JP2011038811A - Measuring method and measuring system of semiconductor pattern - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enter into consideration of an influence of pattern density/position near a pattern to be measured, in a process of estimating a cross-sectional shape of the pattern by matching a SEM signal waveform of the pattern, with a library correlating a cross-sectional shape calculated by a simulation and a signal waveform. <P>SOLUTION: An area size where existence of surrounding patterns is to be considered, based on an internal diffusion length of incident electron beam is decided, three-dimensional information of patterns in the area size is formed using design layout data, and a library correlating a cross-sectional shape of a SEM signal waveform and a cross-sectional shape is formed using the three-dimensional information. Conventionally, a difference of pattern density/position near a pattern to be measured has been an error factor, however, this method improves measurement accuracy since the information of the pattern density/position within a necessary and sufficient area is reflected to the library. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a method for measuring a pattern dimension of a semiconductor device, and more particularly to a method for measuring a three-dimensional shape of a semiconductor pattern and a system for implementing the method.

  The most popular pattern dimension management tool in a semiconductor process is a length measurement SEM specialized for semiconductor measurement using a scanning electron microscope (hereinafter abbreviated as SEM). FIG. 2 shows the principle of the length measurement SEM. The electron beam 910 emitted from the electron gun 901 is narrowed down by a converging lens 902, and 2 on the sample 900 by a deflector 903 controlled by a control device 906 while being focused on the surface of the sample 900 by an objective lens 904. Dimensionally scanned. The secondary electron 920 generated from the sample 900 by the irradiation of the electron beam 910 is captured by the detector 905 and subjected to signal processing by the control device 906, whereby an electron beam image as displayed on the CRT 907 is obtained. Since more secondary electrons are generated at the pattern edge portion, the electron beam image becomes a bright image at a portion corresponding to the pattern edge as displayed by the CRT 907. In the length measurement SEM, the dimension is obtained by multiplying the distance l (pixel) between edges on the electron beam image by the pixel size p (nm / pixel) (1 × p).

  An example of the length measurement process in the length measurement SEM is described in Patent Document 1. In the disclosed example of Patent Document 1, a projection waveform is created by adding and averaging the signal waveform of the wiring in the longitudinal direction of the wiring from the local region in the image obtained by imaging the wiring to be measured, and the left and right wiring edges detected in this waveform Wiring dimensions are calculated as the distance between them.

  Various methods have been proposed as edge position detection methods for automatically calculating the inter-edge distance l. Here, (a) threshold method and (b) model-based measurement are representative methods. The law is described.

  The idea of the threshold method is disclosed in Patent Document 2, for example. As shown in FIG. 3, in the detection waveform 940 of the secondary electrons detected by scanning the electron beam with respect to the pattern 930 formed on the sample, signal amounts corresponding to the left and right edges 931 and 932 of the pattern 930 The large peak portions are referred to as a left white band (left WB) 941 and a right white band (right WB) 945, respectively. The threshold method calculates Max values 942 and 946 and Min values 943 and 947 for the left WB 941 and the right WB 945, respectively, and calculates threshold levels 944 and 948 that internally divide them by a predetermined ratio th (%). Then, the intersection positions 951 and 952 between the threshold values 944 and 948 and the signal waveform 940 are detected as edge positions, and the distance l between the left and right edges (between 951 and 952) is obtained.

  On the other hand, the model-based measurement method is a method disclosed in Non-Patent Document 1, for example. As shown in the upper diagram of FIG. 4, a pattern cross-sectional shape model 970 is defined (in this example, the pattern height H, bottom width W, side wall inclination angle SWA, bottom roundness BR, and top roundness TR are set to five parameters. The relationship between the cross-sectional shapes of various patterns (changing the parameters representing the pattern shape) and the SEM signal waveform is calculated in advance by simulation, and a simulated SEM waveform library 980 is created (FIG. 4). In the example below, SWA and TR are changed). In the model-based measurement method, a cross-sectional shape to be measured is estimated by referring to a library calculated and stored by simulation, and edge positions are defined on the left and right based on the result.

