Particle analysis apparatus and method for analyzing particles

Abstract

To accurately identify particles of interest (for example, asbestos particles) on the basis of a plurality of intensity distributions obtained from within a beam scanning range.SOLUTION: An angle calculator 42 calculates an angle representing the orientation of a surface on the basis of an intensity distribution for each coordinate within a beam scanning range. A normalizer 48 multiplies an angle array generated by the angle calculator 42 by a numerical value corresponding to a shape of interest (for example, a needle shape or a string shape). A particle-of-interest analyzer 52 analyzes whether a candidate particle is a particle of interest on the basis of a group of normalized angles corresponding to the candidate particle.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to particle analysis devices and methods, and more particularly to techniques for identifying particles having a particular shape. [Background technology]

[0002] Known particle measurement systems include electron microscope systems, laser microscope systems, and optical microscope systems. For example, an electron microscope system for measuring particles is composed of a scanning electron microscope equipped with a backscattered electron detector and an information processing device equipped with particle analysis software. The latter information processing device can also be called a particle analysis device.

[0003] For example, when measuring and analyzing asbestos particles (the fibrous elements that make up asbestos) in a dust sample, particle analysis software calculates the aspect ratio for each candidate particle and identifies whether the candidate particle is an asbestos particle based on the aspect ratio. Asbestos particles have an extremely thin shape (needle-like or string-like), it is not easy to distinguish them from other elongated materials or elongated scratches. When multiple asbestos particles overlap and intersect, it is difficult to accurately calculate the aspect ratio.

[0004] Counting asbestos particles while visually identifying individual asbestos particles contained in images generated by a scanning electron microscope places a heavy burden on the inspector, and the analysis results are likely to vary depending on the inspector. Note that particle analysis devices can also analyze particles other than asbestos particles.

[0005] Patent Documents 1 and 2 disclose particle measurement systems. Patent Document 3 discloses a system for analyzing the surface shape of a sample. Patent Document 4 discloses a system for measuring diffraction patterns. Patent Documents 1 to 4 do not disclose a technique for identifying particles of interest having a specific shape, in particular a technique for identifying particles of interest having a specific shape by utilizing rotational symmetry. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Laid-Open No. 2007-155515 [Patent Document 2] Patent Publication No. 2021-165657 [Patent Document 3] Japanese Patent Publication No. 2022-185757 [Patent Document 4] Japanese Patent Application Publication No. 2019-121588 Summary of the Invention [Problem to be solved by the invention]

[0007] An object of the present invention is to accurately identify a particle of interest, or to accurately identify a particle of interest having a needle shape or a string shape, or to identify an individual particle of interest when multiple particles of interest are overlapping. [Means for solving the problem]

[0008] The particle analysis device according to the present invention is characterized by including: a calculator that calculates, for each coordinate within a beam scanning range on a sample, an angle representing the orientation of a surface at that coordinate based on an intensity distribution obtained by detecting a signal emitted from that coordinate using a detection area array; a normalizer that calculates multiple normalized angles by applying normalization corresponding to a shape of interest to multiple angles corresponding to multiple coordinates within the beam scanning range; and an analyzer that analyzes, for each candidate particle within the beam scanning range, whether the candidate particle is a particle of interest based on the group of normalized angles corresponding to the candidate particle.

[0009] The particle analysis method according to the present invention is characterized by comprising the steps of: calculating, for each coordinate within a beam scanning range on a sample, an angle representing the orientation of a surface at that coordinate based on an intensity distribution obtained by detecting a signal emitted from that coordinate using a detection area array; calculating a plurality of normalized angles by applying normalization corresponding to a shape of interest to a plurality of angles corresponding to a plurality of coordinates within the beam scanning range; and analyzing, for each candidate particle within the beam scanning range, whether the candidate particle is a particle of interest based on the group of normalized angles corresponding to the candidate particle. [Effects of the Invention]

