JP5389840B2 - Pattern shape evaluation method and pattern shape evaluation apparatus - Google Patents

Description

  The present invention belongs to a lithography process, an etching process, and other semiconductor process fields where a pattern faithful to design data is required to be resolved and processed in microfabrication. The present invention relates to a technique suitable for improvement and improvement of a semiconductor process and the like.

  A resist pattern created on a wafer by lithography may not be formed according to the design circuit due to optical problems, topologies on the wafer, foreign matters, or the like. In recent years, it has become possible to predict this type of pattern deformation to some extent by applying simulation technology or the like. However, there are still many cases where the pattern on the wafer does not match the simulation result. Therefore, in order to judge the quality of the pattern resolved on the wafer, the pattern is imaged using a CD-SEM (Critical-Dimension Scanning Electron Microscope), and the pattern is measured through size measurement. Evaluation is good or bad.

  However, pattern deformation and defects are often two-dimensional patterns having information in the X and Y directions, and a method for quantitatively evaluating them has not been established. In addition, the roughness of the edge portion becomes obvious as the pattern becomes finer, and it is difficult to distinguish it from the deformation of the systematic pattern.

  As a result of intensive studies to solve the technical problem, the present inventor has come up with the idea that the use of the following method is effective in solving the problem.

  In the present invention, a plurality of formed patterns formed to have the same shape and the same size are observed with an electron microscope, and a plurality of microscopic images are acquired. Next, the plurality of microscopic images are integrated to generate an integrated image. On the other hand, a reference image corresponding to the microscopic image is acquired from the storage area. Next, the pattern width of the predetermined pattern included in the integrated image is measured at multiple points in the one-dimensional direction. On the other hand, for the same pattern as the integrated image, the pattern width of the reference image is measured at multiple points at the same point as the integrated image. After that, a cross-correlation coefficient is calculated between the multipoint measurement result sequence acquired for the integrated image and the multipoint measurement result sequence acquired for the reference image, and the correlation degree between the integrated image and the reference image is calculated from the calculation result. Determine.

  According to the present invention, by using an integrated image as an evaluation target image, the influence of non-stationary elements such as roughness can be removed from the determination result for the degree of correlation. As a result, pattern evaluation can be performed quantitatively.

  Problems, configurations, and effects other than those described above will become apparent from the following description of embodiments.

The figure explaining the basic composition of the scanning electron microscope for measurement (CD-SEM). The figure explaining the integrated image which synthesize | combined the image (captured image) imaged with CD-SEM, and the said captured image. The figure explaining an example of a simulation image and a design circuit image. The figure explaining a mode that the pattern width | variety of an integrated image is measured at multiple points. The figure which graphed the result of having measured an integrated image (real pattern) and a simulation image (reference image) at multiple points with respect to the measurement point. The figure which graphed the change of the cross correlation coefficient along a measurement point. The flowchart explaining the flow of pattern evaluation.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. The embodiment of the present invention is not limited to the embodiments described later, and various modifications can be made within the scope of the technical idea.

<Example 1>
(1) Configuration of CD-SEM Here, a case where a resist pattern formed by a lithography process is evaluated will be described. Of course, the present invention can also be used for pattern evaluation in an etching process and other semiconductor processes.

  In the lithography process, a CD-SEM is used for confirming the sizes of various patterns transferred onto the wafer. FIG. 1 shows a configuration example of the CD-SEM 100. The CD-SEM 100 includes an electron beam source 101 that emits an electron beam, a converging lens 102 that converges the electron beam, and a deflector (deflection coil) that deflects the electron beam so that the electron beam scans the surface of the wafer (sample) 105. ) 103, an objective lens 104 that irradiates the fine pattern 106 on the wafer 105 with an electron beam, a detector 107 that detects secondary electrons or reflected electrons emitted from the fine pattern 106, and secondary electrons or reflected electrons And a computer 108 for measuring the pattern width from the acquired image. Here, the computer 108 has an arithmetic unit, a storage area, and an interface. The computer 108 stores the image acquired through the detector 107 in a storage area, and measures the pattern width based on the image. Further, as will be described later, the computer 108 calculates the degree of cross-correlation between the actual measurement pattern and the reference pattern, and displays an interface screen on a display device (not shown) as necessary.

