JP6356551B2 - Pattern height measuring device and pattern height measuring method - Google Patents

Description

  The present invention relates to a pattern height measuring apparatus and a pattern height measuring method, and more particularly to a pattern height measuring apparatus and a pattern height measuring method for measuring a pattern height by irradiating a surface of a sample with an electron beam.

  In recent years, with the progress of miniaturization of semiconductor devices, mask patterns have been miniaturized and thinned. Therefore, not only the measurement of the two-dimensional mask pattern shape such as the line width but also the measurement of the height of the mask pattern that affects the transfer characteristics and the like is important.

  As a technique for measuring the height of the pattern, the surface of the sample is imaged with a scanning electron microscope (SEM), and the height of the pattern is determined based on the length of the shadow appearing in the obtained SEM image (secondary electron image). An estimation method has been proposed.

  In this method, a predetermined threshold value is set for the brightness value around the pattern, and the shadow length is obtained based on the intersection of the threshold value and the profile curve indicating the distribution of the brightness value of the secondary electron image. Yes.

JP 2012-177654 A JP 2005-310602 A JP 2000-146558 A

  However, the actual profile curve often varies due to the influence of noise in the SEM image. For this reason, if the shadow length is obtained based on the intersection of the threshold and the profile curve as in the prior art, it is likely to be affected by noise, and the reproducibility and measurement accuracy of the measurement results are reduced.

  Then, an object of this invention is to provide the pattern height measuring apparatus and pattern height measuring method excellent in the reproducibility and measurement accuracy of a measurement result.

  According to one aspect of the present invention, an electron beam scanning unit that scans while irradiating the surface of a sample with an electron beam, and a disk formed around the optical axis of the electron beam and divided into a plurality of regions in the circumferential direction. A secondary electron detector, a signal processing unit for generating a secondary electron image obtained by capturing the sample surface from a plurality of detectors in different directions based on a detection signal of the secondary electron detector, and a secondary electron image An image processing unit for calculating a height of a pattern formed on the sample surface based on the line, the image processing unit intersecting the pattern from the secondary electron image Extracting a line profile representing a luminance distribution along the line, a candidate for an approximate curve of a line profile in a region outside the edge of the pattern, a distribution function representing the distribution of secondary electrons, and the pattern The step of obtaining based on the weight and the step of obtaining candidates for the approximate curve of the line profile while changing the spread of the distribution function of the secondary electrons are repeated a plurality of times, and the error value with the line profile is minimized. A step of obtaining an approximate curve candidate, a step of obtaining a distribution function of a secondary electron distribution function that gives the candidate of the approximate curve having the smallest error value from the line profile, and a calibration whose pattern height is known in advance And a step of calculating the height of the pattern based on the correlation between the height of the pattern obtained from the sample for use and the spread of the distribution function.

  According to another aspect of the present invention, an electron beam scanning unit that scans while irradiating the surface of a sample with an electron beam and a disk shape around the optical axis of the electron beam are divided into a plurality of regions in the circumferential direction. A secondary electron detector, a signal processing unit for generating a secondary electron image obtained by capturing the sample surface from a plurality of detectors in different directions based on a detection signal of the secondary electron detector, and the secondary electron A pattern height measuring method performed in a pattern height measuring apparatus including an image processing unit that calculates the height of a pattern formed on the sample surface based on an image, wherein the electron beam scanning unit Scanning the surface of the sample while irradiating the surface, detecting the intensity of secondary electrons emitted from the surface of the sample by the irradiation of the electron beam with the secondary electron detector, and the signal processing. Generating a secondary electron image of the sample surface based on a detection signal from the secondary electron detector, and the image processing unit along the line intersecting the pattern from the secondary electron image. Extracting a line profile representing the distribution of brightness values, and the image processing unit, as a distribution function representing the distribution of secondary electrons, a candidate for an approximate curve of a line profile in a region outside the edge of the pattern; The line profile is obtained by repeating a step of obtaining based on a weight with the pattern and a step of obtaining an approximate curve candidate of the line profile while the image processing unit changes a spread of a distribution function of the secondary electrons. A step of obtaining a candidate for the approximate curve that minimizes an error value with respect to the line profile, and the image processing unit A step of obtaining a spread function of a distribution function of secondary electrons that gives the candidate of the approximate curve that becomes the smallest, and the image processing unit calculates the pattern height obtained from a calibration sample whose pattern height is known in advance and the pattern height Calculating a height of the pattern based on a correlation with a spread of a distribution function.

  According to the pattern height measuring apparatus and the pattern height measuring method of the above aspect, the line profile of the region outside the edge of the pattern is not detected based on one point such as the intersection of the threshold and the line profile. The height of the pattern is calculated based on the above. Therefore, it becomes difficult to be influenced by noise, reproducibility is improved, and the height of the pattern can be measured with an accuracy comparable to atomic force microscopy (AFM).