[0010] According to the present invention, a particle of interest can be identified with high accuracy. Alternatively, according to the present invention, a particle of interest having a needle shape or a string shape can be identified with high accuracy. Alternatively, according to the present invention, when a plurality of particles of interest are overlapping, each particle of interest can be identified. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a block diagram illustrating a particle measurement system according to an embodiment. [Figure 2] 1 is a flowchart illustrating a particle analysis method according to an embodiment. [Figure 3] FIG. 1 is a diagram showing a backscattered electron detector. [Figure 4] FIG. 10 is a diagram illustrating an example of an intensity distribution. [Figure 5] FIG. 10 is a diagram showing another example of the intensity distribution. [Figure 6] FIG. 2 is a diagram illustrating a detection coordinate system. [Figure 7] FIG. 10 is a diagram showing vectors calculated from the intensity distribution. [Figure 8] FIG. 1 is a diagram illustrating a color space. [Figure 9] FIG. 10 is a diagram showing a convex region and a concave region. [Figure 10] FIG. 10 is a diagram showing two vectors obtained from two slopes in a first shape. [Figure 11]FIG. 10 is a diagram showing three vectors obtained from three slopes in the second shape. [Figure 12] FIG. 10 is a diagram showing four vectors obtained from four slopes in the third shape. [Figure 13] FIG. 1 shows asbestos particles. [Figure 14] FIG. 10 is a diagram illustrating a resultant vector. [Figure 15] FIG. 10 is a diagram illustrating extraction of a convex region. [Figure 16] FIG. 10 is a diagram illustrating shape analysis based on variance information. [Figure 17] FIG. 10 is a diagram showing a histogram. [Figure 18] FIG. 1 illustrates histogram-based shape analysis. [Figure 19] FIG. 10 is a diagram illustrating division of candidate particles. [Figure 20] 1A and 1B are diagrams showing a backscattered electron image and an image of a particle of interest. [Figure 21] FIG. 10 is a diagram illustrating an example of image analysis. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, an embodiment will be described with reference to the drawings.

[0013] (1) Overview of the embodiment A particle analysis device according to an embodiment includes a computing unit, a normalizer, and an analyzer. The computing unit calculates, for each coordinate within a beam scanning range on a sample, an angle representing the orientation of a surface at that coordinate based on an intensity distribution obtained by detecting a signal emitted from that coordinate using a detection area array. The normalizer calculates multiple normalized angles by applying normalization corresponding to a shape of interest to multiple angles corresponding to multiple coordinates within the beam scanning range. The analyzer analyzes, for each candidate particle within the beam scanning range, whether the candidate particle is a particle of interest based on the group of normalized angles corresponding to the candidate particle.

[0014] The normalization is a mathematical operation that generates uniformity in the group of angles obtained from the feature of interest and generates diversity in the group of angles obtained from features other than the feature of interest. Given this preprocessing, the normalized group of angles corresponding to a candidate particle is evaluated to determine whether the candidate particle is a particle of interest. A particle of interest is a particle to be analyzed that has a feature of interest.

[0015] For example, whether a candidate particle is a particle of interest may be identified based on information indicating the degree of variation in the normalized angle group corresponding to the candidate particle. Examples of information indicating the degree of variation include variance information, a histogram, etc. Candidate particles may also be analyzed based on other evaluation values.

[0016] In an embodiment, the normalizer calculates multiple normalized angles by multiplying each of the multiple angles by a coefficient corresponding to the shape of interest. This configuration selectively normalizes the multiple angles obtained from the shape of interest by utilizing the rotational symmetry of the shape of interest. More specifically, through normalization, the multiple angles obtained from the candidate particles are aligned to a specific angle or within a specific angle range. In an embodiment, the shape of interest is a needle or string shape. In that case, the coefficient is 2. A method other than coefficient multiplication may be adopted as the normalization method.

[0017] The particle analysis device according to the embodiment includes a classifier that determines convex regions within the beam scanning range as candidate particles (candidate particle regions) based on multiple angles corresponding to multiple coordinates. With this configuration, concave regions (e.g., scratches, dents) can be excluded from the analysis target.

[0018] In an embodiment, the determiner extracts convex regions by applying a divergence calculation to a plurality of angles corresponding to a plurality of coordinates. In this configuration, the angle array is regarded as a vector field and a vector calculation is applied to the angle array. A set of positive divergences corresponds to a convex region.

[0019] In an embodiment, the analyzer calculates dispersion information based on the set of normalized angles corresponding to the candidate particle. The analyzer analyzes whether the candidate particle is a particle of interest based on the dispersion information. The dispersion information indicates the degree of uniformity of the set of normalized angles.

