JP2012173028A - Method for measuring pattern shape and apparatus therefor - Google Patents

Abstract

【Task】
Even if a change in pattern height such as a decrease in resist film occurs in the measurement using the SEM image taken from directly above the pattern, the cross-sectional shape can be accurately measured.
[Solution]
In a method of measuring the shape of a pattern formed on a sample using a charged particle beam device, secondary charged particles generated from the sample are detected by irradiating and scanning the focused charged particle beam on the sample. The charged particle beam image of the pattern formed on the sample surface is acquired, and the charged particle beam image of the pattern acquired based on the relationship between the height of the pattern obtained in advance and the information of the charged particle beam image of this pattern is obtained. The height information of the pattern is obtained from the information, and the dimension of the pattern is calculated from the information of the charged particle beam image of the pattern using the obtained height information of the pattern.
[Selection] Figure 1

Description

  The present invention relates to a pattern shape measuring method and apparatus for evaluating the quality of a processed shape of a circuit pattern formed on a wafer in a semiconductor manufacturing process using an electron microscope image of the circuit pattern.

  Pattern measurement technology for judging whether or not the patterns formed in multiple layers on a wafer are rapidly miniaturized in the manufacturing process of semiconductor wafers and whether these patterns are formed on the wafer as designed. The importance of is increasing. In particular, wiring patterns such as transistor gate wiring are strongly related to the wiring width and device operating characteristics, and in order to determine stable manufacturing conditions and monitor the wiring manufacturing process, the measurement technology of the pattern shape and dimensions is used. Is very important.

  As a length measurement tool to measure the line width of fine wiring on the order of several tens of nanometers, a scanning electron microscope for measuring the line width (length measuring SEM) that can image these wirings at a magnification of 100,000 to 200,000 times (Scanning Electron Microscope) or CD (Critical dimension) SEM) has been conventionally used. An example of a length measurement process using such a scanning electron microscope is described in JP-A-11-316115 (Patent Document 1). In the disclosed example of Patent Document 1, a projection profile is created by averaging the signal profile of the wiring in the longitudinal direction of the wiring from the local region in the image obtained by imaging the measurement target wiring, and the left and right wiring edges detected in this profile Wiring dimensions are calculated as the distance between them.

However, as disclosed in FIG. 1 of Non-Patent Document 1, in the signal waveform of the SEM, when the shape of the measurement target changes, the signal waveform also changes accordingly, causing a measurement error. With the miniaturization of semiconductor patterns, the influence of these measurement errors on the determination of manufacturing process conditions and the management during mass production is increasing. Techniques for reducing such measurement errors are disclosed in Patent Document 2, Non-Patent Document 1, and Non-Patent Document 2. In this method, the relationship between the pattern shape and the SEM signal waveform is calculated in advance by simulation, and high-precision measurement independent of the target shape is realized by using the result.
Patent Document 3 discloses a method of measuring a relative film reduction amount with respect to a reference height using a correlation between a variation in brightness due to roughness on the upper surface of a pattern on a SEM image and a film reduction amount of the pattern. Are listed.

  Further, in Patent Document 4, a library including height fluctuations is created, and the height information measured in advance by another means is used to influence the influence of height fluctuations using only necessary portions of the library. A method of elimination is described.

  On the other hand, in Non-Patent Document 3, in a high NA exposure technique that has been introduced to realize pattern miniaturization, a slight process variation tends to cause a reduction in resist height called film reduction. Are listed.

  Further, Non-Patent Document 4 describes that height information is obtained using a stereo vision method in which images acquired from two or more different directions are processed to obtain the height and shape of a pattern.

Japanese Patent Laid-Open No. 11-316115 JP 2007-218711 A JP 2009-198340 A JP 2009-198339 A

J. S. Villarrubia, A. E. Vladar, J.A. R. Lowney, and M. T. Postek, “Scanning electron microscope analog of scatterometry” Proc. SPIE 4689, pp. 304-312 (2002) J. S. Villarrubia, A. E. Vladar, M.C. T. Postek, “A simulation study of repeatability and bias in the CD-SEM” Proc. SPIE 5038, pp. 138-149, 2003. N. Yasui, M .; Isawa, T. Ishimoto et al. , “Application of Model-Based Library approach to resist pattern shape measurement in advanced lithography” Proc. SPIE 7638, 76382O, 2010. B. Su, R. Oshana et al. “Shape Control Using Sidewall Imaging” Proc. SPIE 3998, 232-238, 2000.

  As shown in the background art, when measuring the dimensions of a semiconductor pattern by a length measurement SEM, there is a problem that a measurement error depending on the shape of the target pattern occurs. On the other hand, in the methods disclosed in Non-Patent Document 1 and Non-Patent Document 2, the relationship between the pattern shape and the SEM signal waveform is calculated in advance by SEM simulation, and the result is used to depend on the target shape. High-precision measurement that does not occur. It is possible to accurately estimate the shape and dimensions by digitizing the pattern shape using parameters, storing SEM simulation results of various shapes as a library, and comparing them with actual waveforms. Hereinafter, this method is referred to as a model-based measurement or library matching method. In such a model-based measurement method, accurate simulation and appropriate modeling of the measurement target shape change are important for realizing stable and highly accurate measurement. In particular, in shape modeling, it is necessary to consider how a shape change to be measured appears as a change in an image signal in a measurement image to be used (SEM image in the example of the present invention).