FIG. 1 is a block diagram showing a pattern height measuring apparatus according to the first embodiment. FIG. 2 is a schematic diagram showing the arrangement of secondary electron detectors of the pattern height measuring apparatus of FIG. FIG. 3 is a diagram illustrating an example of a secondary electron image and a line profile generated by the pattern height measuring apparatus of FIG. FIG. 4 is a diagram showing the movement of electrons near the sample surface. FIG. 5A is a diagram showing the distribution of secondary electrons emitted from the sample surface, and FIG. 5B is a diagram showing changes in the irradiation position of the electron beam 31 and the detected amount of secondary electrons. . FIG. 6 is a diagram showing how the variance σ varies with the pattern height. FIG. 7 is a diagram illustrating calculation results of line profiles and approximate curves of patterns having various heights. FIG. 8 is a flowchart showing the pattern height measuring method according to the first embodiment. FIG. 9 is a flowchart showing a method for obtaining the value of the variance σ from the difference profile. FIG. 10 is a flowchart showing how to obtain the correlation between the pattern height and the variance σ. FIG. 11 is a diagram illustrating a difference profile when the interval between adjacent patterns is narrow. FIG. 12 is a flowchart showing how to obtain the approximate curve of the difference profile and the variance σ when the interval between adjacent patterns is narrow. FIG. 13 is a diagram illustrating a result of obtaining an approximate curve in consideration of the influence of two edges of adjacent patterns. FIG. 14 is a flowchart showing a method for reproducing a cross-sectional shape according to the third embodiment. FIGS. 15A to 15D are diagrams illustrating examples of a difference profile, an approximate curve, a correction profile, and a cross-sectional shape, respectively.

  Hereinafter, embodiments will be described with reference to the accompanying drawings.

(First embodiment)
FIG. 1 is a block diagram showing a pattern height measuring apparatus 100 according to this embodiment.

  As shown in the figure, the pattern height measuring apparatus 100 includes a chamber 2 that accommodates the sample 8, an electron beam scanning unit 1 that irradiates the sample 8 with the electron beam 31, and control and measurement data of each part of the pattern height measuring apparatus 100. And a data processing unit 10 that performs the above processing.

  Among these, the chamber 2 is provided with a stage 7 for holding a sample 8 such as a wafer or a photomask via a support 7a. The stage 7 operates based on a control signal from the data processing unit 10 and moves the observation region of the sample 8 to an electron beam irradiation range by the electron beam scanning unit 1.

  The electron beam scanning unit 1 has an electron gun 3 and emits an electron beam 31 from the electron gun 3 at a predetermined acceleration voltage. The electron beam 31 is converged by the condenser lens 4, focused by the objective lens 6, and irradiated on the surface of the sample 8. Then, the irradiation position of the electron beam 31 is scanned by the deflector 5.

  When the electron beam 31 is irradiated in this way, secondary electrons 32 are emitted from the surface of the sample 8. The emitted secondary electrons 32 are detected by the secondary electron detector 9 disposed above the sample 8.

  FIG. 2 is a schematic diagram showing the secondary electron detector 9 of the pattern height measuring apparatus 100.

  As shown in FIG. 2, the secondary electron detector 9 includes a circular hole 9f provided in the vicinity of the optical axis of the electron beam 31 and scintillators 9a to 9d formed in a disk shape around the hole 9f. . The disk-like scintillators 9a to 9f are divided into four at an equal angle (90 °) in the circumferential direction, and each outputs an independent detection signal. Here, each region of the secondary electron detector 9 divided into four is referred to as first to fourth detectors 9a to 9d, respectively.

  The scanning of the electron beam 31 is performed in a region called a rectangular observation region 81 set on the surface of the sample 8. The first to fourth detectors 9 a to 9 d are arranged in the directions of the four corners of the observation region 81. Since the secondary electron images obtained from the detectors 9a to 9d appear in different directions of the pattern shadow, three-dimensional information of the pattern can be obtained by using these secondary electron images.

  As shown in FIG. 1, the signal of the secondary electron detector 9 is processed by the data processing unit 10 and used for generating a secondary electron image, measuring the height of the pattern, and the like.

  Signals ch1 to ch4 transmitted from the first to fourth detectors 9a to 9d (see FIG. 2) are input to the signal processing unit 11 of the data processing unit 10. The signal processing unit 11 converts the input detection signal into digital luminance value data. Further, the signal processing unit 11 generates a secondary electron image by associating the luminance value data with the irradiation position coordinate data of the electron beam 31 and causes the display unit 20 to display the secondary electron image. Further, the secondary electron image generated by the signal processing unit 11 is stored in the storage unit 13 so as to be used for a pattern height measurement process described later.

  The image processing unit 12 reads out the generated secondary electron image from the storage unit 13, generates various images necessary for pattern height measurement, extracts a luminance value distribution (line profile), and measures the pattern height. Process.