[0020] In an embodiment, the analyzer creates a histogram based on the normalized angles corresponding to the candidate particles. The analyzer analyzes whether the candidate particle is a particle of interest based on the histogram. The histogram has an angle axis and a frequency axis. The normalized angles can be further evaluated through analysis of the histogram.

[0021] In an embodiment, the analyzer determines that the shape of a candidate particle is a combination of multiple shapes of interest based on the histogram. If the shape of a candidate particle is a combination of multiple shapes of interest, the analyzer separates the candidate particle into multiple particles of interest. This configuration makes it possible to separate and identify individual particles of interest when multiple particles of interest overlap.

[0022] A particle analysis method according to an embodiment includes a first step, a second step, and a third step. In the first step, for each coordinate within a beam scanning range on a sample, an angle representing the orientation of a surface at that coordinate is calculated based on an intensity distribution obtained by detecting a signal emitted from that coordinate using a detection area array. In the second step, multiple normalized angles are calculated by applying normalization corresponding to a shape of interest to multiple angles corresponding to multiple coordinates within the beam scanning range. In the third step, for each candidate particle within the beam scanning range, it is analyzed whether the candidate particle is a particle of interest based on a group of normalized angles corresponding to the candidate particle.

[0023] The particle analysis method can be realized, for example, by software. A program for executing the particle analysis method is installed in an information processing device via a network or a portable storage medium. The information processing device has a non-transitory storage medium that stores the program.

[0024] (2) Details of the embodiment 1 shows a particle measurement system 10 according to an embodiment. The particle measurement system 10 measures one or more asbestos particles contained in a sample 24. Asbestos particles are very small, needle-like or string-like particles. Other particles may also be measured by the particle measurement system 10.

[0025] The particle measurement system 10 includes a scanning electron microscope 12 and an information processing device 14. The scanning electron microscope 12 includes a measurement unit 16 and an arithmetic and control unit 18. The measurement unit 16 includes a lens barrel. The lens barrel includes an electron gun 19, an objective lens 20, a backscattered electron detector 26, a sample chamber 27, and the like. A movable stage 22 is provided within the sample chamber 27. The movable stage 22 holds a sample 24. The sample 24 is, for example, dust containing asbestos. In FIG. 1, the sample 24 is exaggerated.

[0026] The electron beam is two-dimensionally scanned over a two-dimensional beam scanning range set on the sample 24. Specifically, the beam scanning range consists of a plurality of coordinates (a plurality of measurement points), and the electron beam is sequentially irradiated onto the plurality of coordinates. The backscattered electrons emitted from each coordinate are detected by the backscattered electron detector 26.

[0027] The backscattered electron detector 26 is provided between the objective lens 20 and the sample 24. Specifically, the backscattered electron detector 26 is disposed near the lower end surface of the objective lens 20. The backscattered electron detector 26 is composed of a plurality of detection regions 26a arranged in a ring shape. The plurality of detection regions 26a can be referred to as a detection region set. An opening that allows the electron beam to pass is formed in the center of the backscattered electron detector 26. A plurality of detection signals are output in parallel from the plurality of detection regions 26a. The plurality of detection signals can be referred to as a detection signal set. A plurality of detection signals may be output from the backscattered electron detector 26 in a time-division manner. The number of detection regions 26a constituting the backscattered electron detector 26 is, for example, 4, 6, 8, 12, or 16. All numerical values ​​given in the specification of this application are merely examples.

[0028] A secondary electron detector (not shown) is also arranged inside the column, and may be a secondary electron detector having a plurality of detection areas.

[0029] The arithmetic and control unit 18 includes a control unit 28, a signal processing unit 30, and an SEM image generator 32. The control unit 28 controls the operation of the measurement unit 16. The SEM image generator 32 forms an SEM image based on a detection signal output from the secondary electron detector or a set of detection signals output from the backscattered electron detector. The formed SEM image is sent to the information processing device 14 as necessary.

[0030] The signal processing unit 30 comprises a plurality of signal processing circuits 30a that process a plurality of detection signals output from a plurality of detection regions 26a that constitute the backscattered electron detector 26. Each signal processing circuit 30a has, for example, a current-voltage converter, an amplifier, an AD converter, etc. The plurality of detection data output from the signal processing unit 30 is sent to the information processing device 14. The plurality of detection data can be referred to as a detection data set. The detection data set obtained from each coordinate indicates the intensity distribution on the detection surface of the backscattered electron detector 26. In other words, the detection data set is data that represents the intensity distribution.

