Defect localization method for patterned silicon wafers on appearance equipment

Abstract

This invention provides a defect location method for patterned silicon wafers on surface treatment equipment. For small-size die products, a defect map is obtained by merging dies, where each measurement area is composed of p*q merged die areas. The defect map is observed using measurement equipment, and calibration alignment points A(x1,y1) and B(x2,y2) are marked on it. These calibration alignment points include multiple potential origins. Geometric relationships are used to determine whether each potential origin at calibration alignment points A and B is the true origin. This method allows for the reuse of low-precision equipment, reducing equipment upgrade costs and lowering company operating costs. It also improves accuracy and efficiency, eliminating the need for defect location to confirm the origin of merged dies.

Description

Technical Field

[0001] This invention relates to the field of semiconductor technology, and in particular to a method for locating defects in patterned silicon wafers on surface treatment equipment. Background Technology

[0002] When the single-sided die dimension of a product is less than 1500µm, due to the limitation of the number of dies per row on the testing machine, it is necessary to perform testing using merged dies (a multiple of the actual die size). See the example. Figure 1 After merging dies, the die size in the (2x2) map increases, but under a microscope, all dies look the same. In applications, it's difficult to determine which four neighboring dies are the origin of the merged die. Figure 2 It cannot be confirmed whether there are four yellow dies or four red dies; manual positioning can only be guessed.

[0003] Common microscope positioning methods in existing technologies:

[0004] 1. Please refer to Figure 3 The method involves roughly aligning the origin (with permissible deviation) and using large-size defects to correct the coordinates. However, this method has poor accuracy, large deviations, and certain requirements for the applicable wafers (large-size defects). When the defects are small, this method is not applicable.

[0005] 2. Please refer to Figure 4 The microscope adds an image processing module, which takes pictures and stores the origin image, and then calls the image to search for positioning. However, the machine needs to be modified, which is costly. Files need to be created for different wafers, resulting in low reuse rate. Image search will also fail when the image contrast is low.

[0006] To address the aforementioned issues, a novel method for defect localization of patterned silicon wafers on appearance devices is required. Summary of the Invention

[0007] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide a defect location method for patterned silicon wafers on appearance equipment, which solves the problem that in the prior art, the die size becomes larger in the defect image after merging dies, but the dies all look the same under a microscope, making it difficult to determine which of the merging adjacent dies is the origin of the merged die in application.

[0008] To achieve the above and other related objectives, the present invention provides a method for locating defects in a patterned silicon wafer on an appearance device, comprising:

[0009] Step 1: Defect detection is performed using the merging die method to obtain a defect map. Each measurement region on the defect map is composed of p*q merged die regions.

[0010] Step 2: Use measurement equipment to observe the defect map and mark the calibration alignment points A(x1,y1) and B(x2,y2) on the defect map. The calibration alignment points include multiple suspected origins.

[0011] Step 3: Use geometric relationships to determine whether each suspected origin point at calibration alignment points A and B is the true origin point.

[0012] Preferably, the single die in step one has a single side size of less than 1500um.

[0013] Preferably, p and q in step one are positive integers greater than or equal to 2.

[0014] Preferably, the measuring device in step two is a microscope.

[0015] Preferably, the calibration alignment points A(x1,y1) and B(x2,y2) in step two are located at 1 / 4 and 3 / 4 of the original diameter of the wafer, respectively.

[0016] Preferably, the method for determining whether each suspected origin is the true origin in step three includes: the distance between any merged die angular coordinate and the initial die angular coordinate is a multiple of die; the deviation value between the suspected origin and the initial die angular coordinate is obtained; and the magnitude of the deviation value is used to determine whether the suspected origin is the true origin.