  3A to 3C are diagrams illustrating examples of secondary electron images and line profiles generated by the image processing unit 12.

  FIG. 3A is a secondary electron image obtained by adding the detection signals of the lower left first detector 9a and the upper left second detector 9b. Since this secondary electron image is equivalent to the secondary electron image obtained from the secondary electron detector disposed on the left side of the paper, it is called a left image here. In the left image, the detection amount of secondary electrons decreases near the right edge of the pattern 91, and a decrease in luminance value appears as a shadow.

  FIG. 3B is a secondary electron image obtained by synthesizing the detection signals of the upper right third detector 9c and the lower right fourth detector 9d. This secondary electron image is equivalent to the secondary electron image obtained when the secondary electron detector is arranged on the right side of the paper, and is referred to as a right image here. In the right image, the detection amount of secondary electrons decreases near the left edge of the pattern 91, and a decrease in luminance value appears as a shadow.

  FIG. 3C shows a luminance value distribution (line profile) when the pattern 91 is irradiated with an electron beam, and corresponds to a line segment B in FIGS. 3A and 3B.

  In the figure, a line profile L is a line profile extracted from the left image, a line profile R is a line profile extracted from the right image, and a curve L + R is a line obtained by adding the line profile L and the line profile R. It is a profile.

  The line profile LR is a line profile (difference profile) obtained by subtracting the value of the line profile R from the line profile L.

  In the difference profile, the luminance value of the flat portion is canceled out, and the luminance value changes according to the unevenness of the sample surface. As shown in the figure, in the difference profile (LR), the change in luminance value is emphasized on the positive side near the left edge of the pattern 91, and the luminance value is emphasized on the negative side near the right edge. . As described above, the difference image is suitable for measuring the height of the pattern 91 because an image in which the unevenness information of the pattern 91 is emphasized is obtained. In the following description, an example in which the pattern height is obtained using this difference image will be described.

  Hereinafter, the principle of pattern height measurement according to this embodiment will be described.

  FIG. 4 is a diagram showing the movement of electrons near the left edge 91a of the pattern 91. As shown in FIG.

  Focusing on the line profile R of the right image, in the region I on the left side of the lower end portion of the edge 91a of the pattern 91, the detection amount of the secondary electrons 32 of the right-side secondary electron detector decreases, and the line profile has a recess. appear. This is because a part of the secondary electrons 32 emitted from the sample surface is blocked by the edge 91a (side wall) of the pattern 91, so that it is difficult to be detected by the secondary electron detector on the right side. In the region II on the side wall of the pattern, since the area of the secondary electron emission portion increases in the vicinity of the edge 91a, the amount of secondary electron emission increases and a peak appears in the line profile. In addition, in the region III at the upper end of the pattern, the luminance value settles to a constant value.

  When attention is paid to the line profile L of the left image, since secondary electrons are detected without being blocked in the region I, no concave portion appears in the line profile L. In the region I, secondary electrons and reflected electrons that hit the edge 91a go to the secondary electron detector on the left side, so that the detection amount increases and gradually as the position of the electron beam 31 approaches the edge 91a of the pattern 91. The brightness value increases.

  Such a change in the line profile in the region I is reflected in the difference profile (LR) and appears as an increase in luminance value.

  FIG. 5A is a diagram showing the distribution of secondary electrons emitted from the sample surface, and FIG. 5B is a diagram showing changes in the irradiation position of the electron beam 31 and the detected amount of secondary electrons. .

  As shown in FIG. 5A, most of the secondary electrons emitted from the sample surface and reaching the detector have a cone shape with a predetermined angle θ with respect to the electron beam 31 with the irradiation position of the electron beam 31 as a vertex. Are released into the release region 33 of

  When attention is paid to the cross section at the height H of the pattern to be measured in the emission region 33, the distribution of secondary electrons is represented by a Gaussian distribution function centered on the irradiation position of the electron beam 31 as shown by a curve 34. The value of the variance σ representing the spread of the Gaussian distribution function changes according to the height H of the pattern.

  As shown in FIG. 5B, when the electron beam 31 is scanned toward the pattern 91, the conical emission region 33 overlaps with the pattern 91, so that the amount of secondary electrons reaching the right detector is increased. The luminance value of the difference profile changes with decreasing. The amount of change in the luminance value of the difference profile changes according to the overlap of the distribution function of secondary electrons in the emission region 33 and the pattern 91. That is, secondary electrons existing in the overlapping region 35 between the emission region 33 and the pattern 91 strike the side wall of the pattern 91 and are prevented. In the difference profile, the decrease in the luminance value of the right detector is inverted and added, so that the luminance value increases at the left edge in the difference profile.

  The luminance value of the difference profile at this time is obtained by integrating (Gaussian integration) a Gaussian distribution function representing the distribution of secondary electrons within the overlapping region 35. By calculating the Gaussian integral value while moving the position of the center of the Gaussian distribution function along the line shape of the difference profile, an approximate function (Gaussian integral profile) of the difference profile is obtained.

