JP2010085376A - Method for measuring pattern using scanning electron microscope - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method of removing blur of a pattern image acquired by an SEM and easily and accurately measuring the three-dimensional shape of a pattern of a sample surface based on information of the height direction of a pattern acquired by another measuring machine such as an AFM. <P>SOLUTION: A converged electron beam is radiated to a pattern of the sample surface to scan it, an SEM image is acquired, occurrence efficiencies of secondary electrons of the pattern is calculated by deconvolution operation based on the contrast shape of the SEM image created from the SEM image and the intensity distribution of the primary electron beam, occurrence efficiencies of the secondary electrons are compared and collated with each other using the occurrence efficiencies of the secondary electrons of the pattern predicted based on the information of the depth direction previously acquired by measuring the shape of the pattern with a scanning probe microscope, and the three-dimensional shape of the pattern is determined. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a pattern measurement method using a scanning electron microscope, and more particularly to a pattern measurement method for acquiring a highly accurate image of a pattern using a scanning electron microscope and measuring a three-dimensional shape of the image.

  Depth, height, and line width of patterns formed on semiconductor and semiconductor photomasks, nanoimprint templates, MEMS, optical-related elements, etc. as patterns become finer and more precise Also, minute changes in the pattern shape, such as the inclination angle of the side walls and the roundness of the corners, greatly affect the product characteristics. Therefore, a technique for measuring these pattern dimensions and shapes with high accuracy is required. Furthermore, when the pattern size is on the order of 100 nm or less, not only the planar shape of the pattern (the shape seen from directly above) but also the three-dimensional shape such as the cross-sectional shape is measured, and high accuracy is achieved. There is a growing need to control pattern manufacturing processes and manage product quality.

  Nowadays, with the progress of pattern miniaturization, a scanning electron microscope (SEM) is generally used for dimension measurement of micropatterns, and a scanning electron microscope for length measurement is also called a CD-SEM. (Hereinafter, simply referred to as SEM in the present invention). Edge detection methods from SEM images are various for each SEM model in order to improve the length measurement reproducibility, and the calibration of the dimensions of the SEM images is performed on samples with a guaranteed pitch.

The generation efficiency of secondary electrons generated from the sample in the SEM is determined by a combination of the material, the incident angle of the primary electrons on the irradiation surface, the edge effect, shadowing, and the like. When the spread of the primary electron beam is sufficiently smaller than the change in the generation efficiency of the secondary electrons, a contrast image corresponding to the generation efficiency can be obtained.
However, in the pattern length measurement using SEM, the primary electron beam of SEM has a spread, and the region where secondary electrons are emitted due to the scattering of electrons emitted from the surface near the irradiation point. The secondary electron signal is electrically and numerically modulated due to various causes such as variations in the beam scanning signal (see Non-Patent Document 1). For this reason, the signal at the edge portion of the pattern by SEM actually spreads in a portion (referred to as a white band) having a large signal amount corresponding to the left and right pattern edges. As a result, in the measurement by the conventional SEM, for example, when the opening pattern is viewed in the differential profile, the inside of the bright line indicating the opening pattern shape is detected as an edge, and several nm to several nm from the actual pattern edge It has also occurred that a location that is 10 nm inside is defined as an edge. These phenomena also vary depending on the height and width of the pattern, its cross-sectional shape, the irradiation conditions of the primary electron beam, the pattern material, and the like. In particular, in the measurement of patterns such as semiconductors and photomasks that are becoming finer, the dimension measurement position in the cross-sectional shape of the pattern is important.

Therefore, in order to obtain a high-resolution, high-accuracy SEM image, for example, by deconvolution from the SEM image, the blur caused by the size of the primary electron beam and the size of the image signal generation region is removed, and high-resolution An image processing method for displaying an image is disclosed (see Patent Document 1). In addition, a pattern measurement method has been proposed in which SEM electron beam simulation waveforms are stored in advance as a library, a waveform having the highest degree of coincidence with an actual SEM image is selected, and a pattern shape is estimated (see Patent Document 2).
C. G. Frace, E .; Buhr, K .; Dirshert, PTB Meas. Sci. Technol. , Vol. 18, pp510-519 (2007) Japanese Patent Laid-Open No. 4-341471 JP 2007-218711 A

  However, Patent Document 1 only describes a high-resolution image from which blur is removed by deconvolution, and does not show anything about the three-dimensional shape of the image. Further, the pattern measurement method of Patent Document 2 requires a considerable amount of time to create a library, and requires a lot of time to measure many types of samples and patterns. There was a problem of being limited. As described above, the conventional pattern measurement method using a scanning electron microscope is a simple and highly accurate method for measuring the three-dimensional shape of a pattern that has become increasingly necessary as the pattern becomes finer. There was no problem.