  As shown in Non-Patent Document 3, in the high NA exposure technology that has been introduced in recent years in order to realize pattern miniaturization, a decrease in resist height called film reduction is caused by slight process fluctuations. Prone to occur. However, it is difficult to obtain height information directly in a method using an observation image from directly above. As will be described later, since the change in the SEM image due to the height change is smaller than the change in other shapes (side wall inclination angle, etc.), the conventional model-based measurement is used for the pattern whose height varies as described above. With this method, highly accurate shape measurement is difficult.

  In the conventional SEM library matching method, since the height is assumed to be known, a shape estimation error occurs if the height change amount is not known. The realization of high-accuracy measurement corresponding to such height fluctuation is a problem to be solved by the present invention. In addition, as shown in Non-Patent Document 4, it is also possible to obtain height information by using stereo vision that processes two or more images acquired from different directions to obtain the height and shape of the pattern. is there. However, in the case of using stereo vision, since it is necessary to acquire a plurality of SEM images, there is a problem that it takes time to acquire the image and the throughput of the measuring apparatus is reduced. There is also a problem that the pattern is damaged by multiple times of electron beam irradiation.

  In the present invention, pattern shape information obtained by the library matching method with respect to the above-described problems caused by the measurement method using the simulation as shown in Patent Literature 2, Non-Patent Literature 1, and Non-Patent Literature 2 Is corrected by using pattern height information measured in advance by another method, and a method for reducing the shape estimation error is disclosed. In particular, by using a method using an actual SEM image as a pattern height measurement means, an increase in measurement time is prevented and high-speed and high-precision measurement is realized.

  That is, in the present invention, in order to solve the above-described problem, in a method for measuring the shape of a pattern formed on a sample using a charged particle beam apparatus, the focused charged particle beam is irradiated onto the sample. The secondary charged particles generated from the sample are detected by scanning to acquire a charged particle beam image of the pattern formed on the sample surface, and the height of the pattern obtained in advance and the information of the charged particle beam image of this pattern The pattern height information is obtained from the information of the charged particle beam image of the pattern acquired based on the relationship with the pattern, and the dimension of the pattern is calculated from the information of the charged particle beam image of the pattern using the obtained pattern height information. I tried to do it.

  Further, in the present invention, in order to solve the above-described problems, an apparatus for measuring the shape of a pattern formed on a sample is generated from the sample by irradiating and scanning the focused charged particle beam on the sample. A charged particle beam means for detecting secondary charged particles to acquire a charged particle beam image of a pattern formed on the sample surface, and storing the relationship between the pattern height and the charged particle beam image information of this pattern A pattern based on the charged particle beam image information obtained by the charged particle beam means based on the relationship between the storage means and the pattern height stored in the storage means and the charged particle beam image information of the pattern. Pattern height information extraction means for obtaining the height information of the pattern, and a charged particle beam image of the pattern obtained by the charged particle beam means using the pattern height information obtained by the pattern height information extraction means It was constructed and a pattern dimension calculating means for calculating a dimension of the pattern from the information.

  According to the present invention, it is possible to reduce the shape estimation error due to the pattern height in the model-based measurement method. In addition, by measuring the pattern height using the same SEM image used for model-based measurement, high-speed and high-precision measurement is possible without increasing the data acquisition and processing time for height measurement. It becomes. Further, since processing can be performed with only one SEM image, damage to the measurement target pattern due to electron beam irradiation (for example, shrink of a resist pattern) can be minimized.

It is explanatory drawing of the 1st Example of this invention. It is explanatory drawing of the means to obtain the pattern height information in 1st Example of this invention. It is a figure which shows the subject of the conventional library matching method. It is explanatory drawing of the SEM image signal in 1st Example of this invention. It is explanatory drawing of library construction in the 1st Example of this invention. It is explanatory drawing of the shape correction method in 2nd Example of this invention. It is explanatory drawing of the film reduction index | index calibration method in the 3rd Example of this invention. It is explanatory drawing of the 4th Example of this invention.

  The present invention can be applied to various types of charged particle beam apparatuses (SEM, ion microscope, etc.). In the following examples, a case where an SEM is used as a representative example will be described.

[Basic type]

  In the first embodiment, a method for reducing an error in SEM measurement caused by a difference in height of a measurement target pattern will be described with reference to FIGS. The present invention improves the pattern shape measurement accuracy by correcting the library waveform matching result based on the pattern height measurement result measured in advance.

  FIG. 3 is a diagram for explaining a measurement error that occurs depending on the pattern height, which is a measurement problem of the conventional method. The pattern 201 and the pattern 202 shown in FIG. 3A have the same top edge position 203 and bottom edge position 204 in the left and right x directions (in the plan view) when viewed from above, and have different heights and sidewall inclination angles. . With respect to the height H of the pattern 201 and the side wall inclination angle θ, the pattern 202 has a height H ′ and the side wall inclination angle becomes θ ′ by changing the height while maintaining the edge position. When the SEM waveform 205 is observed directly above by SEM, the waveform hardly changes even if there is a difference between the two patterns 201 and 202. Due to a phenomenon called the tilt angle effect, the signal amount of the SEM image changes depending on the tilt of the sample surface, but in the case of a height change as shown in FIG. The change in the tilt angle is about several degrees, and the width of the portion of the side walls 206 and 207 that can be observed from directly above is narrow (the projection length on the x-axis of the side wall portion is short with respect to the electron diffusion length). Therefore, the signal amount change between these patterns is relatively small.

