JP2007184364A - Inspection device and method of pattern defect - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device and method which can carry out accurate inspection of pattern defects even with respect to a test piece created by changing exposure conditions such as an FEM sample. <P>SOLUTION: The defect inspection device inspects a test piece created by changing exposure conditions such as an FEM sample. The device includes a function which selects a specific die, then stores image data from which defect information other than defects from the die due to a change in exposure conditions has been removed inside the inspection equipment, and detects defects by comparing the image data with the die to be inspected. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an apparatus and method for inspecting pattern defects on the surface of semiconductor wafers, photomasks, magnetic disks, liquid crystal substrates and the like.

  As semiconductor devices are miniaturized, reticle manufacturing accuracy when transferring a pattern to a wafer and exposure condition margin of an exposure apparatus such as a stepper / scanner are reduced, and more severe condition setting is required. In general, the evaluation of the reticle and the confirmation of the margin for the amount of change in exposure conditions are performed by forming a pattern on the wafer by changing the focus and the amount of exposure, and monitoring the change in the shape of the pattern. A sample created for such a purpose is usually called a FEM (Focus Exposure Matrix) sample.

  The shape change is monitored by a length measuring SEM (Scanning Electron Microscope) or the like, and the pattern is normally formed from the information on the shape and line width of a specific point on the pattern, and a pattern having a desired line width is formed. The range of the exposed exposure conditions is confirmed. In the reticle evaluation, if a pattern is not formed with a desired line width, or if the exposure condition margin when creating with a desired line width is small, the reticle is remade. At the site where the product is actually manufactured, the above-described measurement is performed once a day to check the exposure conditions.

In recent years, with further miniaturization and higher integration, there are places that are easily exposed and difficult to be exposed due to the positional relationship of patterns within a single product. There is a problem that the pattern cannot be created. Therefore, in consideration of the influence of advance proximity, pattern correction such as to pre-thickening the line width of the exposed susceptible locations (OPC: O ptical P roximity C orrection) uses a reticle made Thus, a method of finally creating a pattern as designed is taken.

  However, since the pattern shape becomes more complicated after the OPC process, the exposure margin is further reduced, defects are likely to occur due to fluctuations in conditions, and the occurrence itself is difficult to predict. For this reason, the occurrence of a defect cannot be confirmed only by measuring the line width of a fixed point using a conventional length measuring SEM, making it difficult to evaluate the reticle and determine the exposure margin. Therefore, as described in Patent Document 1, using an inspection apparatus, a defect inspection is performed on the entire surface of the FEM sample, a defect generated by a change in exposure conditions is monitored, and an exposure condition in which no defect occurs is confirmed. Was devised.

JP 2001-194323 A

  However, in the conventional inspection apparatus as described in Patent Document 1, a pattern is formed in a memory mat of adjacent dies having the same type of pattern formed on the wafer or a memory device having the same repetitive pattern. An image is acquired, and the same pattern is compared and detected as a different pattern as a defect. In this method, as for the pattern created by continuously changing the exposure conditions as in the FEM sample, since the change of the exposure conditions is small in the adjacent pattern, the increase behavior of the defect due to the change of the exposure conditions cannot be confirmed. In other words, when inspecting a die in which the exposure conditions deviate significantly from the optimum values and the pattern is not formed, if there is no pattern formed on the die adjacent to this die as well, there is no difference in the pattern image comparison process. It cannot be detected, and an inspection result that the amount of defects increases as the expected exposure condition deviates from the optimum value is not obtained.

  Therefore, a method of using CAD data or the like and using the CAD data image as a reference image and detecting a different part as a defect while performing comparison processing with the actual pattern image is conceivable. Only shape information is included, and brightness gradation information cannot be obtained like a pattern image obtained from an inspection apparatus. In other words, it becomes a comparison in the binarized image between the part with the pattern and the part without the pattern, and the image of the part with the pattern is also subjected to the processing for enhancing the edge of the pattern by binarization processing etc. It is necessary to compare the images. However, a defect with a size of 1 pixel or less depends on the image acquisition conditions, but its size is represented by a change in gradation. If the defect size is small, the signal amount is small and the defect becomes dark. In comparison, it is difficult to detect a defect having a size of one pixel or less existing in a pattern image.