  On the other hand, an atomic force microscope (hereinafter referred to as an AFM) that scans the surface of a sample with a probe called a cantilever in which a probe is provided at one end of a cantilever is used as a method for measuring a fine pattern size and shape on the sample surface. And the like) are known. However, in the shape measurement method by AFM, the tip of the tip has a certain size, and therefore the measured value depends on the shape of the tip of the tip used for measurement, particularly the radius of curvature. However, a fine pattern or a steep step structure has a problem in that it cannot express an accurate dimension or shape of the pattern. Further, in the measurement of the pattern dimension and shape, there is a problem that the measurement value by SEM and the measurement value by AFM are different.

  Therefore, the present invention has been made in view of the above problems. That is, the object of the present invention is to remove the blur of the pattern image obtained by the SEM, and combine it with the information on the height direction of the pattern obtained by other measuring machines such as AFM, and the three-dimensional shape of the pattern on the sample surface. It is in providing the method of measuring easily and accurately.

  In order to solve the above problems, a pattern measuring method according to the invention of claim 1 is a pattern measuring method for measuring a shape of a pattern on a sample surface using a scanning electron microscope, and converges on the pattern on the sample surface. The electron beam is irradiated and scanned, secondary electrons generated from the pattern are detected to obtain an SEM image, and the contrast shape of the SEM image created from the SEM image and the primary electron intensity of the electron beam From the distribution, the secondary electron generation efficiency of the pattern is calculated by a deconvolution operation, and the secondary of the pattern predicted based on information in the depth direction obtained by measuring the pattern with a scanning probe microscope in advance. Based on the generation efficiency of the secondary electrons of the pattern calculated by the deconvolution operation and information on the depth direction. By collating and comparing the the generation efficiency of secondary electrons of the pattern, it is characterized in that to determine the three-dimensional shape of the pattern.

  A pattern measurement method according to a second aspect of the present invention is the pattern measurement method according to the first aspect, wherein the scanning probe microscope is an atomic force microscope.

  A pattern measurement method according to a third aspect of the present invention is the pattern measurement method according to the first or second aspect, wherein the pattern image obtained by the scanning electron microscope inclines the electron beam or the sample, The incident angle to the sample is a plurality of pattern images obtained at two or more different angles.

  A pattern measurement method according to a fourth aspect of the present invention is the pattern measurement method according to any one of the first to third aspects, wherein the Fourier transform is performed at high speed in order to interpolate measurement data in the deconvolution operation. It is characterized by a Fourier transform and a finer sampling period of data.

  According to the pattern measurement method of the present invention, a three-dimensional shape of a fine pattern on a sample surface can be measured easily and accurately. A semiconductor having a fine pattern, a photomask for semiconductor, a template for nanoimprint, MEMS, and optical related There is an effect that contributes to quality improvement of products such as elements.

  Hereinafter, a pattern measurement method according to the best embodiment of the present invention will be described in detail with reference to the drawings.

  About the pattern measurement method of the present invention for measuring the line width and cross-sectional shape of the pattern on the sample surface using a scanning electron microscope, as a sample, a trench structure in which a steep groove is provided on an optically polished flat quartz glass substrate A case where the line pattern is formed will be described as an example.

  First, a scanning electron microscope is used to scan the pattern on the sample surface to be measured by irradiating the focused electron beam (primary electrons) and detect secondary electrons generated from the pattern. Then, the detected electron quantity is converted into a luminance signal to obtain a SEM image of the pattern. FIG. 1 is a planar SEM image of a trench structure pattern, and the black portion in the vertical direction at the center of the screen is the groove of the pattern. A white line portion in the horizontal direction (x direction) perpendicular to the groove indicates a cross-sectional shape measurement portion of the pattern to which the pattern measurement method of the present invention is applied.

