JP2014099382A - Charged particle beam device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform efficient defect detection by creating an optimum-sized inspection template because in the prior art, in the case where a defect is detected by comparing a template image and an inspection subject image, if the template image is small, an effective inspection region will become small due to reduction in overlapping parts of the inspection subject image and the template because of, for example, imaging position deviation, and reversely if the template image is too large, the calculation time will become long by taking time for alignment processing of the template image and the inspection subject image.SOLUTION: A charged particle beam device synthesizes multiple images including regions that should have the same pattern at each of multiple dies, generates a template image with a size encompassing all the regions included in the images used for synthesis, and uses this template image and an inspection subject image of the pattern to detect a defect.

Description

  The present invention relates to a charged particle beam apparatus provided with defect inspection means, and more particularly to a defect inspection apparatus and review apparatus using an electron beam.

  A semiconductor device such as a memory or a microcomputer used in a computer or the like is manufactured by repeating a process of transferring a pattern such as a circuit formed on a photomask by exposure processing, lithography processing, etching processing, or the like. In the manufacturing process of a semiconductor device, the presence or absence of defects such as the quality of lithography processing, etching processing, other processing results, and the generation of foreign matter greatly affects the yield of semiconductor devices. Therefore, in order to improve the yield, the pattern on the semiconductor wafer is inspected at the end of each manufacturing process, and the occurrence of abnormality or defect is detected early or in advance.

  The inspection apparatus used in the process as described above is required to perform high-throughput and high-precision inspection following the wafer diameter increase and circuit pattern miniaturization. In order to inspect a large number of areas in a short time while maintaining inspection accuracy, one of the semiconductor device inspection methods is to specify the area on the wafer that has been specified in advance, and to perform multiple inspections at a time. There is a method of taking an image and inspecting the image. For example, a patent shown in Patent Document 1 uses a template image obtained based on an image of a sample to detect a defect candidate by comparing an inspection image with a comparative image.

JP 2011-158439 A

  In recent years, with the miniaturization of semiconductor circuits, application examples of fixed point inspection in which a region is specified and detection sensitivity is increased are increasing. In fixed point inspection, imaging is performed after the movement of the stage is completed and stopped, so that a noise component accompanying the movement of the stage is reduced and a clearer image can be obtained. On the other hand, distortion due to stage position accuracy and thermal expansion of the wafer support, charged particle beam distortion due to sample charging, wafer position distortion, etc. occur, and fixed points for comparison with the template image. In the case of inspection, there is a problem that the effective inspection area decreases if the deviation from the template image is large. Also, if the template is made too large to avoid this, there is a problem that it takes a calculation time for the alignment processing and the like.

  Patent Document 1 proposes a method for obtaining a template by superimposing images, but uses a large image picked up by continuously moving the stage, and cuts out and superimposes a part having a characteristic necessary for alignment. Yes. However, the difference in imaging position when the stage is stopped due to machine differences such as stage position accuracy and charging unevenness that varies from wafer to wafer is not taken into consideration, and a template having an optimum size cannot be obtained.

  Accordingly, an object of the present invention is to provide an apparatus that generates a template image having an optimum size for use in fixed point inspection and performs inspection at high speed.

  In order to solve the above-described problem, the present invention captures a plurality of images including a region that should have the same pattern in each of a plurality of dies, combines the images, and includes all of the images used for the combination. A template image having a size including the region is generated, and a defective portion is detected by comparing the template image with the same portion on each acquired image. The images used for the composition are shifted from each other by a positional deviation of a size depending on the position of the inspection area, and a plurality of images are superimposed while the positional deviation is maintained to synthesize a template image. And

  According to the present invention, even if there is stage accuracy or wafer variation, it is possible to generate a template image having a size that does not easily reduce the effective inspection area and does not increase the calculation load, so that inspection can be performed at high speed.

  Problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.

The whole block diagram of a charged particle beam apparatus. FIG. 3 is a functional block diagram of an image processing unit according to the present embodiment. The process flowchart of a present Example. Operation screen used to input measurement points on the wafer. Explanatory drawing of template image and mask area | region creation. The flowchart of a defect detection process. Explanatory drawing of alignment. Explanatory drawing of an output screen. Explanatory drawing of a recipe edit screen.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  In this embodiment, as an example of a charged particle beam apparatus, a defect inspection apparatus for detecting defects on the surface of a semiconductor wafer on which a plurality of dies are formed and an embodiment of defect inspection will be described. This is merely an example of the present invention. However, the present invention is not limited to the embodiments described below. For example, the defect inspection apparatus referred to here widely refers to a device that acquires an image and determines the presence or absence of a defect from the image. As an example of the charged particle beam apparatus, an inspection apparatus using a scanning electron microscope, a review apparatus, and a pattern measurement apparatus can be given. Some charged particle beam devices such as review devices are equipped with defect inspection means such as means for inspecting a predetermined position on a sample for fixed-point observation and process monitoring. Of course, the present invention can also be applied to a line apparatus, and the defect inspection apparatus described below includes these.