Japanese Patent Laid-Open No. 11-316115 JP-A-55-72807

J. S. Villarrubia, A. E. Vladar, J. R. Lowney, and M. T. Postek, “Scanning electron microscope analog of scatterometry,” Proc. of the SPIE, Vol. 4689, pp. 304-312 (2002)

With the miniaturization of semiconductor patterns, there is a need for a technique for measuring a pattern more accurately, that is, a technique for reducing a measurement bias (difference between an actual dimension and a measured value).
However, in the length measurement SEM, as shown in FIG. 5A, the electron beam 550 irradiated on the pattern 930 on the sample is diffused inside the pattern 930, so that secondary electrons 552 generated from the pattern 930 are detected. The peak (WB) of the SEM signal waveform 553 obtained by this has a broadening. As shown in FIGS. 5B and 5C, the way in which the SEM signal waveforms 554 and 555 spread varies depending on the cross-sectional shape of the patterns 556 and 557. Therefore, the threshold method changes depending on the cross-sectional shape of the pattern. Measurement bias will occur. For example, when edge detection is performed under the condition of a threshold value of 50%, the edge position 558 detected by the pattern 556 having a large taper as shown in FIG. 5B is 0.5 nm inside the bottom position 560 to be measured. On the other hand, the edge position 559 detected by the pattern 557 having a small taper as shown in FIG. 5C is 2.5 nm outside the bottom position 560 to be measured, and the measurement bias values 561 and 562 of both are clearly different.

  On the other hand, in the case of the model-based measurement method described above, since measurement is performed based on the relationship between the cross-sectional shape of the pattern obtained by calculation and the SEM signal waveform, in principle, the measurement bias shifts depending on the cross-sectional shape of the pattern. Dimension measurement can be realized without being affected by

  However, the factor that changes the SEM signal waveform is not only the cross-sectional shape of the pattern. The SEM signal waveform also changes depending on the pattern density / arrangement. FIG. 6 shows the situation reproduced by simulation. An SEM signal waveform 601 indicated by a solid line indicates a signal waveform of an isolated line pattern 603 having a rectangular cross-sectional shape having a width of 22 nm and a height of 50 nm indicated by a broken line. The signal waveform 602 shows the signal waveform of the line & space pattern 604 having the same cross-sectional shape but a pattern pitch of 44 nm, and the signal waveforms of the two are different. Here, the calculation was performed under the condition that the pattern and substrate material were silicon and the acceleration voltage was 800V. As described above, the difference between the SEM signal waveforms 601 and 602 between the isolated line pattern 603 and the line & space pattern 604 is that reflected electrons (backscattered electrons) and secondary electrons emitted from the sample move to neighboring patterns. This is because the degree to which secondary electrons generated by re-incidence contribute to the final secondary electron intensity (= signal waveform) is different.

  In semiconductor pattern measurement, it is generally required to measure patterns having various densities / arrangements. As described above, even if the pattern cross-sectional shape is the same, the signal waveform changes if the pattern density / arrangement is different. Therefore, even if the model-based measurement method is applied, the measurement bias may not be sufficiently reduced. .

  One of the objects of the present invention is to provide a semiconductor pattern measurement method and system capable of reducing the measurement bias regardless of the pattern density / arrangement.

  In addition, the refinement of semiconductor patterns has a relatively high effect on device performance due to subtle changes in cross-sectional shape such as side wall inclination and bottom rounding. Needs are growing. Since the model-based measurement method estimates the cross-sectional shape of the measurement target based on the relationship between the cross-sectional shape of the pattern and the SEM signal waveform, it can meet the above needs in principle. However, the difference in pattern density / arrangement causes an error in the cross-sectional shape estimation result as described above.

  Another object of the present invention is to provide a semiconductor pattern measuring method and system capable of measuring a three-dimensional shape of a pattern without the influence of pattern density / arrangement.

