JP2012099754A - Design template generating device, recipe generating device and microscopic image matching device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve such a problem that the success rate of pattern matching may be low because a low reliability pattern and a high reliability pattern are evaluated uniformly.SOLUTION: A design pattern or a pattern of design template for use in pattern matching is weighted so that a region of low reliability pattern and a region of high reliability pattern can be discriminated.

Description

  The present invention relates to a design template generation technique suitable for pattern matching processing in a semiconductor process, a recipe generation technique, and a matching technique between a generated design template and a microscopic image.

  In the semiconductor process, pattern matching processing is performed between each image and the design template when acquiring an SEM (scanning electron microscope) image, an OM (Optical Microscope) image, or the like or in signal processing for the acquired image.

  However, the design template includes a pattern with low reliability. For example, highly repetitive patterns and dummy patterns necessary for semiconductor processes are included. For this reason, a decrease in the success rate of the pattern matching process is a problem. Further, the conventional pattern matching process uniformly evaluates a pattern with low reliability and a pattern with high reliability. For this reason, the success rate of the pattern matching process is not necessarily high.

  Usually, an algorithm based on the uniqueness of a pattern is used in the pattern matching process. However, this algorithm does not eliminate the influence of non-unique patterns (highly repeatable patterns) included in the design template, and even if the design template includes highly unique patterns, the correct position In some cases, pattern matching cannot be realized.

  In addition, there are problems inherent to dummy patterns. The dummy pattern includes a pattern that is not actually resolved in the semiconductor process, and the accuracy of the pattern shape is not required. For this reason, when a dummy pattern is erroneously used as a matching reference, there is a problem that pattern matching cannot be performed at a correct position.

  The present invention has been made in consideration of the above points, and avoids pattern matching to a low-reliability pattern, which causes a decrease in the success rate of pattern matching, and improves the success rate of pattern matching. With the goal.

  Therefore, the present inventors weight each design pattern or pattern constituting the design template used for pattern matching so that a pattern area with low reliability and a pattern area with high reliability can be distinguished by weight.

  According to the present invention, the success rate of pattern matching between a microscopic image (charged particle beam image or OM image) and a design pattern can be improved. Problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.

The figure which shows the example of a form of a microscope system. The figure explaining the production | generation image of a recipe. The figure which shows the example of a form of a matching process. The figure explaining the weighting process which considered the reliability. The figure explaining the image of a weighting process. The figure which shows the example of a setting of the attention area | region using the weighting result. The figure explaining the mask process using the weighting result. The figure which shows the process example with respect to an ultra-fine dummy pattern. The figure explaining the weighting process with respect to the dummy pattern for polishing processes. The figure which shows the example of the weight setting screen and the confirmation screen of a setting result. The figure which shows the other example of a form of a microscope system. The figure which shows the example of a form of a matching process.

  Hereinafter, based on the drawings, exemplary embodiments of a recipe generating device (method) and a design template generating device (method) according to the present invention will be described.

<Example 1>
(System configuration)
FIG. 1 shows a system configuration example of a microscope system according to an embodiment. The microscope system shown in FIG. 1 includes a recipe generation computer (recipe generation apparatus) 1 and a scanning electron microscope 11.

  The recipe generation computer 1 inputs the measurement coordinate information 2 and the design data 3 from a storage device (not shown), and generates a recipe 4 and a design template 5. The recipe 4 includes commands, settings, and parameters for the scanning electron microscope 11. The recipe 4 includes coordinate information of an area imaged by the scanning electron microscope 11. The design template 5 is a partial area of the design data 3 corresponding to the imaging area of the SEM image. In the case of this embodiment, the design template 5 is given a weight for pattern matching to each region. The recipe 4 is automatically generated by the recipe generation computer 1. However, the recipe 4 may be generated through a user's manual operation on the recipe generating computer 1. The recipe 4 and the design template 5 are transferred to the scanning electron microscope 11.

  The scanning electron microscope 11 loads the recipe 4 into a control device (not shown), and executes the following process based on the recipe 4. A control device (not shown) controls the overall operation of the scanning electron microscope 11. For example, a deflector 14 and a stage 19 which will be described later are controlled to control the alignment of the SEM image acquisition region.