  Further, in this specification, “defects” are not limited to pattern defects, but include a wide range of observation objects such as foreign matters, pattern dimension abnormalities, and structural defects.

  In the first embodiment, a defect inspection apparatus that extracts defect candidates from a comparative image corresponding to an image to be inspected and displays defect candidates included in a designated area as a defect will be described. Hereinafter, the comparison image corresponding to the inspection image is referred to as a template image or a reference image. In addition, the “corresponding portion” referred to here is a region having a pattern to be compared with the inspection image, in other words, a region that should have the same pattern as the inspection region. For example, a region that should have the same pattern in other dies formed on the wafer.

  FIG. 1 is a schematic diagram illustrating an example of the overall configuration of a defect inspection apparatus. The defect inspection apparatus includes a charged particle optical system, a sample chamber, a control unit, and the like. The charged particle optical system includes an electron gun 101 that emits an electron beam 107, a lens 102 that converges the electron beam 107, a deflector 103 that controls deflection of the electron beam 107, an objective lens 104 that converges the electron beam 107, and an electron beam 107 as a sample. It consists of a secondary electron detector 122 and a reflected electron detector 123 that detect secondary electrons 108 and reflected electrons 109 generated by irradiating 105. The defect inspection apparatus includes a scanning electron microscope (SEM) including the charged particle optical system, a sample stage 106 on which the sample 105 is placed, a moving stage 124 that moves the sample stage 106, and the like. The backscattered electron detector 123 is installed at a position facing each other on a straight line in order to capture a dual shadow image. These are then placed in a column (not shown) and can be maintained in a vacuum by a vacuum pump (not shown). The apparatus configuration is not limited to this, and may include a lens, a diaphragm, a detector, and the like other than those described above, and the arrangement may be different.

  The electron beam 107 radiated from the electron gun 101 is converged by the lens 102, scanned and deflected two-dimensionally by the deflector 103, converged by the objective lens 104, and irradiated onto the sample 105. When the sample 105 is irradiated with the electron beam 107, secondary electrons 108 and reflected electrons 109 corresponding to the shape and material of the sample 105 are generated. The secondary electrons 108 and the reflected electrons 109 are detected by the secondary electron detector 122 or the reflected electron detector 123, amplified by an amplifier (not shown), and then an analog / digital (A / D) converter 113. Is converted to a digital value. The signal from the backscattered electron detector 123 is used to form an L image and an R image that are backscattered electron images, and the signal from the secondary electron detector 122 is used to form an S image that is a secondary electron image. . Hereinafter, signals obtained from samples such as secondary electrons and reflected electrons are collectively referred to as secondary charged particles. Unless otherwise specified, image processing may be performed using an L image, an R image, an S image, or a composite image thereof, and these are collectively referred to as an image in the present specification.

  The data converted into the digital value is stored in the image memory 115. At this time, the address control circuit 114 generates an address synchronized with the scanning signal of the electron beam 107 as the address of the image data stored in the image memory 115. Further, the image memory 115 transfers the stored image data to the image processing unit 119 as needed.

  The image processing unit 119 sends the sent image data to the display unit 117 such as a display via the control unit 118 and performs arithmetic processing based on the image data to perform processing such as defect extraction. The defect extraction (detection) process here is performed by comparing and calculating the sent image data and other image data obtained from the pattern corresponding to the image data. The image processing unit 119 will be described later.

  The lens 102, the deflector 103, and the objective lens 104 are controlled by control signals from the lens control circuit 110, the deflection control circuit 111, and the objective lens control circuit 112, respectively, and the focal position and deflection amount of the electron beam 107 are controlled. Thereby, it can adjust so that the electron beam 107 may be irradiated to an appropriate position with respect to the sample 105. The moving stage 124 on which the sample stage 106 is placed can be translated in two dimensions by a control signal from the mechanism control circuit 116. For this reason, the sample 105 held by the sample stage 106 can also be translated in two dimensions, whereby the position at which the electron beam 107 is scanned with respect to the sample 105 can be controlled. The lens control circuit 110, the deflection control circuit 111, the objective lens control circuit 112, and the mechanism control circuit 116 are all controlled by signals from the control unit 118.