  On the other hand, since the projection lengths on the x-axis of the portions of the side walls 206 and 207 are the same in the patterns 201 and 202, the waveform does not change much. Thus, although the heights of the patterns 201 and 202 change, the SEM signal waveform observed from directly above the pattern does not change much. In conventional model-based measurement, it is difficult to discriminate such changes, so the height is assumed to be known and fixed to an arbitrary value (design value or the like). That is, since the pattern 202 having the pattern height H ′ is also measured using the library with the height H, the estimated value of the sidewall inclination angle is θ, and the actual inclination angle θ ′ is A different value will be output.

  When a fine resist pattern created by a high NA exposure technique is used as a measurement target, as shown in a litho simulation output example 210 in FIG. 3B, the height decreases due to film reduction, and FIG. As shown in FIG. 4, a difference occurs in the sidewall inclination angle θ. For example, if the height is reduced to 60 nm without changing the edge position of the pattern with a height of 80 nm and a side wall tilt angle θ of 87 °, the side wall tilt angle θ becomes about 86 °, and the change is about 1 °. Arise. That is, when library matching is performed with the pattern having a height of 60 nm and a height of 80 nm, this change in the side wall inclination angle (1 °) directly becomes a measurement error. From this, it can be seen that it is important to consider the influence of the pattern height in order to improve the measurement accuracy. Non-Patent Document 3 shows an example in which tilt angle estimation cannot be performed correctly due to such a problem.

  As a means for solving this problem, FIG. 1B shows a measurement procedure by the SEM of the first embodiment of the pattern measurement method according to the present invention.

  Before the start of measurement, an SEM simulation is performed in which the pattern shape and dimensions are set to various values, and the SEM simulation waveform is stored in a library. The setting range of the simulation condition is set to an appropriate value according to the manufacturing process of the measurement target pattern.

  At the time of measurement, first, an SEM image of a sample is acquired with the SEM001 shown in FIG. 1A under preset imaging conditions (magnification, acceleration voltage of irradiation beam, etc.). Specifically, the electron beam 102 emitted from the electron gun 101 of the SEM001 is converged by the focusing lens 103, and scanned by the deflector 104 in the X direction and the Y direction (directions orthogonal to x and z in FIG. 1). The objective lens 105 scans and irradiates the surface of the sample 106 with the focus of the electron beam aligned with the surface of the sample 106 on which the measurement target pattern is formed. Although not shown in FIG. 1A, the sample 106 is placed on a moving stage and can move within a plane, and a desired region on the surface of the sample 106 becomes an irradiation region of the electron beam 102. It is controlled to be located.

  A part of the secondary electrons generated from the surface of the sample 106 irradiated with the electron beam 102 is detected by the detector 107, converted into an electrical signal, sent to the overall control / image processing unit 108, and an SEM image is created. The calculation unit 109 processes the SEM image to calculate the dimension of the pattern, and the result is displayed on the screen of the output unit 110. The overall control / image processing unit 108 also controls the entire SEM 001 including the stage on which the sample 106 (not shown) is placed. A data server 111 stores information from the output unit 110. A database unit 112 refers to information stored in the database unit 112 as necessary when the calculation unit 109 processes an SEM image.

  A processing procedure in the calculation unit 109 is shown in FIG. First, as described above, the SEM 001 is controlled by the overall control / image processing unit 108 to acquire the SEM image of the measurement target pattern (S0001).

  Next, the SEM image acquired by the overall control / image processing unit 108 is received, the SEM image is processed by the image processing unit 1091 of the calculation unit 109, and the pattern height is measured (S0002). The measurement in step (S0002) can be performed by a method of quantifying the roughness change of the upper surface of the pattern accompanying the film reduction on the SEM image and calculating it as a film reduction index value as described in Patent Document 3, for example. it can. That is, in Patent Document 3, a relative film reduction amount (= pattern height) with respect to a reference height is utilized by utilizing a correlation between brightness variation due to roughness of the pattern upper surface on the SEM image and a pattern film reduction amount. It describes a technique for measuring By using this method, the height of the measurement target pattern can be directly measured on the actual SEM image used for library matching, and the height as shown in FIG. 3 is used by using the pattern height in the pattern. High-accuracy measurement in consideration of the influence of fluctuations becomes possible. Further, since the same image as that used for library matching is used, accuracy can be improved only by adding data processing without spending new time for acquiring additional data and the like.

  In Patent Document 4, a library including height fluctuations is created, and height information measured in advance by another means is used to eliminate the influence of height fluctuations using only necessary portions of the library. Although a method is disclosed, there is a problem that calculation of a library including a height variation requires a lot of time. As described above, the influence of the difference in pattern height on the SEM signal waveform is small. For this reason, even in library matching, the position of the pattern edge can be obtained correctly even if libraries having different heights are used. Therefore, in this embodiment, a highly accurate pattern shape estimation is realized by correcting the estimated pattern shape by combining the edge position measurement result obtained by the conventional library matching method and the separately obtained height information. To do.