The Gaussian integration profile I (x) can be obtained by the following equation using an error function.

  Here, x is the irradiation position of the electron beam, μ is the position of the upper end of the edge of the pattern 91, m is a coefficient, and σ is a dispersion value of the normal distribution.

  The value of the dispersion σ representing the spread of the distribution function of secondary electrons changes according to the height H of the pattern 91 to be measured.

  FIG. 6 is a diagram showing a change in the dispersion σ depending on the pattern height H. FIG.

As shown, in the case of relatively high height H ₁ pattern is interrupted secondary electrons generated in a wider range. For this reason, the value of the variance σ ₁ representing the spread of the distribution function of secondary electrons that affects the luminance distribution becomes large.

On the other hand, in the case of a pattern having a relatively low height H ₂ , the range for blocking secondary electrons is narrowed, and the value of the variance σ ₂ that affects the luminance distribution is small.

  FIG. 7 is a graph showing difference profiles near edges of patterns having various heights and calculation results of approximate curves.

  As shown in the figure, it can be seen that the approximate curve obtained based on the overlap between the Gaussian distribution function of the secondary electrons and the pattern agrees well with the difference profile in the region outside the edge of the pattern. It can also be seen that the value of the variance σ of the approximate curve changes in proportion to the pattern height.

The height H of the pattern 91 can be expressed by a linear function of variance σ as shown in the following equation (2).

  Therefore, in the present embodiment, a coefficient k indicating a correlation between the variance σ and the height H of the pattern 91 is experimentally obtained in advance.

  Hereinafter, a pattern height measuring method performed by the pattern height measuring apparatus 100 (see FIG. 1) will be described.

  FIG. 8 is a flowchart showing a pattern height measuring method according to this embodiment.

  First, in step S11, the pattern height measuring apparatus 100 irradiates and scans the electron beam 31 with the electronic scanning unit 1 to observe the sample surface. Secondary electrons emitted from the sample surface are detected by a secondary electron detector 9. The detection signal is converted into luminance value data of a digital signal by the signal processing unit 11 of the data processing unit 10, and the secondary electron image is generated by associating the irradiation position of the electron beam 31 with the luminance value data, and the storage unit 13.

  Next, the process proceeds to step S12, and the image processing unit 12 of the data processing unit 10 generates image data necessary for measurement of the pattern height. For example, when the edge of the pattern extends in the vertical direction, after generating the left image and the right image, a difference image is generated by subtracting the luminance value of the right image from the luminance value of the left image. Thereafter, the generated difference image is stored in the storage unit 13.

  Next, the process proceeds to step S13, and the image processing unit 12 extracts a difference profile of the pattern to be measured. Here, a line that intersects the edge of the measurement target pattern is set. Next, a difference profile that is a distribution of luminance values of the difference image along the line is extracted.

  Next, the process proceeds to step S14, and the data processing unit 10 obtains the value of variance σ from the difference profile.

  FIG. 9 is a flowchart showing a method for obtaining the variance σ from the difference profile in the present embodiment.

First, the process proceeds to step S21 in FIG. 9, and the image processing unit 12 cuts out the difference profile. Here, the peak position of the luminance value is detected from the difference profile. Next, a differential profile is obtained by differentiating the difference profile in the vicinity of the peak, and the position (X _max ) of the upper end of the edge is detected from the minimum value of the differential profile. The position of the upper end of this edge is set as one end X _max of the cutout range of the difference profile.

Also, the luminance value is detected on the opposite side across the peak of the difference profile, and the portion where the luminance value is near 0 is detected as the other end X _min of the cutout range of the difference profile. Thereafter, the difference profile in the range between the start end and the end is extracted and stored in the storage unit 13.

In step S22, the image processing unit 12 sets the value of the counter n to 1 and inputs a predetermined initial value σ _min to the variance σ _n of the Gauss integral of secondary electrons.

Next, the process proceeds to step S23, and the image processing unit 12 obtains an approximate function I (x) of the luminance value in a range from X _min to X _max . Here, calculation is performed by substituting the value of variance σ _n into equation (1) and substituting X _max for μ. The coefficient m is determined by searching for a condition that minimizes the difference from the difference profile.

Then control proceeds to step S24, the image processing unit 12 calculates the error value E _n of the approximate curve and differential profile obtained in step S23. Here, obtaining an error value E _n by integrating in a range excised square sum of the difference between the luminance value and the luminance value of the difference profile of approximation function I (x).

Thereafter, in step S25, the error value E _n and the variance σ _n are stored in the storage unit 13.

In step S26, the image processing unit 12 determines whether the value of the variance σ _n is larger than the maximum value σ _max of a predetermined search range.