[0031] The intensity distribution reflects the shape of the measurement point where the backscattered electrons are emitted. That is, the intensity distribution changes depending on the orientation of the microsurface (sample surface, particle surface) at the measurement point. Specifically, the angle of the main axis (usually the central axis) of the intensity distribution changes depending on the tilt direction of the microsurface. Therefore, the orientation of the surface at the measurement point can be estimated from the angle of the main axis of the intensity distribution.

[0032] The information processing device 14 is a particle analysis device. The information processing device 14 is configured by a computer equipped with particle analysis software. Specifically, the information processing device 14 has a processor 34, a memory unit 36, an input unit 38, and a display unit 40. In FIG. 1, multiple functions performed by the processor 34 are represented by multiple blocks.

[0033] The angle calculator 42 calculates the angle of the main axis of the intensity distribution based on the detection data set, i.e., the intensity distribution, obtained from each coordinate on the sample. The angle calculator 42 may calculate the intensity as well as the angle. In this case, the angle calculator 42 can be considered as a vector calculator. The angle calculator 42 calculates multiple angles, i.e., an angle array, corresponding to multiple coordinates within the beam scanning range.

[0034] The particle image generator 43 generates particle images based on the angle array, where angles may be converted to hues, or angle and intensity combinations may be converted to hue and brightness combinations.

[0035] The divergence calculator 44 regards the angle array as a vector field and applies vector operations to the angle array. Specifically, the divergence calculator 44 applies an operation to the angle array to find the divergence. This results in multiple divergences (divergence arrays) corresponding to multiple coordinates. A set of positive divergences corresponds to a convex region, and a set of negative divergences corresponds to a concave region.

[0036] The convex region determiner 46 determines one or more convex regions within the beam scanning range based on the divergence array. In other words, the convex region determiner 46 excludes concave portions such as scratches and dents from the analysis target. Each convex region is treated as a candidate region.

[0037] The normalizer 48 applies a mathematical operation (a mathematical operation utilizing rotational symmetry) that selectively operates on the shape of interest to the angle array. Specifically, the normalizer 48 generates a normalized angle array by multiplying the angle array by a coefficient corresponding to the shape of interest. In this embodiment, the particle of interest is an asbestos particle, as described above. The shape of interest is an elongated shape (needle-shaped, string-shaped). In this case, the normalization coefficient is 2.

[0038] The angle group extractor 50 extracts a normalized angle group corresponding to each candidate particle (candidate particle region) that is a convex region from the angle array. That is, the angle group extractor 50 extracts a normalized angle group belonging to or obtained from the candidate particle. If the candidate particle has a shape of interest, the normalized angle group corresponding to the candidate particle will be uniform. If the candidate particle has a shape other than the shape of interest, the normalized angle group obtained from the candidate particle will be diverse.

[0039] The particle of interest analyzer 52 evaluates the normalized angle group corresponding to each candidate particle, and thereby analyzes whether the candidate particle is a particle of interest. The analysis methods include a first analysis method and a second analysis method. In the first analysis method, dispersion information is calculated based on the normalized angle group, and the shape of the candidate particle is evaluated based on the dispersion information. In the second analysis method, a histogram (angle histogram) is created based on the normalized angle group, and the candidate particle is evaluated based on the histogram.

[0040] The particle of interest image generator 54 images one or more particles of interest. The generated particle of interest image is displayed on the display unit 40. The particle of interest corresponds to a convex region having a needle-like or string-like shape.

[0041] The computing unit 56 counts the number of particles of interest within the beam scanning range, calculates the aspect ratio of each particle of interest, etc. The analysis results of the particles of interest are displayed on the display unit 40. The display unit 40 is configured by, for example, an LCD.

[0042] The input unit 38 is composed of a keyboard, a pointing device, etc. The user specifies particle analysis conditions and the like using the input unit 38. The memory unit 36 ​​stores parameters and the like to be referenced in particle analysis. The memory unit 36 ​​also stores particle analysis software. The image of the particle of interest and the SEM image may be displayed side by side or superimposed.

