JP5463334B2 - Pattern measuring method and pattern measuring apparatus - Google Patents

Description

  The present invention relates to a pattern measurement method, a pattern measurement device, and a program, and more particularly to a method, device, and program for measuring a pattern in an acquired image.

  It is known to measure a pattern on a semiconductor integrated circuit using CAD (Computer Aided Design) data. Since design data such as CAD data indicates an ideal shape that a semiconductor element should originally be, it is possible to evaluate a semiconductor manufacturing process by comparing CAD data with an actually formed pattern. . Patent Documents 1 and 2 disclose that the pattern deformation amount with respect to the design data is detected by performing edge detection of the inspection target pattern and the reference pattern and comparing the detected edges.

JP 2001-338304 A JP 2002-31525 A

  As described above, the actually formed pattern exhibits a shape different from the design data due to the influence of the manufacturing process. Many patterns of various shapes are formed on the semiconductor wafer, but there is no definite standard for alignment between the design data and the actual pattern. Which pattern is to be measured against the ideal pattern indicated by the design data? It was not possible to measure the degree of deformation or the degree of deformation based on some criteria.

  In order to solve the above problems, the reference pattern is formed based on the position information of the reference pattern on the image formed by a scanning electron microscope or the like, and is detected based on the design data. Based on the positional relationship information between the pattern and the measurement location, a reference location of the measurement location by a scanning electron microscope or the like is set.

  According to the above configuration, the position of the reference pattern is detected from the image obtained by a scanning electron microscope or the like, and the measurement reference position is set based on the design data using this point as a reference. Measurement of how much the actual pattern has been formed or changed from the actual pattern formation location or pattern shape based on the positional information based on the SEM image and the design data Can be done.

The figure which shows the outline of a scanning electron microscope. The figure which shows the example which superimposed the circuit design data and the pattern. Flowchart from length measurement recipe creation centering on circuit design data to evaluation. Examples of image acquisition points. The figure which shows an example of a box in box pattern. The figure explaining where the design data corresponds to the chip | tip on a wafer. The figure which shows an example of the pattern used for defining the distance of design data and the position on a wafer. The figure which shows the positional relationship of a 1st addressing point and an image acquisition point. The figure which shows the example which searches a 2nd addressing point. The figure which shows the positional relationship of the searched 2nd addressing point and image acquisition point. The figure which shows the specific example on the circuit design data of FIG. The figure which shows the example which united the position of SEM image and circuit design data in the 2nd addressing point. The figure which shows the example which cut out the image acquisition point from FIG. FIG. 14 is a diagram showing an example of determining a length measurement location from the image obtained in FIG. The figure which shows the example which measured the distance between the virtual line segment which does not exist on a wafer, and a pattern. The figure which shows the place which performed length measurement in FIG. 15, and the secondary electron signal strength in there. The flowchart explaining the flow of the detailed information setting of a length measurement SEM recipe. It is a conceptual diagram of the created recipe.

  The outline of a scanning electron microscope (hereinafter referred to as Scanning Electron Microscope: SEM) will be described below. The electron optical system of FIG. 1 focuses the charged particle beam (electron beam) 2 emitted from a charged particle source (electron gun) 1 that emits electrons as charged particles onto a sample 4 by a lens 3 in an arbitrary order. Can be scanned. A secondary particle (for example, secondary electrons) 5 generated on the surface of the sample 4 by the electron beam irradiation is detected by a secondary particle detection system 6, and a control system 7 (control) having a function of image calculation control as image data. Processor). The sample 4 can be moved in all three-dimensional directions by the XYZ stage 8. The control system 7 also controls the charged particle source (electron gun) 1, the lens 3, the secondary particle detection system 6, the XYZ stage 8, and the image display device 9.

  In the case of this example, the electron beam 2 is scanned two-dimensionally (XY direction) on the sample 4 by a scanning coil (not shown). The signal detected by the secondary electron detector in the secondary particle detection system 6 is amplified by a signal amplifier in the control system 7 and then transferred to the image memory and displayed on the image display device 9 as a sample image. . The secondary signal detector may detect secondary electrons or reflected electrons, or may detect light or X-rays.

  An address signal corresponding to the memory location of the image memory is generated in the control system 7 or in a separately installed computer, converted to analog, and then supplied to the scanning coil. The address signal in the X direction is a digital signal that repeats from 0 to 512, for example, when the image memory has 512 × 512 pixels, and the address signal in the Y direction reaches the address signal in the X direction from 0 to 512. It is a digital signal repeated from 0 to 512 that is sometimes incremented by one. This is converted into an analog signal.

  Since the address of the image memory corresponds to the address of the deflection signal for scanning the electron beam, a two-dimensional image of the deflection region of the electron beam by the scanning coil is recorded in the image memory. Note that signals in the image memory can be sequentially read out in time series by a read address generation circuit synchronized with a read clock. A signal read corresponding to the address is converted into an analog signal and becomes a luminance modulation signal of the image display device 9.

  The control system 7 is provided with an input device (not shown), and can specify an image capturing condition (scanning speed, the total number of images), a field of view correction method, and output and storage of images.

  The apparatus described in this example has a function of forming a line profile based on detected secondary electrons or reflected electrons. The line profile is formed based on the amount of detected electrons when the primary electron beam is scanned one-dimensionally or two-dimensionally or the luminance information of the sample image. The obtained line profile is, for example, on a semiconductor wafer. It is used for measuring the dimension of the formed pattern.

  In the description of FIG. 1, the control system 7 is described as being integrated with or equivalent to the scanning electron microscope. However, the control system 7 is of course not limited thereto, and is a control processor provided separately from the scanning electron microscope body. Processing as described may be performed. In that case, the detection signal detected by the secondary signal detector is transmitted to the control processor, the transmission medium for transmitting the signal from the control processor to the lens or deflector of the scanning electron microscope, and the transmission medium. An input / output terminal for inputting / outputting a received signal is required.

  Furthermore, the apparatus of this example stores in advance, for example, conditions for observing a plurality of points on a semiconductor wafer (measurement locations, optical conditions of a scanning electron microscope, etc.) as a recipe, and measures and observes according to the contents of the recipe. The function to perform.

  Further, a program for performing the processing described below may be registered in a storage medium, and the program may be executed by a control processor that supplies necessary signals to the scanning electron microscope. That is, the example described below is also an explanation as a program or a program product that can be employed in a charged particle beam apparatus such as a scanning electron microscope equipped with an image processor.

  Further, the control system 7 includes a design data management unit 10 that stores design data of the pattern on the semiconductor wafer and converts it into data necessary for SEM control. The design data management unit 10 has a function of creating a recipe for controlling the SEM based on design data of a semiconductor pattern input by an input device (not shown) or the like. Further, it has a function of rewriting a recipe based on a signal transmitted from the control system 7. In the description of this example, the design data management unit 10 is described as a separate unit from the control system 7. However, the present invention is not limited to this, and the control system 7 and the design data management unit 10 may be integrated. good.

  In this example, the sample 4 was a wafer in the process of manufacturing a semiconductor product. A resist pattern formed on the wafer by a lithography process was used. As the comparison object, semiconductor circuit design data (CAD data) that is the basis of the pattern was used. The semiconductor circuit design data used here is an ideal pattern shape finally desired as a semiconductor circuit pattern. In the following description, the inspection object is a semiconductor wafer, but the present invention is not limited to this as long as the design data and the object to be evaluated are paired. The circuit design data may be of any type as long as the software that displays the circuit design data can display the format and handle it as graphic data.

〔Example〕
Conventionally, in the length measurement SEM, the length measurement point is manually designated by the observer. Therefore, it is necessary to find the location of the measurement point on the wafer, but it is very difficult to find the designated location on the wafer. After the measurement points are specified, the observer needs to set the conditions necessary for creating the measurement SEM recipe such as addressing points and auto focus points for each measurement point. I had to depend on the degree.

  Further, since there is a restriction that a length measurement SEM and a wafer must be present in order to create a length measurement recipe, an efficient operation cannot be performed. For example, even if a pattern is selected as an addressing point, it is not known if it is actually appropriate unless it is actually matched with the pattern on the wafer. By repeating such trial and error, a measurement recipe using a length measurement SEM is created. However, the time required for the measurement recipe itself itself means a reduction in efficiency, and the length measurement SEM must also be used during recipe creation. It is also inefficient in that other measurements cannot be performed during that time.

  Also, as the manufacturing design rules for semiconductor devices have become smaller in recent years, the optical proximity effect (Optical Proximity Effect) inevitably increases because an exposure wavelength shorter than the pattern width is used, and the optical proximity effect correction (Optical Proximity Correct) corrects it. Technology is becoming important. The optical proximity effect is a phenomenon that even in a pattern having the same mask width, the pattern width obtained on the wafer differs in the environment where the pattern is placed.

  The environment here is, for example, a pattern pitch. In a pattern in a dense environment, an isolated pattern in a rough environment has a small dimension under the condition that the pattern width on the mask and the pattern width on the wafer have the same dimension. Such a phenomenon is called an optical proximity effect, and a technique for correcting this effect is optical proximity effect correction. By increasing the mask width of the isolated pattern, the pattern width on the wafer can be adjusted to be the same as the width of the dense pattern.

  In order to evaluate this simple optical proximity effect correction, it was possible to measure if the pattern width could be measured, but in recent years we have also proposed an evaluation method for measuring the shape of the pattern to perform more accurate optical proximity effect correction Has been. As one specific example, the position of the pattern with respect to a certain reference is measured.

  In this case, FIG. 2 shows an example in which the design data and the pattern on the wafer can be accurately superimposed. In this case, it is of course possible to measure both dimensions, but it is also possible to measure the positional relationship between the two. If the position can be measured, information greater than the pattern width can be obtained, and it is expected that the information can be used for more accurate optical proximity correction. Since there is no such reference on the wafer, it is difficult to measure the position of the pattern, but this measurement can be performed by superimposing the circuit design data on the wafer pattern on the image acquired by the length measurement SEM. It becomes possible.

  Hereinafter, this example will be roughly described with reference to the flowchart shown in FIG. First, evaluation points to be inspected are determined using circuit design data. For each evaluation point, a recipe for measuring with the SEM is created. Using the created recipe, an image of an evaluation point on the wafer, which is an inspection object, is acquired by a length measurement SEM. Evaluation is performed by comparing the acquired image with circuit design data. In this example, this comparison is performed with circuit design data. However, the present invention is not limited to this, and is not limited to this as long as it is a reference pattern that can be compared. For example, a simulation image obtained based on the design data may be compared with an image acquired by the length measurement SEM.

  Hereinafter, each step of the flowchart shown in FIG. 3 will be described step by step. The design data used for the evaluation in this example was created using a GDSII format that imitated a circuit design pattern for a pattern having a minimum line width of 90 nm. In the step of determining evaluation points in the pseudo circuit design pattern, a lithography simulator was used to inspect the wafer after the lithography process.

  In the case of inspecting a wafer obtained after another process, an inspection location may be obtained by using a corresponding simulator. For example, an inspection location can be determined using an etching simulator for a pattern after an etching process. In this example, Solid-C manufactured by Sigma-C was used as a lithography simulator. This simulator can directly handle GDSII format data, which is a circuit design pattern, and if the process conditions of the lithography process are used from the circuit design pattern, what kind of circuit design pattern is applied to the wafer after the lithography process? It can be given as a pattern.

As the conditions for the simulation, an exposure wavelength of 193 nm was used, the reduction ratio of the projection optical system was 1/4, the exposure wavelength was 193 nm, NA: 0.73, coherence factor σ: 0.75, annular shielding rate ε: 0.67, The condition of a set exposure amount of 28 mJ / cm ² was used. This is the same condition as a wafer to be created later. A design rule checker was used to select a location that could cause a defect from the pattern shape obtained by simulation. This is a tool for automatically detecting a part whose dimension is shortened from a pattern formed by simulation. In this example, the design rule checker built in Caliber of Mentor Graphics was used. With respect to 90 nm which is the design rule, the design rule checker detects a part where the line width is 80 nm or less and 100 nm or more as a dangerous part. This time, only the line width is extracted as the dangerous part, but the dangerous part may be extracted in other items. The detected dangerous location was output as coordinates on the circuit design pattern.