  Next, as a result of measuring the film reduction index value from the SEM image by the method described in Patent Document 3 in step (S0002), the film reduction index value and the pattern height stored in the database 112 in step S0003 are obtained. Based on the correlation data 003, the pattern height information is converted (S0003).

  Here, the construction method of the database 003 will be described with reference to FIG. As shown in FIG. 2, first, an SEM image is obtained on the entire wafer surface (S1001) for a wafer whose exposure conditions are arbitrarily changed (for example, FEM: Focus Exposure Matrix whose exposure amount and focus amount are changed). The film thickness reduction index value is measured from the acquired SEM image (S1002). As the measurement of the film reduction index value, for example, since it is known that there is a correlation between the roughness of the pattern surface and the amount of film reduction, the processing target area shown in FIG. The longitudinal (y) direction of 302 is set so as to be widely surrounded, and a waveform 0052 in the y direction on the surface of the pattern 302 as shown in FIG. 4C is acquired. Since the roughness of the surface of the pattern 302 is a variation in brightness in the SEM image, the roughness of the surface of the pattern 302 can be measured by quantifying the variation in the waveform 0052. Roughness can be used as an index value for film reduction.

  Next, height measurement is performed at the same point where the SEM image was acquired, for example, height measurement by AFM (S1003), and the correlation between these results, that is, the relationship between the film reduction index value and the pattern height is clarified. It is obtained by creating a database (S1004) and storing it in the database 112 as correlation data 003. In addition to the AFM, other devices such as a cross-section SEM or STEM may be used for the height measurement as long as the height can be measured. In the example of FIG. 4, the wiring-like pattern has been described. However, any pattern may be used as long as the film reduction phenomenon occurs due to a difference in exposure conditions. Of course, if a sufficiently large (wide) pattern can be used, calibration may be performed using an optical film thickness measurement method such as an ellipsometer. In the example of FIG. 2, an example using an FEM wafer is shown. However, if a variation of the resist film reduction amount can be created by a practical change in the amount of exposure light, only the exposure amount, only the focus, or otherwise Of course, a calibration pattern may be created by changing the exposure parameters.

  As described above, the pattern height is calculated by the step (S0003) of FIG. 1 using the database 112 storing the correlation data 003 of the relationship between the film thickness reduction index value and the pattern height obtained by the procedure of FIG. To do.

  In parallel with the height measurement (S0002, S0003), the cross-sectional shape of the pattern is estimated by a conventional model-based measurement method using the SEM image captured in S0001 (step S0004). At this time, only the library created at a reference height that is assumed in advance (for example, the resist height when no film reduction has occurred) is used, and the actual pattern height variation is not taken into consideration. . Details of the library matching method will be described later.

Next, the cross-sectional shape estimated in step S0004 is corrected using the height information measured in step S0003 (step S0005). First, a correction method in the case of a simple trapezoidal shape will be described with reference to FIG. As described above, in the shape change shown in FIG. 3A, the SEM image is not very sensitive to the height change. That is, even when the actual height becomes H ′ due to film reduction and the pattern shape is in the state of 202, when library matching is used using a library having the reference height H, the obtained cross-sectional shape is the pattern 201. It becomes. Therefore, if the shape estimation result estimated by the library matching method: height H, sidewall inclination angle θ and height H ′ measured in step S0003 is corrected from the geometric relationship as follows: Good.
Hmeas = H '(Formula 1)
θmeas = θ '= atan (H · tan (θ) / H') (Formula 2)
Here, Hmeas and θmeas are corrected values, that is, pattern shape estimation results output by the present embodiment. Since the edge position does not change, it is not necessary to correct the line width measurement results at the top and bottom. Although FIG. 3 has been described using only the left side wall inclination angle, the right side wall may be similarly corrected.

  FIG. 3 shows an example in which the target pattern shape is a simple trapezoid. A correction method in the case of a complicated shape such as a two trapezoid will be described with reference to FIG. 6 in the second embodiment.

  Based on the edge position and shape estimation result of the measurement target pattern obtained by the correction, the dimension of the pattern corresponding to the height designated in advance by the user is calculated (S0006). The result obtained at the end is displayed on the SEM image or displayed on the screen of the output unit 110 as numerical data (S0007). The output from the output unit 110 is sent to the data server 111 and stored. It can also be transmitted to another data processing device or storage device not shown.

[Build Library]
Next, a library construction method used in this embodiment will be described with reference to FIG. This library is constructed by the simulation unit 1092 of the calculation unit 109.

  FIG. 5 shows an example of the simulation waveform library 002 used in this embodiment. The simulation waveform library 002 is stored in the database 112, and the simulation input cross-sectional shape 009 and the SEM simulation waveform 010 corresponding thereto are recorded in the simulation waveform library 002.