  Here, the acquired SEM image will be described in more detail. The contrast shape of the SEM image corresponding to the cross section of the pattern (secondary electron cross-sectional intensity distribution of the SEM image) is obtained by simulation calculation. The contrast shape (indicated by h (t)) of the SEM image at the position t, as shown in the following formula (1), is the generation efficiency f (x) of the secondary electrons and the intensity distribution g (tx of the primary electron beam. ) And convolution by convolution integral. Here, it is assumed that the primary electron beam has a Gaussian distribution. FIG. 3 is a diagram for explaining convolution. The horizontal axis indicates the position (nm) of the cross section in the x direction of the pattern image, and the vertical axis indicates the normalized intensity. The assumed cross-sectional shape of the pattern is described in the figure.

  Next, the generation efficiency of secondary electrons in the present invention will be described in more detail. The generation efficiency of secondary electrons emitted from the sample surface depends on the material and shape of the measurement sample. As for the shape, as shown in the schematic diagram of the pattern cross-sectional shape and the generation efficiency of secondary electrons in FIG. 4, the incident angle of primary electrons (FIG. 4 (a)), the distribution angle of secondary electrons and shadowing (see FIG. 4). 4 (b)), depending on the edge effect (FIG. 4) c)).

  As for the incident angle of primary electrons shown in FIG. 4A, the depth in the substance in which electrons that can escape from the surface as secondary electrons are generated is about 10 nm, and how much primary electrons are in that range. The generation efficiency of secondary electrons is determined depending on whether or not it can stay. From this, it is known that the generation efficiency of secondary electrons is proportional to 1 / cos θ, where θ is the incident angle of primary electrons (e) with respect to the normal of the surface.

  Regarding the secondary electron emission angle distribution and shadowing shown in FIG. 4 (b), the electrons are random within the sample after the electrons incident on the sample have received so much scattering that the memory of the incident direction is lost. Has an angular distribution. In this case, the distribution angle distribution of secondary electrons from the surface is a cosine distribution, which is known to be proportional to cos θ with the normal to the surface being an angle (θ) of 0 degrees. In addition, it is considered that only secondary electrons emitted from the bottom of the concave pattern to the opening are emitted to the outside by shadowing, and are emitted in a triangular range connecting the primary electron incident point and the opening. By integrating the angular distribution, the generation efficiency of secondary electrons at each position can be obtained.

As for the edge effect shown in FIG. 4C, the edge effect occurs because secondary electrons produced by incident primary electrons (e) are likely to be emitted outside the sample when they are close to the edge. This effect is considered to be determined exponentially with respect to the distance (x) from the edge. Therefore, it can be added to the generation efficiency as being proportional to e ^{−x / a} (a is a constant related to the spread of secondary electrons).
As shown in FIGS. 4A to 4C, the generation efficiency of secondary electrons corresponding to the pattern shape can be obtained by combining these.

Further, the intensity distribution g (tx) of the primary electron beam is obtained from the following formula (2). Here, a ₀ is a value indicating the dispersion of the primary electrons in the Gaussian distribution, which is usually about 7 nm, and the full width at half maximum (FWHM: FWHM) indicating the width of the location taking half the peak intensity of the primary electrons. Full Width at Half Maximum) is approximately 12 nm.

  FIG. 5 shows the contrast shape h (x) of the SEM image at the position x obtained by the convolution. In FIG. 5, the horizontal axis indicates the position (nm) of the cross section in the x direction of the pattern image, the vertical axis indicates the normalized electron intensity, and the calculated value (solid line) and the actual measurement value (dotted line) by simulation as the contrast shape of the SEM image. Both values are shown. Moreover, the cross-sectional shape which the pattern assumed is described. In FIG. 5, the curves of both the calculated value and the actual measurement value indicating the contrast shape of the SEM image are almost overlapped, indicating that the reliability of the calculated value is high. FIG. 5 also shows an assumed cross-sectional shape and secondary electron generation efficiency f (x).

に対して、記号Fを用いて以下の数式（４）のように表す。

(Deconvolution)
Next, from the contrast shape of the SEM image created from the acquired SEM image and the intensity distribution of the primary electron beam, the generation efficiency of secondary electrons caused by the shape of the pattern is calculated by deconvolution operation.
Here, Formula (3) which is the definition of Fourier transform used to calculate the generation efficiency of secondary electrons

On the other hand, the following expression (4) is expressed using the symbol F.