In step S26, when the value of variance σ _n is equal to or _smaller than a predetermined maximum value σ _max (NO), the process proceeds to step S27 and the value of counter n is incremented by one. Thereafter, after the value of variance σ _n is increased by Δσ in step S28, the process returns to step S23. Thereafter, the processes in steps S23 to S28 are repeated.

On the other hand, _if the value of the variance σ _n is larger than the predetermined maximum value σ _max in step S26 (YES), it is determined that the calculation of the approximate curve has been completed for all of the variance σ in the search range, and steps S23 to S23 are performed. After exiting the loop of S28, the process proceeds to step S29.

In step S29, the image processing unit 12 reads the array containing the error value E ₁ to E _n and variance σ ₁ ~σ _n from the storage unit 13, extracts the value of the variance sigma corresponding to the smallest error value E .

  The value of the variance σ extracted here is used as the value of the variance σ in the target pattern in the process of step S15 in FIG.

  After that, the process proceeds to step S15, and the image processing unit 12 substitutes the value of the variance σ in step S27 into the expression (2) representing the correlation between the pattern height H and the variance σ to obtain the pattern height H. Ask.

  With the above processing, the pattern height measurement by the pattern measuring apparatus 100 is completed.

(How to find the correlation between pattern height and variance)
FIG. 10 is a flowchart showing how to obtain the correlation between the pattern height H and the variance σ.

  First, calibration samples having patterns with different heights are prepared. In the calibration sample used here, for example, a pattern whose height is obtained with an atomic force microscope is formed. As a calibration sample, it is preferable to prepare a plurality of samples in which patterns having different heights are formed.

  Next, in step S31, the calibration sample is observed with the measuring apparatus 100, and a differential profile is extracted. The processing up to the extraction of the difference profile is the same as steps S11 to S13 in FIG.

  Next, the process proceeds to step S32, and the image processing unit 12 obtains the value of the variance σ from the difference profile. The method of obtaining the variance σ is performed by the method described in steps S21 to S29 in FIG.

  Next, the process proceeds to step S33, and it is determined whether the measurement has been completed for all the calibration samples.

  In step S33, when the measurement of the dispersion σ for all the calibration samples is not completed (No), the process returns to step S31, and the measurement of the dispersion σ for the other calibration samples is repeated. On the other hand, when the measurement of all the calibration samples is completed, the process proceeds to step S34.

  In step S34, the coefficient k ((2)) representing the correlation between the pattern height H and the variance σ by the least square method from the measurement result of the pattern height H and the variance σ value of the calibration sample. Request).

  Thus, the coefficient k representing the correlation between the pattern height H and the variance σ is obtained. Since the correlation between the pattern height H and the value of the dispersion σ varies depending on the material of the sample, it is preferable to obtain in advance according to the material.

  As described above, in the measurement of the height of the pattern according to the present embodiment, the spread (distribution σ) of the distribution function of secondary electrons is obtained using the difference profile of the cut-out range, and the pattern is determined based on the spread. Find the height of. Therefore, it is possible to perform pattern height measurement that is not easily affected by noise and that has excellent measurement accuracy and reproducibility. Moreover, in the pattern height measurement of this embodiment, it has been confirmed that the pattern height can be measured with an accuracy of about 1 nm to 2 nm, and the pattern height can be measured with an accuracy close to that of an atomic force microscope or a cross-sectional observation method. You can ask for it.

(Second Embodiment)
In the present embodiment, a method for measuring the height of a pattern when the space between adjacent patterns is narrow using the pattern height measuring apparatus 100 shown in FIG. 1 will be described.

  FIG. 11 is a diagram illustrating an example of a difference profile of a sample in which the interval between adjacent patterns is narrow.

  As shown in FIG. 11, when the interval between the adjacent patterns 91 is narrow, the influence of adjacent edges is exerted, so that a flat portion of the difference profile does not appear in the space 92 between the patterns.

  In such a case, the difference curve cannot be approximated well with the approximate curve obtained by the method described with reference to FIG. 9, and the exact height cannot be obtained.

  Therefore, in the present embodiment, the approximate curve of only the right edge of the space 92 and the approximate curve of only the left edge are separately calculated and then combined to obtain an approximate curve of the sparse 92 portion. Ask.

  Hereinafter, a specific procedure will be described.

  FIG. 12 is a flowchart showing how to obtain an approximate curve when the pattern interval is narrow.

First, in step S41, a difference profile targeted by the image processing unit 12 is cut out. Here, a difference profile is extracted for a region starting from the upper end position X _min of the left edge of the space 92 and ending at the upper end position X _max of the right edge. The positions of the upper ends of the left and right edges can be detected by obtaining the steepest slope in the difference profile.

Next, in step S42, the image processing unit 12 sets the value of the counter n to 1, and inputs a predetermined initial value σ _min to the variance σ _n of the Gaussian distribution of secondary electrons.

