JP5241697B2 - Alignment data creation system and method - Google Patents

Description

  The present invention relates to a technique for reducing a burden on a user when creating an imaging recipe for use in a scanning electron microscope.

  A device pattern or wiring pattern on a semiconductor wafer is generally formed by the following procedure. First, a coating material called a resist is applied to the surface of the semiconductor wafer. Next, an exposure mask (reticle) processed into a pattern shape is placed on the resist so as to overlap, and exposed with visible light, ultraviolet light, or the like. Thereafter, a predetermined pattern is formed on the surface of the semiconductor wafer based on the photosensitive pattern formed on the resist.

  By the way, the shape of the pattern formed by these procedures varies depending on the intensity of the irradiated electron beam and the aperture. For this reason, in order to form a highly accurate pattern, it is necessary to inspect the quality of the pattern. For this inspection, a critical dimension scanning electron microscope (CD-SEM) is widely used. In general, an image of an arbitrary evaluation point on a semiconductor pattern that requires inspection is observed with a CD-SEM, and the pattern width and shape data are acquired from the observed image, thereby evaluating the quality of the pattern.

  In order to evaluate the shape of the pattern formed on the semiconductor wafer by the CD-SEM, it is necessary to create a recipe that specifies the position, magnification, and image quality of the evaluation point. Conventionally, recipes have been created by CD-SEM. However, due to a significant increase in evaluation points due to recent high integration, techniques for automatically generating recipes from CAD data are becoming mainstream recently.

  By the way, in the lithography process, a method is adopted in which an exposure area of a semiconductor wafer is divided into a plurality of shots, and a reticle is exposed on the corresponding area of each shot. At the time of exposure, alignment marks (alignment marks) provided in each shot area of the semiconductor wafer are used to achieve highly accurate alignment between the semiconductor wafer and the reticle.

  On the other hand, when a recipe created from CAD data is executed by a CD-SEM, there is a possibility that a positional deviation occurs between the CAD data and the SEM image due to a deviation of the origin of the semiconductor wafer or a rotational deviation of the semiconductor wafer. Therefore, in order to correct misalignment, a design template created by cutting out image data around the alignment mark from CAD data is transferred to a CD-SEM and matched with an OM image optically imaged by the CD-SEM. Is used.

  By the way, the alignment mark is a pattern that is not used for a semiconductor element. Therefore, from the viewpoint of effectively using the semiconductor wafer for element formation, the alignment mark is formed in a scribe region other than the element formation region. The scribe region is a region that separates the element formation region. At the time of exposure, positioning is performed so that scribe portions of a plurality of adjacent shots overlap each other, and a composite image of partial patterns of different shapes divided and arranged in a plurality of shots is completed as an alignment mark in the overlapping region. Thus, a method of examining the positional deviation of each shot from the shape of the completed alignment mark is employed.

  However, in the conventional method, when one alignment mark is created as a composite pattern of partial patterns divided and arranged in a plurality of shots (hereinafter referred to as “one alignment mark is created across a plurality of shots”). Also, a recipe is created based only on CAD data for one shot. That is, a recipe is created without considering CAD data of other shots that are exposed in duplicate. For this reason, it is impossible to display an accurate alignment mark as a design template. For the same reason, the global alignment point cannot be registered with reference to the CAD data. Even if the coordinates of the global alignment point are known and the global alignment point can be registered, an accurate alignment mark corresponding to the design template cannot be formed from the CAD data. For this reason, there is a problem that the difference between the design template and the OM image optically captured by the CD-SEM is large, and the matching accuracy is lowered.

  As one of the methods for solving this problem, the user himself / herself uses image editing software or the like to cut out and synthesize corresponding data images from a plurality of CAD data of a plurality of shots related to one alignment mark. Then, a method of creating a design template can be considered. However, this method requires the user himself to edit the CAD data, and creating all the global alignment points is laborious and impractical. Further, in the conventional method, when the global alignment point is registered, the user himself / herself has to refer to the CAD data and directly input a target coordinate value.

  In the present invention, when a partial pattern constituting one alignment mark is divided into a plurality of shots at the time of overlapping exposure, CAD data of a plurality of shots related to each alignment mark is synthesized according to the mutual positional relationship. A method of automatically creating synthetic CAD data including a completed alignment mark is proposed.

  According to the present invention, even when a partial pattern constituting one alignment mark is divided into a plurality of shots at the time of overlapping exposure, synthetic CAD data including a completed alignment mark can be automatically created without burden on the user.

The figure explaining the example of a schematic structure of a length measurement scanning electron microscope (CD-SEM). The figure explaining the image detected with the irradiation aspect of an electron beam with respect to a semiconductor wafer, and an electron beam. The figure explaining an imaging sequence and the figure which shows an example of the various visual field position on a low magnification image. The figure which shows the example of superimposition of several shots. The flowchart explaining the creation sequence of synthetic | combination CAD data. The flowchart explaining the registration sequence of a global alignment point.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In addition, the content of the apparatus configuration and processing operation to be described later is merely an example, and other embodiments can be realized by combining or replacing the embodiments with known techniques.