Deconvolution is a technique that removes blur such as spectra caused by instrument functions. When the observed spectrum is given by convolution of the true spectrum and instrument function, the Fourier transform of the observed spectrum is true. It is the product of the Fourier transform of the spectrum and the Fourier transform of the instrument function. Therefore, a standard deconvolution can be calculated by a procedure in which “Fourier transform of measurement data” is divided by “Fourier transform of instrument function” and further inverse Fourier transform is performed. In the present invention, the generation efficiency f (x) of secondary electrons indicating the contrast shape of the pattern is expressed by the following equation (5) using the measurement data h (x) and the device function (here, the beam shape) g (x). ). Here, F ⁻¹ represents an inverse Fourier transform.

  FIG. 6 shows measurement data of the contrast shape h (x) of the SEM image used for deconvolution. The horizontal axis indicates the position (nm) of the cross section in the x direction of the pattern image, and the vertical axis indicates the normalized electron intensity. FIG. 7 shows an apparatus function g (x) used for deconvolution. In the present invention, the beam shape (intensity distribution) of the primary electron beam is shown, and the horizontal axis indicates the position (nm) of the cross section in the x direction. The vertical axis represents the normalized electron intensity.

  In the present invention, when the generation efficiency of secondary electrons changes abruptly, in order to accurately determine the position where the abrupt change occurs, the line profile is differentiated to calculate a differential profile, and the differential signal amount The maximum peak and the minimum peak are obtained, and the width of the line pattern can be obtained as a distance between the position of the maximum peak and the position of the minimum peak.

(Measurement data interpolation)
In the present invention, it is also preferable to interpolate measurement data in order to increase the accuracy of the secondary electron generation efficiency f (x) of the pattern in the above deconvolution operation. Therefore, the Fourier transform is a fast Fourier transform. For example, in the case of a fast Fourier transform algorithm, N = 1024 or 2048 can be used as 2 to the Nth power. Further, as a sampling cycle for digitizing continuous signals at a constant cycle interval, a method in which a data sampling cycle (pixel interval) is made finer than a cycle that is normally used can be used. Both fast Fourier transform and sampling cycle miniaturization are effective, but by performing both of them, the time can be shortened and the generation efficiency f (x) of secondary electrons with high accuracy can be obtained. .

  FIG. 8A shows, as an example, a planar SEM image of a pattern to be measured, and FIG. 8B shows the pattern contrast shape when the image area in the range indicated by the double arrow at the bottom of the SEM screen is sampled at 256 pixels. Show. FIG. 9 shows a pattern contrast shape when the same image region is measured at 1024 pixels with a sampling period of 1/4, and the pattern contrast shape shown in FIG. 9 is more accurate than FIG. 8B. I can read. Since the miniaturization of the sampling period is contrary to the data acquisition time, it is preferable to make the sampling period fine within a range of about 1/2 to 1/4 of a period normally used (for example, 256 pixels).

(Deconvolution result)
The result of deconvolution is shown in FIG. FIG. 10 shows the cross-sectional intensity distribution of the pattern calculated by the deconvolution operation with a thick solid line, the horizontal axis indicates the position (nm) of the cross section in the x direction of the pattern image, and the vertical axis indicates the normalized electron intensity. In FIG. 10, for reference, the SEM cross-sectional intensity distribution (contrast shape of the SEM image) and the device function (beam shape) used for deconvolution are indicated by thin solid lines.

(Pattern shape by scanning probe microscope)
On the other hand, the depth of the pattern of the sample is measured in advance with a scanning probe microscope. FIG. 2 is a cross-sectional shape diagram of the trench structure pattern shown in FIG. 1 measured by AFM, and shows that a groove (trench) having a depth of 100 nm is formed. FIG. 11 is a schematic diagram of a pattern cross section in the pattern measurement method of the present invention. Parameters used when calculating the generation efficiency of secondary electrons from the pattern shape are the trench width (Top width) at the top of the trench pattern, The trench depth (Depth) and the side wall angle (Side Wall Angle: θ).

  In the present invention, a scanning probe microscope that obtains information in the pattern depth direction (height direction) scans the sample surface with a probe called a cantilever provided with a probe at one end of a cantilever, A microscope that measures the size and shape of the pattern is used, and the AFM that is generally widely used is most preferable, but a cantilever type scanning probe microscope (Scanning Probe) such as a scanning tunneling microscope (STM). It can also be used with other microscopes collectively called Microscope (SPM). In the following description, a case where the information is used for AFM to obtain information in the depth direction of the pattern will be described as an example.