Next, the process proceeds to step S43, and the image processing unit 12 obtains an approximate function I _R (x) of the difference profile considering only the right edge in the range from the start end X _min to the end end X _max . Here, by substituting the value of the variance σ _n set in step S42 into equation (1) and substituting X _max for μ, the value of the luminance value with respect to the X coordinate is calculated, thereby obtaining the approximate function I _R (x ) Is obtained.

In step S44, the image processing unit 12 obtains an approximate function I _L (x) of the difference profile considering only the right edge. The approximate function I _L (x) may be calculated using the equation (1). However, when the left and right patterns have the same height, the approximate function I _R (x) may be obtained by inverting the approximate function I _L (x). . That is, paying attention to the fact that the difference profile has a symmetrical shape at the left and right edges, the approximate function I _R (x) is inverted in the X direction around the middle part between X _min and X _max , The approximate function I _L (x) is obtained by inverting the sign of the luminance value.

In step S45, the image processing unit 12 adds the approximate function I _R (x) considering the right edge and the approximate function I _L (x) considering the left edge. I (x) is obtained.

  FIG. 13 shows an example of an approximate curve of a difference profile between adjacent patterns.

By adding the approximate function I _R (x) and the approximate function I _L (x) considering the left edge, an approximate function I (x) as shown by a solid line in FIG. 13 is obtained.

Next, in step S46 in FIG. 12, the image processing unit 12 calculates the error value E _n of the approximation function I (x) and differential profile obtained in step S24. Here, an error value E between the difference profile and the approximate function I (x) is obtained by obtaining the sum of squares of the difference between the brightness value of the approximate function I (x) and the brightness value of the difference profile at each x coordinate. _{Find n} .

Thereafter, in step S47, the error value E _n and the variance σ _n are stored in the storage unit 13.

In step S48, the image processing unit 12 determines whether the value of the variance σ _n is larger than a predetermined maximum value σ _max .

In step S48, when the value of variance σ _n is equal to or _smaller than a predetermined maximum value σ _max (NO), the process proceeds to step S49, and the value of counter n is incremented by one. Then, after increasing the value of variance σn by Δσ in step S50, the process returns to step S43.

On the other hand, _if the value of the variance σ _n is larger than the predetermined maximum value σ _max in step S48 (YES), it is determined that the calculation of the approximate curve has been completed for all variances σ in the set range, and steps S43 to S43 are performed. Exit the loop of S50 and proceed to step S51.

In step S51, the image processing unit 12 reads the array containing the error value E ₁ to E _n and variance σ ₁ ~σ _n from the storage unit 13, extracts the value of the variance sigma corresponding to the smallest error value E .

  With the above processing, the variance σ when the pattern interval is narrow is obtained. Thereafter, as described in step S15 of FIG. 8, the pattern height H can be obtained based on the equation (2) representing the correlation between the pattern height H and the variance σ.

(Method for reproducing the cross-sectional shape of the pattern)
In the following, a description will be given of a pattern cross-sectional shape reproduction function that is additionally performed using the difference profile in the pattern height measuring apparatus 100 of FIG.

  Conventionally, a method of reproducing a cross-sectional shape of a pattern by integrating a difference profile has been proposed (see Patent Document 1).

  However, there is a problem that the cross-sectional shape of the sample surface cannot be correctly reproduced because the luminance value of the shadow at the edge of the pattern is accumulated in the integrated value.

  Therefore, in the pattern height measuring apparatus 100, the cross-sectional shape is more accurately reproduced by correcting the difference image in consideration of the influence of the shadow of the edge of the pattern. That is, the correction profile from which the influence of the shadow is removed is obtained by subtracting the approximate curve obtained by the method described with reference to FIG. 9 from the difference profile. Then, the cross-sectional shape of the pattern is reproduced by integrating the correction profile.

  Specific processing contents will be described below.

  FIG. 14 is a flowchart showing a method for reproducing the cross-sectional shape of the pattern according to this embodiment. 15A shows an example of a difference profile, FIG. 15B shows an approximate curve, FIG. 15C shows a correction profile obtained by subtracting the approximate curve from the difference profile, and FIG. ) Shows the cross-sectional shape of the pattern reconstructed from the difference profile.

  First, in step S61 of FIG. 14, the image processing unit 12 extracts a pattern difference profile. By this processing, a difference profile as shown in FIG.

  Next, the process proceeds to step S62, and the image processing unit 12 obtains an approximate curve of the left edge of the pattern. This approximate curve is obtained by the method described with reference to FIG.

  Next, in step S63, the image processing unit 13 obtains an approximate curve of the right edge of the pattern. The approximate curve of the right edge of this pattern is obtained by inverting the approximate curve obtained in step S52 and placing the left end of the approximate curve at the upper end position of the right edge.

  The approximate curve shown in FIG.15 (b) is obtained by the process of step S62 and step S63.