In order to achieve any of the above objects, the outline of a representative one of the inventions disclosed in the present application will be briefly described as follows.
(1) A method of measuring a semiconductor pattern, the step of calculating a neighborhood region in which a peripheral pattern arrangement should be considered for each of a plurality of measurement positions designated based on input design data; Determining the identity of the pattern arrangement of the plurality of measurement positions based on the pattern arrangement within a range including the vicinity region of each of the measurement positions, grouping and classifying for each measurement position determined to be the same, and For each group of the classified measurement positions, obtaining a cross-sectional shape information from the design data and creating a library; obtaining a SEM signal waveform for each group of the classified measurement positions; and acquiring the SEM Model-based measurement is performed using the signal waveform and the created library, and a cross-sectional shape at each of the plurality of measurement positions is estimated. It is a measurement method of a semiconductor pattern and having a flop, a.
(2) Method of estimating the cross-sectional shape of a sample by applying the electron beam signal waveform of the pattern to be measured to a library that correlates the cross-sectional shape of the sample and the electron beam signal waveform in advance (model-based measurement method) , The position of the measurement target pattern is designated on the design data, and the size xs in the x direction of the neighboring region where the pattern arrangement should be considered based on the internal diffusion length of the incident electron beam determined by the acceleration voltage of the incident electron beam The size ys in the y direction is automatically determined, and within the xs × ys area centered on the position of the pattern to be measured, the design data is used to create a reference three-dimensional shape with a reference three-dimensional shape and a predetermined expected variation range. Model three-dimensional shape information with various shapes changed with numerical data, and input the modeled three-dimensional shape information into an electron beam simulator to simulate SE It generates a signal waveform is obtained so as to store a library in combination with simulated SEM signal waveform shape information. Furthermore, the design data is used to determine the identity of the pattern arrangement within the xs × ys area centered on the position of the pattern to be measured, and for those determined to be the same, use a common library. It is a thing.

According to the present invention, it is possible to provide a semiconductor pattern measurement method and system capable of reducing the measurement bias regardless of the pattern density / arrangement.
Furthermore, according to the present invention, it is possible to provide a semiconductor pattern measurement method and system capable of measuring a three-dimensional shape of a pattern without being affected by the pattern density / arrangement.

FIG. 3 is a diagram showing a dimension measurement flow of the first embodiment according to the present invention. It is a figure which shows the structural example of length measurement SEM. It is a figure which shows the edge point detection method (threshold method) in length measurement SEM. It is a figure which shows the principle of a model base measurement method. It is a first explanatory diagram of a shape-dependent measurement bias. It is the 2nd explanatory view of shape dependent measurement bias. It is a 3rd explanatory drawing of a shape-dependent measurement bias. It is a figure which shows that a SEM signal waveform changes with pattern densities. It is a figure which shows the example of GUI for performing three-dimensional shape modeling of a measurement position. It is explanatory drawing of the concept of the identity determination of a pattern layout. It is a figure which shows the flow of the library preparation for model base measurement. 1 is a diagram illustrating an example of a system configuration according to a first embodiment. FIG. 3 is a diagram showing another system configuration example according to the first embodiment. It is a figure which shows the example of an output of a dimension measurement result. It is a figure which shows the dimension measurement flow of 2nd Embodiment which concerns on this invention. It is a figure which shows the system configuration example which concerns on 2nd Embodiment. It is a figure which shows the example of an output of the dimension measurement result which concerns on 2nd Embodiment. It is explanatory drawing when it is necessary to consider a virtual pattern. It is explanatory drawing in case a measurement object pattern is a hole. It is explanatory drawing in case it is not a simple line pattern.

  A first embodiment of a dimension measurement flow according to the present invention and a system configuration for carrying out the same will be described with reference to FIGS. 1 and 10. In addition, it is shown in the balloon of FIG. 10 which component of FIG. 10 performs each step shown in FIG.

  First, as shown in FIG. 10, an example of a system according to the present invention includes a LAN 300, length measurement SEMs 301, 302, and 303 connected thereto, a SEM recipe database 304 that stores a measurement SEM recipe, A computer 305 that models a three-dimensional shape to be measured using design data, a database 306 that stores semiconductor pattern design information, a database 307 that stores an SEM waveform library, and a length measurement SEM. A database 308 that stores SEM images, a computer 309 that performs electron beam simulation, and a computer 310 that performs model-based measurement are appropriately used. Here, the length measurement SEM does not need to be three, and the number is not limited to this. Further, each computer may be integrated into one unit as necessary, or may be configured by a plurality of units.

Next, an example of a dimension measurement flow executed by this system will be described with reference to FIG.
According to the dimension measurement flow according to the present invention, first, in the above system, the computer 305 for modeling the three-dimensional shape includes the design data information of the wafer to be measured selected by the user, and the selected Position information to be measured on the wafer design data is received (s10, s11). Here, for example, the design data of the wafer selected by the user may be displayed on the GUI so that the user can specify the position to be measured on the screen while viewing the displayed design data. Further, the computer 503 accepts input of information regarding the stacking information of the position to be measured, that is, information on the pattern and the material and thickness of the substrate, as well as imaging conditions of the electron microscope (acceleration voltage of the incident electron beam, etc.) (s12, s13). When information input for all the positions to be measured is received, the computer 305 subsequently refers to the design data and executes the pattern layout identity determination process (s14) at each measurement position.