When library matching measurement is performed to correctly measure the change in the sidewall inclination angle θ, a simulation is performed for a plurality of shapes having different inclination angles θ as shown in FIG. In this example, the case where there are three types of sidewall inclination angles θ is shown for the sake of simplicity, but in practice, the range covering the pattern shape that may occur due to process fluctuations is simulated with a fineness that matches the accuracy to be measured. The simulation waveform library 002 is created in advance separately from the measurement of FIG. At this time, the pattern height H at the time of creating the library is set to a constant value according to the actual pattern. In the case of resist pattern measurement, there is a possibility that the height may fluctuate due to film reduction, but the library is set to a height when no film reduction occurs, for example. Or, when height fluctuations are likely to occur, such as exposure condition evaluation, it is of course possible to use the median value of the range in which fluctuations are expected.
In the present invention, an SEM simulation is performed for such a combination of shape parameters, and the shape information 009 and the simulation waveform 010 are associated with each other and stored in the database 112 as a simulation waveform library 002. In the example of FIG. 5, for the sake of simplicity, only the pattern height and the inclination angle are shown. However, the library 002 includes data in which the rounding of the top corner Rt and the bottom corner Rb of the pattern is changed. Of course, it may be (in that case, it becomes a multidimensional space of two or more dimensions).

[Library matching]
Next, the waveform having the highest degree of coincidence in the simulation waveform library 002 is selected by comparing the simulation waveform library 002 formed as described above with an image of the actual measurement target pattern (library matching). This library matching is performed by the matching waveform selection unit 1093 of the calculation unit 109. In pattern dimension measurement, the distance between the edges on both sides of the pattern varies depending on the measurement object (the distance between the edges corresponds to the dimension to be measured). Therefore, the entire SEM waveform 0082 (FIG. 4B) in the SEM signal waveform processing area 0051 corresponding to the measurement processing window 303 arranged on the measurement target pattern 302 in the SEM image 301 of FIG. When matching processing is performed, a simulation waveform in which the distance between edges is also changed is required, and a large amount of calculation is required for matching.

  Therefore, in the measurement method of this embodiment, as shown in FIG. 5, the SEM signal waveform 0083 in the processing region 012 around the processing target edge is locally cut out from the SEM signal waveform 0082 in the evaluation region 005 of one pattern. Thereafter, waveform matching processing S0005 with the simulation waveform library 002 is performed for each edge. The waveform matching is performed by searching the limited simulation waveform library 002 for a simulation waveform in which the actual SEM signal waveform 0083 in the evaluation target window and the waveform of the edge portion most closely match. At this time, by estimating the x coordinate of the simulation waveform at the same time, not only the shape but also each edge position can be obtained correctly.

  Since the simulation waveform has a clear positional relationship between the cross-sectional side wall shape and the waveform based on the relationship with the input data at the time of simulation calculation, it is possible to accurately determine which coordinate of the simulation waveform is the desired edge position. For example, in FIG. 5, assuming that the bottom edge is measured, where the bottom edge position 0084 corresponds to the simulation waveform is added to the simulation result. At the time of measurement, after matching the simulation waveform 010 and the actual SEM waveform, the position on the actual SEM image 0083 corresponding to the edge position 0084 of the simulation waveform can be determined as the edge position. If the cross-sectional shape information estimated as the bottom edge position obtained in this way is used, the edge position at an arbitrary height can be determined. As the degree of coincidence of waveforms at the time of matching, for example, a simulation waveform that minimizes this may be selected using the sum of squares of differences between waveforms.

  In the example of FIG. 5, for simplicity, the example in which the side wall inclination angle θ is three kinds of discrete values is shown. Even during actual measurement, simulation conditions that can be prepared in advance are limited, and the shape parameters in the simulation waveform library 002 are discrete values. Therefore, there is simulation data of shape parameters that completely match in the simulation waveform library 002. Not necessarily. Therefore, at the time of actual matching, a shape parameter that most closely matches the actual waveform may be selected from the simulation waveform library 002, or may be estimated and used by interpolating waveforms under similar conditions.

  Simulation waveform interpolation can be performed by waveform interpolation. For example, if there is a simulation waveform f_theta80 (x) with a side wall inclination angle of 80 ° and a simulation waveform f_theta82 (x) with a side wall inclination angle of 82 °, the waveform f_theta81 (x) with a side wall inclination angle of 81 ° is obtained by f_theta81 (x ) = f_theta80 (x) + (f_theta82 (x) -f_theta80 (x)) * (82-81) / (82-80). Of course, depending on the data of the simulation waveform library 002, non-linear interpolation using a combination of two or more shape parameters may be performed. By such interpolation processing, it is possible to perform matching for parameters that are not in the simulation waveform library 002. If it can be treated as a continuous function by interpolation processing, it is possible to estimate matching parameters using a nonlinear optimization method such as the Levenberg-Marquardt method.

  By obtaining the simulation conditions having the highest degree of coincidence with the actual waveform by the above procedure, the side wall cross-sectional shape and the position of each edge can be estimated with high accuracy. In the example of FIG. 5, it is estimated from the library matching result that the sidewall inclination angle when the pattern height of the captured SEM image is H is θ2, and the simulation matches the partial waveform in the library matching area 012. The edge position in the SEM image can be obtained from the waveform displacement amount. By separately substituting the height H ′ estimated from the surface state of the pattern and the estimated side wall inclination angle θ = θ2 into (Equation 1) and (Equation 2), the cross-sectional shapes Hmeas and θmeas of the pattern can be obtained. Since the estimated side wall shape of the edge portion is obtained, the pattern dimension at an arbitrary height can be calculated. This process is performed by the pattern dimension calculation unit 1094 of the calculation unit 109.