  First, the scanning electron microscope 11 emits an electron beam 13 from an electron gun 12. The electron beam 13 is deflected by the deflector 14 and scans a specified area on the wafer 15. Reflected electrons are generated from the region irradiated with the electron beam 13. The reflected electrons are detected by the reflected electron detector 16 and amplified by the amplifier 17. By this amplification, an SEM image is generated. Note that secondary electrons may be detected to form an SEM image. The image processing unit 18 measures the length of the designated area on the SEM image. The wafer 15 is placed on a stage 19 that is movable in the X-axis and Y-axis directions. Here, the length measurement process by the image processing unit 18 is limitedly performed on the entire design data 3 or the area surrounding the area designated as the length measurement coordinate.

(Recipe automatic generation process)
FIG. 2 shows a procedure for generating the recipe 4 by the recipe generating computer 1. As shown in FIG. 2, in the case of this embodiment, the recipe generating computer 1 generates a weighted design template 5 before generating the recipe 4.

  First, the recipe generation computer 1 inputs original design data 3. FIG. 2 shows a pattern area 24 cut out from the original design data 3 (the part corresponding to the designated area of the length measurement coordinate information 2). At this stage, all patterns are treated uniformly in the pattern area 24 regardless of the difference in reliability of each pattern. For this reason, all the patterns in the pattern region 24 are represented by the same density.

  Next, the recipe generating computer 1 weights each pattern in the pattern area 24 based on a rule to be described later (step 21). Basically, a low weight is given to a pattern with low reliability, and a high weight is given to a pattern with high reliability. As a result, weighted design data 22 (design template 5) is generated. In FIG. 2, the weighting effect is expressed by the difference in the density of individual patterns constituting the design template 5. As shown in the figure, a pattern with lower reliability is expressed at a lower density. The weighting method includes a method of expressing as a difference in density (convenient for confirmation by an operator) and a method of embedding or associating each pattern as numerical data. In the case of FIG. 2, the weight is assigned in consideration of not only the number of edges but also the pattern width, pattern area, and other pattern uniqueness.

  When the weighted design data 22 is obtained, the recipe generation computer 1 executes an automatic generation process of the recipe 4 in consideration of the weighting (step 23). Here, the recipe generation computer 1 uses, for example, the number of edges of each pattern in the design data 22 as an evaluation value. Conventionally, only the number of edges is obtained without considering high / low reliability, and this is used as an evaluation value. However, in the case of the recipe generation computer 1 according to this embodiment, the edge evaluation value is calculated by multiplying the number of edges by the weighting coefficient (n). In FIG. 2, the multiplication value of the number of edges and the weighting coefficient (n) is represented by a general formula that adds the number of types of weights. The recipe generation computer 1 determines a pattern to be used for pattern matching based on the calculated edge evaluation value, and registers the coordinates in the recipe 4.

  In the case of this embodiment, the recipe generating computer 1 determines the pattern with the highest calculated edge evaluation value as the pattern matching pattern. However, the calculation of the edge evaluation value does not cover all patterns, but the edge evaluation value is calculated only for the pattern matching pattern candidates determined by the conventional method, and the most of the plurality of pattern candidates. A candidate with a high evaluation value may be determined as the final solution.

(Matching process)
FIG. 3 shows a matching processing procedure for the SEM image executed in the image processing unit 18. In the case of this embodiment, the image processing unit 18 executes a process for matching a SEM image and a design template without weight and obtaining a plurality of candidates having a high degree of coincidence (step 31).

  In the case of FIG. 3, the image processing unit 18 assigns a weight based on reliability to each pattern in the region corresponding to the SEM image in the original design template 3 to generate a weighted design template 5 (step 32). ). The contents of step 32 are the same as the contents of step 21 executed by the recipe generating computer 1. In this case, although not shown in FIG. 1, the original design template 3 is input to the image processing unit 18. However, as shown in FIG. 1, since the weighted design template 5 is input to the scanning electron microscope 11 from the recipe generating computer 1, the design template 5 may be used as it is.

  Thereafter, the image processing unit 18 uses the weighted design template 5 to calculate the degree of coincidence for each of the plurality of pattern candidates (step 33). The degree of coincidence is calculated by, for example, the same arithmetic processing as that of the edge evaluation value. In step 33, the image processing unit 18 searches for a candidate with the highest degree of coincidence and sets it as the final solution.