[0017] Preferably, the method for obtaining the deviation value e in step three includes: merging the dimensions of the die in the x and y directions as (u, v), the dimensions of a single die in the x and y directions as (o, m), the coordinates of the initial die as (z, c), and the deviation between the measuring device and the defect map as p. Then, the deviation multiple t of the suspected origin in the x-axis direction from u is │(distance between the suspected origin and the initial die in the x-axis + or - p) / u)│, and the deviation multiple t of the suspected origin in the y-axis direction from u is │(distance between the suspected origin and the initial die in the y-axis + or - p) / v)│. If the deviation value e of the suspected origin is │t - [t]│, where [t] is the value rounded up or down; when the deviation values ​​e of the suspected origin in both the x and y directions are less than the error threshold, the suspected origin is defined as the true origin.

[0018] Preferably, the error threshold in step three is 0.1.

[0019] Preferably, step three further includes, after calculating the true origin at one calibration alignment point, using the method of obtaining horizontal points to calculate the ordinate position of the true origin at another calibration alignment point. When the deviation value e of the suspected origin at the other calibration alignment point in the x-axis direction is less than the error threshold, the suspected origin is defined as the true origin.

[0020] Preferably, the method for obtaining the horizontal points in step three includes: calibrating the slope m = │(y2 - y1) / (x2 - x1)│ of the alignment points A(x1, y1) and B(x2, y2). When the measuring device moves, there is a deviation δ: m = │[(y2 + δ) - (y1 + δ)] / (x2 - x1)│. After calculating the true origin at a calibration alignment point, a slope threshold is set. If the maximum error slope is less than the slope threshold, the two points are defined to be on the same horizontal line.

[0021] Preferably, the method for determining whether each suspected origin is a true origin in step three includes: within a circle of known size, when the distance between any combined die corner coordinate and the starting die corner coordinate is a multiple relationship of the die, the number of valid dies within the circle drawn with this combined die corner coordinate is equal to the number drawn with the starting die corner coordinate; conversely, there is an error tangent to the circumference, resulting in inconsistent numbers of valid dies; the number of valid dies within the circle drawn with the suspected origin is used to determine whether it is a true origin.

[0022] Preferably, the method for determining whether it is a true origin by the number of valid dies within the circle drawn with the suspected origin in step three includes: calculating the radius r of the circle based on the wafer size; calculating the distance from the outermost angle to the starting die corner coordinate using the Pythagorean theorem according to the starting die corner coordinate and the die size. If d > r, it is an invalid chip outside the circle; if d < r, it is a valid chip inside the circle, and the number of valid chips n is calculated; by positioning different suspected origins, the number of valid chips mx generated by different suspected origins is calculated. When mx = n, it indicates that this point is the true origin.

[0023] Preferably, a notch for rough alignment is formed in the lower half of the wafer in step three, and the number of valid dies above the center point is calculated.

[0024] Preferably, if the deviation between the measuring device and the defect map in step three is p, the number of valid chips generated by non - deviation, x - axis + p deviation, x - axis - p deviation, y - axis + p deviation, and y - axis - p deviation of different suspected origins is calculated respectively.

[0025] Preferably, in step three, it also includes determining whether it is a true origin according to the absolute deviation sum and data range of multiple groups of valid chip numbers of each suspected origin.

[0026] As described above, the method for defect location of the chip of the present invention on the appearance device has the following beneficial effects:

[0027] The method of the present invention enables the reuse of low - precision devices, reduces equipment upgrade expenses and lowers the company's operation costs; improves precision and efficiency, and the confirmation of the merged die origin no longer requires defect location. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 The diagram shows a merged die diagram of the existing technology;

[0029] Figure 2 The diagram shown is an illustration of a merged die in existing technology that is difficult to determine.

[0030] Figure 3 The diagram shows a defect location method in the prior art.

[0031] Figure 4 This is a schematic diagram of another defect location method according to the present invention;

[0032] Figure 5 The diagram shown is a schematic representation of the process flow of the present invention.

[0033] Figure 6 The diagram shown illustrates the process of finding a horizontal point according to the present invention.

[0034] Figure 7 The diagram shows a defect image and a microscope image of a deviation identification method according to an embodiment of the present invention.

[0035] Figure 8 The diagram shows a defect image and a microscope image of a counting and identification method according to an embodiment of the present invention.