  For the dimensions, if the user sets the sample height (top, bottom, center, position of 10% of the pattern height from the bottom, etc.) to be measured in advance, the cross-sectional shape obtained can be used at the corresponding location. The dimension can be obtained from the difference between the edge position and the facing edge position (step S0006). Thereby, dimension data at an arbitrary position of the pattern and information on the pattern cross-sectional shape can be displayed on the screen of the output unit 110 and provided to the user. In addition, information from the output unit 110 can be stored in the data server 111 via communication means.

  In the first embodiment, a method of measuring the pattern height variation using the brightness change of the SEM image calibrated in advance is disclosed, but the height measurement in steps S0002 and S0003 uses a probe such as AFM. Of course, a step measurement method or the like may be used. In this case, although extra measurement time is required, there is an advantage that prior calibration (the database 003 and the calibration procedure of FIG. 2) is unnecessary. Further, since the height is directly measured, it can be applied to materials other than resist. Since the AFM measurement only needs to know the height, it is not necessary to scan the side wall in detail, and a relatively high-speed measurement can be realized. When measuring the height and shape using only the AFM, it is difficult to obtain a stable measurement value unless sufficient data is acquired in consideration of the shape / dimension fluctuation (roughness) of the pattern. By combining a wide range measurement using an SEM image, it is possible to obtain a measurement value that is more stable than the case where measurement is performed with an AFM alone.

  As described above, by using the first embodiment of the present invention, it is possible to accurately measure the dimensions and shape of a pattern regardless of the shape and height of the pattern. Compared with the method disclosed in Patent Document 4, it is possible to obtain the same level of accuracy with a significantly smaller amount of library calculation.

[Output GUI]
An example in which the pattern shape estimation result obtained based on the above-described method is displayed on the screen of the output unit 110 in S0007 will be described with reference to FIG. FIG. 6A is a display example of the output result. A signal waveform 0082 is displayed on the measurement target image 301. If the estimated cross-sectional shape 0113 is displayed together with the displayed signal waveform, the estimation result can be easily confirmed. The cross-sectional shape 0113 displays a shape in which the pattern height and the shape estimation result are applied to a simple model (in this case, a trapezoid). At this time, the x coordinate of the edge obtained from the matching result is displayed so as to match the image 301 and the waveform 0082. Of course, the simulation waveforms selected by matching may be displayed in an overlapping manner. The estimated result is displayed as a numerical value in the result display area 0114. Of course, these numerical data may be output to a text file or the like.

  In addition, when measurement reproducibility evaluation or the like is performed in advance and variation in shape measurement is clear, an error range (shape estimation) of the shape estimation result is displayed as in the display example of the cross-sectional shape 0113 shown in FIG. (Accuracy) 0115 may be displayed so as to overlap the cross-sectional shape 0113, or may be written together in the result display area 0114 as 0116. By repeatedly measuring the same pattern and evaluating variation in other shape parameters, measurement errors can be quantitatively evaluated and used as numerical values for height measurement accuracy. . In measurement of a semiconductor pattern, local shape / dimension variation of the pattern called roughness occurs. For example, by evaluating the variation in one or a plurality of SEM images, in addition to the measurement accuracy, It is also effective to display the roughness components together as a result.

[Application to two trapezoidal models]
As a second embodiment, a measurement result correction method when the pattern shape model becomes more complicated will be disclosed. As shown in FIG. 3, the first embodiment shows an example of a simple trapezoidal model. A change in the shape of the resist pattern may not be expressed well by a simple trapezoid. In such a case, better measurement accuracy can be expected by using a model in which two or more trapezoids are stacked.

  FIG. 6 shows an example of a two trapezoid model. 7A and 7B, the left side shows the shape model used for the library, and the right side shows the correction result using the height measurement result. In both FIGS. 7A and 7B, the height H of the library model is the case of the lower Hl and the upper Hu, and the inclination angle estimated values θl and θu by the library matching and the line widths of the upper and lower parts of the pattern are Wtop. , Wbottom, the width of the trapezoidal connection part is Wmiddle. Even in the case of the two trapezoids, as in the case of FIG. 3, even when the height of the library does not completely match the actual height, the library matching method can be used to determine the top, bottom, and two trapezoid connection portions. Edge positions are measured correctly.

Therefore, FIG. 7A shows an example in which correction is performed with the ratio of the heights of the two trapezoids being constant. The corrected values, that is, the pattern shape estimation results Hu_meas, Hl_meas, θu_meas, and θl_meas output in this embodiment are obtained by the following calculation.
Hmeas = H '(Formula 3)
Hu: Hl = Hu ': Hl'
Hu_meas = Hu '= H' ・ Hu / H (Formula 4)
Hl_meas = Hl '= H' · Hl / H (Formula 5)
θu_meas = θu '= atan (Hu · tan (θu) / Hu_meas) (Formula 6)
θl_meas = θl '= atan (Hl · tan (lθ) / Hl_meas) (Formula 7)
FIG. 7B shows a correction method under the assumption that the height of the lower trapezoid is constant. When film loss occurs due to focus variation, the shape of the upper part of the pattern is likely to change, so the shape model in FIG. 7B may be appropriate.