  Next, in step S64, the image processing unit 12 subtracts the approximate curve obtained in steps S62 and S63 from the difference profile in step S52. Thereby, the influence of the shadow is removed from the difference profile, and a correction profile as shown in FIG. 15C is obtained.

  Thereafter, in step S65, integration processing is performed on the difference profile obtained by subtracting the approximate curve.

  As a result, an accurate cross-sectional shape of the pattern can be reproduced without including the influence of the shadow.

  FIG. 15D is a diagram showing the cross-sectional shape of the pattern reconstructed based on the difference profile. The solid line in the figure shows the cross-sectional shape reconstructed by removing the influence of the shadow in the vicinity of the edge by the method of this embodiment, and the broken line is reconstructed by integrating the difference profile in FIG. The cross-sectional shape made is shown. As shown in the figure, when the cross-sectional shape is reconfigured using the difference profile as it is, a bulge that should not exist in the actual pattern appears at the bottom of the edge, and the accurate cross-sectional shape cannot be reproduced.

  On the other hand, in the cross-sectional shape reconstructed by removing the influence of the shadow near the edge, it was confirmed that the bulge near the edge can be removed and a more accurate cross-sectional shape can be reproduced. Further, according to the present embodiment, by removing the bulge in the vicinity of the edge, it is possible to obtain knowledge about the fine unevenness in the vicinity of the edge, which is difficult with the conventional method.

(Other embodiments)
In the above embodiment, the content of measuring the height of the pattern based on the difference profile from the image captured by the detectors in the two opposing directions has been described, but the present invention is not limited to this.

  The height of the pattern may be obtained based on the image captured by the detector in one direction.

  That is, as described with reference to FIGS. 4 and 5, the line profile of the shadow portion is obtained by integrating the overlapping region 35 of the secondary electron emission region 33 and the pattern 91 distributed in a conical shape. This value is reflected in the distribution of luminance values in the region I of the line profile R on the shadow side in FIG.

  Therefore, the region I of the line profile R in FIG. 4 is extracted, and a profile in which the sign of the value is inverted is obtained. Then, by performing the processing with reference to FIGS. 8 to 10 for the extracted profile, an approximate function can be obtained and the height of the pattern can be calculated from the value of the σ value.

  As described above, the height of the pattern can be obtained based on the image captured by the detector in one direction by extracting the profile in the vicinity of the shadow side edge from the line profile.

  DESCRIPTION OF SYMBOLS 1 ... Electron beam scanning part, 2 ... Chamber, 3 ... Electron gun, 4 ... Condenser lens, 5 ... Deflection coil, 6 ... Objective lens, 7 ... Stage, 7a ... Support body, 8 ... Sample, 9, 9a-9d ... Secondary electron detector, 10 ... Data processing unit, 11 ... Signal processing unit, 12 ... Image processing unit, 13 ... Storage unit, 20 ... Display unit, 31 ... Electron beam, 32 ... Secondary electron, 33 ... Emission area, 35 ... Overlapping area, 81 ... Observation area, 91 ... Pattern, 91a ... Edge, 92 ... Space, 100 ... Pattern height measuring device.

Claims (14)