  An input unit 120 including a keyboard and a mouse is used for GUI (Graphical User Interface) operations such as device operations and parameter settings displayed on a display unit 117 including a display. The external input / output unit 121 is used for exchanging electronic files between an external storage device such as an HDD or a USB memory and this apparatus. Also, an input / output port of a communication network such as a LAN can be used.

  The image processing unit 119 is a part that performs template image generation, defect candidate detection, and defect determination, which will be described below, and each of these functional blocks is configured by combining arithmetic processing circuits that execute processing of each unit (so-called hardware implementation). Alternatively, a memory that stores a program corresponding to the processing of each unit is provided in the image processing unit 119, and the program is executed by a processor that is also provided in the image processing unit 119. The functional blocks shown may be implemented.

  Each functional block diagram included in the image processing unit 119 will be described with reference to FIG.

  At the time of template generation, a plurality of images are input to the image alignment function unit in the image processing unit 119, and coordinates that match the position are output to the average / noise removal function unit for each image. As the alignment, for example, the normalized correlation value may be obtained at each location while shifting the coordinates of the input image, and the portion having the highest value may be regarded as the location where the location was.

  The average / noise removal function unit performs an averaging process or a noise removal process for each overlapping pixel based on the coordinates where the positions of the input images match. Averaging is performed for each pixel, and the averaged image data is output to the template image generation function.

  The template image generation function unit stores the input image data in a recipe (inspection condition setting file).

  At the time of defect detection, the recipe and the inspection target image are input to the defect candidate detection function unit, the image alignment function unit aligns the template image and the inspection target image, obtains a difference for each pixel, and is determined by the recipe. A portion exceeding the threshold value becomes a defect candidate and is output to the defect feature amount calculation function unit.

  The defect feature amount calculation function unit obtains a feature amount that is an amount that defines the feature of the defect such as a size, an average pixel value, and a shape for each detected defect candidate, and outputs the feature amount to the defect determination function.

  A defect candidate, a feature amount of the defect, and a recipe are input to the defect determination function, and a defect candidate that meets the feature amount condition set in the recipe is finally determined as a defect. The determined defect is displayed on the display unit 117 including a display or the like.

  The configuration of the system including the image processing unit is not limited to this, and some or all of the devices constituting the system may be a common device.

  The processing in the system can be realized by either hardware or software. When configured by hardware, it can be realized by integrating a plurality of arithmetic units for executing processing on a wiring board or in a semiconductor chip or package. When configured by software, a program for executing desired arithmetic processing by a central processing unit (CPU) mounted on a device constituting the system or a general-purpose CPU mounted on a general-purpose computer connected to the system. It can be realized by executing. Also, some functional blocks may be realized by a dedicated processing circuit, and the remaining functional blocks may be realized by software using a program and a processor. It is also possible to upgrade an existing apparatus with a recording medium in which this program is recorded.

  FIG. 3 shows a flowchart of template image generation in the present embodiment. First, at step 301, a location for acquiring an image on the wafer is designated. A plurality of dies are formed on the wafer, and the die reference coordinates indicating the reference position of each die and the in-die coordinates that are relative coordinates indicating the position of the pattern or region in the die from the die reference coordinates. All the coordinates above are defined. Therefore, if the coordinates in the same die are the same, a location having the same pattern in different dies is represented.

  Determining the presence or absence of defects using multiple in-die or one representative die image using the same in-die coordinate image is called fixed point inspection, and also observing the workmanship of wiring width, hole diameter, etc. Call it.

  The means for instructing the image acquisition location may input coordinates directly from the input device 120, or it is equipped with a GUI for input assistance so that a chip can be selected and a plurality of locations can be registered at once as shown in FIG. This reduces the burden of user input. FIG. 4 shows a coordinate input GUI, shows a die configuration on the wafer, a wafer map 402 that is also used for selecting a die for acquiring an image, an in-die coordinate display unit 403 for actually inputting the coordinates inside the die, and a registration button 404. , A cancel button 405, and a measurement location list 406 indicating a list of registered measurement locations. The user inputs the coordinates inside the die to the in-die coordinate display unit 403 by using the input device, and selects the die for image acquisition by directly clicking the die from the wafer map 402. The wafer map 402 displays the selected die in gray, and the points inside the selected die indicate actual image acquisition locations. If there is no problem with the input content, clicking the registration button will register it in the image acquisition location list 406. Click the cancel button 405 to delete the input content. The image acquisition location list 406 displays a list of input image acquisition locations. In this display example, the X and Y coordinates inside the die are shown in parentheses, and the number of image acquisition locations is shown next to it. Such an input auxiliary GUI may be prepared and input by the user, or a file describing coordinates created by another device or a computer may be read from the external input unit 121.