(Pattern weighting process taking reliability into account)
FIG. 4 shows a specific example of a pattern weighting process taking reliability into consideration. In the case of this embodiment, the pattern weighting process is executed in each of the recipe generating computer 1 and the image processing unit 18. Therefore, hereinafter, the execution subject is simply referred to as a computer.

  The computer includes a plurality of parameter sets (a set of weights) that combine the original design data 3 or the design template 5, parameters for detecting a pattern with low reliability specified in advance by the user, and weights for the pattern. ) Is input (step 41). Here, the parameters for detecting a specific pattern as a pattern with low reliability include, for example, a pattern width unique to dummy data.

  Next, the computer detects a corresponding pattern from the design data 3 or the design template 5 in accordance with a parameter designated as a low-reliability pattern, and adds a designated weight to the pattern (step 42). .

  The computer repeatedly executes the processing of step 42 until there is no parameter designated as a pattern with low reliability (step 43).

  If a negative result is obtained in step 43, the computer outputs design data or a design template in which a designated weight having a designated pattern with low reliability is attached to a storage device (not shown) (step 44). The input / output design data or design template may be an image. Also, the pattern can be deleted by designating the weight as zero.

(Specific examples of unreliable patterns)
FIG. 5 shows a specific example of a pattern with low reliability. In the case of this embodiment, a pattern with high repeatability is treated as a pattern with low reliability, and a small weight is given. In this embodiment, the shorter the repetition period or the greater the number of repetitions, the lower the reliability.

  For this reason, the computer acquires a profile for each of the vertical direction (y direction) and the horizontal direction (x direction) of the pattern region 51. And a computer detects a pattern with high periodicity in each profile as a pattern with low reliability, and weights it. The pattern region 52 represents a state in which the weights of a plurality of patterns extending in the vertical direction (y direction) are set low as a result of acquiring the profile in the horizontal direction (x direction). The pattern area 53 represents a state in which the weights of a plurality of patterns extending in the horizontal direction (x direction) are set low as a result of acquiring the profile in the vertical direction (y direction).

  Next, the computer preferentially combines the weights with lower weights in each of these two pattern areas 52 and 53 to generate a pattern area 54. As can be seen from the pattern region 54, the pattern is expressed in three different densities. This is because the appearance interval of the pattern extending in the horizontal direction (x direction) is wider than the appearance interval of the pattern extending in the vertical direction (y direction). Therefore, the reliability of the pattern extending in the vertical direction (y direction) with the smallest appearance interval is set to the lowest, and the reliability of the pattern extending in the horizontal direction (x direction) with the next smallest appearance interval is set to the next lowest. . In FIG. 5, the reliability of the cross pattern is the highest.

  In the case of FIG. 5, the example of the design template (pattern region 54) in which all the patterns in the pattern region are weighted has been described. It may be erased from within. An example is shown in FIG. In this case, the two patterns excluding the cross pattern are deleted from the pattern area 61, and the pattern area 62 is the final output. Here, the cross pattern may be set as a region of interest (ROI). If the pattern region 62 including only one attention region is used for the pattern matching processing in this way, the possibility of erroneous detection during pattern matching is significantly reduced. However, when a plurality of attention areas appear adjacent to each other in the pattern area 62, the effect of reducing false detection may not always be realized. Thus, the effect of reducing false detection varies depending on other patterns located around the attention area.

  In addition, a mask can be set in an area with low reliability so that it is not used for pattern matching. An example is shown in FIG. The pattern area 71 corresponds to the original design data, and the pattern area 72 corresponds to the weighted design data. Here, when a mask (shaded area in the figure) is set so as to hide only a pattern lower than a predetermined threshold, a pattern area 73 is obtained. By using this pattern region 73 for the design template 5, the matching accuracy of the pattern matching process can be increased.

  Note that the smaller the pattern size, the lower the reliability, and the larger the size, the higher the reliability. In addition, the greater the number of repetitions of unevenness, the lower the reliability, and the smaller the number of repetitions, the higher the reliability. In addition, the larger the margin for specifying connectivity, the lower the reliability, and the smaller the margin for specifying connectivity, the higher the reliability.

