Cell-to-cell comparison method

Abstract

A unit-to-unit comparison method according to an embodiment of the present invention is used to inspect a test object periodically arranged with unit cells that can be represented by two-dimensional vectors, and includes the steps of: obtaining an image of the test object including an array of the unit cells; generating a block of pixels corresponding to the same positions of a plurality of unit cells in the array from the image; determining the gray levels of the plurality of pixels in the block; and determining whether the plurality of pixels in the block have defects based on the determined gray levels of the plurality of pixels.

Description

Technical Field

[0001] The present invention relates to a unit-to-unit comparison method, and more specifically, to a method for inspecting defects in an object of inspection which is periodically arranged with unit cells that can be represented by a two-dimensional vector. Background Technology

[0002] Generally, semiconductor devices are manufactured by repeatedly performing unit processes such as film formation, etching, and metal wiring to form micro-patterns with electrical properties on a semiconductor substrate.

[0003] With the increasing integration of semiconductor devices and the higher speed of manufacturing processes, defect inspection during the manufacturing stage has become more thorough. This is because localized defects in micropatterns are directly related to defects in semiconductor devices.

[0004] Defects in semiconductor devices are detected by establishing gray levels within the wafer, then comparing the gray level of the defective area with that of adjacent areas of the same gray level, and using the difference. In other words, defect inspection is achieved by comparing the gray levels of two areas.

[0005] In this type of defect inspection method that relies on comparison, there is a problem that noise, such as pattern discoloration or pattern flashing, which does not affect semiconductor yield, may also be detected when detecting defects. Therefore, in order to calculate the signal-to-noise ratio (SNR) that distinguishes between grayscale caused by defects and grayscale caused by noise, noise inference work needs to be performed in the defective inspection area.

[0006] However, noise inference can be inaccurate, potentially negatively impacting the reliability of defect detection results. Therefore, more reliable comparison methods need to be developed for defect inspection of semiconductor devices performed through comparison. Summary of the Invention

[0007] The present invention provides a unit-to-unit comparison method for inspecting defects in an object of inspection in which the inspection area of ​​the unit of the object to be inspected is compared with the corresponding areas of the same position of multiple unit units.

[0008] An example of the present invention relates to a unit-to-unit comparison method for inspecting defects in an object of inspection that is periodically arranged with unit units that can be represented by a two-dimensional vector. The method may include the following steps: obtaining an image of the object of inspection containing an array of the unit units; generating from the image a set of pixels, i.e., a block, corresponding to the same position of a plurality of unit units in the array; determining the gray level of each of the plurality of pixels in the block; and determining whether the plurality of pixels in the block are defective based on the determined gray levels of the plurality of pixels.

[0009] In the step of determining whether multiple pixels within the block are defective, the determination of whether a pixel is defective can be made by comparing the degree to which the grayscale of the pixel currently being determined to be defective deviates (discretes) from the average grayscale of the pixels other than the pixel currently being determined to be defective with a specified threshold.

[0010] In the step of determining whether multiple pixels within the block are defective, the grayscale of the currently defective pixel in the block, and the average and standard deviation of the grayscale of the pixels in the block other than the currently defective pixel can be used.

[0011] The step of determining whether multiple pixels within the block are defective may include calculating the average grayscale value and standard deviation of the pixels within the block other than the pixel currently being determined to be defective.

[0012] The step of determining whether multiple pixels within the block are defective may further include the step of calculating the absolute value of the difference between the grayscale of the current defective pixel among the multiple pixels within the block and the average value.

[0013] The step of determining whether multiple pixels within the block are defective may also include the step of dividing the absolute value by the standard deviation.

[0014] The step of determining whether multiple pixels within the block are defective may further include comparing the value obtained by dividing the absolute value by the standard deviation with a specified threshold value.

[0015] This may include the step of displaying the defect results of the aforementioned multiple pixels as determined above.

[0016] In the step of determining the grayscale of each of the multiple pixels within the aforementioned block, the determined grayscale may include noise.

[0017] The unit-to-unit comparison method involved in this invention improves the reliability of defect inspection of the inspected object by comparing the inspection area of ​​the inspected unit with the corresponding areas of the same position of multiple unit units through a one-to-many correspondence.

[0018] Furthermore, the unit-to-unit comparison method of the present invention compares the inspection area of ​​the inspected unit with the statistical values ​​of the corresponding areas of multiple unit units at the same position, thereby omitting the extra work of inferring noise when determining defects in the inspected object.

[0019] The effects that can be obtained by the present invention are not limited to those mentioned above, and any other effects not mentioned can be clearly understood by those skilled in the art based on the following description. Attached Figure Description

[0020] Figure 1 This is an example of a defect inspection object, i.e., an inspected object, performed by a unit-to-unit comparison method according to an embodiment of the present invention.

[0021] Figure 2 This is a flowchart of a unit-to-unit comparison method according to an embodiment of the present invention.

[0022] Figure 3 This is a diagram illustrating the principle of block generation in a unit-to-unit comparison method according to an embodiment of the present invention.

[0023] Figure 4 This is a flowchart of a method for determining defects in a unit-to-unit comparison method according to an embodiment of the present invention.

[0024] Figure 5 This is a simplified structural diagram of a defect detection system for a unit-to-unit comparison method according to an embodiment of the present invention. Detailed Implementation

[0025] Throughout this manual, the terms "unit," "device," and "system" refer to a unit that performs actions that combine one or more functions, and can be implemented through hardware, software, or a combination of hardware and software.

[0026] As used in this specification, the terms "part," "device," and "system" can be considered equivalent to computer-related entities, namely hardware, combinations of hardware and software, software, or running software. Furthermore, the application programs running in this invention can be constructed in units of "parts," and can be recorded in a readable, writable, and erasable form in a single physical memory, or can be distributed and recorded across two or more memory locations or recording media.

[0027] The terminology used in this specification is defined in consideration of the functionality of the invention and may vary depending on the intent or convention of the user. Therefore, the definition of such terminology should be based on the entire contents of this specification.

[0028] Furthermore, the embodiments disclosed below do not limit the scope of the present invention, but are merely illustrative of the constituent elements suggested in the scope of the claims of the present invention. Embodiments of constituent elements that are included in the technical concept throughout the specification of the present invention and that are included in the constituent elements of the claims and can be substituted as equivalents may be included in the scope of the present invention.

[0029] In addition, the terms “first,” “second,” “one side,” “the other side,” etc., used in the embodiments disclosed below are used to distinguish one constituent element from other constituent elements, and the constituent elements are not limited by the above terms.

[0030] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings. In this description, detailed explanations of well-known techniques that may obscure the essence of the invention are omitted.

[0031] Figure 1 This is an example of a defect inspection object, i.e., an inspected object, performed by a unit-to-unit comparison method according to an embodiment of the present invention.

[0032] Reference Figure 1 The defect inspection object, namely the workpiece 100, performed by the unit-to-unit comparison method according to an embodiment of the present invention has a structure in which unit cells 110, which can be represented by a two-dimensional vector 120, are arranged periodically. As shown in the figure, the workpiece 100 may have a structure in which unit cells 110 with the same micro-pattern are arranged in a matrix shape on a wafer. Examples of such workpieces 100 include semiconductor wafers, image sensors, and display panels.

[0033] The inspected item 100 is primarily a semiconductor device, inspected for defects during the manufacturing process or before shipment. The inspected item 100 includes micropatterns formed on a substrate or wafer through various processes such as doping, thermal oxidation, vapor deposition, etching, and exposure. Because the inspected item 100 features a repetitive arrangement of identical micropatterns, by comparing unit cells (where identical micropatterns are bundled as a unit) with adjacent unit cells, it is possible to detect any abnormalities or defects in the micropatterns within the inspected item 100.

[0034] When detecting whether there are abnormalities and defects in the micro-patterns within the test object 100, the present invention compares one unit within the test object 100 with multiple other unit units within the test object 100, excluding the one unit unit.

[0035] As shown in the figure, the object under inspection 100 includes unit cells 110 with identical micro-patterns, which are periodically arranged in a plane. Each unit cell 110 is represented by a two-dimensional vector having position values ​​on a first axis and a second axis perpendicular to the first axis. The first axis and the second axis are in the same plane.

[0036] Figure 2 This is a flowchart of a unit-to-unit comparison method according to an embodiment of the present invention. Figure 3 This is a diagram illustrating the principle of block generation in a unit-to-unit comparison method according to an embodiment of the present invention.

[0037] Reference Figure 2 The unit-to-unit comparison method of the present invention is used to inspect defects in an object of inspection that is periodically arranged with unit units represented by two-dimensional vectors, and includes the following steps: obtaining an image of the object of inspection (S100); generating a block (S200); determining the grayscale of multiple pixels in the block (S300); determining whether the multiple pixels in the block have defects (S400); and displaying the determination result (S500).

[0038] This invention involves moving a wafer through a process in a semiconductor device manufacturing apparatus. During this process, a camera from a defect inspection system photographs the wafer. The processing of the photographic results, or the steps of this invention, is automatically performed by the image processing hardware of the defect inspection system or by a processor executing a built-in program. Furthermore, the results obtained in each step of this invention are automatically stored for use in the next step. These considerations should be taken into account even if not explicitly stated below.

[0039] In step S100, an image of the object under inspection 100 is obtained. In one embodiment, the image of the object under inspection 100 can be obtained using the imaging device of the defect inspection system described later. Figure 3 As shown, the image of the object being inspected obtained in step S100 can be an actual image displayed to the outside, but when the basic unit constituting the image, namely the pixel, is regarded as forming a two-dimensional matrix, it can also be a collection of image data including the position information of each pixel and the brightness information at that pixel.

[0040] Reference Figure 3 The image of the inspected object 100 reflects the structure of the periodically arranged unit cells 110, which can be represented by a two-dimensional vector 120, i.e., the inspected object including an array of unit cells, and also reflects the defects within the inspected object.

[0041] In step S200, a set of pixels corresponding to the same position of multiple unit cells within the unit cell array is generated from the image obtained in step S100, namely a tile. Generating a tile refers to a classification operation that groups pixels corresponding to the same coordinate points of the two-dimensional vectors of each unit cell in order to detect defects, and to compare the unit cells with each other.

[0042] Among them, the pixels corresponding to the same coordinate points of the two-dimensional vectors of each unit contained in the image of the inspected object are the elements that constitute the block, and at the same time, they are the basic units for unit-to-unit comparison.

[0043] In one embodiment, the number of blocks generated in step S200 is the same as the number of pixels in the image that constitutes a unit. However, the number of blocks can be greater than or less than the number of pixels in the image that constitutes a unit, as needed.

[0044] Reference Figure 3 The illustrated diagram exaggerates the combination of pixels in the image of the inspected object corresponding to the same coordinate points of the two-dimensional vectors of multiple unit cells, i.e., block 130. In the illustrated embodiment, a block 130 includes 20 pixels separated from 20 unit cells.

[0045] In step S300, the grayscale of each of the multiple pixels within block 130 generated in step S200 is determined. That is, the multiple pixels within block 130 have positional and brightness information of the image of the object being inspected, and the grayscale is determined by the brightness information.

[0046] At this point, the intensity of light reflected or scattered at the corresponding position of the object being inspected is converted into an analog image signal of current. The analog image signal is then converted into a digital image signal by an ADC, thereby forming the brightness information of multiple pixels, i.e., grayscale.

[0047] In addition, since the size of the two-dimensional vector of a unit cell is not necessarily an integer multiple of the pixel, the gray level can be calculated by interpolation of adjacent pixels when determining the gray level of multiple pixels within a block.

[0048] It can be expected that the multiple pixels contained within a block 130 all contain the same image information of the same position within the unit cell of the inspected object. Therefore, it can be expected that the multiple pixels contained within a block 130 have the same grayscale. However, the grayscale of each of the multiple pixels contained within a block 130 depends not only on the surface shape of the corresponding position of the inspected object, but also on defects in the inspected object, differences in exposure during shooting, noise generated during the conversion into electrical signals, and noise generated by discoloration or grains unrelated to semiconductor yield.

[0049] In step S400, the presence or absence of defects in multiple pixels within a block is determined based on the grayscale values ​​of the multiple pixels within the block as determined in step S300.

[0050] The grayscale of the pixel corresponding to the defect area can differ significantly from the grayscale of other pixels within the block. Based on this, in step S400, the grayscale of the pixel currently being judged as defective is compared with the average grayscale of the pixels other than the pixel currently being judged as defective within the block by a specified threshold value to determine whether the pixel currently being judged as defective is defective.

[0051] More specifically, in step S400, the grayscale of the pixel currently being judged as defective among multiple pixels within the block, and the average and standard deviation of the grayscale of pixels other than the pixel currently being judged as defective among multiple pixels within the block, are used. Therefore, when determining whether a unit cell is defective, a unit cell and multiple unit cells with the same surface shape as the aforementioned unit cell can be compared with each other, thereby improving the reliability of defect detection. Furthermore, when determining whether a unit cell is defective, the standard deviation of the grayscale of pixels other than the pixel currently being judged as defective is used, thereby eliminating the noise inference operation used to exclude increases in grayscale reflecting noise.

[0052] Figure 4 This is a flowchart of a method for determining defects in a unit-to-unit comparison method according to an embodiment of the present invention.

[0053] Reference Figure 4 In an embodiment of the present invention, the step S400 of determining whether multiple pixels in a block are defective in the unit-to-unit comparison method includes the following steps: calculating the average value and standard deviation of the gray levels of pixels other than the currently defective determination target pixel (S410); calculating the absolute value of the difference between the gray level of the currently defective determination target pixel and the average value (S420); dividing the calculated absolute value by the standard deviation (S430); and comparing the value obtained by dividing the absolute value by the standard deviation with a specified threshold value (S440).

[0054] In step S410, the average grayscale value and standard deviation of the pixels in the block, excluding the pixel that is currently being judged to be defective, are calculated.

[0055] In step S420, the absolute value of the result is calculated by subtracting the average grayscale of all pixels in the block other than the currently defective pixel, obtained in step S420, from the grayscale of the pixel currently being judged for whether it is defective. This is because if the brightness of the currently defective pixel is significantly higher or lower than the brightness of other pixels in the same block, a defect is expected to exist.

[0056] In step S430, the absolute value calculated in step S420 is divided by the standard deviation of the gray levels of the pixels within the block, excluding the pixel currently being judged as having a defect, as determined in step S410. That is, the ratio of the degree to which the gray level of the pixel currently being judged as having a defect deviates from the average gray level is obtained relative to the standard deviation of the gray levels of the pixels within the block, excluding the pixel currently being judged as having a defect.

[0057] In step S440, the value obtained by dividing the absolute value in step S430 by the standard deviation is compared with a specified threshold. If the value obtained by dividing the absolute value by the standard deviation is above the specified threshold, the area of ​​the inspected object corresponding to the currently detected defect-determining pixel is determined to be defective; if the value obtained by dividing the absolute value by the standard deviation is less than the specified threshold, the area of ​​the inspected object corresponding to the currently detected defect-determining pixel is determined to be normal. That is, through the comparison in step S440, a defect is determined when the grayscale of the currently detected defect-determining pixel deviates significantly from the average grayscale, and a normal grayscale is determined when the deviation is not significant.

[0058] The specified threshold value can be assigned as a predefined value before running step S400 and stored in the storage medium for use in defect determination. Furthermore, the specified threshold value can be assigned different values ​​depending on the block.

[0059] Step S500 displays the results of the determination of whether multiple pixels have defects in step S400. Through this step, the operator can determine whether the inspected object has defects and the location of the defects.

[0060] Figure 5 This is a simplified structural diagram of a defect detection system for a unit-to-unit comparison method according to an embodiment of the present invention.

[0061] Reference Figure 5 A defect detection system for implementing the unit-to-unit comparison method according to an embodiment of the present invention includes an imaging device, an image processing device, and a display device. The defect detection system detects whether an object 101, which is periodically arranged with unit cells represented by two-dimensional vectors, has defects and the location of the defects, and displays the defect results.

[0062] The imaging device includes a light source 210, a platform 220, and a camera 230. Light emitted from the light source 210 illuminates a portion of the object 101 that moves with the platform 220. The camera 230 collects the light reflected or scattered from the object and performs detection at the focal plane to obtain an image of the object. That is, the imaging device performs the above-described step S100.

[0063] The image processing device includes an ADC 310, a data processing module 320, a judgment module 330, and a storage medium 340. The ADC 310 converts the intensity of light reflected or scattered at the corresponding position of the object under inspection into an analog image signal of current through the camera 230, and then into a digital image signal.

[0064] The data processing module 320 receives the digital image signal and uses the digital image signal to perform a portion of the above steps S200 to S400.

[0065] The data processing module 320 generates a set of pixels corresponding to the same position in multiple unit cells within the unit cell array of the image obtained from the imaging device, i.e., a tile. At this point, the data processing module 320 can then process the digital image signal in parallel for each of the multiple tiles.

[0066] The data processing module 320 determines the grayscale of each pixel within the generated block 130. That is, the pixels within block 130 contain positional and brightness information of the image of the object being inspected, and the grayscale is determined using this brightness information. Since the size of a two-dimensional vector of a unit cell is not necessarily an integer multiple of the pixel size, the data processing module 320 can calculate the grayscale of multiple pixels within the block by interpolating the grayscale of adjacent pixels.

[0067] In addition, the data processing module 320 can perform the following operations: calculate the average grayscale value and standard deviation of the pixels within the block other than the pixel currently being judged as having a defect; calculate the absolute value of the difference between the grayscale value of the pixel currently being judged as having a defect and the aforementioned average value; and divide the calculated absolute value by the standard deviation. That is, the data processing module 320 automatically executes the above steps S410 to S430.

[0068] In addition, the data processing module 320 sends the data generated during image processing to the storage medium 340, and sends the value obtained by dividing the absolute value calculated above by the standard deviation above to the judgment module 330.

[0069] The judgment module 330 can compare the value received from the data processing module 320 with a specified threshold value stored in the storage medium 340 to determine whether the inspected object has a defect and the location of the defect. That is, when the value received from the data processing module 320 is above the specified threshold value, the judgment module 330 will determine the area of ​​the inspected object corresponding to the currently detected defect judgment pixel as a defect, and when the value received from the data processing module 320 is less than the specified threshold value, the area of ​​the inspected object corresponding to the currently detected defect judgment pixel will be determined as normal.

[0070] In addition, the judgment module 300 sends the result of whether multiple pixels of the inspected object are defective to the display device 350, which displays information on whether the inspected object is defective and the location of the defect. Through the display device 350, the operator can know whether the inspected object is defective and the location of the defect.

[0071] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, those skilled in the art will understand that the present invention can be implemented in other specific forms without changing its technical concept or essential features. Therefore, it should be understood that the embodiments described above are merely illustrative and not restrictive in all respects.

Claims

1. A unit-to-unit comparison method for inspecting defects in an inspected object that is periodically arranged with unit units that can be represented by a two-dimensional vector, wherein, Includes the following steps: Obtain an image of the object under inspection containing an array of the unit cells; Blocks are generated by extracting the pixels that make up each unit of the image from the obtained image and grouping the extracted pixels into separate sets; The block is formed as a set of pixels extracted from the image of multiple unit units within the array at the corresponding same position relative to each unit unit; Determine the grayscale of each pixel within the block; and Based on the determined grayscale of each of the multiple pixels, it is determined whether the multiple pixels within the block have defects. In the step of determining whether multiple pixels within the block are defective, the defect determination is performed by comparing the degree to which the grayscale of the currently defective pixel deviates from the average grayscale of all pixels in the block other than the currently defective pixel with a specified threshold. In the step of determining whether multiple pixels within the block are defective, the grayscale of the currently defective pixel and the average and standard deviation of the grayscale of the pixels in the block other than the currently defective pixel are used.

2. The unit-to-unit comparison method according to claim 1, wherein, In the step of determining whether multiple pixels within the block are defective This includes the step of calculating the average grayscale value and standard deviation of the pixels in the block, excluding the pixel currently being judged as having a defect.

3. The unit-to-unit comparison method according to claim 2, wherein, In the step of determining whether multiple pixels within the block are defective It also includes the step of calculating the absolute value of the difference between the grayscale of the currently defective pixel in the block and the average value.

4. The cell-to-cell comparison method according to claim 3, wherein, In the step of determining whether multiple pixels within the block are defective It also includes the step of dividing the absolute value by the standard deviation.

5. The cell-to-cell comparison method according to claim 4, wherein, In the step of determining whether multiple pixels within the block are defective It also includes the step of comparing the value obtained by dividing the absolute value by the standard deviation with a specified critical value.

6. The cell-to-cell comparison method according to claim 1, wherein, This includes the step of displaying the determined defect results of the plurality of pixels.

7. The cell-to-cell comparison method according to claim 1, wherein, In the step of determining the grayscale of each pixel within the block The grayscale value being determined includes noise.

Images