Here, an example of a GUI (graphic user interface) for receiving each input information is shown in FIG. In this GUI, for example, a number input unit 401 for designating the number of design data to be measured, a design data input unit 402 for inputting the number of design data designated by the number input unit 401, and information related to design data are displayed. In the case where the positional relationship with the lower layer pattern is important, or in the case of design data for double patterning, it is necessary to input a plurality of design data. A pattern layout display unit 405 is displayed that receives the input of 404 and the x / y coordinate input unit 404 and displays the pattern layout of the corresponding portion. Here, the pointing device 406 may be used to appropriately adjust the measurement position 104 displayed on the pattern layout display unit 405. In this case, the x / y coordinate input unit 404 is interlocked with the movement of the pointing device 406. Furthermore, the pattern selection unit 407 for selecting a desired pattern type from choices such as lines and holes, the pattern material stacking number input unit 410, and the pattern material thickness as items for inputting film information of the pattern unit and the substrate unit A material / material input unit 411, a substrate material base total number input unit 412, and a substrate material thickness / material input unit 413 are displayed. Here, the information on the film material and the film thickness is information necessary for executing the SEM simulation, and the normal general design data does not include such information. Input is required. If TCAD (Technology / CAD) data is available, necessary information may be taken from the data. Further, if necessary, the imaging condition input unit 414 such as the applied voltage, the scanning direction, and the detection mode may be displayed to receive such information.
As described above, based on the input information and design data that have been input, the GUI displays the cross-sectional shape 415 of the measurement pattern for confirmation.

  Here, the pattern position identity determination processing method at the measurement position performed in step s14 shown in FIG. 1 will be described with reference to FIG. Measurement positions 101 to 106 indicated by solid lines indicate measurement positions input by the user, and areas indicated by broken lines around each measurement position indicate neighboring areas in which pattern arrangement should be considered. As shown in FIG. 6 described above, the SEM signal waveform changes under the influence of neighboring patterns, but the range in which the influence should be considered is finite. This range depends on the diffusion length of incident electrons as shown in FIG. 5A. This diffusion length varies mainly depending on the acceleration voltage of the incident electrons and the sample material. Therefore, in the present invention, based on the imaging condition (414) and the film material information (410 to 413) input in the GUI of FIG. 7, the range of the neighboring region is calculated with reference to a table created in advance. . If the layout such as the pattern width and the pattern interval is the same and the imaging conditions are the same within the range of the calculated neighborhood area (the area surrounded by the broken lines 101 to 106), it is determined to be the same in step s14. Is done. For example, if the imaging conditions of 101 to 106 in FIG. 8 are the same, they are line patterns having the same line width in design, but it is determined in steps s14 that 102, 104, and 103 are the same. 105, and these are collected as the same group. 103 and 105 are the same if the range including the neighboring area is reversed left and right (rotated 180 degrees). As shown in the lower table of FIG. 8, the data structure after the determination in step s14 is the library number (the same number for those determined to be the same) and rotation angle information for each measurement position. Will be given.

  After the above-described pattern layout identity determination step s14, the computer 305 executes a modeling target three-dimensional shape modeling step s15 as shown in the dimension measurement flow shown in FIG. As described above, in the model-based measurement method, the relationship between the cross-sectional shapes of various patterns and the SEM signal waveform needs to be calculated in advance by electron beam simulation. Therefore, in this step, creation of input cross-sectional shape data used when performing electron beam simulation is executed for each group made identical in the identity determination step s14.

Here, a method of creating input cross-sectional shape data will be described with reference to FIG. In the figure, the case of the measurement position 103 in FIG. 8 is shown as an example. The three-dimensional shape is calculated using planar pattern layout information (pattern size, arrangement) obtained from design data and film thickness information obtained by input from the user. What is calculated is a three-dimensional shape obtained by variously changing the reference three-dimensional shape (three-dimensional shape as designed) within a predetermined variation range. FIG. 9 shows an example of the reference cross-sectional shape 201 and the modified cross-sectional shape 202 in one cross-section (AA) of the three-dimensional shape. The modified cross-sectional shape 202 is obtained by changing the line width and the side wall inclination angle of the reference cross-sectional shape 201. In practice, however, various parameters representing the cross-sectional shape as shown in FIG. Create multiple 3D shape information. In FIG. 4, the cross-sectional shape is expressed by the pattern height H, bottom width W, side wall inclination angle SWA, bottom roundness BR, and top roundness TR, but some of them may be fixed. It is also possible to use another expression method such as stacking a plurality of trapezoids. These parameters may be added and deleted as appropriate.
As described above, the steps (s10 to s15) of the dimension measurement flow shown in FIG. 1 described so far are performed by the computer 305 of the system shown in FIG. 10 and the result is stored in the database 306.