[0043] FIG. 2 shows a flow chart of a particle analysis method according to an embodiment. In S10, a sample is measured using a scanning electron microscope. This generates an intensity distribution array 60. The intensity distribution array 60 consists of a plurality of intensity distributions corresponding to a plurality of coordinates within a two-dimensional beam scanning range. The beam scanning range has an x-axis and a y-axis. FIG. 2 shows signal intensities A, B, C, and D that make up the intensity distribution obtained from the coordinate P(x1, y1).

[0044] In S12, an angle operation is applied to the intensity distribution array consisting of a plurality of intensity distributions corresponding to a plurality of coordinates, thereby generating an angle array 62. In Fig. 2, the angle θ corresponding to the coordinate P is shown. In S13, a color particle image is generated based on the angle array 62, if necessary.

[0045] In S14, a calculation to determine the divergence is applied to the angle array 62. This generates a divergence array 64. FIG. 2 shows the divergence (▽·w1) corresponding to the coordinate P. The divergence (▽·w1) will be described in detail later. In S16, one or more convex regions are identified based on the divergence array. At this time, processes such as binarization and labeling are sequentially applied to the divergence array. FIG. 2 shows two extracted convex regions 66-1 and 66-2. Each of the convex regions 66-1 and 66-2 is treated as a candidate particle.

[0046] In S18, normalization is applied to the angle array 62. Specifically, each angle constituting the angle array 62 is multiplied by a coefficient of 2. FIG. 2 shows the normalized angle θ associated with the coordinate P.

[0047] In S20, for each candidate particle, a set of normalized angles corresponding to the candidate particle is extracted from the normalized angle array. In Fig. 2, two sets of normalized angles 70-1 and 70-2 corresponding to two convex regions 66-1 and 66-2 are shown.

[0048] In S22, each of the normalized angle groups 70-1 and 70-2 is evaluated. Specifically, the shape of each of the candidate particles is evaluated based on each of the normalized angle groups 70-1 and 70-2. As a result, a particle of interest (particle of interest region) 72 having a needle-like or string-like shape is selected. In S24, the particle of interest 72 is imaged to generate a particle of interest image 76. In S26, measurement is performed on the particle of interest 72.

[0049] As described above, according to the embodiment, a particle of interest can be extracted with high accuracy by using both convex region determination and normalization on the premise that an intensity distribution array is obtained. Note that a segmented backscattered electron detector may be replaced by a segmented secondary electron detector.

[0050] The particle analysis method according to the embodiment will be described in more detail below.

[0051] 3 shows the backscattered electron detector 26. The backscattered electron detector 26 is composed of multiple detection regions 26a arranged to surround the optical axis. A signal emitted from a coordinate P on the sample 24, i.e., backscattered electrons 80, is detected in each detection region 26a. The signal intensity observed in each detection region 26a depends on the topography of the sample, specifically, the orientation of the microfacet at the coordinate P.

[0052] FIG. 4 shows an example of an intensity distribution. In the illustrated example, the backscattered electron detector 26 comprises four detection regions 26a1-26a4 arranged on the xy plane. Each of the detection regions 26a1-26a4 has a fan-like shape. In FIG. 4, the central axis of each of the detection regions 26a1-26a4 is represented by a dashed line. In FIG. 4, the signal intensities AD observed in the four detection regions 26a1-26a4 are plotted on the four central axes. The signal intensities AD form an intensity distribution 84 resembling a radar chart. The orientation of the intensity distribution 84 (the orientation of the main axis of the intensity distribution 84 or the extension direction of the intensity distribution 84) is indicated by reference numeral 86. Note that, for each central axis, the inner end 82a corresponds to the minimum intensity, and the outer end 82b corresponds to the maximum intensity.

[0053] Another example of the intensity distribution is shown in Figure 5. The backscattered electron detector 88 consists of eight detection areas 88a1-88a8 arranged in a ring. Eight signal intensities AH observed in the eight detection areas 88a1-88a8 form an intensity distribution 89. The direction (angle) of the intensity distribution 89 is indicated by reference numeral 87.

[0054] A detection coordinate system is shown in Figure 6. The intensity distribution can be converted into a vector as follows. Conversion methods include a first conversion method and a second conversion method. First, the first conversion method will be described.