Similarly to FIG. 7A, the pattern shape estimation results Hmeas, Hu_meas, Hl_meas, θu_meas, and θl_meas output in this embodiment in the case of the film reduction model of FIG. 7B are obtained by the following calculation.
Hmeas = H '(Formula 8)
Since Hl = Hl 'constant,
Hu_meas = Hu '= H'-Hl (Formula 9)
Hl_meas = Hl '= Hl (Formula 10)
θu_meas = θu '= atan (Hu · tan (θu) / Hu_meas) (Formula 11)
θl_meas = θl '= atan (Hl · tan (lθ) / Hl_meas) (Formula 12)
As for the models shown in FIGS. 7A and 7B, an appropriate model may be used according to the process to be measured. In FIG. 7, only the left side wall inclination angle has been described. However, the right side wall may be similarly corrected. With the above processing, even in the case of the two trapezoid model, accurate pattern shape measurement can be performed by performing correction using the entire height measurement result.

  The output of the measurement result in the second embodiment is the same as that in the first embodiment described with reference to FIG.

[Supplement of calibration method]
As a third embodiment, the details of the calibration method for pattern height measurement shown in FIG. 2 will be described with reference to FIG. FIG. 8A shows the relationship between the film thickness reduction index value and the actual film thickness reduction amount. The film reduction index value is calculated by using the roughness of the surface due to exposure of the pattern surface as an index value.If the film reduction amount exceeds a certain amount, the pattern cannot be formed well, and the surface roughness and film reduction are reduced. The index value increases in dispersion and loses sensitivity (right side from “measurement limit 400” in FIG. 8A). For this reason, even in the calculation of the calibration straight line shown in FIG. 2 of the first embodiment, data processing in consideration of this measurement limit is required. In a region having measurement sensitivity, although a certain variation due to noise in the SEM image or the like occurs, the film decrease amount and the film decrease index value calculated from the SEM image have a correlation. Therefore, referring to FIG. 8B, the measurement limit 400 using the film reduction index value and the calibration straight line 403 are obtained by using the actual film reduction amount measured by AFM and the film reduction index value obtained from the SEM image. Will be described (corresponding to step S1004 in FIG. 2).

  First, data obtained by measurement with the AFM is sorted in order of measurement values of the AFM (S2001). Next, out of the data obtained by measurement with AFM, using the data of the pattern in which no film reduction has occurred, the film reduction index value is calculated from the SEM image of the pattern in which no film reduction has occurred, An average value 401 and a variation 402 of the film loss index values of the calculated pattern in which no film loss occurs are calculated (S2002). Since the index value variation 402 is mainly a measurement variation caused by SEM image noise, reliable data can be selected based on the variation. In addition, the data of the pattern where film reduction has not occurred is stable even if the detailed exposure condition is not known by using a sufficiently large pattern that the surface is not exposed even if the exposure machine defocuses. A large amount can be acquired.

  Next, a measurement limit is set based on the variation in the film reduction index value of the pattern in which no film reduction occurs calculated in S2002 (S2003). For example, the variation is evaluated for each data with the same amount of actual film loss, and takes a value (for example, the variation is larger than 150%) that is significantly larger than the index value variation 402 when there is no film loss. May be the measurement limit. The determination of the measurement limit may be performed by a human using the graph of FIG. 8A without being automatically performed.

  Next, if the regression line of the data below the measurement limit is obtained by excluding data in the region exceeding the set measurement limit, the calibration line 403 with less noise can be calculated (S2004). At this time, by using the variation 402 of the index value of the pattern in which no film loss has occurred, it is possible to eliminate the large data 404 and to perform the calculation again. By eliminating abnormal values, a more accurate calibration line can be obtained. As described above, by using the procedure shown in FIG. 8B, it is possible to determine the limit of height measurement using the film thickness reduction index value and to quantify the expected accuracy and the calibration straight line. Using these results, it is determined whether or not the height correction is performed in the first embodiment, and if the height information is not reliable, the result is output to thereby provide higher accuracy and higher reliability. Shape measurement is possible.

  In the example described with reference to FIG. 8, an example in which the actual film reduction amount and the index value for film reduction are linear is shown. However, when the relationship cannot be expressed well by a straight line due to the detection characteristics of the SEM or the like. An appropriate curve model may be introduced.

  The display example of the output of the pattern shape estimation result obtained by the third embodiment is the same as the content shown in FIG. However, the trapezoidal shape shown in FIG. 6A is displayed as a two trapezoid as shown in FIG. 7A or 7B. Further, in the case where the pattern shape estimation result is output and displayed in the third embodiment, as a numerical value of the height measurement accuracy in the display example of the cross-sectional shape 0113 shown in FIG. 6B, for example, the calibration straight line shown in FIG. And the value of the variation from the straight line is displayed.

  In the above-described embodiments, the case where the cross-sectional shape of the pattern is one trapezoid and the case where one trapezoid is overlapped have been described. However, the present invention is not limited to this, and more complicated shapes are used. Similarly, by displaying the estimated cross-sectional shape on the screen, the measurement result can be easily confirmed.