  That is, for the design rule of 90 nm, 80 nm and 100 nm are respectively set as upper and lower threshold values, and those points that are equal to or higher than the threshold value are extracted as dangerous points, and are used as design data. Based on this, the coordinates of the part are output. In this example, the information is displayed on the display device provided in the design data management unit 10 or the image display device 9. If such display is performed, it is easy to determine where the length measurement point (evaluation point) should be set on the semiconductor wafer in which a plurality of patterns are mixed. Further, the design data management unit 10 or the control system 7 is provided with an input device such as a pointing device, and a desired location is selected as a length measurement location from locations extracted as dangerous locations. It is also good. And if this recipe is configured so that the recipe can be automatically rewritten, the operator can specify the location to be measured from the dangerous locations based on his own rule of thumb. It is possible to achieve both the appropriate selection of the dangerous place and the efficiency of the evaluation time. The detected dangerous location may be automatically recorded in the recipe as a length measurement target location, and unnecessary length measurement locations may be deleted later.

  Next, a recipe for inspecting the output coordinates with the length measuring SEM is created. In order to measure the evaluation points with the length measurement SEM, not only the coordinates of the evaluation points, but also the global alignment points that correct the stage coordinates and the chip coordinates, and the addressing points that are used to correct the position step by step until it is reached. It is necessary to change the setting conditions depending on the environment where the evaluation points are set, such as where to perform auto focus and auto stigma.

  In this example, among these, the addressing point, auto focus, and auto stigma point are automatically determined from the design data.

  In addition, since the design data can be handled as an addressing template, it is possible to create a measurement recipe without a wafer to be measured. The length measurer specified the measurement conditions other than the above coordinates (magnification at the addressing point and length measurement point, acceleration voltage, sample current, electron optical system condition, contrast condition). This is because these settings are not determined from the circuit design data, but are information to be determined by the length measuring person.

  As a recipe for the length measurement SEM, first, global alignment is performed, and coordinate correction between a wafer and a chip formed on the wafer is performed. As a result, the coordinates on the chip can be used on the length measurement SEM. Next, an image of each dangerous place is acquired. In the present invention, addressing is performed in two stages. In the first addressing, an error due to the stage accuracy of the length measurement SEM is corrected, and in the next addressing, the position of the pattern to be measured is corrected. Since the range in which the stage stop accuracy of the length measuring SEM used in the present invention is guaranteed is 4 μm, the first addressing point is fixed at a magnification of 20 K at which one side of the visual field area (square) is 6.75 μm. Of course, addressing may be performed at a magnification smaller than this, that is, in a visual field region where one side is 4 μm or more, but this magnification is used as the magnification at the first addressing point this time.

  Thus, the first addressing point depends on the stage stop accuracy of the length measuring SEM to be used. For the second addressing point, select the one in the same field of view as the final length measurement point. In the following description, an example in the case of saving with 2048 × 2048 pixels will be described. If the area per pixel in the SEM image is the same, an SEM image having an area four times as large as one side of the field of view and 16 times larger in area than the case of storing an image with 512 × 512 pixels. You can get it. Of course, the length measurement point is included in this 2048 × 2048 region, but in addition to that, the second addressing point is also included. The image to be finally acquired was set to 900 nm on a side with 512 × 512 pixels. This corresponds to 150K in terms of magnification, but the magnification is not limited to this, and is not limited to this as long as the desired resolution can be obtained.

  In this example, since it was known in advance that a desired resolution could be obtained with this magnification, this magnification was used. Under this condition, the second addressing point includes a square region having a side of 900 nm, and one side thereof is four times 900 nm, that is, a square region of 3600 nm. This is because the number of pixels per side of the acquired image is quadrupled, so the image area is quadrupled as a side. If an image of this area is acquired, and the circuit design data and the pattern acquired by the SEM can be matched at an appropriate place / appropriate area in this area, the finally obtained image is at a position corresponding to the design data. It will be.