  In addition, a pattern with low reliability includes a dummy pattern necessary for a semiconductor process. The dummy pattern includes a SRAF (sub resolution assist features) pattern 81 as shown in FIG. The SRAF pattern 81 is a dummy pattern that is very thin and cannot be resolved into a sample. Further, the dummy pattern includes a tile-like dummy pattern 91 added to disperse stress when a sample as shown in FIG. 9 is subjected to a polishing process (CMP: chemical mechanical polishing).

  For example, if the width of the SRAF dummy pattern is registered as a parameter in step 41 and a low weight (including zero) is added to a pattern having a width equal to or smaller than the parameter, the design template 5 used for pattern matching can be used. The pattern to be removed can be removed.

  In the case of this embodiment, for example, a low weight is given to the pattern at the position of the pattern that disappears when the design data 3 or the design template 5 is expanded or contracted to the specified line width. . In general, however, for very thin patterns such as SRAF patterns, it is preferable to set the weight to zero from the beginning and delete the corresponding patterns.

  If a tile-like dummy pattern for CMP or the like is included in the SEM image, the vertical and horizontal sizes are registered as parameters in step 41, and a small value is weighted to the pattern that matches the parameter. desirable. These weighting processes are executed based on, for example, detection of a dummy pattern by computer image processing.

(GUI for weight setting)
Here, an example of a GUI used by a user for setting weights is shown. The GUI display function may be realized as a program executed on a computer (not shown), or may be executed by the recipe generation computer 1 or the image processing unit 18. Incidentally, a display device is connected to each computer, and a GUI is displayed on the screen of the display device. The computer here functions as a display control device.

  FIG. 10 shows a GUI display example. The parameter setting window 101 shown in FIG. 10 includes a weighting confirmation window 102, a repeated pattern setting window 111, an ultra fine pattern setting window 121, and a tiling pattern setting window 131.

  The weight confirmation window 102 includes an input pattern field 103 and an output pattern field 104. In the output pattern column 104, the result of applying the set parameters to the image in the input pattern column 103 is indicated by, for example, a difference in pattern density. That is, the output pattern column 104 constitutes a weighting result confirmation screen. Examples 105, 106 and 107 of images appearing due to the difference in the set parameters are shown. You can see how the displayed pattern changes due to the difference in parameters. In the case of FIG. 10, a pattern with low reliability is represented not by shading but by low-density shading. This is because the visibility in the patent application is taken into consideration.

  The repetitive pattern setting window 111 is a parameter setting screen for each repetitive pattern. In the case of FIG. 10, setting screens are prepared for three repeating patterns. In each setting screen, input fields for weights to be given to the corresponding patterns, input fields for the period of the corresponding pattern, the number of repetitions, size, margin, and the like are arranged. In this example, a numerical value is directly entered in each input field. However, it may be possible to select and input from a plurality of candidates prepared in advance in a pull-down format or a slide format. In addition, the number of registered repetitive patterns can be increased or decreased by using an “add button” or a “delete button”.

  The extra fine pattern setting window 121 is a parameter setting screen for removing extra fine dummy patterns such as SRAF. In the case of FIG. 10, the extra-fine pattern setting window 121 has an input column for the weight to be given to the corresponding pattern and an input column for the size used for determination. Also in this case, it may be possible to select and input from a plurality of candidates prepared in advance in a pull-down format or a slide format.

  The tiling pattern setting window 131 is a setting screen for tiling pattern parameters for CMP. In the case of FIG. 10, setting screens are prepared for two tiling patterns. In each setting screen, an input field for weights to be given to the corresponding pattern and an input field for the size used for determination are arranged. Also in this case, it may be possible to select and input from a plurality of candidates prepared in advance in a pull-down format. Further, the number of registered repetitive patterns can be increased or decreased by operating the “add button” or “delete button”.

  If the GUI shown in FIG. 10 is used, the user can adjust the weight for each pattern while confirming the effect of the setting. The user can also use it to check whether the currently used weight is appropriate.

  In each setting window, the size and weight can be input independently. However, a mechanism may be adopted in which the weight is automatically set based on other conditions when the size is determined.

(Summary)
As described above, by using the microscope system according to the present embodiment, it is possible to optimize the alignment coordinates registered in the recipe 4 and the alignment coordinates at the time of pattern matching of the captured SEM image. As a result, the possibility of using a pattern with low reliability during pattern matching processing can be reduced, and the accuracy of pattern matching can be increased.

<Example 2>
In the case of the above-described embodiment, the case where the design template 5 after the weighting process is used in each of the recipe generation computer 1 and the image processing unit 18 has been described. However, the design template 5 after the weighting process described above may be used only in either the recipe generating computer 1 or the image processing unit 18.

  FIG. 11 illustrates a case where the design template generation computer 152 is mounted only on the scanning electron microscope 151. In FIG. 11, the same reference numerals are given to the portions corresponding to FIG. 1. In FIG. 11, for convenience of explanation, the design template generation computer 152 is shown as a separate device from the image processing unit 18, but the functions of the design template generation computer 152 may be executed by the image processing unit 18.

<Example 3>
In the description of the above-described embodiment, as shown in FIG. 3, a case where the weighted design template 5 is applied to a pattern after narrowing down to a plurality of candidates by a conventional method in advance and one pattern is finally specified. explained. However, as shown in FIG. 12, the design template 5 after weighting and the SEM image may be directly compared to determine a pattern to be used for matching (step 161).

<Other forms>
In the above-described embodiment, the case where the present invention is applied to the scanning electron microscope has been described. However, the present invention can also be applied to recipe creation and pattern matching for captured images in a microscope system using charged particle beams other than electrons or an optical microscope system.

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

  Each of the above-described configurations, functions, processing units, processing means, and the like may be realized by hardware by designing a part or all of them with, for example, an integrated circuit. Further, each of the above-described configurations, functions, and the like may be realized by software by interpreting and executing a program that realizes each function by the processor. Information such as programs, tables, and files for realizing each function can be stored in a recording device such as a memory, a hard disk, or an SSD (Solid State Drive), or a recording medium such as an IC card, an SD card, or a DVD.

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

1: Recipe generation computer 2: Measurement coordinate information 3: Design data (original)
4: Recipe 5: Design template (weighted)
11: scanning electron microscope 12: electron gun 13: electron beam 14: deflector 15: wafer 16: backscattered electron detector 17: amplifier 18: image processing unit 19: stage

Claims (9)

In an apparatus that generates a design template used for pattern matching with a microscopic image,
Means for reading design data or a design template from the first storage area;
Means for reading from the second storage area a parameter that defines at least one pattern with low reliability and a weight associated with each pattern;
Means for detecting the low reliability pattern included in the design data or the design template based on the parameters;
And a means for adding the corresponding weight to the detected pattern. In the design template production | generation apparatus of Claim 1,
The design template generation apparatus, wherein the pattern having low reliability is a pattern having repeatability. In the design template production | generation apparatus of Claim 1,
The design template generation apparatus, wherein the pattern with low reliability is a dummy pattern. In the design template production | generation apparatus of Claim 3,
The design template generation apparatus, wherein the dummy pattern is a pattern for a semiconductor process. In the design template production | generation apparatus of Claim 1,
A design template generation apparatus comprising: display control means for outputting an operation screen for inputting the parameter and / or weight value to a display apparatus. In the design template production | generation apparatus of Claim 5,
The display control means displays a confirmation screen for a weighting result based on the input parameter and / or weight value on the operation screen. In an apparatus for generating a recipe used for acquiring a microscopic image,
Means for reading design data or a design template from the first storage area;
Means for reading from the second storage area a parameter that defines at least one pattern with low reliability and a weight associated with each pattern;
Means for detecting the low reliability pattern included in the design data or the design template based on the parameters;
Means for adding the corresponding weights to the detected pattern;
A means for determining a pattern to be used for acquiring a microscopic image by using, as an evaluation value, a multiplication processing result of the number of edges of each pattern after weighting and a corresponding weight,
And a means for registering the coordinates of the determined pattern in the recipe. In a device that matches design templates and microscopic images,
Means for reading design data or a design template from the first storage area;
Means for reading from the second storage area a parameter that defines at least one pattern with low reliability and a weight associated with each pattern;
Means for detecting the low reliability pattern included in the design data or the design template based on the parameters;
Means for adding the corresponding weights to the detected pattern;
A microscopic image matching apparatus comprising: a weighted design template and a microscopic image matching unit. In the microscopic image matching apparatus according to claim 8,
The microscopic image matching apparatus, wherein the microscopic images are a plurality of microscopic images narrowed down as candidates by matching with an unweighted design template.

Images