  After modeling of the three-dimensional shape to be measured (s15), the computer 309 shown in FIG. 10 calculates SEM signal waveforms for various three-dimensional shapes by electron beam simulation (Monte Carlo simulation) and combines them with shape information. Library creation is executed (s16). In step s16, a library is created for each group determined to be the same in step s14. Information input through the GUI shown in FIG. 7 is applied to imaging conditions such as acceleration voltage required for the simulation. The three-dimensional shape information requires a range that covers the neighboring region, but the range in which the SEM signal waveform is calculated (the range in which electrons are injected in the simulation) may be the range of the measurement target pattern. For example, in FIG. 9, the three-dimensional shape information is necessary for the entire area within the broken line 103, but the SEM signal waveform may be calculated for the area range 210 or 211. SEM signal waveforms calculated for various cross-sectional shapes are stored in combination with cross-sectional shape information (= a group of parameters expressing the cross-sectional shape) in the calculation range, and become a library for performing model-based measurement. Individual numbers are assigned to the libraries as shown in the table of FIG.

  This step s16 is performed by the computer 309 shown in FIG. 10, and the created library is stored in the database 307 together with the library number. Since the electron beam simulation takes time (for example, when one hour is required to calculate one SEM signal waveform, 100 hours are required if the number of variations of the three-dimensional shape is 100). It is desirable to provide a dedicated system composed of

After each of the above steps s10 to s16, an electron beam image is acquired by a length measurement SEM (301, 302, 303 in FIG. 10) (s17). The acquired SEM image is stored in the database 308, and the computer 310 calls a library corresponding to each group of measurement positions and executes model-based measurement (s18).
Note that the model-based measurement step s18 may be performed by a length measurement SEM as shown in FIG. Prior to the start of imaging, a necessary library is read from the database 307 according to the table of FIG. 8, and model-based measurement is executed together with normal dimension measurement each time an image is captured.

  An output example of the dimension measurement result measured in each step described above is shown in FIG. In this example, the cross-sectional shape is modeled as a trapezoid, and the dimensions of the bottom, middle, and top and the left and right side wall inclination angles (left SWA, right SWA) are output. Here, for example, by writing together the result of the conventional dimension measurement method (threshold method in this example) and the result of the model base measurement on the same screen, the user can obtain the cross-sectional shape of the measurement target pattern, It is possible to grasp the difference from the result obtained by the conventional dimension measurement method.

  As described above, according to the present invention, pattern density / arrangement information, which has not been taken into consideration in the conventional model-based measurement, is reflected in the library, so that more accurate cross-sectional shape measurement is possible. Specifically, in the conventional model-based measurement, as shown in the table of FIG. 8, the library is not used properly as a measurement error factor, whereas in the present invention, the SEM signal waveform is changed. Since a library that takes into account the pattern layout information in the affected range is used, more accurate cross-sectional shape estimation is possible.

  In addition, according to the present invention, since the range that affects the SEM signal waveform is automatically determined, even if there is no knowledge about electron beam image formation, the signal waveform changes if the pattern density / arrangement is different. Therefore, it is possible to create an appropriate library that does not cause a measurement error, and as a result, it is possible to accurately estimate the cross-sectional shape.

  In addition, according to the present invention, since a library can be shared between a plurality of length measuring SEMs on the same system (a library can be reused), an electron beam simulation requiring a great amount of time can be performed to a minimum. It becomes possible to carry out at a calculation cost.

A dimension measurement flow according to the present invention and a second embodiment of a system configuration for carrying out the same will be described with reference to FIGS.
As shown in FIG. 13, the input information (s10 to s13) received by the computer 305 and the pattern layout identity determination processing step (s14) are the same as those in the first embodiment. In the present embodiment, after the pattern layout identity is determined using the design data, the process for automatically generating the imaging recipe of the length measurement SEM is executed using the design data (s19). This is different from the first embodiment. In order to acquire the electron beam image of the measurement target pattern in the length measurement SEM, in addition to the position coordinates of the measurement pattern, information on the alignment pattern and the automatic focus detection pattern is required. In step s19, using the design data, automatic selection of a pattern suitable for alignment and automatic focus detection is performed by the computer 305 and registered in the SEM recipe database 304 of FIG. In the step (s17) of acquiring the electron beam image by the length measurement SEM, the registered imaging recipe is read from the database 304 and executed.

In the present embodiment, design data is used in steps s10 to s14, and the imaging conditions necessary for the imaging recipe are already input in s14. Therefore, it is efficient to automatically generate an imaging recipe. It can be said. Moreover, since the recipe is created on the computer, it is not necessary to use a length measuring SEM for creating the imaging recipe, which is advantageous in terms of COO (Cost of Ownership).
Furthermore, by superimposing design data on the captured SEM image, as shown in FIG. 15, (a) SEM image, (b) design data, (c) superimposition thereof, (d) measurement target position It is also possible to output an estimated sectional shape 1501, (e) a dimension value of the estimated sectional shape, and the like. It is also possible to superimpose with a contour line (not shown) extracted from the SEM image.
As described above, in the first and second embodiments described above, the measurement target pattern is a single-layer line pattern. However, it is necessary to consider the positional relationship with the lower layer pattern as shown in FIG. In some cases. In the figure, 1601 is an upper layer pattern and 1602 is a lower layer pattern. In this case, two design data (upper layer and lower layer) are read in the dimension measurement location 1604 where two layers overlap among the dimension measurement location shown in FIG. 16 and its neighboring areas 1603 and 1604, and this is performed in step s14. In determining identity, the film material of each layer is also used as a criterion for determining identity.

  FIG. 17 shows a library creation method when the measurement target pattern is a hole or a pillar (column). Although these patterns are rectangular in the design data, if holes or pillars are selected in the GUI pattern type column 407 shown in FIG. A circular concave pattern and, if a pillar, a circular convex pattern.

  FIG. 18 shows a case where the three-dimensional shape cannot be expressed by simple lines or holes. The measurement position 106 in FIG. 8 corresponds to this case. In the example of FIG. 18, a line pattern portion 1802 that is a measurement target and an overhanging portion 1803 that is a non-measurement target are included in the vicinity range. In such a case, the number of SEM waveforms to be calculated will increase by about a factor of 2 if the library also includes the case where the shape of the non-measurement target region changes in various ways (for example, the shape variation of the measurement target region and If there are N shape variations of the non-measurement target parts, the shape variation within the vicinity range is N × N). Only the measurement target part is changed in shape.

  Although the present invention has been specifically described above based on the embodiments of the invention made by the present inventors, the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the scope of the invention. Not too long.

Further, the effects obtained by the representative ones of the inventions disclosed in the present application can be summarized as follows.
(1) According to the present invention, the pattern density / placement information within the necessary and sufficient area range is reflected in the library, so the pattern shape is the same as the problem of the conventional model-based measurement method. However, if the pattern density / arrangement is different, the measurement error due to the change of the signal waveform is improved, and the measurement bias can be reduced regardless of the pattern density / arrangement. In addition, more accurate three-dimensional shape measurement of the pattern is possible.
(2) In addition, according to the present invention, since the region that should consider the existence status of the surrounding pattern is automatically determined, even if there is no knowledge about electron beam image formation, if the pattern density / arrangement differs, It is possible to create an appropriate library that does not cause measurement errors due to changes.
(3) Further, according to the present invention, since the library can be shared (reuse of the library), it is possible to create a library in a shorter time by eliminating duplication of electron beam simulation that requires a lot of time. It becomes possible.

s10 to s19: Steps 101 to 106 of dimension measurement ... Measurement positions 201, 202 ... Modeled three-dimensional cross sections 301 to 310 ... Components 401 to 414 of the dimension measurement system. .. GUI components 501 to 502... Upper layer pattern and lower layer patterns 503 to 504... Dimension measurement points 601, 602... SEM signal waveform 603... Isolated line pattern 604. 901 ... Electron gun 902 ... Converging lens 903 ... Deflector 904 ... Objective lens 905 ... Detector 906 ... Control device 907 ... CRT
920 ... Secondary electrons 930 ... Pattern

Claims (8)

A method for measuring a semiconductor pattern,
Calculating a neighborhood area in which a surrounding pattern arrangement should be considered for each of a plurality of measurement positions designated based on input design data;
Based on the pattern arrangement within the range including the vicinity area of each of the plurality of measurement positions calculated, the identity of the pattern arrangement of the plurality of measurement positions is determined and grouped for each measurement position determined to be the same. A step of classification;
For each group of the classified measurement positions, obtaining a cross-sectional shape information from the design data to create a library;
An SEM signal waveform is acquired for each group of the classified measurement positions, model-based measurement is performed using the acquired SEM signal waveform and the created library, and a cross-sectional shape of each of the plurality of measurement positions is estimated. And steps to
A method for measuring a semiconductor pattern, comprising: A semiconductor pattern measuring method according to claim 1,
Further, the method includes a step of generating a recipe for acquiring the SEM signal waveform based on the design data,
The method for measuring a semiconductor pattern, wherein the acquisition of the SEM signal waveform is executed based on the generated recipe. A semiconductor pattern measuring method according to claim 1,
The method of measuring a semiconductor pattern, wherein the design data includes information on the material and thickness of the pattern at the measurement position. A semiconductor pattern measurement method for estimating a cross-sectional shape of a sample by applying the electron beam signal waveform of a measurement target pattern to a library that correlates the cross-sectional shape of the sample with the electron beam signal waveform,
Designating the position of the measurement target pattern on the design data;
Automatically determining a size xs in the x direction and a size ys in the y direction of a neighboring region in which pattern arrangement should be considered based on the internal diffusion length of the incident electron beam;
For the xs × ys area centered on the position of the measurement target pattern, using the design data, the reference three-dimensional shape and the three-dimensional shape in which the reference three-dimensional shape is variously changed within a predetermined expected variation range. A pattern shape modeling step for modeling shape information with numerical data;
A step of creating a library that inputs the modeled three-dimensional shape information to an electron beam simulator to generate a simulated electron beam signal waveform, and stores the simulated electron beam signal waveform and the shape information in combination as a library; A method for measuring a semiconductor pattern. A semiconductor pattern measurement method for estimating a cross-sectional shape of a sample by applying the electron beam signal waveform of a measurement target pattern to a library that correlates the cross-sectional shape of the sample with the electron beam signal waveform,
Designating the position of the measurement target pattern on the design data;
A step of automatically creating a recipe for acquiring an electron beam image of a pattern to be measured using design data;
Automatically determining a size xs in the x direction and a size ys in the y direction of a neighboring region in which pattern arrangement should be considered based on the internal diffusion length of the incident electron beam;
For the xs × ys area centered on the position of the measurement target pattern, using the design data, the reference three-dimensional shape and the three-dimensional shape in which the reference three-dimensional shape is variously changed within a predetermined expected variation range. A pattern shape modeling step for modeling shape information with numerical data;
A step of creating a library that inputs the modeled three-dimensional shape information to an electron beam simulator to generate a simulated electron beam signal waveform, and stores the simulated electron beam signal waveform and the shape information in combination as a library; A method for measuring a semiconductor pattern. A semiconductor pattern measuring method according to claim 4 or 5,
In the pattern shape modeling step, measurement is performed using design data prior to modeling the reference three-dimensional shape and the three-dimensional shape information obtained by variously changing the reference three-dimensional shape within the expected variation range with numerical data. The pattern arrangement identity in the neighboring area where the pattern arrangement between the target patterns should be considered is determined, and the common three-dimensional shape information is assigned to the measurement target patterns determined to be the same by assigning common three-dimensional shape information. A method for measuring a semiconductor pattern, comprising: inputting information to an electron beam simulator to generate a simulated electron beam signal waveform; and storing the information as a common library in combination with the common three-dimensional shape information. A semiconductor pattern measuring method according to claim 4 or 5,
The step of determining the size of the neighboring region that should take into account the pattern arrangement further includes a step of designating an imaging condition of the electron beam image, and the pattern arrangement is performed based on the internal diffusion length of the incident electron beam in the imaging condition. A method for measuring a semiconductor pattern, comprising: determining a size of a neighboring region to be considered. A semiconductor pattern measuring method according to claim 4 or 5,
The step of designating the position of the measurement target pattern on the design data further includes a step of inputting information on the pattern, the film material of the substrate, and the film thickness.
In the pattern shape modeling step, the pattern shape is modeled using the information on the film material and the film thickness, and the semiconductor pattern measuring method is characterized in that:

Images