(Comparison of shapes)
Next, the generation efficiency of the secondary electrons of the pattern predicted based on the information in the depth direction obtained by measurement with the AFM is obtained, and the contrast shape of the pattern by the SEM calculated by the deconvolution operation and the AFM described above. The three-dimensional shape of the pattern is obtained by comparing and collating with the contrast shape of the pattern predicted based on the above information.
FIG. 12 shows the generation efficiency of secondary electrons of a pattern obtained by calculation based on the assumed cross-sectional shape, the horizontal axis indicates the position (nm) of the cross section in the x direction, and the vertical axis indicates the normalized electron intensity. Indicates.

  FIG. 13 shows the generation efficiency of secondary electrons (thin three solid lines) obtained by calculation based on the assumed cross-sectional shape of the trench structure pattern, and the generation efficiency of secondary electrons obtained by deconvolution based on experiments. (Bold one solid line) is a figure to compare and collate, both show the left side wall portion of the trench structure pattern, the horizontal axis is the position (nm) of the cross section in the x direction of the pattern image, and the vertical axis is normalized Represents strength. The generation efficiency of secondary electrons is shown by setting the side wall angles at three angles of 84 °, 85 °, and 86 °. When compared with the generation efficiency of secondary electrons by deconvolution, The closest angle is 85 °. Therefore, the side wall angle of the above pattern is required to be 85 °.

  Next, the trench width (Top Width) is obtained. 14 and 15 show the generation efficiency of secondary electrons obtained by calculation based on the assumed cross-sectional shape (5 thin solid lines) and the generation efficiency of secondary electrons obtained by deconvolution based on experiments ( 14 (a) is a diagram showing the generation efficiency of secondary electrons in the left side wall portion of the trench structure pattern shown in FIG. 1 and the generation efficiency of secondary electrons determined by deconvolution. FIG. 15A shows the secondary electron generation efficiency of the right side wall portion of the trench structure pattern shown in FIG. 1 and the secondary electron generation efficiency obtained by deconvolution. In FIG. 14 (a) and FIG. 15 (a), the horizontal axis represents the pattern position coordinate (nm), and the vertical axis represents the normalized intensity (nm). Yes.

  14 (b) and 15 (b) show the peak position of the secondary electron generation efficiency in FIGS. 14 (a) and 15 (a) on the vertical axis, and the trench width (Top width) (nm). ) On the horizontal axis, and the oblique straight line indicates the trench width at the peak position of the generation efficiency of secondary electrons. 14B and 15B, the horizontal straight line (experimental value) indicates the value of the trench width at the peak position of the pattern obtained by deconvolution based on the experiment. The intersection with the oblique straight line indicating the trench width at the peak position of the secondary electron generation efficiency is 62 nm. Therefore, the trench width (Top width) of the above pattern is required to be 62 nm.

  From the above results, according to the pattern measurement method of the present invention, the fine pattern having the trench structure on the sample has a trench width (Top width) of 62 nm, a trench depth (Depth) of 100 nm, and a side wall angle (Side Wall Angle). ) It was required to be a three-dimensional pattern of 85 °.

  In the above example, a single pattern is described as a pattern. However, the pattern measurement method of the present invention can be applied even if the pattern is a line / space pattern, a hole pattern, or the like having a predetermined pitch. In the above example, the pattern in the depth direction of the trench structure has been described. However, the pattern in the height direction can be applied, and the information in the depth direction of the present invention includes the information in the height direction. It is a waste.

(Other embodiments of the present invention)
In the pattern measurement method of the present invention, it is effective to use information of an SEM image obtained by observing a sample from an oblique direction in observing a side wall of a pattern or estimating a three-dimensional profile.

  In the present embodiment, the sample shape calculated by deconvolution operation from two or more pattern images obtained by SEM measurement at two or more different beam incidence angles and the intensity distribution shape of the primary electron beam. Using the generation efficiency of secondary electrons caused by the above and the generation efficiency of secondary electrons predicted based on the depth information measured in advance, and comparing them, the cross-sectional shape can be made more accurate It can be measured.

  As a method of acquiring an SEM image observed from an oblique direction, for example, a method in which an electron beam irradiated from an electron optical system is deflected, and a direction in which the electron beam is irradiated to the sample is inclined to capture an inclined image. A method of inclining and imaging the stage on which the sample is placed so that an arbitrary position can be observed, a method of mechanically inclining the SEM electron optical system itself, and the like are applied.