  In addition, the CD-SEM also includes a sample stage (not shown) on which the wafer 105 is mounted, an XY stage (not shown) for positioning the sample stage, and a Z-axis stage. FIG. 1 also shows an enlarged view of the peripheral region 110 including the fine pattern 106, and an image 130 obtained by observing FIG. 120 from the incident direction of the electron beam.

  The measurement of the fine pattern 106 is performed by scanning the formation area of the fine pattern 106 transferred onto the wafer 105 with an electron beam, and measuring the size of the fine pattern 106 with respect to an acquired image (CD-SEM image) of the area. Realized. Note that by preparing a recipe for managing length measurement points, wafer maps, and the like, the CD-SEM 100 can automatically measure the pattern width at any length measurement point. The measured data is used for evaluation of device manufacturing conditions and process management. In particular, in the evaluation of device manufacturing conditions, thousands of imaging and measurement are performed.

(2) Current Status and Problems of Current Apparatus The current CD-SEM (hereinafter referred to as “current apparatus”) has the same configuration as the CD-SEM 100 shown in FIG. The current apparatus is also used for state observation and measurement of the pattern formed on the wafer 105, and is used for evaluation of appropriate process conditions, optical conditions, and pattern formation accuracy after processing. In the current apparatus, the correlation between the pattern width and the length of the gap between patterns and the process condition is used as an evaluation index. However, with the miniaturization of patterns, quantitative measurement has become difficult with respect to the influence of optical systems and the deformation of patterns due to mask defects. In addition, as the miniaturization progresses, the influence of pattern roughness increases, and it becomes difficult to distinguish between deformation and defects.

(3) Outline of Evaluation Method According to Embodiment In order to solve the technical problem described above, the computer 108 implements the following two processes through execution of a computer program.
(a) A plurality of CD-SEM images (microscopic images) are obtained for a plurality of patterns 106 formed to have the same shape and size on the wafer 105, and the plurality of CD-SEM images are aligned with each other. An integrated image is generated by combining (integrating).
(b) The pattern width of the same pattern appearing in the integrated image and the reference image is measured in a one-dimensional direction, and the pattern width measurement result sequence acquired for the integrated image and the pattern width measurement result sequence acquired for the reference image A cross-correlation coefficient representing the degree of correlation between the two is calculated for each measurement point.

  In the case of this embodiment, the integrated image refers to an image obtained by superimposing a plurality of CD-SEM images acquired from the same position (same in-chip coordinates) in a plurality of chips existing on the same wafer. This example corresponds to a case where CD-SEM images acquired from different shots on the same wafer are integrated.

  However, an integrated image may be generated by integrating CD-SEM images at the same shot position on different wafers, and CD-SEM images corresponding to patterns having the same shape and size within the same shot are integrated. You may do it. The exposure conditions for forming the pattern 106 on the wafer 105 are not necessarily the same. As long as the shape and size formed on the wafer 105 are the same, the exposure conditions (for example, acceleration voltage) may be different.

  FIG. 2 shows an example of generating an integrated image. In the case of FIG. 2, the CD-SEM 100 acquires CD-SEM images 1, 2, and 3 of the three patterns 106 formed on the wafer 105 under the same conditions, that is, CD-SEM images 201, 202, and 203. In the case of this description, the same condition means that the optical conditions and process conditions at the time of exposure and imaging are the same. Note that the positions in the imaging field of the pattern 106 to be acquired of the CD-SEM images 201, 202, and 203 are determined to be the same as much as possible. By aligning the positions in the imaging field of view, it is possible to avoid the difference in distortion amount depending on the position in the field of view from being superimposed on the integrated image.

  The contours of the patterns appearing on the three CD-SEM images 201, 202 and 203 shown in FIG. 2 are unclear, and the contours are variously deformed. This is because the individual CD-SEM images 201, 202 and 203 are superimposed with randomly generated deformations such as roughness.

  On the other hand, in the integrated image 204 generated by integrating the three CD-SEM images 201, 202, and 203, randomly generated pattern deformation is canceled out and is more intrinsic than the individual CD-SEM images before synthesis. The pattern deformation is remarkable and clear.

(4) Reference Image Example In this embodiment, an image to be compared with the integrated image 204 is referred to as a reference image. The reference image includes a simulation image, a circuit design image, a CD-SEM image in which it has been confirmed that no defect exists (hereinafter referred to as “defect-free CD-SEM image”), and the like. FIG. 3 shows a simulation image 301 and a design circuit image 302 as examples of the reference image. The simulation image 301 is an image generated from the design circuit image 302, for example, and is formed by partially expanding the width of two patterns located near the center of the imaging field of view, so that the gap between the patterns is partially It shows how it is formed narrowly.

(5) Pattern Width Multipoint Measurement Processing Here, pattern width multipoint measurement processing by the computer 108 will be described. FIG. 4 shows a pattern width multipoint measurement model 402 for the integrated image 401. FIG. 4 illustrates the integrated image 401, but the same method is applied to the above-described multipoint measurement processing of the pattern width of the reference image.

  In this specification, the pattern width includes not only the width of an arbitrary pattern but also the width of a gap between patterns. In the case of FIG. 2 and FIG. 3, the white portions correspond to the gaps between the patterns, and the black areas or shaded areas correspond to the pattern areas. Therefore, the measurement of the pattern width in FIG. 4 means the measurement of the length of the gap between patterns. Hereinafter, the shape of a gap between patterns to be measured is also referred to as a pattern.

  In the integrated image 401 shown in FIG. 4, the pattern to be measured extends linearly from the upper end to the lower end in the imaging field having a rectangular shape, and the gap between the patterns is narrow near the center of the pattern. A multipoint measurement model 402 shown in the figure corresponds to each measurement point of the pattern width. FIG. 4 shows how the pattern width is measured over n points at regular intervals from the upper end side to the lower end direction in the imaging field of view.

  Note that the current apparatus uses a value obtained by averaging the pattern widths (widths of gaps between patterns) measured at multiple points (1 to n points) for one CD-SEM image. However, when the pattern width is averaged, information on the pattern deformation portion is lost.

  Therefore, in this embodiment, the pattern width value measured for each point is used as it is without averaging. That is, as shown in FIG. 5, the computer 108 according to the present embodiment, as shown in FIG. The degree of correlation between the integrated image and the reference image is obtained as a cross-correlation coefficient for each measurement position.

  In the figure, the vertical axis represents the pattern width 501 and the horizontal axis represents the measurement position 502. In the figure, the light line indicates the locus of the pattern width measured for the integrated image, and the dark line indicates the locus of the pattern width measured for the reference image. The computer 108 is equipped with a function for displaying the screen shown in FIG. 5 on the display device as one of the interface screens as necessary. By displaying the screen, the operator can visually observe and determine the degree of pattern shape matching between the accumulated image and the reference image. For example, it can be used to extract a problem part. However, in this embodiment, the degree of matching of the pattern shape between the accumulated image and the reference image is determined through signal processing described below.

(6) Evaluation by calculating cross-correlation coefficient for each measurement point In the case of this embodiment, the computer 108 measures the pattern width measured for n measurement points (measurement positions) for each of the integrated image and the reference image. This point sequence (measurement result sequence) was regarded as a waveform (function), and the degree of shape matching (correlation degree) between the two at each measurement point was verified by the cross-correlation coefficient. Various formulas have been conventionally proposed as formulas (functions) suitable for calculating cross-correlation between two time series data.

In this embodiment, treats each measurement result columns as the n discrete time-series data, the correlation coefficient r _k between two measurements column is calculated by the following equation.

In (Expression 1), x represents time-series data corresponding to the accumulated image, and y represents time-series data corresponding to the reference image. However, in (Expression 1), m1 _k and m2 _k are given by the following expressions, respectively.

The value of cross-correlation coefficients r _k represents the similarity of the whole between two time series data (waveform). The value of cross-correlation coefficients r _k expresses the degree of correlation when comparing two time-series data (sequence) together with values from +1 to -1. If all the comparison results between two time series data corresponding to the same time point (measurement point) are all +1, it is determined that there is a correlation, if all are zero, it is determined that there is no correlation, and all are -1. It is determined that there is a negative correlation. Therefore, the value of the cross-correlation coefficient r _k is closer to 1, the two time-series data (waveform) can be said degree of matching is high.

  FIG. 6 shows an example of the calculation result of the correlation coefficient for each position when one time series data (waveform) is fixed and the position of the other time series data (waveform) is shifted at equal intervals. In the case of FIG. 6, it can be seen that the cross-correlation coefficient 602 is 0.9 or more at the maximum, that is, a very strong correlation is shown at the approximately 40th measurement point.

  The computer 108 calculates the cross-correlation coefficient in this way, and compares the maximum value with a threshold value (for example, 0.6). If the maximum value of the cross-correlation coefficient exceeds the threshold value, the computer 108 determines that the degree of coincidence between the integrated image and the reference image is high. On the other hand, when the maximum value of the cross-correlation coefficient does not exceed the threshold value, the computer 108 determines that the degree of coincidence between the integrated image and the reference image is low. Note that the threshold value suitable for the determination differs depending on the type of device to be determined and the process conditions.

  Therefore, when the output result as shown in FIG. 6 is obtained, it can be seen that the simulation image 301 has a high degree of coincidence in shape with the corresponding portion (measurement portion) of the integrated image 204 (that is, the actual pattern). From this, it can be quantitatively confirmed that the simulation image 301 highly simulates the actual pattern.

  When the reference image to be compared with the integrated image 204 is the design circuit image 302, it can be quantitatively confirmed that the actual pattern is formed as designed.

(7) Summary FIG. 7 schematically shows the flow of the processing described above. The computer 108 reads the reference image 711 from the storage area, and multipoints the pattern width of the pattern area to be measured (712). Next, the computer 108 graphs the time-series data of the pattern width (713). In parallel with this processing, the computer 108 acquires a plurality of CD-SEM images 721 from a plurality of points on the wafer, aligns them, and generates an integrated image 722 that is integrated. Thereafter, the computer 108 reads the integrated image 722 from the storage area, and performs multipoint measurement on the pattern width of the pattern area to be measured (723). Next, the computer 108 graphs the time-series data of the pattern width (724).

  Thereafter, the computer 108 calculates a cross-correlation coefficient between the two time-series data (731), and converts the cross-correlation coefficient calculated for each measurement point to the initial value of the simulation program or the initial value of the appropriate circuit design program. Provide feedback (732-735). This feedback processing makes it possible to create a highly accurate process window (736). That is, the process window can be expanded.

  If necessary, the computer 108 compares the calculated cross-correlation coefficient with a threshold value and determines the degree of correlation (737). The computer 108 determines whether the degree of correlation is high, low, or the reverse degree of correlation is high from the comparison result between the cross-correlation coefficient and the threshold value, and outputs the determination result to the feedback process. Further, the computer 108 outputs the determination result to the user interface (738). For example, together with the graphs shown in FIG. 5 and FIG. 6, the contents of the determination result are notified to the operator through image, sound, and lighting control of the notification lamp.

(8) Effect Realized by Form Example By adopting this form example, it is possible to quantitatively confirm the degree of correlation between patterns that conventionally had to be determined visually. For example, it is possible to quantitatively calculate a deviation or coincidence between a design circuit image or simulation image due to a hot spot or a mask defect and an actual pattern shape on the wafer, and to evaluate pattern formation. In addition, by using an integrated image of a plurality of CD-SEM images at the time of evaluating the degree of correlation, it is possible to perform evaluation without removing the influence of roughness.

<Other forms>
In the case of the above-described embodiment, the case where the deformation or defect of a linearly formed pattern is measured has been described. However, this method can also be applied to verification of the formation accuracy of a two-dimensional pattern such as an L shape. . In the case of the above-described embodiment, the case where the formed resist pattern is evaluated has been described. However, the present embodiment is also applicable to the case where the pattern after etching processing is evaluated by comparing the resist pattern with the pattern after etching processing. Techniques can be applied.

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

  Moreover, you may implement | achieve some or all of each structure, a function, a process part, a process means, etc. which were mentioned above as an integrated circuit or other hardware, for example. Each of the above-described configurations, functions, and the like may be realized by the processor interpreting and executing a program that realizes each function. That is, it may be realized as software. Information such as programs, tables, and files for realizing each function can be stored in a memory, a hard disk, a storage device such as an SSD (Solid State Drive), or a storage medium such as an IC card, an SD card, or a DVD.

  Control lines and information lines indicate what is considered necessary for the description, and do not represent all control lines and information lines necessary for the product. In practice, it can be considered that almost all components are connected to each other.

Claims (10)

In a method for evaluating the shape of a formed pattern by an electron microscope,
Observing a plurality of formed patterns formed to have the same shape and the same size by the electron microscope, and obtaining a plurality of microscopic images;
Integrating the plurality of microscopic images to generate an integrated image;
Obtaining a reference image corresponding to the microscopic image from a storage area;
A step of measuring the pattern width of the predetermined pattern included in the accumulated image in a multipoint manner in a one-dimensional direction;
For the same pattern as the integrated image, for the same point as the integrated image, the step of measuring the pattern width of the reference image at multiple points;
A cross-correlation coefficient is calculated between the multipoint measurement result sequence acquired for the integrated image and the multipoint measurement result sequence acquired for the reference image, and the correlation between the integrated image and the reference image is calculated from the calculation result. A pattern shape evaluation method comprising: determining a degree. In the pattern shape evaluation method according to claim 1,
The pattern shape evaluation method, wherein the reference image is a simulation image, a pattern layout design image, or a defect-free microscopic image. In the pattern shape evaluation method according to claim 1,
The determining step includes calculating a correlation level at each measurement point as a cross-correlation coefficient, comparing a maximum value of the cross-correlation coefficient with a threshold value, and determining a correlation level based on the comparison result. And a pattern shape evaluation method characterized by comprising: In the pattern shape evaluation method according to claim 1,
The pattern shape evaluation method, wherein the formed pattern is a resist pattern or a pattern after process processing. In the pattern shape evaluation method according to claim 1,
The plurality of microscopic images are acquired from the same shot, between shots, or between wafers. In an apparatus for evaluating the shape of a formed pattern,
Means for observing a plurality of formed patterns formed to have the same shape and the same size by an electron microscope, and acquiring a plurality of microscopic images;
Means for integrating the plurality of microscopic images and generating an integrated image;
Means for obtaining a reference image corresponding to the microscopic image from a storage area;
Means for measuring the pattern width of a predetermined pattern included in the integrated image in a multipoint manner in a one-dimensional direction;
For the same pattern as the integrated image, for the same point as the integrated image, means for measuring the pattern width of the reference image at multiple points;
A cross-correlation coefficient is calculated between the multipoint measurement result sequence acquired for the integrated image and the multipoint measurement result sequence acquired for the reference image, and the correlation between the integrated image and the reference image is calculated from the calculation result. A pattern shape evaluation apparatus comprising: means for determining a degree. The pattern shape evaluation apparatus according to claim 6,
The pattern shape evaluation apparatus, wherein the reference image is a simulation image, a pattern layout design image, or a defect-free microscopic image. The pattern shape evaluation apparatus according to claim 6,
The determination means determines a correlation degree based on a process for calculating a correlation degree at each measurement point as a cross-correlation coefficient, a process for comparing the maximum value of the cross-correlation coefficient with a threshold value, and the comparison result. The pattern shape evaluation apparatus characterized by performing the process to perform. The pattern shape evaluation apparatus according to claim 6,
The formed pattern is a resist pattern or a pattern after process processing. The pattern shape evaluation apparatus according to claim 6,
The pattern shape evaluation apparatus, wherein the plurality of microscopic images are acquired from the same shot, between shots, or between wafers.

Images