  In addition, the detected defects include not only defects caused by changes in exposure conditions but also defects caused by, for example, foreign matter, etc., so the work of separating defects caused by changes in exposure conditions from other defects However, the amount of work increases in proportion to the number of detected defects. Furthermore, since the detected defect changes depending on the detection capability of the inspection apparatus and the inspection conditions at the time of defect inspection, there is a possibility that the defect due to the change of the exposure condition to be inspected may be overlooked.

  The present invention has been made in view of such circumstances, and an apparatus and method capable of inspecting a pattern defect with high accuracy even for a sample prepared by changing exposure conditions such as an FEM sample. Is to provide.

  As a result of diligent research in view of the above problems, the inventor selected a specific die in a defect inspection apparatus for inspecting a specimen prepared by changing exposure conditions such as an FEM sample, The inventors have conceived of providing a function of storing image data from which defect information other than defects caused by fluctuations is removed in an inspection apparatus, and detecting defects by comparing the image data with the inspection target die.

  As described above, according to the pattern defect inspection apparatus and inspection method of the present invention, defects generated due to variations in exposure conditions can be detected with high sensitivity for samples prepared by changing the exposure conditions. Therefore, the inspection result can be verified accurately and simply, and the reticle can be evaluated and the optimum exposure condition of the exposure apparatus can be confirmed. Also, by providing inspection means using an appropriate reference image and a simple evaluation function for inspection results, a defect inspection apparatus that has been used only for defect inspection of semiconductor products in the past can be used for reticle evaluation and exposure condition confirmation. It can also be applied to applications. Further, it can be used for applications such as reticle evaluation and confirmation of the optimum exposure conditions of the exposure apparatus without the need for significant changes and improvements to the conventional defect inspection apparatus.

  The best mode for carrying out a pattern defect inspection apparatus and inspection method of the present invention will be described below in detail with reference to the accompanying drawings. 1 to 8 are diagrams illustrating embodiments of the present invention. In these drawings, the same reference numerals denote the same components, and the basic configuration and operation are the same. To do.

  FIG. 1 is a diagram showing a semiconductor production line to which a pattern defect inspection apparatus according to the present invention is applied. As shown in FIG. 1, the semiconductor production line is usually in a clean room 010 in which a normal environment is maintained. In the clean room 010, an exposure apparatus 001 for imaging a pattern on a resist-coated wafer, a length measurement SEM002 for confirming whether the exposed pattern is formed according to a design value, and a pattern defect from the appearance of the product wafer. A defect inspection apparatus 003 for performing detection and a review apparatus 004 for observing defects detected by the defect inspection apparatus 003 manually or automatically are installed. The exposure apparatus 001, the length measurement SEM002, the external appearance apparatus 003, and the review apparatus 004 can transmit and receive various data to and from the data processing apparatus 006 connected via the communication line 005. Based on the CAD data after the OPC process of the data processing device 006, the exposure device 001 adjusts the exposure conditions using the focus and exposure amount as parameters to form a predetermined pattern on the wafer, whereby the FEM wafer 099 is Created.

  FIG. 2 is a diagram showing a measurement example of the FEM wafer 099 by the length measurement SEM002 shown in FIG. In the exposure apparatus 001, a pattern is created on the FEM wafer 099, in which the exposure amount is changed in units of 3 mJ on the horizontal axis and the focus value is changed in units of 0.05 μm on the vertical axis. As shown in FIG. 2, on the FEM wafer 099, a die 201 in which a pattern is normally created and a die 202 in which a pattern is not normally created are formed depending on exposure conditions. The created FEM wafer 099 is conveyed to the length measurement SEM002, where the length measurement of a predetermined portion of the pattern on the wafer is performed, and the length measurement result 032 is output. The length measurement result 032 is managed by the data processing device 006 together with data such as a lot number, wafer number, inspection date and time.

  Based on the length measurement result 032, a pivotal curve 210 indicating the relationship between the line width and the focus value at each exposure amount is obtained. For example, when it is desired to create a pattern with a line width of 130 nm from the pivotal curve 210, if the focus value is in the range of −0.05 μm to 0.05 μm at an exposure amount of 12 mJ, such a pattern is formed using this reticle. It can be determined that it can be created. If a pattern with a line width of 90 nm is to be created, it is possible to create such a pattern using this reticle under the condition that the exposure value is 18 mJ and the focus value is 0 μm. Since the line width changes greatly, the tolerance of focus fluctuation is narrow, and it can be determined that it is difficult to manufacture a stable product.