  FIG. 16A shows the contrast shape of the SEM image and the generation efficiency of secondary electrons when the tilt angle (tilt angle) of the sample 161 with respect to the electron beam 162 is 0 °, as shown in FIG. It is a figure which shows the assumed pattern cross-sectional shape, and the horizontal axis | shaft shows the position (nm) of the cross section of the x direction of a pattern image, and the vertical axis | shaft shows the intensity | strength normalized. FIG. 17A shows the contrast shape of the SEM image cut off and the generation efficiency of secondary electrons when the tilt angle (tilt angle) of the sample 171 with respect to the electron beam 172 is 3 ° as shown in FIG. It is a figure which shows the assumed pattern cross-sectional shape. Since the generation efficiency of secondary electrons is sensitive to the incident angle (tilt angle) of the electron beam, the measurement accuracy is further improved by using the method of this embodiment.

  As described above, the pattern measurement method of the present invention includes the depth, height, pattern width, sidewall inclination angle, corner of a fine pattern of a semiconductor, a photomask for semiconductor, a template for nanoimprint, MEMS, an optical element, and the like. The pattern shape such as roundness can be measured easily and accurately in three dimensions.

It is a plane SEM image of the pattern of a trench structure. FIG. 2 is a cross-sectional shape diagram of the trench structure pattern shown in FIG. 1 measured by AFM. It is a figure explaining the convolution with the intensity distribution of a primary electron beam, and the generation efficiency of a secondary electron. It is a schematic diagram for demonstrating pattern cross-sectional shape and the generation efficiency of a secondary electron. It is the contrast shape of the SEM image calculated | required by the convolution. It is the measurement data of the contrast shape of the SEM image used for deconvolution. This is a device function used for deconvolution. It is the planar SEM image of the pattern to be measured and the contrast shape when sampled at 256 pixels. This is the contrast shape when the same image area as in FIG. 8 is sampled at 1024 pixels with a sampling period of 1/4. It is the generation efficiency of secondary electrons of the pattern calculated by deconvolution. It is a schematic diagram of the pattern cross section which shows the item measured in this invention. This is the secondary electron generation efficiency of the pattern obtained by calculation based on the assumed cross-sectional shape. It is a figure which compares and collates the generation efficiency of the secondary electron which calculated | required the side wall angle of the trench structure pattern, and the generation efficiency of the secondary electron calculated | required by deconvolution. It is a figure which compares and collates the generation efficiency of the secondary electron calculated | required by calculation of the trench width in the left side wall part, and the generation efficiency of the secondary electron calculated | required by deconvolution. It is a figure which compares and collates the generation efficiency of the secondary electron which calculated | required the trench width by calculation in the right side wall part, and the generation efficiency of the secondary electron calculated | required by deconvolution. It is a figure which shows the contrast shape of the SEM image in case the tilt angle (tilt angle) of a sample is 0 degree, the generation efficiency of a secondary electron, and the assumed pattern cross-sectional shape. It is a figure which shows the contrast shape of the SEM image in case the tilt angle (tilt angle) of a sample is 3 degrees, the generation efficiency of a secondary electron, and the assumed pattern cross-sectional shape.

Explanation of symbols

161,171 Sample 162,172 Electron beam

Claims (4)

A pattern measuring method for measuring the shape of a pattern on a sample surface using a scanning electron microscope,
Scanning by irradiating the electron beam converged on the pattern on the sample surface, detecting secondary electrons generated from the pattern, obtaining an SEM image,
From the contrast shape of the SEM image created from the SEM image and the primary electron intensity distribution of the electron beam, the generation efficiency of secondary electrons of the pattern is calculated by a deconvolution operation,
Using the generation efficiency of secondary electrons of the pattern predicted based on the information in the depth direction obtained by measuring the pattern with a scanning probe microscope in advance,
By comparing and collating the generation efficiency of secondary electrons of the pattern calculated by the deconvolution operation with the generation efficiency of secondary electrons of the pattern predicted based on the information in the depth direction, A pattern measurement method characterized by obtaining a three-dimensional shape.   The pattern measurement method according to claim 1, wherein the scanning probe microscope is an atomic force microscope.   The SEM image obtained by the scanning electron microscope is a plurality of SEM images obtained by tilting an electron beam or a sample and obtaining an incident angle of the electron beam on the sample at two or more different angles. The pattern measuring method according to claim 1 or 2.   The pattern according to any one of claims 1 to 3, wherein, in the deconvolution operation, in order to interpolate measurement data, Fourier transform is fast Fourier transform, and a data sampling cycle is made fine. Measurement method.

Images