  FIG. 4 schematically shows a dangerous area output by the lithography simulator and the design rule checker, that is, a square area with a side of 900 nm centered on a place where an image is to be acquired and a square area with a side of 3600 nm including it. However, since the second addressing point is not fixed at this stage, it is not described. The process of specifying the second addressing point will be described later.

  For global alignment points performed in the first stage, alignment accuracy correction marks arranged on the chip are used. When manufacturing a semiconductor device, the intended device is created by superimposing it on the layer immediately before it, but it was placed and superposed to verify that the overlay was within specification. In the subsequent process, the overlay accuracy is confirmed by a dedicated device. As a typical example of this mark, there is a pattern called box-in-box as shown in FIG. 5, but the two squares are in different layers, and the deviation of the center coordinate between them is the alignment accuracy. Detect as.

  Since this pattern is placed at a position determined for each device layer, coordinates can be designated from design data. The template image used for alignment was design data. Of course, global alignment may be performed using a pattern other than the box-in-box pattern.

  Next, it is necessary to match the position of the image, design data, and SEM image acquired from the length measurement SEM. In general, there is no particular agreement regarding the origin of coordinates of design data, and the design data and the coordinates of the SEM image are not related to each other unless a correlation value between them is given. In FIG. 6, it is assumed that the shaded portion is design data and the outer frame is a wafer chip. The circuit design data is very difficult to handle in the same area as the chip area of the wafer because of its data size. In this example, circuit design data was prepared only for the area where the location to be measured exists. By doing so, the data can be easily handled.

  In order to match the positions of the design data and the SEM image as the respective image data, the coordinates of both specific positions may be obtained in the respective coordinate systems, and both may be set to the same coordinates. In this example, as shown in FIG. 7, since an H-shaped character pattern having an appropriate size is arranged in the design data, the coordinates of those same locations are used. Unlike the design data, the corner of the SEM pattern is a curve. Therefore, the line between the two straight portions is extended, and the portion where they intersect is the corner.

  Based on the above conditions, detailed information on the length measurement SEM recipe is defined. FIG. 17 is a flowchart for explaining the flow of the detailed information setting. First, the final length measurement point and its magnification are designated (S0001). In this example, the length measurement point and the magnification are specified first, but the present invention is not limited to this.

  Next, in order to determine a point for correcting the position of the SEM image, that is, an addressing point, an area to be a candidate is first determined. Specifically, it may be within a range in which the field of view can be changed by moving the electron beam path. In the case of a device in which a 15 μm region vertically and horizontally around the evaluation point corresponds to the region, the addressing point is within this range. Determine. In other words, it is only necessary to set the addressing points within a region having one side of 30 μm.

  The magnification and the field of view of the first addressing point are matched and cannot be corrected unless the length measurement SEM stage stops within the field of view at least, and this depends on the stage stop accuracy of the length measurement SEM. The size of the template as an addressing point is also determined from a square area having a side of 30 μm. In this example, as shown in FIG. 8, the size of the template image of the addressing point is limited to a square area having a side of 6750 nm from a square area having a side of 30 μm, and a place suitable for addressing is searched. (S0002) (S0003).

  In the case of this example, since the magnification of 20K was set as the magnification that gives a wider area than the stage stop accuracy, the length of one side of this field of view was 6750 nm. Of course, if the area is wider than this, there is no problem in the system because the template image is within the field of view even if there is an error in the stage movement of the length measurement SEM. In this way, this time, it was automatically obtained by applying the normalized correlation method using a square region having a side of 6750 nm as a template from a range of a square region having a side of 30 μm.

  In this example, an area suitable for addressing was automatically determined using the method described in “Iwanami Lecture Multimedia Informatics 5 Image and Spatial Information Processing p.56”. It was determined excluding the area.

  In this example, a cross-shaped pattern as shown in the above figure is defined as the first addressing point. Subsequently, the auto focus point and auto stigma point were automatically determined from the design data, and this was determined in the same manner as the method for determining the first addressing point. However, the auto focus point and the auto stigma point are different in that an appropriate region is defined in the same visual field region (a square region of 900 nm square in the present invention) as the final image acquisition magnification.

  Next, it is determined whether or not the area where the first addressing point is set may overlap the length measurement area (S0004). If a pattern exists in the addressing area, auto-focusing can be performed in that area. The operator determines whether to search for an addressing point in an area including the length measurement area (S0005) or to search for an addressing point in an area not including the length measurement area (S0006). Finally, the first addressing is performed. The position / area of the point is determined (S0007).

  Next, a place to be a second addressing point is determined. This is because the image area that is finally required is a square area of 3600 nm on a side including a square area of 900 nm square, from the area other than the image acquisition point. The addressing point may be determined by this method. However, since one pixel of the SEM image is about 13 nm square at the first addressing point, an addressing error inevitably occurs within this size range.

  In order to prevent misalignment at the second addressing point due to this error, instead of the 3600 nm square region, this time as a region 13 nm or more inside, a location that can be the second addressing point is searched from the 3550 nm square region. did. Since the image to be finally acquired only needs to be included in this square area of 3550 nm on one side, the second addressing point is set from within the square area of 3550 nm * 2-900 nm = 6200 nm on one side with the image acquisition point as the center as shown in FIG. searched. In the case of this example, the magnification (field area) of the second addressing point is set to the same magnification as that of the length measurement area (S0008). In this example, when creating a length measurement recipe, the second addressing point search area is set in the same manner as the first addressing point (S0013) before the position and area of the second addressing point are determined (S0013). The operator determines whether the second addressing area can overlap the length measurement area (S0010). The operator determines whether to search for an addressing point in a region including the length measurement region (S0011) or to search for an addressing point in a region not including the length measurement region (S0012), and then determines the second addressing point. Determine the position and area of.

  As a result of the search, when the second addressing point is determined as shown in FIG. 10, the midpoint of the line segment connecting the image acquisition point and the center of the second addressing point is set as the center point of the final acquired image (S0014). Since the final image acquisition region is a square region having a side of 3600 nm, this region includes the second addressing point and becomes the region of the image finally acquired by the length measurement SEM. Further, by performing addressing in this region and cutting out only a square region having a side of 900 nm, an image in which the positional relationship between the two is completely matched can be obtained by superimposing the design data and the SEM image.

  A recipe for acquiring an SEM image for each danger point output by the process simulator is created in this way. As a result of automatically obtaining the second addressing point from the square area of 6200 nm on the side centering on the image acquisition point this time, the result is as shown in FIG. In other words, this area has the most characteristic shape and is suitable as an addressing point. An area of the image finally acquired from the second addressing point and the image acquisition point is determined.

When forming a pattern on a wafer, for example, the following steps are performed. First, a coating type antireflection film having a thickness of about 60 nm was spin-coated on the wafer, and a chemical amplification type positive resist film having a thickness of about 200 nm was further spin-coated thereon. For this wafer, using a photomask generated from the design data used in this example, the projection optical system reduction ratio 1/4, exposure wavelength 193 nm, NA 0.73, coherence factor σ 0.75, and annular shield The exposure was performed under the same conditions as those used in the process simulator with a rate ε0.67 and a set exposure amount of 28 mJ / cm ² . After the exposure, the wafer was post-exposure baked (PEB) at 100 ° C. for about 90 seconds, and further immersed in a 0.21 N alkaline developer for about 60 seconds, followed by development to produce a transfer pattern on the wafer.

  An image of the dangerous part of the wafer is obtained by the length measuring SEM, and the recipe used is a recipe created based on the design data by the above method. Since the recipe is based on the design data, the addressing point does not function unless the design data and the position of the image acquired by the length measurement SEM can be associated with each other. In this example, a technique described in Japanese Patent Laid-Open No. 2002-328015 is used as the matching technique.

  By using this method, design data can be used as a template for a length measurement SEM, and an SEM image of a dangerous point is obtained. FIG. 12 shows a schematic diagram of the result of matching the acquired SEM image and the design data in a square area of 3600 nm on one side including the second addressing point and the dangerous part in the lower figure. This is a diagram in which the SEM image in the image area including the second addressing point and the design data corresponding to the position are overlapped, and both are the results of matching within this field of view.

  In this example, the second addressing point separated from the dangerous part is used as the reference pattern, and the position information on the image of the reference pattern and the design data of the part including the reference pattern and the dangerous part (the part to be measured) are included. Based on this, a reference position for measurement (for example, a line portion of a line pattern arranged on the left side of FIG. 12) is set. As shown in FIG. 12, the SEM image and the design data are superimposed by matching the second addressing point with the position on the SEM image corresponding to the second addressing point. By such superposition, the actual position information obtained from the actual SEM image and the ideal positional relationship between the reference pattern obtained from the design data and the measurement location are used to set the reference position for measurement. Can do.

  Particularly in the case of this example, there is a problem that it is difficult to accurately set the measurement reference position in the vertical direction of the drawing because the measurement location is a line pattern extending in the vertical direction of the drawing. However, by using the pattern extending in both the vertical and horizontal directions of the drawing as the second addressing point, it is possible to ensure the alignment accuracy in both the vertical and horizontal directions of the drawing, so that the measurement reference position in the vertical direction of the drawing can be accurately arranged. Become. For example, it is a very effective measurement standard setting method for evaluating how much the tip of the line pattern is reduced with respect to the design data.

  A specific length measurement method using the measurement reference position set by the above method will be described below.

  FIG. 13 is an example of an SEM image obtained by cutting out only the acquired image region from the acquired SEM image and design data superimposed thereon. In this visual field region, the positional relationship between the design data and the SEM image seems to be shifted. However, as a positional relationship between the two, since a part of the result of matching in a large region including this region is cut out. This is the correct answer.

  The SEM image acquired using this figure is evaluated against the design data. First, assuming that the evaluation is performed from the viewpoint of the line width detected by the design rule checker, the line width is measured with the SEM image in the field of view.

  In the case of the example of FIG. 13, it is possible to evaluate how much the actual pattern is formed with respect to the design data with reference to the design data.

  In the example of FIG. 14, the difference from the design data for the SEM image is used as the evaluation value. By comparing with the design data, it is possible to evaluate other than the line width. The distance between the line segment of the design data and the edge portion of the SEM image can be used as an evaluation item. In the case of the example of FIG. 14, it is possible to evaluate how much the vertical direction of the pattern is reduced, or the degree of left / right deviation with reference to design data. Although not shown in the figure, in the case of a pattern such as a contact hole, it is possible to propose a new contact hole evaluation method in which a measurement reference position, a contact hole upper part, and a contact hole bottom part are compared. In this case, the measurement reference positions of both the upper part of the contact hole and the bottom part of the contact hole may be settable.

  Furthermore, the data obtained in this way may be displayed as a wafer map. A specific method for displaying as a wafer map is described in, for example, Japanese Patent Application Laid-Open No. 2001-110862 (USP 6,765,204).

  These distances are values that can be obtained only when the SEM image and the corresponding design data can be correctly superimposed. In addition, the design data corresponding to the next process of the device pattern can be further overlapped to enable evaluation from another viewpoint.

  The square design data indicated by the dotted line in FIG. 15 indicates the next position as the semiconductor device manufacturing process, that is, the ideal position of the pattern formed thereon. Since there is no error in the overlay for each design data, the degree of overlap with the dotted line data may be an evaluation criterion. In this example, the distance between the dotted line design data and the pattern edge was measured. A new length measurement value could be calculated by comparing the pattern of the SEM image acquired in this way with the design data. The parameters required for these length measurements are not determined at the circuit design data stage, but are determined to be optimal from the SEM image after the SEM image is acquired. In this example, the linear approximation method was used for length measurement from the SEM image. The threshold value was 50% of the maximum secondary electron signal intensity.

  FIG. 16A shows the length measurement location in the SEM image. Secondary electron signals from the top to the bottom of the figure were detected. FIG. 16B shows the secondary electron signal intensity. The position on the right side of the figure is identified by a linear approximation method with a threshold value of 50%. Since the position on the left side of the figure has already been identified by matching the circuit design data and the SEM image, the dimension can be measured from the values of both. It was possible to measure the length at other locations by the same method.

  FIG. 18 is a conceptual diagram of the created recipe. By displaying such a conceptual diagram on a display device or the like during the recipe creation process, the recipe creator (operator) can create a recipe while determining whether the image acquisition area or the like is appropriate. it can. In the case of this example, by displaying the area to be set in the recipe creation process on the circuit design data, it is possible to confirm whether there is no overlap in the area scanned with the electron beam. Also, these areas can be confirmed easily by color-coding them and displaying their roles.

  For example, in the case of AP1 / AF, it can be seen that the area is the first addressing area and is the area where auto-focusing is performed. This confirmation is effective when performing electron beam scanning for addressing or auto-focus before length measurement, since the length measurement value may be different from that when there is no electron beam scanning.

  DESCRIPTION OF SYMBOLS 1 ... Charged particle source (electron gun), 2 ... Charged particle beam (electron beam), 3 ... Lens, 4 ... Sample, 5 ... Secondary particle, 6 ... Secondary particle detection system, 7 ... Control system, 8 ... X -YZ stage, 9 ... image display device, 10 ... design data management unit.

Claims (6)

A method of acquiring a pattern image of a sample and measuring a deviation between a measurement target pattern on the image and design data of the pattern,
With respect to the measurement target pattern on the sample, a reference pattern arranged at a spaced position and an image of an area including the measurement target pattern are obtained,
By performing template matching with the acquired image using a design data area including a reference pattern separated from the measurement target pattern as a template, the measurement target of the acquired image is obtained along with the reference pattern of the acquired image. Move the pattern relative to the design data including the reference pattern and the measurement target pattern to perform overlay ,
Measuring a deviation between an edge of the measurement target pattern of the acquired image moved relative to the design data with the reference pattern and an edge of the measurement target pattern of the design data, Pattern measurement method. In claim 1,
An image is acquired so that the reference pattern and the measurement target pattern are arranged on the same image. In claim 2,
A pattern measuring method, wherein the center of the image is positioned at an intermediate point between the reference pattern and the measurement target pattern. In claim 1,
The pattern measurement method, wherein the image is an image acquired by a scanning electron microscope. A pattern measuring device comprising a processing device for measuring a deviation between a measurement target pattern on a pattern image of a sample and design data of the pattern,
The processing unit, with respect to the measurement target pattern on the specimen, and obtains the reference pattern arranged at a position separated, an image of an area including the said measurement target pattern,
By performing template matching with the acquired image using a design data area including a reference pattern separated from the measurement target pattern as a template, the measurement target of the acquired image is obtained along with the reference pattern of the acquired image. Move the pattern relative to the design data including the reference pattern and the measurement target pattern to perform overlay ,
Measuring a deviation between an edge of the measurement target pattern of the acquired image moved relative to the design data with the reference pattern and an edge of the measurement target pattern of the design data, Pattern measuring device. From the acquired pattern image of the sample, a computer program that causes a computer to measure a shift between the measurement target pattern on the image and the design data of the pattern,
The program is stored in the computer.
With respect to the measurement target pattern on the sample, a reference pattern arranged at a spaced position and an image including the measurement target pattern are acquired,
By using the design data region including the reference pattern separated from the measurement target pattern as a template and performing template matching with the acquired image, the measurement target of the acquired image is obtained along with the reference pattern of the acquired image. The pattern is moved relative to the design data including the reference pattern and the measurement target pattern, and superposition is performed .
Characterized in that a shift between an edge of the measurement target pattern of the acquired image moved relative to the design data with the reference pattern and an edge of the measurement target pattern of the design data is measured. Computer program.