  After the length measurement in the length measurement SEM002, the FEM wafer 099 is transported to the defect inspection apparatus 003 for defect inspection.

  FIG. 3 is a diagram showing a configuration example of the defect inspection apparatus 003 shown in FIG. In FIG. 3, the defect inspection apparatus 003 includes an illumination light source 901, a stage 902, a beam expander 903, a coherence reduction optical system 904, a beam splitter 905, a polarization element group 906, an objective lens 907, and an image sensor 908. Yes.

  The illumination light source 901 is a light source composed of, for example, a UV laser light source having a wavelength of 266 nm or a wavelength of 355 nm. The stage 902 is composed of a stage that is movable in the X, Y, Z and θ (rotation) directions, and on which an FEM wafer 099 is placed as a pattern to be inspected. The beam expander 903 expands the UV laser illumination light from the illumination light source 901 to a predetermined size. The coherence reduction optical system 904 reduces the coherence of the UV laser illumination light or the like from the illumination light source 901. The beam splitter 905 is configured by a polarization beam splitter or the like, and reflects the UV laser illumination light or the like from the UV illumination light source 901 and directs it toward the polarization element group 906 and the objective lens 907. The polarizing element group 906 includes UV laser illumination light and the like so that the reflected light from the FEM wafer 099 does not cause uneven brightness when reaching the image sensor 908 due to the pattern shape or density difference of the FEM wafer 099. Control the polarization direction and polarization ratio of reflected light. The objective lens 907 collects UV laser illumination light or the like from the illumination light source 901 on the FEM wafer 099, and it is preferable to use a reflective objective lens in order to reduce the influence of chromatic aberration of UV light and DUV light. . The image sensor 908 has a pixel size of about 0.05 μm to 0.3 μm in terms of the sample, and outputs a grayscale image signal corresponding to the brightness (lightness) of reflected light from the FEM wafer 099 to be inspected. To do.

  With the above configuration, the UV laser illumination light L1 emitted from the illumination light source 901 is expanded by the beam expander 903, and passes through the coherence reduction optical system 904, the lens 922, the beam splitter 905, and the deflection element group 906. The light enters the objective lens 907 and irradiates the FEM wafer 099. Reflected light from the FEM wafer 099 enters the image sensor 908 via the objective lens 907, the polarizing element group 906, the beam splitter 905, and the imaging lens 923 from vertically above the FEM wafer 099. That is, while scanning the stage 902 and moving the FEM wafer 099 to be inspected at a constant speed, the image sensor 908 increases brightness information (grayscale image signal) of the pattern to be inspected formed on the FEM wafer 099. It can be detected with accuracy. The grayscale image acquired by the image sensor 908 is subjected to pattern comparison inspection processing in the image signal processing circuit 932. The pattern comparison inspection process is performed on the basis of wafer information such as die arrangement information and die size previously held in the operation / control unit 931.

  On the other hand, a spatial image of the pupil plane of the objective lens 907 can be detected by the mirror 928, the lens 923, and the observation camera 930 provided in the optical path between the beam splitter 905 and the image sensor 908. With this observation camera 930, the defect detected on the FEM wafer 099 can be observed, or the alignment of the FEM wafer 099 on the stage 902 can be confirmed.

  FIG. 4 is a flowchart showing a flow of defect inspection processing in the defect inspection apparatus 003 shown in FIGS. In FIG. 4, first, a reference die on the FEM wafer 099 is selected (step S01). At this time, a wafer map as shown in FIG. 2 is displayed on a monitor screen (not shown), and the user can designate any die on the screen as the reference die 501 by the operation / control unit 931. And Information such as a wafer map and a die size is acquired in advance by the observation camera 930. Here, in order to avoid the trouble of the subsequent confirmation work, it is preferable to use a die that is normally created as much as possible as a reference die by confirming an exposure margin range determined to be optimum in the past measurement. Further, the reference die may be determined with reference to a history such as a pivotal curve obtained when the FEM wafer 099 is measured by the length measuring SEM002.

