation (chipping) calculation process has steps (1) to (4).

[0065] In step (1), the vicinity of the HLM portion is cut out, and the minimum square surface is obtained using multiple points that do not include the HLM to normalize the thickness deviation. For example, there is step 1. In step 1, an arbitrary area including the HLM is cut out, and when it is a wafer with a diameter of 300 mm, the thickness deviation is normalized using any 9 points inside the radius of 147 mm.

[0066] Here, the reason why the standardized region is set to be inside the radius of 147 mm is to prevent the thickness collapse shape of the wafer outer periphery from being included in the standardized region.

[0067] In step (2), arbitrary regions A and B are cut out.

[0068] In step (3), the statistical values ​​such as the average thickness of arbitrary regions A and B are calculated. For example, there are steps 2 to 4. In step 2, the thickness deviation information of the region inside the radius of 148 mm is deleted, and the thickness deviation of the HLM portion and its adjacent regions is extracted. In step 3, the thickness deviation of the HLM portion from step 2 is deleted (non-HLM portion region: comparison region). In step 4, the thickness deviation of the HLM portion from step 2 is extracted (HLM portion region: evaluation target region).

[0069] Here, the thickness deviation information in the area inside the radius of 148 mm is deleted because, for example, when T7 of HLM is set as the evaluation object, T7 is printed outside the radius of 148 mm.

[0070] In step (4), an arbitrary statistical value difference AB is calculated. For example, [thickness difference = chip size] = [average value of thickness in step 4] - [average value of thickness in step 3].

[0071] Figure 3 This is an explanatory diagram showing an example of a part of the process of calculating the thickness deviation (chips) around the HLM. It shows an example of steps (1) to (3). In step (1), an arbitrary area including the HLM is cut out, and the thickness deviation is standardized at any 12 points inside the radius of 147 mm. In step (2), arbitrary areas A and B are cut out. In step (3), statistical values ​​such as the average thickness of arbitrary areas A and B are calculated.

[0072] (Process (1))

[0073] According to the HLM peripheral thickness deviation (debris) calculation process, first, the thickness data of the laser mark periphery is cut out from the wafer thickness data. Then, the minimum plane is calculated using the thickness data of the inner area of ​​the area printed with the laser mark (at least 3 points), and the thickness deviation of the cut-out area is standardized, wherein the distance between the inner area and the area printed with the laser mark is more than twice the distance from the outer periphery of the wafer to the area printed with the laser mark in the radial direction.

[0074] The upper limit of the number of points of thickness data is not particularly limited, and can be set to, for example, 100,000 points or less.

[0075] (standardization)

[0076] Note that the calculation for obtaining the least squares is not necessarily required, and the statistical values ​​of the regions A and B may be obtained without performing normalization.

[0077] That is, [calculated statistical value] = ([shape data of the area printed by HLM] - [minimum square area]) - ([shape data of the part without HLM] - [minimum square area]) = [shape data of the area printed by HLM] - [shape data of the part without HLM]. Even if the minimum square area is not obtained, the necessary statistical value can be calculated.

[0078] However, if the thickness shape around the HLM is simply cut out to depict the thickness data, the shape of the HLM print area may be buried in the thickness deviation of the wafer itself, and it is difficult to visually identify the thickness shape deviation around the HLM. In order to make the information obtained by calculation consistent with the visual information, it is also preferred to calculate the minimum square based on the cut-out area and standardize the thickness deviation of the cut-out area. This is because by implementing standardization of the cut-out area, it is easy to visually see the HLM debris (protrusions: white in grayscale) (refer to Figure 5 ).

[0079] (Process (2))

[0080] (Calculation of statistical values)

[0081] like Figure 3 As shown in step (2), a boundary line is drawn with the printing width of the laser mark to divide the printing area (area A) and the non-printing area (area B). The boundary line can be drawn with an r-θ coordinate system or an xy coordinate system. Figure 3 where is the r-θ coordinate system.

[0082] (Process (3))

[0083] (Calculation of statistical values)

[0084] like Figure 3 As shown in step (3), the statistical value such as the average thickness of arbitrary regions A and B is calculated. Here, the statistical value of the thickness of regions A and B is set to any data of the average value, standard deviation, mode value, center value, and variance. At this time, the type of statistical value used is the same type of statistical value as the statistical value of regions A and B.

[0085] (Process (4))

[0086] (Determine whether debris is generated)

[0087] The difference between the statistical values ​​of area A and area B is calculated to quantify the thickness variation around the laser mark (difference = [statistical value A] - [statistical value B]).

[0088] A predetermined threshold is set for the "difference in statistical value", and when the threshold is exceeded, it is determined that debris has occurred.

[0089] As for the threshold value at this time, it is preferable to obtain the correlation between the degree of the debris size and the statistical value in advance, and to set the threshold value in advance according to the debris size that should be determined as the generation of debris.

[0090] Example

[0091] Hereinafter, the present invention will be specifically described with reference to Examples, but the present invention is not limited to these Examples.

[0092] (Example 1)

[0093] (Wafer preparation)

[0094] Twelve wafers with a diameter of 300 mm were prepared and designated as samples #1 to 12.

[0095] (Formation of HLM)

[0096] HLM is printed at a specified location on the back side of each wafer, followed by double-sided polishing.

[0097] (Obtain wafer thickness data)

[0098] As a flatness measuring machine, an optical interferometer type LNSW manufactured by COBELCO Research Institute was used for measurement.

[0099] Figure 4 This is a diagram showing an example of an embodiment of the HLM peripheral thickness deviation (chipping) calculation process, which is a diagram of steps 1 to 4 and a diagram of "evaluation area and comparison area coordinates when the notch is the origin = (0, 0)".

[0100] use Figure 4 The figure of "Evaluation area and comparison area coordinates when the notch is the origin = (0,0)" and the figures of steps 1 to 4 are used to illustrate the specific ranges of [printing area = area A] and [non-printing area = area B] and [area C for calculating the minimum square] formulated in the embodiment.

[0101] exist Figure 4 In the embodiment shown, when the wafer is viewed from the back with the notch facing downward, the cutout region is defined as a region divided by x=(-22.6 mm to -5 mm) and y=(0.5 mm to 16.6 mm) with the notch being the origin (0,0).

[0102] Then, the minimum square is obtained using any 9 points inside the cut-out area with a radius of 147 mm, and the thickness of the cut-out area is normalized (refer to Figure 4 Step 1).

[0103] Delete the thickness data of the area inside the radius of 148mm and cut out only the arc-shaped thickness data including the HLM printing area (refer to Figure 4 Step 2).

[0104] Printing area = Area A is the area formed by lines connecting points A1 to A4; non-printing area = Area B is the area formed by lines connecting points B1 to B2 and points A1 to A2, and the area formed by lines connecting points A3 to A4 and points B3 to B4.

[0105] Here, B1-B2 and A1-A2, A3-A4, B3-B4 are connected by straight lines; B1-A1, A1-A3, A3-B3 are connected by a curve formed by a curvature of 148 mm; B2-A2, A2-A4, A4-B4 are connected by a curve formed by a curvature of 149.2 mm.

[0106] The area C used to calculate the minimum square is an area cut out by a line connecting B1 and C1, C1 and C2, C2 and B3, and a line connecting B1 and B3 with a curve having a curvature of 148 mm.

[0107] Here, the coordinates are A1 (x, y) = (-16.4mm, 2.5mm), A2 (x, y) = (-16.4mm, 2.2mm), A3 (x, y) = (-6.2mm, 2.0mm), A4 (x, y) = (-6.2mm, 1.2mm), B1 (x, y) = (-22.6mm, 3.3mm), B2 (x, y) = (-22.6mm, 1.7mm), B3 (x, y) = (-5mm, 1.7mm), B4 (x, y) = (-5mm, 0.5mm), C1 (x, y) = (-22.6mm, 16.6mm), C2 (x, y) = (-5mm, 16.6mm).

[0108] The HLM print area (width about 10mm) is separated from the rest of the area (refer to Figure 4 Steps 3 and 4).

[0109] (Threshold setting)

[0110] Based on information from wafer users indicating the correlation between the frequency of defocusing and manufacturing conditions, the difference between the average values ​​of the thickness of region A and region B was found, and no product exceeding 0.01 μm was detected. Therefore, the threshold was set to 0.01 μm.

[0111] (Calculation of statistical values ​​of thickness variation parameters of evaluation target samples and determination of pass / fail status)

[0112] Next, for the 12 evaluation samples, the region A and the region B were extracted in the same manner as above, the average thickness of the region A and the region B was obtained, and the difference in the average thickness of the region A and the region B was calculated.

[0113] Then, wafers whose difference was less than the threshold value of 0.01 μm were deemed acceptable, and wafers whose difference was 0.01 μm or more were deemed unacceptable.

[0114] The result was that 9 pieces were qualified and 3 pieces were unqualified.

[0115] (Verification of eligibility)

[0116] In order to verify whether the judgment of whether or not the product is qualified is carried out correctly, the HLM printing area is measured using an ultra-high precision 3D measuring machine (UA3P) manufactured by Panasonic. The UA3P can directly measure the surface displacement by destructive inspection. All of the defective products were found to have bulges in the HLM printing area, while none of the qualified products were found to have bulges in the HLM printing area.

[0117] (Standardized Research)

[0118] Regarding Example 1, a comparative study was conducted between a case where normalization was performed and a case where normalization was not performed.

[0119] Figure 5 The diagrams are of steps 1 to 4 when normalization is performed and when normalization is not performed. Even when normalization is not performed, there is no difference in the crumb size of the HLM calculated, but the bulge of the HLM part can be made more obvious when normalization is performed than when normalization is not performed.

[0120] (Comparative Example 1)

[0121] The 12 evaluation target wafers of Example 1 were also used to perform measurement by ESFQR.

[0122] The measurement conditions of ESFQR are as follows.

[0123] Measuring machine: WaferSight2+ manufactured by KLA

[0124] Exclusion edge (Japanese: エッヂエクスクルージョン): 2mm

[0125] Number of parts: 20 parts

[0126] Part length: 10mm

[0127] Exclude laser marking: No

[0128] (Measurement results and qualification determination results)

[0129] Before the 12 evaluation target wafers of Example 1 were subjected to destruction inspection using UA3P, they were measured under the above-mentioned ESFQR conditions. Figure 6 is a table showing the measurement results, Figure 7 : is a graph showing the measurement results. The measurement results of ESFQR are not correlated with UA3P of Example 1, and it is found that it is not possible to determine whether or not debris is generated based on the measurement of ESFQR.

[0130] As described above, according to the embodiments of the present invention, it is possible to correctly detect a local thickness change caused by debris.

[0131] This manual contains the following methods.

[0132] [1]: A debris determination method, which determines whether debris is generated around the hard laser mark after a hard laser mark is formed on the back side of a wafer, or after the back side of the wafer is ground after the hard laser mark is formed, characterized in that the thickness deviation parameters of the wafer are measured by a flatness measuring machine, and then the statistical data of the thickness deviation parameters of the area containing the hard laser mark (hereinafter referred to as area A) are extracted, and the statistical data of the thickness deviation parameters of the area adjacent to the area A (hereinafter referred to as area B) are extracted, and the statistical data of the area A and the area B are compared to obtain the difference, and when the difference is greater than a specified threshold value, it is determined that debris has been generated.

[0133] [2]: The debris determination method of the above-mentioned [1] is characterized in that the statistical data is set to any one of the mean value, standard deviation, mode value, central value, and variance.

[0134] [3]: The debris determination method of [1] or [2] is characterized in that the threshold value is obtained based on the correlation between the difference in statistical data between the area A and the area B and whether debris is generated.

[0135] [4]: A debris determination method according to any one of the above-mentioned methods [1] to [3], characterized in that: in an area (hereinafter referred to as area C) which is closer to the inside of the wafer than the area A, the minimum square is obtained using thickness data of at least 3 points, and the difference in statistical data between the area A and the area B is obtained after standardizing the parameters of the thickness deviation.

[0136] [5]: A debris determination method according to any one of the above-mentioned methods [1] to [4], characterized in that the portion where the hard laser mark is formed is set as a portion along the outer side of the back side of the wafer, and the region B is set as a region adjacent to the region A in the circumferential direction.

[0137] The present invention is not limited to the above-described embodiments. The above-described embodiments are illustrative, and any scheme having substantially the same structure and achieving the same function and effect as the technical concept described in the claims of the present invention is included in the technical scope of the present invention.

Claims

1. A debris determination method, after forming a hard laser mark on the back side of a wafer, or after grinding the back side of the wafer after forming the hard laser mark, to determine whether debris is generated around the hard laser mark, characterized in that: The thickness deviation parameters of the wafer are measured by a flatness measuring machine, and then statistical data of the thickness deviation parameters of the area containing the hard laser mark (hereinafter referred to as area A) are extracted, and statistical data of the thickness deviation parameters of the area adjacent to the area A (hereinafter referred to as area B) are extracted, and the statistical data of the area A and the area B are compared to obtain the difference, and when the difference is above a specified threshold, it is determined that debris has been generated.

2. The debris determination method according to claim 1, characterized in that: The statistical data is set to any one of a mean value, a standard deviation, a mode value, a central value, and a variance.

3. The debris determination method according to claim 1 or 2, characterized in that: The threshold value is obtained according to the correlation between the difference in statistical data between the area A and the area B and whether debris is generated.

4. The debris determination method according to claim 1 or 2, characterized in that: In an area closer to the inside of the wafer than the area A (hereinafter referred to as area C), the minimum square is obtained using thickness data of at least 3 points, and the difference in statistical data between the area A and the area B is obtained after standardizing the parameters of the thickness deviation.

5. The debris determination method according to claim 1 or 2, characterized in that: The portion where the hard laser mark is formed is a portion along the outer side of the back surface of the wafer, and the region B is a region adjacent to the region A in the circumferential direction.

Images