[0055] In Figure 6, the backscattered electron detector is composed of four detection areas, and the four signal intensities observed in the four detection areas are AD. The central axes of the four detection areas are at an angle θ A -θ D It is identified by:

[0056] When considering the synthesis of four signal intensities AD, the x component (Sx) and y component (Sy) after synthesis are expressed by the following equations (1) and (2).

number

number

[0057] When the intensity distribution is expressed as a vector (see FIG. 7), the magnitude S and angle θ of the vector are expressed by the following equations (3) and (4).

number

number

[0058] In the embodiment, when analyzing the particle of interest, of the size S and the angle θ, the angle θ is referred to, but both the size S and the angle θ may be referred to.

[0059] Next, the second conversion method will be described. For example, the second conversion method can be adopted when a segmented backscattered electron detector with five or more segments is used. The second conversion method applies a discrete Fourier transform to the intensity distribution (a waveform consisting of multiple signal intensities arranged in angular order).

[0060] By applying the discrete Fourier transform to the intensity distribution, a complex number F(k) expressed by the following equation (5) can be obtained for each wave number k (i.e., frequency).

number

[0061] A complex number F(k) is expressed as a vector Vk on the complex plane as follows:

number

[0062] Specifically, the complex number F(k) is expressed by the following equation (7).

number

[0063] Here, N is the number of detection areas, n is the number of the detection areas, and k is the wave number as described above. When identifying the shape of the sample surface, it is sufficient to focus on k=1. f(n) represents the signal intensity obtained in the nth detection area.

[0064] When the wave number k is 1, the above complex number F(1) is expressed as follows:

number

[0065] The signal strength observed by the entire reflected signal detector can be calculated by substituting 0 for k, specifically as follows:

number

[0066] The magnitude S and angle of the vector can be calculated as follows, assuming k=1.

number

number

[0067] 8 shows a color space (specifically, an HSV color space). For example, in the illustrated color space, the angle corresponds to the hue, and the radial position corresponds to the brightness. A color particle image can be generated by expressing the combination of the direction and magnitude of each vector obtained as described above as a combination of hue and brightness.

[0068] Next, we will explain how to extract candidate particles (convex regions). Since an angle array is equivalent to a vector array, it can be considered a vector field. By applying a vector operation to find the divergence of the angle array, individual convex regions can be extracted. Conversely, individual concave regions can be excluded from the analysis.

[0069] A more detailed explanation follows. The complex number expressed in equation (5) above can be expressed as a vector as shown in equation (6) above. When considering the shape of the sample surface, it is sufficient to focus on k=1. Here, if m and n are unit vectors, the vector field can be expressed as follows:

number

[0070] The divergence of the vector field is calculated by the following equation (13):

number

[0071] where ∇ is defined as follows:

number

[0072] A divergence array is formed by multiple divergences corresponding to multiple coordinates. The divergence array typically includes multiple convex regions and multiple concave regions. Each convex region corresponds to a set of positive divergences. Each concave region corresponds to a set of negative divergences. In this embodiment, each of the multiple convex regions is a candidate region. Each concave region corresponds to a scratch or dent, and is excluded from the analysis. Figure 9 illustrates one convex region 106 and one concave region 108 within the beam scanning range.

[0073] For example, a binary image is generated by thresholding the divergent array with a positive threshold. Then, individual convex regions are extracted as candidate regions by labeling the binary image. Each candidate region is an independent closed region.

[0074] Normalization will be explained below. FIG. 10 shows a first shape 90. The first shape 90 corresponds to a triangular prism lying on its side. The first shape 90 has two slopes 90a and 90b. When an electron beam is irradiated onto a measurement point on slope 90a and a signal from the measurement point is observed, a first angle (first vector) 92A is calculated. On the other hand, when an electron beam is irradiated onto a measurement point on slope 90b and a signal from the measurement point is observed, a second angle (second vector) 92B is calculated.

[0075] When viewed from the electron beam source, the first shape has two-fold symmetry. As the electron beam scans the first shape 90, a first set of angles and a second set of angles are observed. Note that n-fold symmetry means that when a model is rotated by (360 / n) degrees, the model before and after the rotation overlap.

[0076] To identify a shape with two-fold symmetry, each first angle is multiplied by 2, and each second angle is multiplied by 2. Then, all angles after multiplication become equal. For example, multiplying the first angle θ1 by 2 gives 2×θ1. Multiplying the second angle θ1+180 by 2 gives 2×θ1+360, which is nothing but 2×θ1. In this way, normalization is a process of distinguishing a shape of interest from other shapes by multiplying each angle by a coefficient corresponding to the shape of interest.

[0077] For reference, FIG. 11 shows a second shape 96. The second shape 96 is a triangular pyramid. The second shape 96 has three inclined surfaces 96a, 96b, and 96c. When an electron beam is irradiated onto a measurement point on the inclined surface 96a and a signal from the measurement point is observed, a first angle (first vector) 98A is calculated. When an electron beam is irradiated onto a measurement point on the inclined surface 96b and a signal from the measurement point is observed, a second angle (second vector) 98B is calculated. When an electron beam is irradiated onto a measurement point on the inclined surface 96c and a signal from the measurement point is observed, a third angle (third vector) 98C is calculated. When viewed from the electron beam source, the second shape 96 has three-fold symmetry. As the electron beam scans the second shape 96, a first group of angles, a second group of angles, and a third group of angles are observed.

[0078] When three-fold symmetry is recognized, multiplying each first angle by 3, each second angle by 3, and each third angle by 3 results in all angles being equal. For example, multiplying the first angle θ1 by 3 gives 3 × θ1. Multiplying the second angle θ1 + 120 by 3 gives 3 × θ1 + 360, which is nothing but 3 × θ1. Multiplying the third angle θ1 + 240 by 3 gives 3 × θ1 + 720, which is nothing but 3 × θ1.

[0079] For reference, FIG. 12 shows a third shape 100. The third shape 100 is a quadrangular pyramid. The third shape 100 has four slopes 100a, 100b, 100c, and 100d. When an electron beam is irradiated onto a measurement point on slope 100a and a signal from the measurement point is observed, a first angle (first vector) 102A is calculated. When an electron beam is irradiated onto a measurement point on slope 100b and a signal from the measurement point is observed, a second angle (second vector) 102B is calculated. When an electron beam is irradiated onto a measurement point on slope 100c and a signal from the measurement point is observed, a third angle (third vector) 102C is calculated. When an electron beam is irradiated onto a measurement point on slope 100d and a signal from the measurement point is observed, a fourth angle (fourth vector) 102D is calculated. When viewed from the electron beam source, the third shape 100 has four-fold symmetry. As the electron beam scans the third shape 100, a first set of angles, a second set of angles, a third set of angles, and a fourth set of angles are observed.

[0080] If four-fold symmetry is recognized, then multiplying each first angle by 4, each second angle by 4, each third angle by 4, and each fourth angle by 4 will result in all angles being equal.

[0081] As described above, applying normalization corresponding to the shape of interest to the angle array makes it easier to identify specific particles. A coefficient of 2 is used to identify asbestos particles. Particles having any shape may also be identified. In this case, the normalization coefficient may be specified by the user.

[0082] FIG. 13 shows a schematic diagram of an asbestos particle. (A) shows a background surface (base material surface) on which an asbestos particle 112 is present. (B) shows a cross-section of the asbestos particle (reference numeral 113 indicates the cross-section position). The asbestos particle has a shape similar to the first shape. Therefore, a first set of angles 118 obtained from one surface of the asbestos particle and a second set of angles 120 obtained from the other surface of the asbestos particle have a two-fold symmetry relationship.

[0083] Next, we will explain the methods for analyzing particles of interest, specifically the methods for identifying asbestos particles. First, we will explain the first analysis method, and then we will explain the second analysis method.

[0084] In the first analysis method, for each candidate particle, dispersion information is calculated based on the normalized angle group obtained from the candidate particle, and whether or not the candidate particle is a particle of interest is determined based on the dispersion information.

[0085] The set of normalized angles obtained from the candidate particle is expressed as follows:

number

[0086] When each normalized angle is regarded as a unit vector, the following resultant vector is defined by combining a group of unit vectors corresponding to the group of normalized angles.

number

[0087] Incidentally, angles θ1 to θ N The average value of θ AVE is defined as follows:

number

[0088] 14 shows three unit vectors V1, V2, and V3 corresponding to three angles θ1, θ2, and θ3. Also, in FIG. 14, a resultant vector V is defined by combining the three unit vectors V1, V2, and V3. z The resultant vector V z Argument angle θ z corresponds to the average value of the angle.