(Basic information)
When forming a wiring pattern on a semiconductor wafer, the following processes are executed in order. First, a coating material called a resist is applied to the surface of a semiconductor wafer, and an exposure mask (reticle) corresponding to the wiring pattern is placed thereon so as to overlap. Thereafter, the resist is exposed by irradiating visible light, ultraviolet rays or electron beams from above the reticle to form a photosensitive pattern and a non-photosensitive pattern. Thereafter, the surface of the semiconductor wafer is processed based on the formed pattern to form a wiring pattern.

  The shape of the formed wiring pattern changes depending on the intensity or aperture of the visible light, ultraviolet light, or electron beam to be irradiated. For this reason, in order to form a highly accurate wiring pattern, it is necessary to inspect the quality of the pattern.

  A CD-SEM is used for this inspection. Usually, a critical point on a semiconductor pattern that requires measurement is observed with a CD-SEM as an evaluation point (hereinafter referred to as “EP”), and various dimension values such as a wiring width of the pattern are measured from the observed image. The user evaluates the quality of the pattern based on the measured dimension value.

  In order to capture an EP in a high quality without misalignment, an addressing point (hereinafter referred to as “AP”), an autofocus point (hereinafter referred to as “AF”), an auto stigma point (hereinafter referred to as “AST”). Or a part or all of imaging points of auto brightness / contrast points (hereinafter referred to as “ABCC”), and addressing, auto focus adjustment, auto stigma adjustment, auto brightness / contrast are set at each imaging point. Make adjustments.

  The amount of displacement of the imaging position at the time of addressing is the amount of displacement due to matching between the SEM image of the AP with known coordinates registered in advance as a registered template and the SEM image (actual imaging template) observed in the actual imaging sequence. presume.

  Hereinafter, EP, AP, AFC, AST, and ABCC are collectively referred to as an imaging point. Also, the coordinates, size / shape, imaging sequence, imaging conditions, and registered template of points including part or all of the imaging points are managed as imaging recipes.

  Conventionally, generation of an imaging recipe is manually performed by an operator himself, which requires a lot of labor and time. In addition, it is necessary to actually image a semiconductor wafer at a low magnification in order to determine each imaging point and register a registration template in an imaging recipe, and the generation of the imaging recipe is a cause of a decrease in the operating rate of the SEM. . Furthermore, with the introduction of OPC (Optical Proximity Correction) technology associated with pattern miniaturization, the number of EPs that require evaluation has increased explosively, and manual creation of imaging recipes is becoming unrealistic.

  Therefore, based on circuit design data (hereinafter referred to as “CAD data (Computer Aided Design)”), which is pattern design information on a semiconductor wafer in a low magnification field of view, imaging parameters and AP templates for observing EPs Application of a semiconductor inspection system for registering etc. is desired. Here, the imaging parameters include (a1) number of points (AP, AFC, AST, ABCC), (a2) coordinates, (a3) size / shape, (a4) imaging sequence (imaging order of EP and imaging points) Electron beam normal incidence coordinates), (a5) imaging position changing method (stage shift, beam shift), (a6) imaging conditions (probe current, acceleration voltage, electron beam scanning direction, etc.), (a7) imaging Includes some or all imaging parameters of sequence or template evaluation values or priorities.

  In this semiconductor inspection system, (b1) EP coordinates, size / shape, and imaging conditions as evaluation point information, (b2) CAD data (including layer information) around the EP as design pattern information, as input information. Pattern removal information of mask data, pattern line width information, type of wafer to be imaged, process, pattern and base material information, (b3) imaging points (AP, AF, AST) as processing parameters of the imaging recipe automatic generation engine , ABCC) search range, necessary condition of the selection element index value to be satisfied by the imaging point (given by a threshold value of the index value, etc.), selection element index priority (given by weight between index values, etc.), imaging. Prohibited areas that must not be selected as points, estimated amount of shape divergence between design patterns and actual patterns, equipment conditions (stage simulation) (B4) Estimated error of stage shift / beam shift), (b4) Required positioning accuracy of imaging position for each imaging point as user-specified specifications, required image quality (focus adjustment, stigma adjustment, brightness / contrast adjustment, contamination requirements, etc. Including a request regarding an allowable electron beam incident angle in EP), a required imaging time, and (b5) history information (information on imaging points that have succeeded or failed in the past).

  Further, in addition to the input information (b1) to (b5) described above, a part of the output information (imaging parameters) (a1) to (a7) described above or a default value / settable range is set as the input information. did. That is, any combination of parameters from the output information (a1) to (a7) and the input information (b1) to (b5) described above is used as input information, and the arbitrary parameters of the output information (a1) to (a7) described above are set. It has a function of recording the combination as output information.

  On the other hand, when generating an imaging recipe used when sequentially imaging a plurality of measurement points on a semiconductor wafer using a CD-SEM, the following problems are conceivable. First, in order to inspect the performance of the semiconductor pattern in the EP, an imaging recipe for imaging the EP must be created. However, as the semiconductor pattern becomes finer, the number of EPs that require inspection increases, and the imaging recipe An enormous amount of labor and time is required to generate Although there are documents related to the automatic generation of imaging recipes, the EP coordinates, size, imaging conditions, etc. are premised on the manual operation of the SEM operator, and the work time of the SEM operator has not been reduced.

  Next, a technique for exhaustively arranging EPs to be registered in the imaging recipe within a range where electrical characteristic analysis is possible in order to improve the throughput of the CD-SEM based on a device whose observation location is specified in advance will be described. . A specific example of a technique for setting an EP (FOV for measurement) based on a device whose observation location is specified in advance is as follows.

(I) In the recipe creation method, an imaging recipe for observing an EP is generated based on a device whose observation location is previously selected on CAD data. As the output information, (a1) coordinates (position), (a2) number of points, (a3) size and shape of the evaluation point EP in which the imaging range (FOV) for observing the EP includes all device shapes are automatically calculated. And register in the imaging recipe.

(Ii) Input information includes: (b1) device configuration information (coordinates, connection information, graphic information), contour extraction parameters as length measurement point information, (b2) EP size as default as user request specification, (b3 ) Whether or not the EP size is optimized is a combination of arbitrary parameters as input information, and the combination of any of the parameters (a1) to (a3) described above is output information.

By applying the specific example described above, the following effects can be obtained.
(I) Since high signal transmission accuracy is required for circuit operation, the verification of the electrical characteristics of the highly paired device and critical path part, which are required for the pattern finish, is difficult. At first glance, it is possible to detect defects that cause abnormal operation at an early stage in the manufacturing process even though the shape is close to the design data.

(Ii) It is possible to suppress the occurrence of contamination in EP without involving an SEM operator, and a highly accurate imaging recipe can be generated in a short time. (Configuration example of CD-SEM)

  FIG. 1 shows a schematic configuration of a critical dimension scanning electron microscope (CD-SEM). CD-SEM is one application example of the alignment data creation system according to the present invention. The CD-SEM is one of electron beam apparatuses that acquires a secondary electron image (Secondary Electron: SE image) or a reflected electron image (Backscattered Electron: BSE image) of a semiconductor wafer (sample). Hereinafter, the SE image and the BSE image are collectively referred to as an SEM image. In addition, the SEM image includes a top-down image obtained by observing the measurement object from the vertical direction and a part or all of a tilt image obtained by observing the measurement object from an arbitrary inclination angle direction.

  The CD-SEM 100 includes an electron optical system 102, a stage drive system, and a signal processing system. The primary electrons 104 generated by the electron gun 103 pass through the capacitor 105, the deflector 106, and the objective lens 108 in this order as an electron beam and irradiate the semiconductor wafer 101 as a measurement sample placed on the stage 117. The deflector 106 and the objective lens 108 control the irradiation position and the aperture of the electron beam so that the electron beam is irradiated at an arbitrary position on the semiconductor wafer 101. Secondary electrons and reflected electrons are emitted from the irradiation position of the electron beam. Note that the trajectories of secondary electrons and reflected electrons are separated from the trajectories of primary electrons 104 by the E × B deflector 107. Among these, secondary electrons are detected by the secondary electron detector 109, and reflected electrons are detected by the reflected electron detectors 110 and 111. The backscattered electron detectors 110 and 111 are installed in different directions. Secondary electrons and backscattered electrons detected by the secondary electron detector 109 and the backscattered electron detectors 110 and 111 are converted into digital signals by the corresponding A / D converters 112, 113, and 114, respectively, and imaged as SEM images. Stored in the memory 122. Note that the image memory 122 in this embodiment is not limited to SEM images, but also CAD data corresponding to each shot, registered coordinate data of global alignment points, imaging recipes and other data related to alignment (hereinafter also referred to as “alignment data”). .) Is also memorized. The CPU 121 performs image processing on the digital signal in the image memory 122 according to the purpose. The processing of the CPU 121 includes alignment data creation processing and global alignment point registration processing, which will be described later.

FIG. 2 shows a method for imaging the signal amount of electrons emitted from the surface of the semiconductor wafer 207 when the surface of the semiconductor wafer 207 is scanned with the electron beams 201 to 206. The perspective view 200 shows the electron beams 201 to 203 scanned in the x direction and the electron beams 204 to 206 scanned in the y direction. The deflection direction (scanning direction) of the electron beam can be changed by the deflector 106. In perspective view 200, a location on the semiconductor wafer 207 to which the electron beam 201 to 203 which is scanned in the x direction is irradiated with G ₁ ~G ₃ respectively. Similarly, it is shown in perspective view 200, the electron beam 204-206 scanned in the y direction and a location on the semiconductor wafer 207 irradiated by G ₄ ~G ₆ respectively. The signal amount of the electrons emitted in G _{1 to} G ₆ becomes the brightness 6 values of the pixels H _{1 to} H ₆ in the image 209 shown in the plan view 210 (the subscripts 1 to 6 of G and H correspond to each other). To do). The plan view 210 also shows a coordinate system 208 indicating the x and y directions on the image.

  The CPU 121 and the image memory 122 described above constitute a processing / control unit 115. An input device (not shown) is connected to the processing / control unit 115 for cursor operation and information input. The processing / control unit 115 functions as a computer system and controls the stage controller 119 and the deflection control unit 120 so that a predetermined imaging point is imaged based on an imaging recipe, or an arbitrary imaging point on the semiconductor wafer 101. Various types of image processing and control are performed on the captured image acquired. The imaging point here includes a part or all of AP, AFC, AST, ABCC, and EP. The processing / control unit 115 also has a GUI (Graphic User Interface) display function for displaying a captured image, a recipe setting image, and the like on the screen of the display 116.

  The XY stage 117 is moved in the direction of each axis of xyz or is driven to rotate around each axis based on the control of the stage controller 119. By controlling the driving of the XY stage 117, the observation position and the observation direction of the semiconductor wafer 101 by the primary electrons 104 can be changed. That is, the observation position on the semiconductor wafer 101 can be arbitrarily changed. Note that the change of the observation position by the XY stage 117 is called a stage shift, and the change of the observation position by the deflector 106 is called a beam shift. In general, the stage shift has a wide movable range, but the positioning accuracy of the imaging position is low. On the contrary, the beam shift has a property that the movable range is narrow but the imaging position is highly accurate.

  Although FIG. 1 illustrates an embodiment including two reflected electron image detectors, the number of reflected electron image detectors may be one or three or more. Further, the processing / control unit 115 in this embodiment executes control of the SEM apparatus and imaging based on the automatic generation of the imaging recipe and the created imaging recipe according to the processing procedure described later. However, some or all of these processes and controls can be assigned to a plurality of computer systems for execution.

  In order to obtain a tilt image obtained by observing the measurement object from an arbitrary tilt angle direction using the SEM 100, (i) the tilt image is obtained by deflecting the electron beam irradiated from the electron optical system and tilting the irradiation angle of the electron beam. (Ii) a method of tilting the XY stage 117 itself for moving a semiconductor wafer (FIG. 1 shows a state in which the XY stage 117 is tilted by a stage tilt angle 118. And (iii) a method of mechanically tilting the electron optical system itself.

(Imaging sequence based on imaging recipe)
FIG. 3 shows a typical imaging sequence 300 executed when observing an arbitrary EP. An imaging point, an imaging order, and an imaging condition in the imaging sequence 300 are specified by an imaging recipe. First, in step 301, the semiconductor wafer 101 as a sample is placed on the XY stage 117. In step 302, a global alignment mark on the wafer is observed with an optical microscope (OM) or the like, thereby correcting the origin deviation of the semiconductor wafer and the rotation deviation of the semiconductor wafer.

  In step 303, the imaging position is moved to the AP by the movement control of the XY stage 117 by the processing / control unit 115 and imaging is performed. The processing / control unit 115 obtains an addressing parameter based on the captured image, and executes an addressing process using the obtained parameter. Here, a description of AP will be added. When observing the EP, if the EP is directly observed by the stage shift, there is a risk that the imaging point is largely shifted due to the positioning accuracy of the XY stage 117.

  Therefore, first, for positioning, an AP where the coordinate value of the imaging point is given as a template (imaging point pattern) is observed. The template is registered in the imaging recipe. For this reason, in the following, it is referred to as a registered template. AP is selected from the vicinity of EP (a range that can be moved by beam shift at the maximum). AP is generally a low magnification field of view with respect to EP. For this reason, even if there is a slight shift in the imaging position, there is a low risk that all the patterns to be imaged are out of the field of view. Therefore, the amount of positional deviation of the imaging point at the AP is estimated by matching the registered template of the AP registered in advance with the SEM image (actual imaging template) of the AP actually captured.

  The coordinate values of AP and EP are known. Therefore, the relative displacement amount between AP and EP can be obtained. In addition, the positional deviation amount of the imaging point in the AP can be estimated by matching. For this reason, the relative displacement amount from the AP imaging position to be actually moved to the EP can be obtained by subtracting the displacement amount from the relative displacement amount. By moving the imaging point by beam shift with high positioning accuracy by the amount of relative displacement, it is possible to image EP with high coordinate accuracy.

  The registered AP is a pattern that exists at a distance that can be moved by beam shift from (i) an EP (in order to suppress the occurrence of contamination in the EP, in the AP imaging range (Field of view: FOV)) (Ii) The imaging magnification of the AP is lower than the imaging magnification of the EP in consideration of the positioning accuracy of the stage, and (iii) the pattern shape or brightness pattern is characteristic. It is desirable that the matching between the registered template and the actual imaging template satisfies a condition such as easy.

  Note that manual selection by the SEM operator is generally used for AP selection, but AP selection and imaging sequence determination may be automated. A CAD image and an SEM image are used as the registered image template of the AP. As disclosed in Japanese Patent Laid-Open No. 2002-328015, CAD data is temporarily registered as an image template in order to avoid imaging only for registration of an image template, and acquired when an AP is actually captured. Variations such as re-registering the SEM image as an image template are conceivable.

  It supplements about the above-mentioned AP selection range. In general, the electron beam normal incidence coordinates are set to the center coordinates of the EP. For this reason, the AP selection range is a beam shift movable range centered on the EP at the maximum, but when the vertical incident coordinates of the electron beam are different from the center coordinates of the EP, the beam shift movable range from the vertical incident coordinates. Is the selection range. Further, depending on the incident angle of the allowable electron beam required for the imaging point, the search range from the normal incident coordinate of the electron beam may be smaller than the beam shift movable range. The same applies to the other templates.

  In the next step 304, imaging is performed by moving the imaging position to AF by beam shift under the control of the processing / control unit 115. The processing / control unit 115 obtains parameters for autofocus adjustment based on the captured image, and executes autofocus adjustment using the obtained parameters. A description of AF will also be added. Autofocus is performed in order to acquire a clear image at the time of imaging. However, if a semiconductor wafer as a sample is irradiated with an electron beam for a long time, a contaminant is attached. Therefore, in order to suppress the adhesion of contamination to the EP, a method is used in which coordinates around the EP are observed in advance as AF, parameters for autofocus are obtained, and the EP is observed based on the obtained parameters. Therefore, the registered AF is (i) a pattern that exists at a distance that can be moved by beam shift from AP and EP, and the FOV at the time of AF imaging does not include the FOV at the time of EP imaging. (Ii) The imaging magnification of AF is about the same as the imaging magnification of EP (however, this is the case of AF for EP, and in the case of AF for AP, the AF is imaged at the imaging magnification that is about the same as the imaging magnification of the AP) The same applies to AST and ABCC, which will be described later), and (iii) a pattern shape that facilitates autofocusing (easy to detect blurring of an image due to focus shift) is desirable.

  In the next step 305, imaging is performed by moving the imaging position to the AST by beam shift under the control of the processing / control unit 115. The processing / control unit 115 obtains parameters for auto stigma adjustment based on the captured image, and executes auto stigma adjustment using the obtained parameters. A description of AST is also added. Astigmatism correction is performed in order to acquire an image without distortion at the time of imaging. However, as in the case of AF, when a sample is irradiated with an electron beam for a long time, contaminants adhere. Therefore, in order to suppress the adhesion of contamination in the EP, a method is used in which coordinates around the EP are observed in advance as AST, parameters for correcting astigmatism are obtained, and the EP is observed based on the obtained parameters. Therefore, the registered AST is (i) a pattern that exists at a distance that can be moved by beam shift from the AP and EP, and the FOV at the time of AST imaging does not include the FOV at the time of EP imaging. (Ii) AST imaging magnification is about the same as EP imaging magnification, and (iii) has a pattern shape that facilitates astigmatism correction (easily detects image blur due to astigmatism) It is desirable.

  In the next step 306, imaging is performed by moving the imaging position to ABCC by beam shift under the control of the processing / control unit 115. The processing / control unit 115 obtains parameters for brightness / contrast adjustment based on the captured image, and executes auto brightness / contrast adjustment based on the obtained parameters. A description of ABCC is also added. The auto brightness / contrast adjustment is performed in order to acquire a clear image having an appropriate brightness value and contrast at the time of imaging. For example, by adjusting a parameter such as a voltage value of a photomultiplier (photomultiplier) in the secondary electron detector 109, for example, the highest part and the lowest part of the image signal become full contrast or a contrast close thereto. Set as follows. However, as in the case of AF, if a sample is irradiated with an electron beam for a long time, contaminants will adhere. Therefore, in order to suppress the adhesion of contamination in the EP, a technique is used in which coordinates around the EP are observed in advance as ABCC, parameters for brightness / contrast adjustment are obtained, and the EP is observed based on the obtained parameters. Therefore, the registered ABCC is a pattern that exists at a distance that can be moved by beam shift from AP and EP, and the FOV at the time of ABCC imaging does not include the FOV at the time of EP imaging, (ii) The imaging magnification of ABCC is about the same as the imaging magnification of EP. (Iii) Since the brightness and contrast of the image captured at the measurement point using the parameters adjusted in ABCC are good, ABCC is at the measurement point. It is desirable that a condition such as a pattern similar to the pattern is satisfied.

  Note that some or all of the imaging of AP, AF, AST, and ABCC in steps 303, 304, 305, and 306 described above may be omitted depending on circumstances, or the execution order of steps 303, 304, 305, and 306 is arbitrary. There are variations such as switching to, or overlapping in the coordinates of AP, AF, AST, ABCC (for example, auto focus and auto stigma are performed at the same location).

  In the final step 307, the imaging point is moved to the EP by the beam shift under the control of the processing / control unit 115 and imaging is performed. For example, the processing / control unit 115 performs pattern length measurement and other processing based on the set length measurement conditions. In EP, matching is performed between a newly captured SEM image and a registered template registered in advance (an image corresponding to the EP registered in the imaging recipe), etc., and the measurement position shifts. May be detected. In the imaging recipe, information such as the coordinates of the imaging points (EP, AP, AF, AST, ABCC), the imaging sequence, and imaging conditions are written. The SEM observes the EP based on the imaging recipe.

  In FIG. 3, regarding the low-magnification image 308, the positional relationship on the template of EP309, AP310, AF311, AST312, and ABCC313 and the difference in magnification are illustrated by dotted line frames.

(Create synthetic CAD data)
As described above, the alignment mark is formed in the scribe region in order to effectively use the semiconductor wafer as an element formation region. In addition, in order to detect a positional deviation between shots, one alignment mark may be arranged so as to extend over a plurality of shots. In this embodiment, a partial pattern that completes one alignment mark when each edge or corner of a plurality of adjacent shots is exposed in an overlapping manner is that each alignment mark extends over a plurality of shots. The state which is divided | segmented and arrange | positioned at the edge part and corner part of. The shape of the partial pattern is different between a plurality of shots corresponding to one alignment mark.

  FIG. 4 shows an example of this type of alignment mark. FIG. 4 shows an example in which one alignment mark 420 is formed by overlapping four partial patterns 411 to 414 arranged at one corner (corner) of four rectangular shots 401 to 404. Note that the shots 401 to 404 in FIG. 4 do not represent the entire shot, but are enlarged representations of four corners related to one alignment mark of interest for convenience of explanation. Accordingly, the shot 401 is an enlarged view of the lower right corner of the entire shot, the shot 402 is an enlarged view of the lower left corner of the entire shot, the shot 403 is an enlarged view of the upper right corner of the entire shot, and the shot 404 is the upper left corner of the entire shot. An enlarged view is shown.

  In the conventional method of creating a recipe using only one shot, alignment is performed using only one of the partial patterns 411 to 414, or these four partial patterns 411 to 414 are synthesized manually by the SEM operator. However, it is necessary to complete the alignment mark 420.

  Hereinafter, an operation example in which the synthetic CAD data 421 including the alignment mark 420 is automatically created will be described. In the following description, a case where one alignment mark is configured by combining four partial patterns will be described, but one alignment mark may be configured by combining two or more partial patterns. Further, in the following description, a case will be described in which four partial patterns are formed, one at each of four corners, on one shot, but the number of partial patterns formed on one shot is one. That's all there is to it. Further, in the following description, it is assumed that partial patterns are arranged at the four corners of each shot. However, the partial pattern is arranged at any position on a scribe area where overlapping exposure of other adjacent shots is possible. May be.

  As described above, the alignment mark is formed in the scribe region for separating adjacent shots. Therefore, it can be seen that when the synthesized CAD data 421 including the alignment mark is generated from the CAD data of the four shots that are adjacent to each other, it may be created as the size up to the scribe area of the adjacent shot.

  There are variations described below in the method of generating the synthetic CAD data, and each generation method can be selected as necessary. When the methods shown in (iii) and (iv) are selected from the four generation methods listed below, no user operation such as inputting an overlap width is required for synthesis of CAD data. Therefore, when these generation methods are employed, all processes related to the generation of the synthetic CAD data can be automated.

(I) The SEM operator (user) individually designates the size of the four corners of the upper, lower, left and right overlapping with the other adjacent shots for each CAD data corresponding to each shot, and based on the input information, Synthesize CAD data.

(Ii) On the screen such as the recipe setting screen on which the CAD data image corresponding to each shot is displayed, each image corresponding to the CAD data (a different partial pattern is formed through a mouse drag operation or the like). ) To complete one alignment mark. When one alignment mark is completed, the CPU 121 specifies the overlapping position based on the position coordinates after the movement of each CAD data, and synthesizes four pieces of CAD data at the overlapping position.

(Iii) When one alignment mark is created by overlapping figures (for example, triangles, quadrangles, etc.) in the scribe area, the CPU 121 calculates the center of gravity or center coordinates of the menopause figure, The CAD data is synthesized using the distance as overlay width information. In other words, the four CAD data are synthesized so that the center of gravity or center coordinates of the menopause figure coincide.

(Iv) An OM image obtained by capturing an alignment mark with a CD-SEM is captured, and a position matching the OM image is searched from the scribe area while gradually adjusting the overlapping width of the scribe area, and a correlation value is obtained. If there is a correlation value that is greater than or equal to a threshold value or higher than a threshold value arbitrarily designated by the user, CAD data is synthesized with the overlap width with the highest evaluation value. Here, it is also conceivable to use a global alignment mark image file created by the user instead of the OM image captured on the CD-SEM.

  FIG. 5 shows an example of an automatic creation sequence of synthetic CAD data including the alignment mark 420. Note that the processing sequence shown in FIG. 5 is realized through a program executed by the CPU 121. First, in step 501, the CPU 121 accepts selection of a CAD data overlay method. As the superposition method, there are (i) to (iv) described above.

  First, when direct input of the overlap width is selected (in the case of a positive result in step 502), the CPU 121 synthesizes CAD data corresponding to four shots based on the input overlap width and aligns the overlap portion. Synthetic CAD data including the completed mark form is created (step 511).

  On the other hand, when the adjustment input of the CAD data overlap width on the GUI is selected (in the case of a positive result in step 503), the CPU 121 determines based on the moving amount of the image corresponding to each CAD data on the screen. The overlap width is calculated (step 521). When the overlap width is calculated, the CPU 121 combines CAD data corresponding to the four shots based on the calculated overlap width, and creates combined CAD data including the completed alignment mark in the overlap portion (step). 511).

  On the other hand, when a method for automatically adjusting the center of gravity of the menopause figure is selected (in the case of a positive result in step 504), the CPU 121 searches for the menopause figure from the CAD data corresponding to the scribe area of each shot ( Step 531), the center of gravity of the menopause figure is calculated (Step 532). Thereafter, the CPU 121 calculates the overlap width based on the distance between the centers of gravity of the menopause figures calculated for each of the four adjacent shots (step 533). When the overlap width is calculated, the CPU 121 combines CAD data corresponding to the four shots based on the calculated overlap width, and creates combined CAD data including the completed alignment mark in the overlap portion (step). 511).

  On the other hand, when the method of automatically calculating the overlap width based on the global alignment mark (OM image) imaged by the optical microscope mounted on the CD-SEM is selected (in the case of a positive result in step 505), the CPU 121 The increment is added to the current value (for example, the initial value is zero) of the overlap width of the scribe area (step 541), and the overlap width of the CAD data is adjusted by the updated overlap width. Instead of an OM image, an image file of a global alignment mark stored in a memory may be used. Thereafter, the CPU 121 executes a matching process with the OM image for the composite image of the scribe area overlapped by the overlapping width (step 542). Next, the CPU 121 determines whether or not the correlation value between the searched area and the OM image is equal to or larger than the threshold value (step 543), and while the negative result is obtained, the step size until the correlation value exceeds the threshold value. Adjust the overlap width one by one. When the correlation value with the OM image is equal to or greater than the threshold value, the overlap width is calculated based on the coordinates of the CAD data at that time (step 544). However, the overlap width may be continuously updated even if a correlation value equal to or greater than the threshold value is obtained, and the overlap width at which the highest correlation value is finally obtained may be determined as the optimum value. When the overlap width is calculated, the CPU 121 combines CAD data corresponding to the four shots based on the calculated overlap width, and creates combined CAD data including the completed alignment mark in the overlap portion (step). 511).

  If none of the above four is present (in the case of a negative result in step 505), the CPU 121 sets the overlap width to zero. Thereafter, the CPU 121 synthesizes CAD data corresponding to the four shots with the overlap width being zero, and generates composite CAD data including the completed alignment mark in the overlap portion (step 511).

(Register global alignment information)
Registration of global alignment points in the conventional method is manual input by the user himself. On the other hand, in the present embodiment, registration of global alignment points can be automated. FIG. 6 shows a processing sequence at the time of automatic registration of global alignment points. The processing sequence shown in FIG. 6 is executed through a program executed by the CPU 121. Incidentally, it is assumed that a global alignment point design template used in the past and parameter information used at that time are output as a file and held in the image memory 122 of the CD-SEM 100 or the like.

  First, in step 601, the CPU 121 acquires the magnification input by the user. In the next step 602, the CPU 121 reads a design template of global alignment points used in the past and its parameter information. The design template used here is obtained by cutting out global alignment points and surrounding images from synthesized CAD data created in the past.

  After that, the CPU 121 determines the synthesized CAD data corresponding to the scribe area of the semiconductor wafer or the area specified by the user according to the acquired magnification, and the design template of the global alignment point used in the past (as previously described, And a pattern with a high degree of similarity is searched (step 603).

  In the next step 604, the CPU 121 determines whether or not the correlation value with the design template is equal to or greater than a threshold value even for one of the detected similar patterns. If a negative result is obtained in this step 604, the CPU 121 returns to the processing of step 602 and executes a similar search operation for another design template. On the other hand, if a positive result is obtained in step 604, the CPU 121 proceeds to step 605, and registers the coordinates of the pattern having the highest correlation value as matching coordinates. That is, it is registered as a global alignment point.

  Thereafter, in step 606, the CPU 121 sets a parameter used in the past design template or another parameter information file as a global alignment point parameter. Thus, in this embodiment, the registration of global alignment points can be automated by using automatically generated synthetic CAD data.

(Create imaging recipe)
As described above, the imaging recipe is managed as the coordinates, size / shape, imaging sequence, imaging conditions, and design template of a point including part or all of the imaging point. In the case of this embodiment, part or all of the generation process of the synthetic CAD data and the registration process of the global alignment point can be automated. At this time, automatic generation of imaging recipes by automatically registering the coordinates, size / shape, imaging sequence, and imaging conditions of the points corresponding to some or all of the automatically generated synthetic CAD data and global alignment points in the imaging recipe Can be realized.

(Other examples)
In the above description, a case has been described in which a length-measuring scanning electron microscope (CD-SEM) is equipped with a function for automatically creating synthetic CAD data, a function for automatically registering global alignment points, and a function for automatically creating an imaging recipe. However, some or all of the function for automatically creating synthesized CAD data, the function for automatically registering global alignment points, and the function for automatically creating imaging recipes may be installed in a recipe creation system or in a so-called CAD system. May be.

  DESCRIPTION OF SYMBOLS 100 ... CD-SEM, 101 ... Semiconductor wafer, 102 ... Electron optical system, 103 ... Electron gun, 104 ... Primary electron, 105 ... Condenser, 106 ... Deflector, 107 ... ExB deflector, 108 ... Objective lens, 109 ... Two Secondary electron detector, 110, 111 ... Backscattered electron detector, 112, 113, 114 ... A / D converter, 115 ... Processing / control unit, 116 ... Display, 117 ... XY stage, 118 ... Stage tilt angle, 119 ... Stage controller 120... Deflection controller 121 121 CPU 122 Image memory 201-206 Electron beam 207 Semiconductor wafer 208 Coordinate system

Claims (10)

In the alignment data creation system,
A memory that stores at least CAD data of a plurality of shots having a partial pattern that completes one alignment mark when the edge or corner of adjacent shots are overlapped and exposed;
A set of CAD data corresponding to a plurality of shots having the partial pattern constituting the alignment mark to be processed is read from the memory, and the set of CAD data is synthesized based on overlay information between the shots. And an arithmetic unit for generating synthetic CAD data including a completed alignment mark. In the alignment data creation system according to claim 1,
When the partial pattern is a menopause figure and the center of gravity or center of each partial pattern coincides to complete the alignment mark,
Based on the calculated distance information, the calculation unit calculates a centroid or center coordinates for each menopause figure included in the set of CAD data, calculates distance information between the centroids or center coordinates, and the calculated distance information. And a process of synthesizing the set of CAC data so as to make the center of gravity or center of the partial pattern coincide with each other. In the alignment data creation system according to claim 1,
The arithmetic unit is a candidate for synthesized CAD data in which a captured image of a global alignment mark or an image file of a global alignment mark held in a memory overlaps a scribe area of the set of CAD data by a set overlapping width. And a process of calculating an overlap width for obtaining a high degree of similarity and a process of synthesizing the set of CAC data based on the calculated overlap width. . The alignment data creation system according to claim 1,
A second memory for storing design template data corresponding to global alignment marks used in the past;
A second process of collating synthesized CAD data created in the arithmetic unit with design template data stored in the second memory and registering an alignment mark having a high similarity with the collated design template as a global alignment point. An alignment data creation system, further comprising: an arithmetic unit. The alignment data creation system according to claim 4,
An alignment data creation system, wherein a pattern around an alignment mark registered as a global alignment point by the second arithmetic unit is cut out as design template data from corresponding synthetic CAD data. In the alignment data creation method,
A portion that constitutes an alignment mark to be processed from a memory that stores at least CAD data of a plurality of shots having a partial pattern that completes one alignment mark when the edges or corners of adjacent shots are overlapped and exposed. A process of reading a set of CAD data corresponding to a plurality of shots having a pattern;
An alignment data creating method comprising: synthesizing the set of CAD data based on superposition information between shots and generating synthesized CAD data including a completed alignment mark. In the alignment data creation method according to claim 6,
When the partial pattern is a menopause figure and the center of gravity or center of each partial pattern coincides to complete the alignment mark,
The process of generating the synthesized CAD data is as follows:
A process of calculating the center of gravity or the center coordinates for each menopause figure included in the set of CAD data;
Processing for calculating distance information between the center of gravity or center coordinates;
A process for synthesizing the set of CAC data so as to make the center of gravity or center of the partial pattern coincide based on the calculated distance information. In the alignment data creation method according to claim 6,
The process of generating the synthesized CAD data is as follows:
A global alignment mark image file stored in a memory or a global alignment mark image file stored in a memory is compared with a combined CAD data candidate in which a scribe region of the set of CAD data is overlapped by a set overlap width. A process of calculating an overlap width that provides a high degree of similarity;
A method of synthesizing the set of CAC data based on the calculated overlap width. The alignment data creation method according to claim 6 is:
A process of matching design template data corresponding to global alignment marks used in the past with newly created synthetic CAD data and registering alignment marks that have a high degree of similarity with the collated design template as global alignment points An alignment data creation method characterized by further comprising: The alignment data creation method according to claim 9 is:
A method for creating alignment data, further comprising: cutting out a pattern around an alignment mark registered as a global alignment point as design template data from corresponding synthetic CAD data.

Images