  As mentioned above, although the invention made by the present inventor has been specifically described based on the embodiments, it is needless to say that the present invention is not limited to the above embodiments and can be variously modified without departing from the gist thereof. Yes.

  The pattern measurement technique disclosed in this specification can be applied as long as an image can be acquired with an electron microscope or a charged particle beam apparatus similar thereto.

001 ... SEM 101 ... electron gun 102 ... electron gun 103 ... focusing lens 104 ... deflector 105 ... objective lens 106 ... sample 107 ... detector 108 ... overall control / signal processing unit 109 ... calculation unit 1092 ... simulation unit 1093 ... matching waveform Selection unit 1094 ... Pattern dimension calculation unit 110 ... Output unit 111 ... Data server 009 ... Simulation input cross-sectional shape 010 ... SEM simulation waveform 002 ... SEM waveform simulation library 003 ... Correlation database of film reduction index value and pattern height 0084 ... Edge position 301 ... SEM image 302 ... Measurement target pattern 303 ... Measurement processing window 0113 ... Estimated cross-sectional shape 0114 ... Result display error Rear

Claims (9)

A method for measuring the shape of a pattern formed on a sample using a charged particle beam device,
By irradiating and scanning the focused charged particle beam on the sample, secondary charged particles generated from the sample are detected, and a charged particle beam image of a pattern formed on the sample surface is obtained,
Obtaining the height information of the pattern from the information of the acquired charged particle beam image of the pattern based on the relationship between the pattern height obtained in advance and the information of the charged particle beam image of the pattern,
A pattern shape measuring method, comprising: calculating the dimension of the pattern from information of a charged particle beam image of the pattern using the obtained height information of the pattern.   2. The pattern shape measuring method according to claim 1, wherein the height of the pattern is obtained in advance using means different from the charged particle beam apparatus.   In the step of measuring the shape of the pattern, the position of the edge of the pattern and the shape of the pattern are estimated from the charged particle beam image of the pattern, and the estimated position and shape of the edge of the pattern are obtained. 3. The pattern shape according to claim 1, wherein the pattern shape is corrected using pattern height information, and the dimension of the pattern is calculated using information on the corrected edge position and shape of the pattern. Measurement method.   In the step of measuring the shape of the pattern, the information of the charged particle beam image of the pattern is information on the roughness of the surface of the pattern, and the height of the pattern obtained in advance and the charged particle beam image of the pattern 4. The pattern shape measuring method according to claim 1, wherein the relationship with the information is information in which the roughness of the surface of the pattern is associated with the height of the pattern.   Estimating the position of the edge of the pattern and the shape of the pattern is compared with the cross-sectional shapes of various patterns having a constant height in which the signal waveform of the charged particle beam image of the pattern is stored in advance in the library. A signal waveform having a high degree of coincidence compared with the signal waveform of the charged particle beam image obtained by simulation is selected, and the pattern shape is determined from the cross-sectional shape of the pattern stored in the library corresponding to the selected signal waveform. The pattern shape measuring method according to claim 3, wherein an edge position and the shape of the pattern are estimated. An apparatus for measuring the shape of a pattern formed on a sample,
Charged particle beam means for detecting a secondary charged particle generated from the sample by irradiating and scanning the focused charged particle beam on the sample and acquiring a charged particle beam image of a pattern formed on the sample surface When,
Storage means for storing the relationship between the height of the pattern and information of the charged particle beam image of the pattern;
Based on the relationship between the height of the pattern stored in the storage means and the information of the charged particle beam image of the pattern, the information of the charged particle beam image of the pattern acquired by the charged particle beam means is used. Pattern height information extraction means for obtaining height information;
Pattern size calculation means for calculating the size of the pattern from the information of the charged particle beam image of the pattern acquired by the charged particle beam means using the height information of the pattern obtained by the pattern height information extraction means; A pattern shape measuring apparatus comprising:   The pattern shape calculation means estimates the edge position of the pattern and the shape of the pattern from the charged particle beam image of the pattern, and extracts the estimated height position and shape of the pattern from the pattern height information 7. The pattern according to claim 6, wherein correction is performed using pattern height information obtained by the means, and the dimension of the pattern is calculated using information on the corrected edge position and shape of the pattern. Shape measuring device.   The relationship between the previously obtained pattern height stored in the storage means and information of the charged particle beam image of the pattern is the height information of the pattern obtained by measuring the pattern with another device. 8. The information according to claim 6 or 7, wherein the information is associated with surface roughness information of the pattern obtained from a charged particle beam image acquired by imaging the pattern with the charged particle beam means. Pattern shape measuring device.   And further comprising library means for storing a signal waveform of a charged particle beam image obtained by simulation for cross-sectional shapes of various patterns having a constant height and cross-sectional shapes of various patterns having a constant height, The calculating means estimates the position of the edge of the pattern and the shape of the pattern, and various patterns having a constant height in which the signal waveform of the charged particle beam image of the pattern is stored in the library. A signal waveform having a high degree of coincidence compared with a signal waveform of a charged particle beam image obtained by simulation with respect to the cross-sectional shape of the pattern, and a cross-section of the pattern stored in the library unit corresponding to the selected signal waveform The size of the pattern is calculated by estimating the position of the edge of the pattern and the shape of the pattern from the shape. 7 pattern shape measuring apparatus according.

Images