An electron beam scanning unit that scans while irradiating the surface of the sample with an electron beam;
Wherein a disk shape on the electron beam around the optical axis, is divided into a plurality of regions in a circumferential direction, the electron beam detecting secondary electrons detecting the intensity of secondary electrons emitted from the sample surface by irradiation of And
A signal processing unit that generates a secondary electron image in which a plurality of detectors in different directions are captured on the sample surface based on a detection signal of the secondary electron detector;
An image processing unit that calculates a height of a pattern formed on the sample surface based on the secondary electron image, and a pattern height measuring device comprising:
The image processing unit
Extracting a line profile representing a luminance distribution along a line intersecting the pattern from the secondary electron image;
Obtaining a candidate for an approximate curve of a line profile in a region outside the edge of the pattern based on a weight of a distribution function representing the distribution of secondary electrons and the pattern;
The step of obtaining the approximate curve candidate of the line profile is repeated a plurality of times while changing the spread of the distribution function of the secondary electrons, and the approximate curve candidate having the smallest error value from the line profile is obtained. ,
Obtaining a spread of a distribution function of secondary electrons that gives a candidate for the approximate curve with the smallest error value from the line profile;
Calculating the height of the pattern based on the correlation between the height of the pattern obtained from a calibration sample whose height is known in advance and the spread of the distribution function;
A pattern height measuring device characterized in that: The distribution function of the secondary electrons is represented by a Gaussian distribution function centered on the irradiation position of the electron beam, and the spread of the distribution function of the secondary electrons is a variance σ of the Gaussian distribution function,
The image processing unit obtains an approximate curve candidate of the line profile by obtaining an overlap between the Gaussian distribution function and the pattern while moving the position of the center of the Gaussian distribution function along the line. The pattern height measurement apparatus according to claim 1, wherein the pattern height measurement apparatus is a pattern height measurement apparatus. The height of the pattern is represented by a linear equation of the variance sigma, the image processing section by assigning the value of variance sigma determined from the candidate of the approximate curve to a linear expression of the variance sigma, the height of the pattern The pattern height measuring device according to claim 2, wherein the pattern height measuring device is obtained.   The image processing unit calculates an error value between the line profile and the approximate curve candidate by taking a sum of squares of differences between the line profile and the approximate curve candidate. The described pattern height measuring device.   The image processing unit detects a portion having the largest falling slope of the line profile as the position of the upper end of the edge of the pattern, and in the region outside the position of the upper end of the edge of the pattern, The pattern height measuring apparatus according to claim 4, wherein a candidate is calculated and the error value is calculated. The image processing unit
Generating two secondary electron images reflecting the sample surface from two directions orthogonal to the edges of the pattern and facing each other;
Generating a difference image taking a difference value of the two secondary electron images;
Extracting a luminance value distribution along a line orthogonal to the pattern from the difference image;
The pattern height measuring apparatus according to claim 5, wherein the line profile is extracted by performing the operation. The image processing unit
Obtaining a correction profile obtained by subtracting the approximate curve from a line profile extracted from the difference image;
Reproducing the height distribution of the sample surface along the line by integrating the correction profile; and
The pattern height measuring device according to claim 6, wherein: An electron beam scanning section for the electron beam is scanned while irradiating the sample surface, and wherein a disk shape on the electron beam around the optical axis, in the circumferential direction is divided into a plurality of areas secondary electron detector, the two A signal processing unit that generates a secondary electron image obtained by detecting a plurality of detectors in different directions based on a detection signal of a secondary electron detector; and a signal processing unit that is formed on the sample surface based on the secondary electron image. A pattern height measuring method performed in a pattern height measuring apparatus including an image processing unit for calculating the height of the pattern,
Scanning while irradiating the surface of the sample with an electron beam in the electron beam scanning unit;
Detecting the intensity of secondary electrons emitted from the surface of the sample by irradiation of the electron beam with the secondary electron detector;
In the signal processing unit, generating a secondary electron image of the sample surface based on a detection signal from the secondary electron detector;
The image processing unit extracting a line profile representing a distribution of luminance values along a line intersecting the pattern from the secondary electron image;
The image processing unit obtaining a candidate for an approximate curve of a line profile in a region outside the edge of the pattern based on a weight of a distribution function representing the distribution of secondary electrons and the pattern;
The image processing unit repeats the step of obtaining a candidate for the approximate curve of the line profile a plurality of times while changing the spread of the distribution function of the secondary electrons, and the approximate curve of the approximate curve having the smallest error value with the line profile is Seeking a candidate,
The image processing unit obtaining a spread of a distribution function of secondary electrons that gives a candidate of the approximate curve that minimizes an error value with a line profile;
The image processing unit calculates the height of the pattern based on the correlation between the height of the pattern obtained from a calibration sample whose pattern height is known in advance and the spread of the distribution function;
A pattern height measuring method characterized by comprising: Distribution function of the secondary electrons is represented by a Gaussian distribution function around the irradiation position of the electron beam, the spread of the distribution function of the secondary electrons is the variance σ of the Gaussian distribution function,
The image processing unit obtains an approximate curve candidate of the line profile by obtaining an overlap between the Gaussian distribution function and the pattern while moving the position of the center of the Gaussian distribution function along the line. The pattern height measuring method according to claim 8, wherein the pattern height is measured. The height of the pattern is represented by a linear equation of the variance sigma, the image processing section by assigning the value of variance sigma determined from the candidate of the approximate curve to a linear expression of the variance sigma, the height of the pattern The pattern height measurement method according to claim 9, wherein the pattern height measurement method is obtained.   The pattern height measurement according to claim 10, wherein an error value between the line profile and the approximate curve candidate is calculated by taking a sum of squares of differences between the line profile and the approximate curve candidate. Method.   A portion having the largest falling slope of the line profile is detected as the position of the upper end of the edge of the pattern, and the approximate curve candidate is calculated and the error value is calculated in a region outside the upper end of the edge of the pattern. The pattern height measurement method according to claim 11, wherein calculation is performed. Generating two secondary electron images reflecting the sample surface from two directions orthogonal to the edges of the pattern and facing each other;
Generating a difference image taking a difference value of the two secondary electron images;
Extracting a luminance value distribution along a line orthogonal to the pattern from the difference image;
The pattern height measurement method according to claim 12, wherein the line profile is extracted by performing the operation. Obtaining a correction profile obtained by subtracting the approximate curve from a line profile extracted from the difference image by the image processing unit;
The image processing unit integrates the correction profile to reproduce the height distribution of the sample surface along the line;
The pattern height measuring method according to claim 13, comprising: