JP2007207763A - Charged-particle beam microscope, image formation method, and computer for evaluating image resolution - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image evaluation method capable of evaluating the image resolution of a microscope image objectively. <P>SOLUTION: In the image evaluation method, resolution in partial regions of an image is obtained over an entire area of the image or a portion of the image, averaging is performed over the entire area of the image or the portion of the image, and the averaged value is established as the resolution evaluation value of the entire area of the image or the portion of the image. This method eliminates the subjective impressions of the evaluator from evaluation of microscope image resolution, so image resolution evaluation values of high precision and good repeatability can be obtained for the evaluation value of the image resolution. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a charged particle beam microscope, an image forming method, and a computer for evaluating image resolution.

  In charged particle microscopes (scanning electron microscope SEM, scanning ion microscope SIM, etc.), there are roughly two types of conventional image resolution evaluation methods. The first method is a gap method in which a sample in which gold particles are vapor-deposited on carbon is observed with a microscope as in Patent Document 1 and evaluated by a gap distance between two minimum points of particles that can be visually confirmed on the observed image. . Another method is the FFT method in which the image resolution is evaluated from frequency analysis of the observed image data by two-dimensional Fourier transform (FFT), as described in Patent Document 2, for the observed image.

JP-A-5-45265 Japanese Patent Laid-Open No. 11-224640

  In the conventional first gap method, the gold particles to be imaged vary in size and shape, and the subjectivity of the evaluator enters between the minimum two points of the particles that can be visually confirmed. There is a problem that high accuracy and reproducibility cannot be expected in the evaluation value.

  Also in the conventional second FFT method, resolution evaluation by visual judgment is performed when creating a resolution judgment curve set on the power spectrum diagram of frequency analysis, and there is still a problem that the subjectivity of the evaluator is not completely removed. is there.

  Especially in the semiconductor manufacturing process, multiple microscopes have been used for quality control of semiconductor devices for a long time, but the image resolution has been evaluated by the conventional method. Machine differences) and time changes cannot be managed with high accuracy, and there is a problem that device manufacturing with small variations in device quality cannot be performed.

  An object of the present invention is to provide an objective image evaluation, and in particular to provide an image evaluation method capable of evaluating the image resolution of a microscope image based on objective judgment, and a microscope having a resolution evaluation function. is there.

  According to the present invention, in order to achieve the above object, a charged particle source, a detector for detecting secondary particles emitted by irradiating a sample with a charged particle beam emitted from the charged particle source, In the charged particle microscope including a control processor that forms an image based on the output of the detector, the control processor calculates the resolution of the plurality of regions from the plurality of regions including the plurality of pixels of the image, The average value of the resolution of the plurality of regions is calculated. Other objects and specific configurations of the present invention will be described in the section of the embodiment of the present invention.

  According to the present invention, since the subjectivity of the evaluator does not enter the image resolution evaluation process in the microscope, high accuracy and good reproducibility can be achieved with respect to the image resolution evaluation value.

  In the image resolution evaluation method using the magnification calibration sample, the optical performance is adjusted with high accuracy using the evaluation value signal, so that desired image resolution, magnification, image noise, and image quality can be maintained and managed with high accuracy.

  Furthermore, when the present invention is applied to a plurality of microscopes used for quality control of semiconductor manufacturing, differences (machine differences) and temporal changes for each microscope in image resolution and the like can be managed with high accuracy. Device manufacturing with low quality variation can be realized.

  An outline of a scanning electron microscope (hereinafter referred to as SEM) as an embodiment of the present invention will be described. The electron optical system of FIG. 1 focuses the charged particle beam (electron beam) 2 emitted from a charged particle source (electron gun) 1 that emits electrons as charged particles onto a sample 4 by a lens 3 in an arbitrary order. Can be scanned. The secondary particles (secondary electrons) 5 generated on the surface of the sample 4 by the electron beam irradiation are detected by the secondary particle detection system 6, and a control system 7 (control processor) having a function of image calculation control as image data. ). The sample 4 can be moved in all three-dimensional directions by the XYZ stage 8. The control system 7 also controls the charged particle source (electron gun) 1, the lens 3, the secondary particle detection system 6, the XYZ stage 8, and the image display device 9.

  In this embodiment, the electron beam 2 is scanned two-dimensionally (XY direction) on the sample 4 by a scanning coil (not shown). The signal detected by the secondary electron detector in the secondary particle detection system 6 is amplified by a signal amplifier in the control system 7 and then transferred to the image memory and displayed on the image display device 9 as a sample image. . The secondary signal detector may detect secondary electrons or reflected electrons, or may detect light or X-rays.

  An address signal corresponding to the memory location of the image memory is generated in the control system 7 or in a separately installed computer, converted to analog, and then supplied to the scanning coil. The address signal in the X direction is a digital signal that repeats 0 to 512, for example, when the image memory is 512 × 512 pixels, and the address signal in the Y direction reaches the address signal in the X direction from 0 to 512. It is a digital signal repeated from 0 to 512, sometimes plus 1. This is converted into an analog signal.

  Since the address of the image memory corresponds to the address of the deflection signal for scanning the electron beam, a two-dimensional image of the deflection region of the electron beam by the scanning coil is recorded in the image memory. Note that signals in the image memory can be sequentially read out in time series by a read address generation circuit synchronized with a read clock. A signal read corresponding to the address is converted into an analog signal and becomes a luminance modulation signal of the image display device 9.

  The image memory has a function of storing (combining) images (image data) in an overlapping manner for improving the S / N ratio. For example, a single completed image is formed by storing images obtained by eight two-dimensional scans in an overlapping manner. That is, a final image is formed by combining images formed in one or more XY scanning units. The number of images (frame integration number) for forming one completed image can be arbitrarily set, and an appropriate value is set in consideration of conditions such as secondary electron generation efficiency. Further, an image desired to be finally acquired can be formed by further overlapping a plurality of images formed by integrating a plurality of sheets. When the desired number of images is stored, or after that, blanking of the primary electron beam may be executed to interrupt information input to the image memory.

  In addition, when the number of integrated frames is set to 8, when the 9th image is input, a sequence may be provided in which the 1st image is deleted and 8 images remain as a result. In addition, when the ninth image is input, it is also possible to perform weighted averaging such that the accumulated image stored in the image memory is multiplied by 7/8 and the ninth image is added thereto. .

  The control system 7 is provided with an input device (not shown), and can specify an image capturing condition (scanning speed, the total number of images), a field of view correction method, and output and storage of images. Further, the control system 7 incorporates a storage medium (not shown) and can store various data.

  The apparatus according to the present invention has a function of forming a line profile based on detected secondary electrons or reflected electrons. The line profile is formed based on the amount of detected electrons when the primary electron beam is scanned one-dimensionally or two-dimensionally or the luminance information of the sample image. The obtained line profile is, for example, on a semiconductor wafer. It is used for measuring the dimension of the formed pattern.

  In the description of FIG. 1, the control system 7 is described as being integrated with or equivalent to the scanning electron microscope. However, the control system 7 is not limited to this, and a control processor provided separately from the scanning electron microscope body is described below. Processing as described may be performed. In that case, the detection signal detected by the secondary signal detector is transmitted to the control processor, the transmission medium for transmitting the signal from the control processor to the lens or deflector of the scanning electron microscope, and the transmission medium. An input / output terminal for inputting / outputting a received signal is required.

  Further, the apparatus of this embodiment stores in advance, for example, a condition (measurement location, optical conditions of a scanning electron microscope, etc.) for observing a plurality of points on a semiconductor wafer as a recipe, It has a function to perform observation.

  Alternatively, a program for performing the processing described below may be registered in a storage medium, and the program may be executed by a control processor that has an image memory and supplies necessary signals to the scanning electron microscope. That is, the embodiment of the present invention described below is also established as an invention of a program that can be used in a charged particle beam apparatus such as a scanning electron microscope equipped with an image processor. In addition, an evaluation sample with an image resolution described below is placed on the stage 8 of the present embodiment.

  A sample obtained by evaporating gold particles having a large atomic number on a deposition material such as carbon is usually used as an image resolution evaluation sample. FIG. 2 shows a representative example of the SEM image. The image is a digital image having a size of 512 × 512 pixels. The image resolution is calculated by the following evaluation algorithm. FIG. 8 shows the evaluation concept of the image resolution together with the gap resolution evaluation method, and FIG. 9 shows a flowchart for calculating the image resolution. Details will be described in the following examples.

[Example 1]
In this embodiment, a microscope image is regarded as an observation image from one direction of a three-dimensional object, and first, the gradient and curvature of the object plane at each pixel position P are obtained. Next, the virtual shortest distance necessary for recognizing the object position from the surroundings, that is, the local resolution is calculated from the minimum contrast necessary for recognition and the gradient. Then, the weighted harmonic average of the local resolution is calculated over the entire image, and the average value is regarded as the image resolution of the image representative. A specific example will be described below with reference to the image resolution calculation flowchart of FIG.

A microscope image is regarded as an image observed from one direction of a three-dimensional object, and geometric features such as a gradient of the object shape and a radius of curvature are calculated by distance image analysis. The image is a digital image composed of n × n pixels. The pixel position of the arbitrary point P on this image is (x, y), and the density is z. Now, in order to calculate the gradient of the object shape and the radius of curvature at the point P (x, y), a local region (partial region) having a size of m × m pixels (normally m = 5) around the point P is determined. Cut out and approximate the shape of the local surface with a multi-order equation z (x, y). Here, a quadratic is adopted as one of the multi-orders and approximated by a curved surface represented by the quadratic equation below (Formula 1).
z (x, y) = ax ² + by ² + cxy + dx + ey + f (Formula 1)

Regarding the matching of the curved surfaces, the coefficients a,..., F are determined so that the sum of squares of the matching error is minimized. If these coefficients are determined, the gradient g (x, y; θ) [θ is the azimuth angle of the gradient] of the local surface can be calculated using the following geometric calculation (Formula 2).
g (x, y; θ) = {(δz / δx) ² + (δz / δy) ² } ^1/2 (Formula 2a)
θ = arctan {(δz / δy) / (δz / δx)} (Formula 2b)

  Similarly, the radius of curvature Rc can be calculated. The fitting of the approximate curved surface has an effect of removing image noise. That is, if an image in which noise is mixed is partially cut out and approximated by a curved surface or a plane, the noise can be partially removed, and the gradient can be calculated while reducing the influence of noise.

  From the viewpoint of noise removal effect and shortening of calculation time, a local area size of 5 × 5 pixels (m = 5) is convenient, and the order of an equation suitable for this is preferably second order. If 3 × 3 pixels (m = 3) is adopted as the size of the local region, the calculation time is shortened. However, in the image pattern, the noise removal effect of frequency components exceeding a period of 3 pixels in particular is a secondary of m = 5 Worse than using equations. On the other hand, in an image that contains little noise in the image itself and contains many frequency components with a period less than 5 pixels in the image pattern, the gradient calculation accuracy is higher when 3 × 3 pixels are used as the size of the local region. In addition, when a plane (not a curved surface) is applied as an approximate surface (the display function formula is linear), the calculation time is faster, but the noise removal effect is worse than when a curved surface represented by a quadratic equation is used.

Here, an azimuth angle representing a specific azimuth angle θ _i −π / n to θ _i + π / n is represented by θ _i (i = 1, 2,..., N; n = 36), and the azimuth angle. A resolution obtained by adding a weight Wp of the local image resolution Rp having θ _i and performing harmonic averaging in the image is regarded as an azimuth image resolution R (θ _i ). The image resolution R of the entire image is obtained by geometrically averaging the azimuth image resolution capability R (θ _i ) at all azimuth angles. Thereby, for example, even if unexpected noise is mixed in a part of the microscope image, the calculated image resolution is hardly affected.
1 / R (θ _i ) = [∫ {Wp (x, y) / R _p (x, y; θ _I )} dxdy]
/ [∫Wp (x, y) dxdy] (Formula 3a)
Rp (x, y; θ _i ) = 2ΔC / | g (x, y; θ _i ) | (Formula 3b)
R = {R (θ ₁ ), R (θ ₂ ),..., R (θ _n )} ^{1 / n} (Formula 4)

Here, ΔC is a threshold contrast for recognizing the gap on the object shape corresponding to the resolution, and is proportional to the maximum amplitude of the density expected value E {z} calculated from (Equation 1). It was set as Kc. Usually, Kc = 0.1.
ΔC = Kc · [E {z (max)} − E {z (min)}] (Formula 5)

The local weight Wp is the gradient g only (expression 6a) or the product of g and [E {z (x, y)} − E {z (min)}] (expression 6b). In the evaluation image, when the gradient g is large and attention is paid to a bright part, the latter is adopted. Thereby, the influence of the brightness change on the resolution can be greatly reduced.
Wp (x, y) = | g (x, y) | (Formula 6a)
[E {z (x, y)} − E {z (min)}] · | g (x, y) |] (Formula 6b)

Here, when the fitting surface is a curved surface, Wp is taken up only when the half value (= Rp / 2) is equal to or smaller than the absolute value of the minimum curvature radius in order to calculate the local resolution with high accuracy. Was calculated from (Equation 6), otherwise Wp = 0 was set. Also, the standard deviation σ of density (corresponding to image noise), which is an image quality evaluation parameter of the image, can be calculated using the local noise σ _p .
σ = {1 / (n−2)} {Σ (σ _p ² )} ^1/2 (when m = 5) (Expression 7a)
σ _p = {1 / m} {Σ [z (x, y) −E {z (x, y)}] ² } ^1/2 (Expression 7b)

  An evaluation example by this resolution evaluation method will be described with reference to FIG. (a) And (b) is the image (512 * 512 pixel) image | photographed changing the optical conditions of SEM from the sample for resolution evaluation. In the image resolution Rcg obtained by this evaluation method and the conventional gap resolution Rgap, the resolution ratio of the conditions (a) and (b) and the errors caused by the plurality of evaluators are also displayed in comparison. Although it shows that R (b) / R (a) exceeds 1 at any resolution, the error by multiple evaluators is about ± 50% in the conventional gap method, but this method is the same. As long as it is image data, the error is zero.

As described above, the resolution evaluation algorithm calculated in this way determines all parameters necessary for the calculation only by information held in the image. Therefore, this evaluation method is an algorithm that does not include the evaluator's subjectivity and has the following characteristics.
1) Evaluator's subjectivity associated with the conventional gap resolution evaluation method can be eliminated. In other words, objective evaluation is possible.
2) Since the calculation is performed in real space, the calculation parameters are easy to understand as physical quantities (calculated in the frequency space in the FFT method).
3) Not only resolution but also standard deviation of density (image noise) can be calculated and evaluated.
4) Furthermore, the SN ratio, image quality, and observation magnification accuracy (or error) of the image can be calculated and evaluated.
5) Less susceptible to changes in brightness and image noise.
6) Applicable to pattern samples with strong directionality (not possible with FFT method). Less susceptible to sample contamination.
7) A resolution evaluation based on the same sample (image) can be performed between a plurality of apparatuses, and machine differences between apparatuses can be easily determined.

  In particular, the feature 1) enables objective comparison of image resolution between the same type of microscopes, and is very useful for user model selection. In this embodiment, the example in which the image resolution is obtained over the entire sample image (image) has been described. However, the present invention is not limited to this. For example, a specific region of the sample image (below the entire image region, the local resolution Rp is calculated). The resolution of the local area or higher) may be calculated. For example, an image area selecting means (pointing device or the like) (not shown) is connected to the control system 7 so that an arbitrary position and size on the image can be selected, and an area corresponding to the image selected by the selecting means is selected. The image resolution may be calculated. As a result, background information that is not directly related to the measurement target can be eliminated, and a realistic image resolution of the measurement target portion can be obtained. In addition, since only a specific part of the calculation is required, it is effective in improving the throughput of the image resolution calculation.

[Example 2]
Next, an embodiment of a scanning electron microscope (SEM) using this image resolution evaluation method will be described. The scanning electron microscope is only an example, and can be applied to all inspection apparatuses that need to evaluate images.

  In a production line for semiconductor products and the like, a plurality of SEMs 701 to 704 are network-connected to a management master computer 705 for length measurement management of semiconductor device patterns and the like as shown in FIG.

  In each SEM, an image resolution calculation function based on the above-described image resolution evaluation method is incorporated in a computer of the control system of the SEM, and the image resolution performance can be self-evaluated by an instruction from the apparatus operator. The resolution evaluation value is displayed using an image display device that has displayed a microscope image.

  In addition, in SEMs that have been used for length measurement management of device patterns and the like over a long period of time, each SEM periodically evaluates image resolution using a sample for image resolution evaluation, and displays and records along with transition information of the evaluation values. To do. This periodic resolution evaluation value is taken up by the master computer 705 and centrally managed together with information from other SEMs. When the image resolution evaluation value deviates from the setting allowable range or the setting value, the abnormality is notified to the apparatus operator both in the microscope and in the master computer. The master computer 705 includes an image display monitor and an image processor as described above, and the fact that the image resolution evaluation value is outside the setting allowable range or the setting value is displayed on the image display monitor. As a specific display form, transition information of evaluation values of a plurality of SEMs may be displayed in a graph, and the set allowable range may be identified and displayed as outside the allowable range, or as shown in FIG. The SEM may be displayed as a model, and a specific SEM model may be blinked when the setting allowable range or setting value is deviated. By displaying in this way, it is possible to take measures to prevent a decrease in measurement accuracy caused by a measurement error (machine difference) between the devices. In this embodiment, when the image output from the inspection apparatus deviates from the allowable range of the resolution evaluation value that should be included, high resolution is maintained by displaying or storing that fact. It is possible to manage not only the machine differences but also the inspection equipment.

  In addition, in the microscope in which the abnormality has occurred, adjustment of the optical system, which is one of the inspection conditions, is instructed by an instruction from the apparatus operator or an instruction from a program incorporated in the control system. With this signal, the lens system is controlled via the control system of the microscope, and adjustment is performed so that the image resolution falls within a desired tolerance. The progress information of this adjustment process is recorded in the abnormal SEM itself and reported to the master computer. In this example, each SEM self-evaluates the image resolution. However, the control master computer may evaluate the image resolution of the image sent from each SEM and send the evaluation value signal back to each SEM. Each SEM controls the lens system via the control system of the microscope in the same manner as described above based on the evaluation value signal, and adjusts the image resolution to fall within a desired allowable value.

  Thus, the resolution difference of a plurality of SEMs is simultaneously measured and managed, and if an abnormality is recognized, adjustment is performed using the above means. Thereby, the difference in the device quality between each manufacturing process is reduced, and the quality can be stabilized. In addition, it communicated with factories in different locations, making it possible to manufacture devices with stable quality between factories as well.

Example 3
Next, examples using the image resolution evaluation sample will be described in more detail. In this example, a dimensional calibration sample disclosed in Japanese Patent Application Laid-Open No. 8-31363 was used as a standard sample having an absolute dimension pattern as an image resolution evaluation sample. FIG. 4 shows the SEM image. The pitch dimension of the lattice pattern formed on the silicon substrate is very high accuracy of 0.240 ± 0.001 μm.

  By using this, the SEM image can be calibrated with high precision. That is, high-accuracy conversion can be performed between the size unit pixel of the digital image of the microscope and the actual size unit (for example, nanometer). Thereby, the resolution evaluation value R of the present invention can be displayed in length units (for example, nanometers). However, since the sample image has a vertical line shape, the evaluation image resolution here is not the image resolution calculated from the average of all the directions in (Expression 4), but the minimum value among the azimuth image resolutions calculated from (Expression 3a). Azimuth | direction image resolution of the azimuth | direction which shows (in the example of FIG. 4 horizontal) is used.

  Regarding the display of the resolution evaluation value R, the actual dimension (length) unit display is a value that does not depend on the observation magnification, and is convenient for actual use. On the other hand, the pixel unit display is convenient when a desired number of pixels is assigned to the image resolution value, that is, when the optimum observation magnification is directly calculated. In the case of display of image resolution, the unit is usually a length unit, but this can also be displayed in parallel in units of pixels or both units, as selected by the apparatus operator. Also, by setting the number of pixels to be assigned to the image resolution value in advance, it has become possible to automatically set the observation magnification for various samples.

The image quality evaluation parameters of the microscope image are an image resolution and a signal noise ratio (SNR). Here, when the average value Zav of the expected density value E {z} is assumed, the signal S can be approximated by S = Zav−E {z (min)}, and the noise N can be approximated by N = σ (see (Equation 7)). Can be calculated from the following equation.
SNR = [Zav−E {z (min)}] / σ (Expression 8)

If the maximum amount of information C (in bits) that can be held per pixel of the structural pattern is adopted as an image quality evaluation parameter by paying attention to a fine structure pattern of the image resolution value, C is the image resolution value R in pixel units. Approximate calculation from the following equation using
C = {log ₂ (1 + SNR)} / R ² (Formula 9)

  As a result, the image here can be evaluated from the viewpoint of image resolution, noise, and image quality. By setting an evaluation threshold value for judging whether the image is good or bad for at least one of these evaluation parameters, it is possible to determine whether the image is good or bad with respect to each of the evaluation values together with the evaluation setting parameter for the captured image. Judgment is now possible.

  On the image display device, different messages are displayed for each image depending on the quality of these evaluation values and determination results (for example, whether or not the threshold value has been exceeded or how much the threshold value has been exceeded). However, only a part of them can be displayed by the selection of the device operator. It may be registered in a storage medium in the control processor so that it can be read later.

Example 4
Next, the accuracy (or error) of the observation magnification will be described. Now, let R be an image resolution evaluation value (unit: μm) of a microscope image obtained by imaging a size calibration sample under a specific observation magnification, and Δp (unit μm) be an absolute pitch dimension L of the size calibration sample. In the example of the dimension calibration sample, L = 0.240 μm and Δx = 0.001 μm. On the surface of the sample, a repetitive pattern having a rectangular wave-shaped cross section is formed by utilizing anisotropic etching of silicon. The rectangular edge is a sharp edge that stands perpendicular to the sample surface.

The calibration of the observation magnification is performed by correcting the image rotation so that the lattice pattern image of the dimension calibration sample is vertical on the display screen, and then the pitch dimensions of a plurality of points (about 10 to 20 points) at a magnification of 50,000 to 100,000. And calibrate so that the average value is 0.240 μm. That is, the scanning range of the electron beam is changed so that the length measurement average value approaches 0.240 μm. The measurement of a plurality of points is performed in order to average and reduce the pitch dimension of the pattern itself because it has an error of several nanometers. At this time, the magnification error E and the magnification accuracy P in% display can be calculated from the following equations, respectively.
E = 100x {(Δx) ² + R ² } ^1/2 / L (Formula 10)
P = 100-E (Formula 11)

  As a result of verification with a general scanning electron microscope, the error E of the observation magnification was ± 2% and ± 0.5% or less, respectively. As a result, not only individual microscopes but also multiple microscopes can manage the accuracy and error of observation magnification with high precision from the viewpoint of differences (machine differences) and temporal changes from microscope to microscope as well as image resolution. became. The time-dependent change data of these evaluation parameters can be displayed on the image display device in the form of a graph or a table as required by the operation of the device operator. FIG. 5 shows an example of a one-month change graph for the image resolution of the SEM apparatuses A and B displayed on the image display apparatus. It can be seen that both devices are within the allowable range of resolution 3.2-3.8. If it is out of this range, equipment maintenance work is required. Further, as can be seen from this graph, the image resolutions of both apparatuses are within the allowable range, but it has been found that there is a machine difference that the apparatus B is 0.1 better. The machine difference correction coefficient between the devices is calculated based on this machine difference for the length measurement value of each device, and the length difference from both devices is corrected to reduce the machine difference in the microscope management of device manufacturing. It has become possible to manage with low.

  FIG. 6 shows an evaluation example of an astigmatism image using azimuth resolution. FIG. 6 a is a microscopic image from a gold particle sample when the correction parameter sp value is 0, 40 and 80 in the image astigmatism control system in the control system 7 of the microscope. The azimuth image resolution at that time is plotted in polar coordinates in FIG. 6b. When the azimuth image resolution curve is sp = 0, it is almost circular. However, as the azimuth resolution is increased to 40, 80, the azimuth resolution does not change much in the direction of the azimuth 8, 16 (arrow direction). It can be seen that there is an increase in the direction of azimuth 17 and 35 (arrow direction). These changes reflect the blur direction and size of the image of FIG. 6a. This image blur corresponds to the blur of the beam of the microscope (see sp = 80 image in FIG. 6a). By incorporating a program that combines this azimuth image resolution evaluation method and an optimization method that minimizes this evaluation value in all directions into a computer in the control system 7, focus adjustment and astigmatism correction are automatically performed with high accuracy. It was possible to realize a microscope to perform the following.

  In the above embodiment, the SEM has been described, but the same effect was obtained in a scanning transmission electron microscope (STEM) and a scanning ion microscope (SIM). In STEM, a sample is a thin film, and luminance signals used for image formation are secondary electrons, primary electrons transmitted through the sample, X-rays excited by primary electrons, and the like. In particular, in SIM, sputtering damage on the surface of the sample due to an ion beam that does not exist in SEM or STEM occurs, so the same part of the sample cannot be observed at high magnification. For this reason, the range of the sample for resolution evaluation has been expanded because it has applicability to the pattern sample having 5) strong directivity described above as a feature of this method. Practically, it was greatly effective in improving operability and improving the resolution of the observed image.

  In the above description, the microscope is taken up as an example device. The image resolution of the present invention is also applied to a device that uses a microscope image for beam alignment or setting of a beam irradiation region, such as a charged particle beam device such as a focused ion beam processing device, an electron beam diagnostic device, or an electron beam exposure device. The evaluation method can be implemented in exactly the same way and has the same effect. The “microscope” mentioned here includes a “charged particle beam device” using a microscope image.

  Further, a microscopic image according to the present invention is regarded as an observation image from one direction of a three-dimensional object, and a curved surface (denoted by a multi-order (or first-order) function z = f (x, y) for each local region ( Alternatively, the method of approximating the plane) is very effective as an image processing method for reducing image noise. The density distribution after image processing is nothing but the E {z (x, y)} distribution in (Expression 7). FIG. 3 shows the effect of image processing using an SEM image of a resolution standard sample. The figure (a) is an original image, and the figure (b) is a processed image. In FIG. 5B, the image is smooth, and it can be seen that the noise removal effect is great.

  The present invention is not limited to SEM and SIM, and any microscope that has conventionally used a conventional gap resolution method or FFT method as an image resolution evaluation method may be, for example, an optical microscope or a scanning probe microscope. However, it can be applied to any microscope. Similarly, the image processing method is effective not only for SEM and SIM but also for all microscope images.

Schematic of a scanning electron microscope. The figure explaining the resolution evaluation result using a gold vapor deposition sample. The figure which shows the noise removal effect in the image processing method of this invention in a SEM image. The SEM image of the dimension calibration sample which is a dimension calibration sample (pitch dimension: 0.240 ± 0.001 μm). The graph with time for one month in the image resolution of SEM apparatus A and B displayed on the image display apparatus. The figure explaining the evaluation example of the astigmatism image using azimuth | direction resolution. The figure which showed an example of the inspection apparatus evaluation system. The schematic diagram of local image resolution. Flow chart for calculating image resolution.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Charged particle source, 2 ... Charged particle beam, 3 ... Lens, 4 ... Sample, 5 ... Secondary particle, 6 ... Secondary particle detection system, 7 ... Control system, 8 ... XYZ stage, 9 ... Image display device.

Claims (22)

A charged particle source, a detector for detecting secondary particles emitted by irradiating the sample with a charged particle beam emitted from the charged particle source, and a control processor for forming an image based on the output of the detector In a charged particle microscope equipped with
The charged particle microscope, wherein the control processor calculates a resolution of the plurality of regions from a plurality of regions including a plurality of pixels of the image, and calculates an average value of the resolutions of the plurality of regions. In claim 1,
The charged particle beam microscope characterized in that the control processor calculates the plurality of resolutions based on concentration gradients of the plurality of regions. In claim 1,
The charged particle beam microscope characterized in that the control processor periodically calculates the average value, and displays or stores the calculated average value together with time transition information. In claim 1,
The charged particle beam microscope characterized in that the control processor calculates at least one of the image noise, the S / N ratio of an image, or an image quality evaluation parameter calculated based on the resolution, in addition to the resolution. In claim 4,
The control processor has a function of setting a threshold value to at least one of the average value, image noise, image SN ratio, or image quality evaluation parameter, and the threshold value is set to the average value, image noise, and image SN. A charged particle beam microscope characterized in that, when the ratio or image quality evaluation parameter exceeds, a message to that effect is displayed or stored. In claim 4,
A charged particle beam microscope characterized by periodically evaluating at least one of the average value, image noise, image signal-to-noise ratio, or image quality evaluation parameter, and displaying or storing the evaluation value together with time transition information. In claim 1,
The charged particle beam microscope characterized in that the control processor displays magnification accuracy or magnification error on a display device. In claim 1,
The charged particle beam microscope is characterized in that a unit of the average value is a length. In claim 1,
The charged particle beam microscope, wherein the control processor obtains the resolution from the concentration gradient for each of the plurality of regions and performs weighted averaging over the entire image or a partial region. In claim 9,
The control processor is a curved surface of a multi-order or linear function z = f (x, y) for each of the plurality of regions, where z is the density of the image and (x, y) is the position of an arbitrary portion in the screen Alternatively, a charged particle beam microscope characterized by applying a plane and obtaining the concentration gradient from a differential value of the function. A charged particle source, a detector for detecting secondary particles emitted by irradiating the sample with a charged particle beam emitted from the charged particle source, and a control processor for forming an image based on the output of the detector In a charged particle microscope equipped with
The control processor approximately applies a curved surface or a plane to a partial region of the image, performs the processing over the entire image or a partial region, and combines the results to form an image. Charged particle beam microscope. In an image forming method of scanning a sample with a charged particle beam emitted from a charged particle source and forming an image based on the charged particles emitted from the scanning region,
The charged particle beam is scanned on a plurality of patterns having equivalent dimensions on the sample, the pattern is measured, and the average value of the measurement results is close to the dimension of the pattern. An image forming method comprising changing a scanning range of a charged particle beam. An image resolution evaluation computer for evaluating the resolution of an image obtained from a plurality of charged particle beam devices via a network,
The image resolution evaluation computer calculates a resolution of the plurality of regions from a plurality of regions including a plurality of pixels of the image, and calculates an average value of the resolutions of the plurality of regions. Calculator. In claim 13,
The image resolution evaluation computer calculates the plurality of resolutions based on density gradients of the plurality of regions. In claim 13,
The computer for evaluating image resolution periodically calculates the average value, and displays or stores the calculated average value together with time transition information. In claim 13,
The image resolution evaluation computer calculates at least one of the image noise, the S / N ratio of the image, or an image quality evaluation parameter calculated based on the resolution, in addition to the resolution. Calculator. In claim 16,
The image resolution evaluation computer has a function of setting a threshold value to at least one of the average value, image noise, image S / N ratio, or image quality evaluation parameter, and the threshold value is set to the average value, image noise, An image resolution evaluation computer which displays or stores information on a display device when an SN ratio of an image or an image quality evaluation parameter is exceeded. In claim 16,
An image resolution evaluation computer periodically evaluates at least one of the average value, image noise, image S / N ratio, or image quality evaluation parameter, and displays or stores the evaluation value together with time transition information. A computer for evaluating image resolution. In claim 13,
The image resolution evaluation computer displays magnification accuracy or magnification error on a display device. In claim 13,
A computer for evaluating image resolution, wherein a unit of the average value is a length. In claim 13,
The computer for evaluating image resolution obtains the resolution from the density gradient for each of the plurality of regions, and performs weighted averaging over the entire image or a partial region. In claim 21,
The computer for evaluating image resolution sets the density of the image to z and the position of an arbitrary part in the screen to (x, y). ), And the density gradient is obtained from the differential value of the function.

Images