  An image acquisition location may be automatically designated, and at that time, a wafer center portion with relatively good imaging position accuracy and a wafer outer peripheral portion with a relatively large imaging position deviation are included.

  Next, the process moves to step 302, where the wafer is moved based on the coordinate value input in step 301, and an image of the registered coordinates is taken. The image is input to the image memory 115. Reference numerals 501 to 507 in FIG. 5 show examples of images acquired in step 302. Reference numerals 501 to 507 in FIG. 5 are images including regions that should have the same pattern acquired by different dies, but the imaging fields of view are shifted from each other by this positional shift due to positional shift during imaging. This misalignment occurs due to variations in imaging position caused by stage position accuracy, distortion due to thermal expansion of the wafer support, charged particle beam distortion due to sample charging, wafer distortion, and the like. In particular, when a positional deviation of a size depending on the position of the non-inspection region occurs, it is difficult to grasp and correct the magnitude of the positional deviation in advance. In practice, this positional deviation is similar to images 501 to 507. The field of view will shift by that amount. In particular, the visual field shift caused by the distortion of the charged particle beam due to the charging of the sample varies depending on the pattern to be inspected because the charging state depends on the pattern.

  Next, in step 303, the acquired images are aligned and then superimposed. For example, a normalized correlation value may be obtained at each location while shifting the coordinates of the image acquired as alignment, and the portion with the highest value may be regarded as the location where the location was. Since there is a possibility that an image in which a pattern as shown in 503 is not formed is included in the image, the images are superposed except for an image having a low image similarity when the images are aligned. For example, a normalized correlation value may be used as a method for obtaining the defect similarity. In addition, even if a pattern is formed, there is a possibility that a defect as shown in 504 is included. Therefore, after alignment, a statistical process is performed for each overlapping pixel, and the pixel value is out of average. Overlapping is performed except for. For example, the standard deviation of the overlapped pixels is obtained, the average value is recalculated by removing pixels that are more than three times the standard deviation from the overlapped pixel average value, and this is performed for all the overlapped pixels to remove the defective portion. Template can be obtained.

  By acquiring an image with the coordinates and conditions for performing the inspection, it is possible to acquire an image set having the same coordinate shift as that during the inspection. Since the images in this image set are superimposed and combined while maintaining the positional deviation of each image set, the size of the created template image is the size that optimally covers the coordinate shift that occurs during the inspection. More specifically, the size of the template image is a size including all the areas included in the image used for the synthesis. By the above method, it is possible to generate a template image that is larger than the image to be inspected “for the amount of deviation generated in the actual inspection of the pattern”.

  Next, in step 304, a rectangular area 508 that circumscribes an area composed of a set of areas included in at least one of the superimposed images is set. The rectangular area may be set during the overlapping process.

  In step 305, areas 510 to 513 that do not include a composite image within the rectangle set in 304 are set as mask areas. The mask area 509 may be set during the overlay process.

  Next, in step 306, the rectangular area 508 circumscribing the superimposed image is stored.

  In step 307, the mask area 509 is saved. At this time, the mask area information may be included in the image information of the rectangular area 508 stored in step 306. According to the present embodiment, an area (effective inspection area) that can be used for comparison inspection in the template image is a place where the areas 510 to 513 masked by the mask area 509 are viewed from the rectangular area 508.

  In addition, a rectangular area inscribed in an area composed of a set of areas included in at least one of the images used for the synthesis is not a template image including all image areas acquired like the rectangular area 508. It may be used with a template image, in which case there is no need to set a mask area. In this case, the entire inscribed rectangular area becomes the effective inspection area.

  FIG. 6 shows a flowchart of defect detection using a template image generated using the above process and registered in the recipe. The defect detection process described below is performed by a defect detection unit included in the image processing unit. First, in step 601, an image on the wafer is acquired. The coordinates registered in step 301 are used as the coordinates to be acquired. In the next step 602, alignment with the template image is performed. For alignment, for example, as shown at 703 in FIG. 7, the template image 701 is scanned on the image to be inspected 702 to obtain a normalized correlation value at each location, and a portion having the highest value is regarded as a place where the pattern matches. And a method of obtaining the image shift amount. The shift amount corresponds to the values indicated as dx and dy in FIG. Of course, any other method may be used as long as the positional relationship between the template image and the image to be inspected can be specified. In step 603, the difference value of each pixel is obtained by shifting the image by the image shift amount obtained by the alignment process and subtracting the images. In step 604, the difference value between the pixels overlapping the mask area of the template image is set to 0 or a value smaller than a certain value. A portion where the difference value is equal to or greater than the certain value, that is, a portion where a difference greater than a certain value is observed with the template image is detected as a defect. In step 605, the detected defect information is stored in a file or the like.

  As described above, in this embodiment, superposition is performed while maintaining the positional deviations of the acquired plurality of inspected images, and a template image larger than the acquired image is created. By making the size of the template image larger than the acquired image by the size of the maximum positional variation of the multiple images that have been acquired, the effective inspection area is not reduced and the throughput is reduced at a wasteful calculation cost. Therefore, the defect detection process can be performed efficiently. In particular, compared to the method of setting the amount of misalignment in advance and enlarging the template image by that amount, the variation in the actual image to be inspected is used. Is not set unnecessarily large, and a template image having an optimum size can be obtained.

  In the present embodiment, since the comparison is performed using a template image having a size different from that of the image to be inspected, if these are displayed as they are on the display unit, it is difficult for the user to make a visual comparison. Therefore, as shown in FIG. 8, when displaying the detected defect, the template image is cut out in the same pixel size and the same size as the image to be inspected and displayed side by side, thereby facilitating visual comparison. At this time, it becomes easier to perform comparison by cutting out from the template image using the location aligned with the image to be inspected in step 603 as an image for comparison display.

  The template image may be cut out not only when outputting to a screen but also when outputting to a file.

  Further, as shown in FIG. 9, a screen for selecting or designating the inspection area in the template image created in the present embodiment at the time of recipe creation may be displayed. At this time, by switching between the entire template display and the enlarged display, it becomes easy to set the inspection region in detail even for a large template. In FIG. 9, the area to be inspected is surrounded by a thick dotted line, but may be filled with a transparent color.

  In addition, this invention is not limited to an above-described Example, Various modifications are included. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described. Further, a part of the configuration of one embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can be added to the configuration of one embodiment. Further, it is possible to add, delete, and replace other configurations for a part of the configuration of each embodiment.

  Information such as programs, tables, and files for realizing each function can be stored in a recording device such as a memory, a hard disk, or an SSD (Solid State Drive), or a recording medium such as an IC card, an SD card, or an optical disk.

Further, the control lines and information lines indicate what is considered necessary for the explanation, and not all the control lines and information lines on the product are necessarily shown. Actually, it may be considered that almost all the components are connected to each other.

101: electron gun, 102: lens, 103: deflector, 104: objective lens, 105: sample, 106: sample stage, 107: electron beam, 108: secondary electron, 109: reflected electron, 110: lens control circuit, 111: Deflection control circuit, 112: Objective lens control circuit, 113: Analog / digital converter, 114: Address control circuit, 115: Image memory, 116: Mechanism control means, 117: Display unit, 118: Control unit, 119: Image processing unit, 120: input unit, 121: external input / output unit, 122: secondary electron detector, 123: backscattered electron detector, 124: moving stage

Claims (6)

A charged particle beam apparatus for inspecting a defect in a pattern of a substrate on which a plurality of dies are formed by irradiation with a charged particle beam,
A charged particle optical system for detecting secondary charged particles by irradiating the sample with the charged particle beam;
An image memory for storing an image of the sample obtained from the secondary charged particles;
An image processing unit that synthesizes a plurality of images including regions that should have the same pattern in each of the plurality of dies and generates a template image having a size including all the regions included in the image used for the synthesis;
A defect detection unit that detects defects using the template image and the pattern image to be inspected,
The images used for the composition are shifted from each other by a positional shift having a size depending on the position of the inspection area,
The charged particle beam apparatus characterized in that the composition is performed by superimposing the plurality of images while maintaining the positional deviation. The charged particle beam apparatus according to claim 1,
The charged particle beam apparatus according to claim 1, wherein the template image is a rectangular image inscribed in a region composed of a set of regions included in at least one of the plurality of images used for the synthesis. The charged particle beam apparatus according to claim 1,
The charged particle beam apparatus according to claim 1, wherein the template image is a rectangular image circumscribing a region formed of a set of regions included in at least one of the plurality of images used for the synthesis. The charged particle beam apparatus according to claim 2,
The charged particle beam apparatus according to claim 1, wherein the displacement is caused by distortion of the charged particle beam due to charging of the sample. The charged particle beam apparatus according to claim 1,
A charged particle beam apparatus that compares and displays an image having the same size as the image to be inspected cut out from the template image and the image to be inspected. The charged particle beam apparatus according to claim 1,
A charged particle beam apparatus comprising: a display unit configured to display a screen for selecting a region to be inspected from a region included in the template image.