[0036] Figure 9 The diagram shown illustrates the counting and identification method of an embodiment of the present invention for calculating the number of dies at different locations. Detailed Implementation

[0037] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention.

[0038] Please see Figure 5 This invention provides a method for locating defects in patterned silicon wafers on an appearance device, comprising:

[0039] Step 1: Defect detection is performed using the merging die method to obtain a defect map. Each measurement region on the defect map is composed of p*q merged die regions.

[0040] In some embodiments, the single die in step one has a single side size of less than 1500um.

[0041] In some embodiments, p and q in step one are positive integers greater than or equal to 2.

[0042] Step 2: Use measurement equipment to observe the defect map and mark the calibration alignment points A(x1,y1) and B(x2,y2) on the defect map. The calibration alignment points include multiple suspected origins.

[0043] In some embodiments, the measuring device in step two is a microscope.

[0044] In some embodiments, the calibration alignment points A(x1,y1) and B(x2,y2) in step two are located at 1 / 4 and 3 / 4 of the original diameter of the wafer, respectively.

[0045] Step 3: Use geometric relationships to determine whether each suspected origin point at calibration alignment points A and B is the true origin point.

[0046] In some embodiments, the method for determining whether each suspected origin is the true origin in step three includes: the distance between any merged die angular coordinate and the starting die angular coordinate is a multiple of die; the deviation value between the suspected origin and the starting die angular coordinate is obtained; and the magnitude of the deviation value is used to determine whether the suspected origin is the true origin.

[0047] In some embodiments, the method for obtaining the deviation value e in step three includes: merging the dimensions of the die in the x and y directions as (u, v), the dimensions of a single die in the x and y directions as (o, m), the coordinates of the starting die as (z, c), and the deviation between the measuring device and the defect map as p. Then, the deviation multiple t of the suspected origin in the x-axis direction from u is │(distance between the suspected origin and the starting die in the x-axis + or - p) / u)│, and the deviation multiple t of the suspected origin in the y-axis direction from u is │(distance between the suspected origin and the starting die in the y-axis + or - p) / v)│. If the deviation value e of the suspected origin is │t - [t]│, where [t] is the value rounded up or down; when the deviation values ​​e of the suspected origin in both the x and y directions are less than the error threshold, the suspected origin is defined as the true origin.

[0048] In some embodiments, the error threshold in step three is 0.1.

[0049] For example, please see Figure 7 Merge die size (2924, 3704); die size (1462, 1852); starting coordinates (-1483, -192);

[0050] Assume the deviation between the microscope stage coordinates 0,0 and the test map is ±50µm; the equipment calibration alignment points are 1 / 4 (-48267, 3512) and 3 / 4 (48225, 3512) of the original diameter;

[0051] The methods for finding locations at Site1 include:

[0052] X①=(48267(±50)-1483(starting point)) / 2924=16.02 (15.98);

[0053] Y①=(3512(±50)+192(starting point)) / 3704=1.013 (0.98);

[0054] X②=(48267(±50)-1462-1483) / 2924=15.52 (15.48);

[0055] Y②=(3512(±50)+192) / 3704=1.013 (0.98), the X-axis deviation at this point is 0.48 die;

[0056] X③=(48267(±50)-1483) / 2924=16.02 (15.98);

[0057] Y③=(3512(±50)+192+1852) / 3704=1.513 (1.487), the Y-axis deviation at this point is 0.48 die;

[0058] X④=(48267(±50)-1462-1483) / 2924=15.52 (15.48);

[0059] Y④=(3512(±50)+192+1852) / 3704=1.513 (1.487), the point deviated by 0.48 in both the X and Y directions;

[0060] When e < 0.1, it is considered to be the error value of the machine tool, and this point is a multiple of the starting origin.

[0061] When e>0.1, the surface deviation is much greater than the machine tool error, and is considered to be a multiple of the starting point.

[0062] The methods for finding locations in Site2 include:

[0063] After identifying site1(X①, Y①) as the origin, according to the principle of horizontal lines, the corresponding points of site2 are ⑤ and ⑥.

[0064] X⑤=(48225(±50)+1483(starting point)) / 2924=17.02 (16.98);

[0065] Y⑤=(3512(±50)+192(starting point)) / 3704=1.01 (0.99);

[0066] X⑥=(48225(±50)+1483+1462) / 2924=17.52 (17.48);

[0067] Y⑥=(3512(±50)+192) / 3704=1.01 (0.99).

[0068] Conclusion: If the deviation value e from the integer of both points is less than 0.1, the origin of site1 is ① and the origin of site2 is ⑤.

[0069] In some embodiments, step three further includes, after calculating the true origin at one calibration alignment point, calculating the ordinate position of the true origin at another calibration alignment point using the method of obtaining horizontal points. When the deviation value e of the suspected origin at another calibration alignment point in the x-axis direction is less than the error threshold, the suspected origin is defined as the true origin.

[0070] In some embodiments, the method for obtaining the horizontal point in step three includes: calibrating the slope m of alignment points A(x1,y1) and B(x2,y2) as m = |(y2-y1) / (x2-x1)|, and considering the deviation δ when the measuring device moves: m = |[(y2+δ)-(y1+δ)] / (x2-x1)|. After calculating the true origin at a calibration alignment point, a slope threshold is set. If the maximum error slope is less than the slope threshold, the two points are defined as being on the same horizontal line.

[0071] For example, please see Figure 6 Due to limitations imposed by the test die within the same row of the defect detection equipment, when a small die has a single side less than 1500um, it will be tested by merging dies (2x2, 2x3, etc.) to generate a defect map. This method is specifically designed for defect localization after merging dies for such small dies.

[0072] The key to microscope positioning is wafer alignment, such as... Figure 6 The defect diagrams shown are site1 and site2. The case is presented in a 2x2 format, and under a microscope, they all show neatly arranged dies.

[0073] Methods for finding horizontal points for site1 and site2 include:

[0074] site1:(-46784,11)site2:(52632,43);die size(1462,1852);

[0075] If the equipment error is set to 50µm, then the error slope range is:

[0076] M1 = │((43 + 50) - (11 - 50)) / (52632 + 46784)│ = 0.00133;

[0077] M2 = │((43 - 50) - (11 + 50)) / (52632 + 46784)│ = 0.00069;

[0078] The slope threshold takes the maximum value: M1 = 0.00133;

[0079] In Figure 6 the right figure of

[0080] M① = (43 + 1852 - 11) / (52632 + 46784) = 0.019;

[0081] M② = (43 - 11) / (52632 + 46784) = 0.0003;

[0082] M③ = │(43 - 1852 - 11) / (52632 + 46784)│ = 0.0183.

[0083] Conclusion: M② < M1, it is considered that the two points are on the same horizontal line, otherwise M① and M③ are not on the same horizontal line.

[0084] In some embodiments, the method for determining whether each suspected origin is a true origin in step three includes: within a circle of known size, when the distance between any combined die corner coordinate and the starting die corner coordinate is a multiple of the die, then the number of valid dies within the circle drawn with this combined die corner coordinate is equal to the number drawn with the starting die corner coordinate; conversely, there is an error tangent to the circumference, resulting in inconsistent numbers of valid dies; the number of valid dies within the circle drawn with the suspected origin is used to determine whether it is a true origin.

[0085] In some embodiments, the method for determining whether a suspected origin is a true origin by the number of valid dies within the circle drawn with the suspected origin in step three includes: calculating the radius r of the circle based on the wafer size; calculating the distance from the outermost corner to the starting die corner coordinate using the Pythagorean theorem based on the starting die corner coordinate and the die size, d > r means the chip is invalid outside the circle, d < r means the chip is valid inside the circle, and calculating the number of valid chips n; by positioning different suspected origins, calculating the number of valid chips mx generated by different suspected origins, when mx = n, it indicates that this point is a true origin.

[0086] In some embodiments, a notch for coarse alignment is formed in the lower half of the wafer, then calculate the number of valid dies above the center point.

[0087] In some embodiments, if the deviation between the measurement device and the defect map in step three is p, then the number of valid chips generated without deviation, with a deviation of +p on the x-axis, with a deviation of -p on the x-axis, with a deviation of +p on the y-axis, and with a deviation of -p on the y-axis for different suspected origins is calculated respectively.

[0088] In some embodiments, step three further includes determining whether each suspected origin is a true origin based on the absolute deviation and the data range (range) of multiple groups of valid chip numbers for each suspected origin.

[0089] For example, please refer to Figure 8 , Merge die size(2924,3704); die size(1462,1852); starting coordinates (-1483,-192);

[0090] The counting identification method includes:

[0091] 1. Calculation of the tested die

[0092] According to the two-dimensional coordinate map automatically generated by the testing machine, calculate the number of valid dies with the center point above (there is a notch in the lower half, resulting in calculation errors);

[0093] Valid die: 1326.

[0094] 2. Discriminating the origin by calculating the valid dies

[0095] Assume that the deviation between the coordinates 0, 0 of the machine and the testing map is ±50um, and the device calibration alignment points are at 1 / 4 (-48267, 3512) and 3 / 4 (48225, 3512) of the original diameter. Given the WaferSize (wafer size), MergeDieSize (merged die size), hypothetical origin coordinates, and the circle radius r (100000um). Use progressive calculation to calculate the distance of the die corner farthest from the center of the circle:

[0096] For example, Die1 = √x^2 + y^2 = √1441^2 + 3517^2 = 3800 < r (inside the circle)

[0097] ……Die34 = √97933^2 + 3517^2 = 97996 < r (inside the circle)

[0098] Die35 = (100857, 3517) > r (outside the circle, not counted)

[0099] Please refer to Figure 9 , according to points ① to ④, calculate the die quantity distribution as follows: The absolute deviation < 1 and range < 2 is the origin.

[0100] It should be noted that the illustrations provided in this embodiment are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0101] In summary, the method of this invention allows for the reuse of low-precision equipment, reducing equipment upgrade expenditures and lowering company operating costs; it improves accuracy and efficiency, and eliminates the need for defect localization to confirm the origin of the merged die. Therefore, this invention effectively overcomes the various shortcomings of existing technologies and has high industrial applicability.

[0102] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A method for locating defects in a patterned silicon wafer on an appearance device, characterized in that, At least including: Step 1: Defect detection is performed using the merging die method to obtain a defect map. Each measurement region on the defect map is composed of p*q merged die regions. Step 2: Use measurement equipment to observe the defect map and mark the calibration alignment points A(x1, y1) and B(x2, y2) on the defect map. The calibration alignment points include multiple suspected origins. Step 3: Determine whether each suspected origin at calibration alignment points A and B is the true origin by calculating the distance deviation between each suspected origin and an initial die angle coordinate. The distance between any merged die angle coordinate and the initial die angle coordinate is a multiple of the die angle. The magnitude of the deviation value is used to determine whether the suspected origin is the true origin.

2. The defect location method for patterned silicon wafers on an appearance device according to claim 1, characterized in that: In step one, the single die has a single side size of less than 1500um.

3. The defect location method for patterned silicon wafers on an appearance device according to claim 1, characterized in that: In step one, p and q are positive integers greater than or equal to 2.

4. The defect location method for patterned silicon wafers on an appearance device according to claim 1, characterized in that: The measuring device mentioned in step two is a microscope.

5. The defect location method for patterned silicon wafers on an appearance device according to claim 1, characterized in that: In step two, the calibration alignment points A(x1, y1) and B(x2, y2) are located at 1 / 4 and 3 / 4 of the original diameter of the wafer, respectively.

6. The defect location method for patterned silicon wafers on an appearance device according to claim 5, characterized in that: The method for obtaining the deviation value e in step three includes: merging the dimensions of the die in the x and y directions as (u, v), the dimensions of a single die in the x and y directions as (o, m), the coordinates of the starting die as (z, c), and the deviation between the measuring device and the defect map as p. Then, the deviation multiple t of the suspected origin in the x-axis direction from u is │(the distance between the suspected origin and the starting die in the x-axis plus or minus the deviation p) / u)│, and the deviation multiple t of the suspected origin in the y-axis direction from u is │(the distance between the suspected origin and the starting die in the y-axis plus or minus the deviation p) / v)│. If the deviation value e of the suspected origin is │t - [t]│, where [t] is the value rounded up or down; when the deviation values ​​e of the suspected origin in both the x and y directions are less than the error threshold, the suspected origin is defined as the true origin.

7. The defect location method for patterned silicon wafers on an appearance device according to claim 6, characterized in that: The error threshold in step three is 0.

1.

8. The defect location method for patterned silicon wafers on an appearance device according to claim 7, characterized in that: Step three also includes calculating the ordinate position of the true origin at one calibration alignment point after calculating the true origin at another calibration alignment point using the method of obtaining horizontal points. When the deviation value e of the suspected origin at another calibration alignment point in the x-axis direction is less than the error threshold, the suspected origin is defined as the true origin.

9. The defect location method for patterned silicon wafers on an appearance device according to claim 8, characterized in that: The method for obtaining the horizontal point in step three includes: calculating the slope between the calibration alignment points A(x1, y1) and B(x2, y2), and calculating the maximum error slope as the slope threshold based on the maximum deviation that exists when the measuring device moves. If the slope is less than the slope threshold, then the two points are defined as being on the same horizontal line.

10. A method for locating defects in a patterned silicon wafer on an appearance device, characterized in that, At least including: Step 1: Defect detection is performed using the merging die method to obtain a defect map. Each measurement region on the defect map is composed of p*q merged die regions. Step 2: Use measurement equipment to observe the defect map and mark the calibration alignment points A(x1, y1) and B(x2, y2) on the defect map. The calibration alignment points include multiple suspected origins. Step 3. Determine whether it is the true origin based on the number of valid dies within the circle drawn with the suspected origin. The method includes: within a circle of known size, when the distance between any combined die corner coordinate and the starting die corner coordinate is a multiple of the die, if the number of valid dies within the circle drawn with this combined die corner coordinate is equal to the number drawn with the starting die corner coordinate, then this point is the true origin; conversely, if there is an error and it is tangent to the circumference, resulting in inconsistent numbers of valid dies, then this point is not the true origin.

11. The defect location method for patterned silicon wafers on an appearance device according to claim 10, characterized in that: The method for determining whether it is the true origin based on the number of valid dies within the circle drawn with the suspected origin in Step 3 includes: calculating the radius r of the circle based on the wafer size; calculating the distance from the outermost angle to the starting die corner coordinate using the Pythagorean theorem according to the starting die corner coordinate and the die size. Dies outside the circle with d > r are invalid dies, and dies inside the circle with d < r are valid dies, and calculating the number of valid dies n; by positioning different suspected origins, calculating the number of valid dies mx generated by different suspected origins. When mx = n, it indicates that this point is the true origin.

12. The defect location method for patterned silicon wafers on an appearance device according to claim 10, characterized in that: In Step 3, a notch for rough alignment is formed in the lower half of the wafer, and then calculate the number of valid dies above the center point.

13. The defect location method for patterned silicon wafers on an appearance device according to claim 10, characterized in that: If the deviation between the measuring device and the defect map in Step 3 is p, then calculate the number of valid dies generated by different suspected origins without deviation, with the deviation p added to the x-axis, with the deviation p subtracted from the x-axis, with the deviation p added to the y-axis, and with the deviation p subtracted from the y-axis respectively.

14. The defect location method for patterned silicon wafers on an appearance device according to claim 13, characterized in that: Step 3 also includes determining whether it is the true origin based on the absolute deviation sum and data range of multiple sets of valid die numbers for each suspected origin.

Images