  Subsequently, a defect such as a foreign substance existing in the reference die is detected and removed. The defect detection here is based on the prior art, and the inspection conditions such as the threshold and the image overlay when performing the inspection are the same as in the conventional case. Hereinafter, a procedure for detecting and removing defects in the reference die will be described with reference to a specific example shown in FIG. First, an image of a reference die and its left and right dies are acquired (step S02), and a difference image is obtained by calculating an absolute value of a gradation difference from the image of the reference die and its left and right dies. Generate (step S03). In the example shown in FIG. 5, a difference image 404 is generated from the image 401 of the reference die 501 and the image 402 of the adjacent die on the left, and the difference image 404 is generated from the image 401 of the reference die 501 and the image 403 of the adjacent die on the right. 405 is generated. Here, as shown in the drawing, it is assumed that a defect 411, a defect 412, and a defect 413 exist in the reference die image 401, the left adjacent die image 402, and the right adjacent die image 403, respectively. . As a result, the generated difference image 404 includes a region 414 corresponding to the defect 411 on the reference die and a region 415 corresponding to the defect 412 on the left adjacent die. Similarly, the generated difference image 405 includes a region 414 corresponding to the defect 411 on the reference die and a region 416 corresponding to the defect 413 on the right adjacent die.

  Subsequently, in each difference image, a portion having a gradation value equal to or greater than a predetermined threshold is determined as a defect candidate in the reference die (step S04). For example, in FIG. 5, the gradation value of each pixel of the difference image 405 is as indicated by 417, a 2 × 2 pixel region 414 having a gradation value of 40, and 1 × 1 having a gradation value of 50. A pixel region 416. In this case, if a value larger than 10 is set as the defect detection threshold value, the region 414 and the region 415 become defect candidates. Thereafter, defect candidates having the same in-die coordinates in the difference image 404 and the difference image 405 are detected as defects on the reference die (step S05). In the example shown in FIG. 5, the region 414 is detected as a defect on the reference die as indicated by 406. Although the number of defects on the reference die to be detected changes depending on the above threshold, the number of detected defects and the positions of these defects in the die are monitored as the threshold changes. It is desirable to be able to check on the screen. At this stage, the FEM wafer 099 is transported to the review device 004, and the coordinate information of the detected defect is output to the review device 004 to confirm the defect, thereby correcting the defect area thereafter. Can be done well. At this time, it is preferable to classify the defects by the review device 004 and perform the above correction only for the defects classified into a specific category.

  Subsequently, the average value of the gradation of the area 414 detected as a defect on the reference die in 406 is calculated, and the gradation value of each pixel in the area 414 is averaged (step S06). The gradation value of each pixel is as indicated by 408. Here, when the average value of the gradation of the area 414 detected as a defect on the reference die is brighter than the image 401 of the reference die, that is, the gradation of the image obtained by subtracting the image 402 or the image 403 from the image 401 of the reference die. Is a negative value, the average value of the gradation of the area 414 is added to the area 414 in 407 in which each pixel of the image 401 of the reference die is represented by gradation. On the other hand, when the average value of the gradation of the area 414 is darker than the image 401 of the reference die, that is, when the gradation of the image obtained by subtracting the image 402 or the image 403 from the reference die image 401 becomes a positive value, The average value of the gradation in the area 414 is subtracted from the area 414 in 407 in which each pixel of the image 401 of the reference die is represented by gradation (step S07). By these processes, the gradation value 409 of each pixel of the image excluding the defect portion other than the defect caused by the change of the exposure condition is obtained. An image 410 obtained by correcting the defect area from the image 401 of the reference die 501 thus obtained is stored in the defect inspection apparatus 003 (step S08). Alternatively, the image 410 may be output to the data processing device 006.

  Next, a comparison inspection with other dies is performed using an image 410 obtained by correcting a defect area from the image 401 of the reference die 501. First, an image of a die in the inspection range is acquired (step S09), and a difference image between each die image and the image 410 is created (step S10). As a result, defect coordinate data 033 is generated (step S11). Here, although the inspection range is basically the entire surface of the wafer, it is assumed that the inspection range can be specified in the same manner as a conventional inspection apparatus. Furthermore, the inspection time is shortened by acquiring information previously inspected by the length measuring SEM002 from the length measuring SEM002 or the data processing device 006 via the communication line 005, and excluding dies having no pattern or dies having no pattern. it can. Further, the threshold setting method, other inspection parameters, and the like can perform high-precision inspection by following the functions of the conventional inspection apparatus. Also, if the exposure conditions differ greatly from the optimum value, the number of detected defects will be enormous. However, if the maximum number of detected defects per die is set in advance and a defect exceeding the set value is detected, the corresponding die will be inspected. By skipping, it is possible to prevent a decrease in inspection speed due to detection of a large number of defects. In this case, it is desirable that the die whose inspection has been skipped is displayed by distinguishing the result from other dies by color coding or the like on the display screen. The defect coordinate data is output to the length measurement SEM002 or review device 004 via the communication line 005.

  Based on the defect coordinate data 033 obtained above, a wafer map indicating the defect position on the wafer is displayed as shown in FIG. 6 (step S12). It is preferable to automatically assign an ID number to the detected defect and display a list of ID numbers, detected die coordinates, in-die coordinates, defect size, and the like. The user designates a specific defect on the defect map 500 using the control / operation unit 930 and confirms the defect by moving the stage 902 so that the defect comes to the center of the visual field of the observation camera 930. be able to. The designation of the defect can be performed not only by designating the defect on the displayed defect map 500 but also by directly inputting the point of the defect ID portion displayed in a list or coordinate information.

  Next, it is confirmed by the inspection apparatus 003 or the review apparatus 004 whether the detected defect is a defect caused by a change in exposure conditions. Here, the defect confirmation means (1) the distinction between a defect caused by a change in exposure conditions and a defect caused by a reticle defect or a suddenly generated defect such as attached foreign matter, and (2) the detected defect coordinates. The main purpose is to confirm whether or not there is actually a defect, and (3) to confirm whether or not the die on which no defect has been detected actually has a defect.

  First, a distinction is made between a defect that has occurred due to a change in exposure conditions and a defect that has occurred suddenly, such as a defect caused by a defective reticle or an attached foreign substance (step S13). In the defect map shown in FIG. 6, regarding the distinction between the defect caused by the change of the exposure condition and the defect caused by the reticle defect or the suddenly generated defect such as attached foreign matter, the defect caused by the change of the exposure condition is: Note that it occurs in a specific pattern whose shape is likely to change due to changes in exposure conditions. In addition, since many semiconductor products have the same pattern repetition or inversion structure, it is possible to predict that a defect occurring at the same coordinate in each die is a defect caused by a change in exposure conditions. Therefore, a margin of coordinates and the number of detections are set for the coordinates of the detected defect, and the defect group 504 detected at the same coordinate in the die can be determined as a defect caused by a change in exposure conditions. Defects extracted under these conditions include defects (repeat defects) generated due to defects in the reticle itself, but repeat defects have substantially the same size and signal amount of defects generated at the same coordinates in the die. On the other hand, since the size and signal amount of the defect caused by the change of the exposure condition change with the change of the exposure condition, in addition to the coordinate margin and the detected number information for the coordinate information in the die, the size and By using information such as the signal amount, the two can be distinguished.

  FIG. 7 is a diagram showing an example of cell areas and defects in a die. The same coordinate 611 on the die 612 with respect to the defect 609 on the die 600 is detected when the defect is detected at the same coordinate position 611 as the defect 609 from the origin 601 of the die 600 and the signal amount is different. It can be determined that the same defect group 504 has occurred. Some semiconductor products have a plurality of cell areas such as memory cell areas 606 and 607 of the same size such as DRAM (Dynamic Random Access Memory) in the die 600. If a defect 609 caused by a change in exposure conditions exists in the cell area 606, the same coordinate 610 from the cell area origin of another cell area 607 has the same pattern, so that a defect may occur under the same exposure condition. It is done. Therefore, the defect 609 generated by the change in the exposure condition from the die origin 601 of the die 600 is replaced with the coordinate from the cell area origin 602, and the defect 610 located in the same cell coordinate position from the cell area origin 603 is made the same defect group. This can reduce the review work. The same coordinate position from each origin of the cell area with respect to the reference defect is the coordinate of the defect 609 on the coordinate axis with the die origin 601 as the origin (Xd1, Yd1), and the coordinate of the cell area origin 602 where the defect 609 exists (Xb1 , Yb1), assuming that the coordinates of the other cell area origin 603 existing in the same die are (Xb2, Yb2), the coordinates (Xd2, Yd2) of the defect 610 are expressed by [Equation 6-1].

  (Xd2, Yd2) = (Xd1-Xb1 + Xb2, Yd1-Yb1 + Yb2) − [Formula 6-1]

  FIG. 8 is a diagram showing an example of defect coordinates existing at symmetrical positions in the cell area. Because semiconductor products often have patterns of the same shape at repeated, inverted, and symmetrical positions, defects that easily occur due to changes in exposure conditions by making one that satisfies one or more of these conditions the same defect The position on the pattern can be grasped. On the coordinate axis with the die origin 601 as the origin, the coordinates of the defect 609 are (Xd1, Yd1), the coordinates of the cell area origin 602 where the defect 609 exists (Xb1, Yb1), and the coordinates of the cell area upper right end point 700 (Xe1, Ye1) If the center of the cell area 606 is the origin 701, the coordinates (Xc1, Yc1) of the origin, the coordinates (a, b) of the symmetric position coordinate 704 of the X coordinate axis 702 passing through the origin 701, and the symmetric position coordinates of the Y coordinate axis 703 The coordinates (c, d) of 705 are represented by [Expression 7-1], [Expression 7-2], and [Expression 7-3], respectively.

(Xc1, Yc1) = {(Xe1-Xb1) / 2, (Ye1-Yb1) / 2} − [Formula 7-1]
(A, b) = {Xd1, 2 (Yc1) -Yd1} − [Formula 7-2]
(C, d) = {2 (Xc1) -Xd1, Yd1}-[Formula 7-3]

  Next, the defect is visually confirmed (step S14). In step 13, the defect group having the appearance feature attributed to the exposure condition is extracted. However, there is a possibility that the generated defect is minute and cannot be detected depending on the inspection condition under the illumination condition where the defect starts to occur. Therefore, regarding the confirmation of whether or not a die in which no defect has been detected actually has a defect, the same coordinate position 505 of another die is extracted from the designated defect group 504, and the same coordinates as the defect group 504 are extracted on the defect map 500. It is displayed separately from the defect 506 which is not above. Further, in addition to displaying information such as a defect number and defect coordinates for each defect as in the conventional inspection apparatus, information such as a defect group number and defect coordinates is displayed for each defect group. In addition, when actually observing defects, it is necessary to move the stage to the defect position. However, as in the case of the conventional inspection apparatus, clicking the defect on the defect MAP 500, specifying the defect number by numerical input, etc. It shall be possible to move to a position.

  If the coordinates of the defect and the defect candidate are added by the above means, the time labor required for the review becomes enormous. Therefore, the function of adding the same coordinate position 505 to the defect coordinate data 033 is provided, the defect group 504 and the image of the same coordinate position 505 are acquired on the inspection apparatus 003, and the defect image list 510 is created. The review performed by moving to each coordinate can confirm the defect in the same die coordinate at a time only by specifying one coordinate, and can easily confirm the area 511 in which no defect has occurred. . When there are several types of defect groups 504, a defect image list 510 is stored for each defect group, and an area 511 in which no defect has occurred can be displayed in an overlay manner. Further, the inspection result 034 in which the coordinate position 505 is added to the defect coordinate data 033 is output, and the automatic review / automatic classification is performed by the review device 004 having a higher resolution than the inspection device 003, so that it is high for simple and fine defects. Can be confirmed with accuracy. The review result 035 can be confirmed by the defect inspection apparatus 003 and further sent to the data processing apparatus 006 so that it can be confirmed outside the clean room 010. Further, even if the reference die 501 includes a defect generated by a change in exposure conditions, it can be easily confirmed in the same manner as described above. In this case, since the normal part becomes a defect, the defect coordinate can be deleted from the defect coordinate data 033.

  Next, the inspection result is fed back. First, a die range where a defect has occurred due to a change in exposure conditions is registered in the inspection apparatus 003 together with defect information 033 (step S15). The defect information 033 is managed by the data processing device 006 together with data such as a lot number, wafer number, inspection date and time. The length measurement information 032 from the length measurement SEM is compared with the defect information 033, and a pattern is created with a desired line width, and an optimum exposure condition range 034 in which no defect is generated due to a change in exposure conditions. Is determined (step S16). In the conventional defect inspection apparatus, since the observation magnification is lower than that of the length measurement SEM, it is impossible to measure the width of the pattern that changes due to the change of the exposure condition with high reliability. If this observation is performed or other length measuring means is used, the exposure conditions can be evaluated only by the defect inspection apparatus of the present invention.

  Based on the information of the optimum exposure condition range 034 obtained in this way, the reticle is evaluated or the semiconductor device is actually manufactured by the exposure apparatus 001 (step S17). The length measurement information 033, the defect information 032, the optimum exposure condition range 034, and the review information 035 are managed by the data processing device 006 together with data such as a lot number, a wafer number, and an inspection date. When exposing an actual product, merge information 034 of optimum exposure conditions is sent from the data processing device 006 to the exposure device 001 using information on the product to be exposed, that is, lot number, wafer number, inspection date and time as key information. Shall be output.

  As mentioned above, although the specific embodiment was shown and demonstrated about the inspection apparatus and inspection method of the pattern defect of this invention, this invention is not limited to these. A person skilled in the art can make various changes and improvements to the configurations and functions of the invention according to the above-described embodiments or other embodiments without departing from the gist of the present invention.

It is a figure which shows the semiconductor manufacturing line to which the inspection apparatus of the pattern defect of this invention was applied. It is a figure which shows the example of a measurement of FEM wafer by length measurement SEM shown in FIG. It is a figure which shows the structural example of the defect inspection apparatus shown in FIG. It is a flowchart which shows the flow of the defect inspection process in the defect inspection apparatus shown in FIG.1 and FIG.3. It is a figure which shows the specific example of the detection and removal of the defect in the reference | standard die | dye performed in a defect inspection process. It is a figure which shows the example of a display of the wafer map which shows the defect position on a wafer. It is a figure which shows the cell area in a die | dye, and the example of a defect. It is a figure which shows the example of the defect coordinate which exists in the symmetrical position in a cell area.

Explanation of symbols

001 Exposure equipment 002 Measuring SEM
003 Defect inspection device 004 Review device 005 Communication line 006 Data processing device 010 Clean room 032 Measurement information 033 Defect information 034 Exposure condition range 035 Review information 099 FEM wafer 201 Die with normal pattern created 202 Pattern produced normally Not die 210 Pivotal curve 501 Reference die 901 Illumination light source 902 Stage 903 Beam expander 904 Coherence reduction optical system 905 Beam splitter 906 Deflection element group 907 Objective lens 908 Image sensor 922 Lens 923 Imaging lens 930 Observation camera 931 Control / operation 932 Image signal processing circuit

Claims (6)

A defect inspection apparatus for inspecting a plurality of patterns formed on a semiconductor substrate,
Means for comparing a predetermined reference image and a plurality of patterns formed on the semiconductor substrate by continuously changing exposure conditions using the same reticle, and detecting a mismatched portion as a defect;
A defect inspection apparatus comprising: means for calculating information for adjusting the exposure condition based on an increase amount of the defect accompanying a change in the condition for forming the pattern.   The defect detection apparatus according to claim 1, wherein the reference image is an image that has been corrected to exclude defects due to changes in the exposure conditions.   With respect to the detected defect, it is determined whether the defect is a defect caused by a reticle defect or a defect caused by a change in the exposure condition based on a margin of coordinates, the detected number, size, and signal amount. The defect inspection apparatus according to claim 1, further comprising means for performing the operation. A defect inspection method for inspecting a plurality of patterns formed on a semiconductor substrate,
By comparing a predetermined reference image and a plurality of patterns formed on the semiconductor substrate by continuously changing the exposure conditions using the same reticle, and detecting a mismatch point as a defect,
A defect inspection method, wherein information for adjusting the exposure condition is calculated based on an increase amount of the defect accompanying a change in the condition for forming the pattern.   The defect detection method according to claim 4, wherein the reference image is an image that has been corrected to exclude defects caused by changes in the exposure conditions.   With respect to the detected defect, it is determined whether the defect is a defect caused by a reticle defect or a defect caused by a change in the exposure condition based on a margin of coordinates, the detected number, size, and signal amount. The defect inspection method according to claim 4, wherein the defect inspection method is performed.

Images