[0089] Composite Vector V zBased on this, the length r of the average resultant vector is calculated as follows:

number

[0090] The absolute value of each complex number is 1, and in the above equation (18), the sum of the complex numbers is divided by the number of angles N, so 0≦r≦1. In the embodiment, the shared information d is defined by the following equation (19).

number

[0091] However, 0≦d≦1. When the group of normalized angles has uniformity, the variance information d becomes small. Specifically, if the candidate particle is an asbestos particle, the variance information d becomes small. When the group of normalized angles has diversity, the variance information d becomes large. As information in place of the variance information, the above r may be used as is, or another index may be used.

[0092] In Figure 15, (A) shows the divergent array obtained by vector calculation. The divergent array includes convex regions 122 and 124 and concave regions 126 and 128. (B) shows the binarization processing result (candidate particle extraction result). As shown in the figure, convex regions 122 and 124 are retained, and concave regions 126 and 128 are discarded. (C) shows the labeling result. Convex region 122 is assigned label #1, and convex region 124 is assigned label #2.

[0093] After the above processing, a group of normalized angles corresponding to the convex region 122 is extracted and evaluated. Specifically, based on the dispersion information obtained from the group of normalized angles, it is determined whether or not the convex region 122 is a particle of interest (specifically, an asbestos particle). Similarly, a group of normalized angles corresponding to the convex region 124 is extracted, and based on the group of normalized angles, it is determined whether or not the convex region 124 is a particle of interest.

[0094] 16 shows an example of evaluation of normalized angle groups. (A) shows the state before evaluation, and (B) shows the state after evaluation. The normalized angle group corresponding to candidate particle 130 has uniformity, and variance information d1 is calculated based on this normalized angle group. On the other hand, the normalized angle group belonging to candidate particle 132 has diversity, and variance information d2 is calculated based on this normalized angle group.

[0095] For example, the variance information d1 and d2 are each compared with a threshold value dx. In the illustrated example, the variance information d1 is smaller than the threshold value dx, and the variance information d2 is larger than the threshold value dx. Accordingly, the candidate particle 130 is determined to be the particle of interest. The candidate particle 132 is excluded.

[0096] Next, the second analysis method will be described. In the second analysis method, a histogram (normalized angle histogram) is generated for each candidate region, and candidate particles are analyzed based on the histogram.

[0097] For example, if the candidate particle has a needle shape, a histogram with one peak will result. If the candidate particle corresponds to the intersection (combination) of two needle shapes, a histogram 134 with two peaks 136, 138 will result, as shown in Figure 17. If the candidate region is a circular region, a histogram with multiple low peaks or no clear peaks will result.

[0098] By evaluating the histograms in this manner, it is possible to analyze the shape of the candidate particles.

[0099] 18 shows a specific example of the second analysis method. In S30, a histogram is created based on a group of normalized angles corresponding to a candidate particle. In S32, it is determined that the number of peaks in the histogram that satisfy a predetermined condition is one. In S34, the candidate particle is determined to be a particle of interest. For example, the predetermined condition is satisfied when the level of the peak is equal to or greater than a certain value. The width of the peak may also be evaluated.

[0100] In S36, it is determined that the number of peaks in the histogram that satisfy the predetermined condition is two. In S38, the candidate particle is divided into two regions (divided regions) corresponding to the two peaks. In S40, each divided region is determined to be a particle of interest. In S42, it is determined that the number of peaks in the histogram that satisfy the predetermined condition is three or more. In S42, the candidate particle is determined to be a particle of non-interest.

[0101] 19 specifically illustrates the processing content of S38 above. A candidate particle 140 is divided into two regions 142, 144. In this case, the regions 142, 144 corresponding to each peak may be identified based on a histogram. A candidate particle may also be divided into three or more regions. From the normalized angle array, the normalized angle or normalized angle range that produces the highest frequency may be identified, and then the normalized angle or normalized angle range that produces the second highest frequency may be identified. Such processing can also be considered to be processing based on a histogram.

[0102] 20 shows a backscattered electron image 146 and a particle of interest image 147. The backscattered electron image 146 is a backscattered electron composition image. In the particle of interest image 147, individual asbestos particles appear as large white streaks.

[0103] According to the particle analysis method of the embodiment, it is possible to accurately identify needle-shaped or string-shaped particles without being affected by scratches or dents. In the particle analysis method of the embodiment, in order to further improve the identification accuracy, the aspect ratio may be calculated for each particle and the aspect ratio may be taken into consideration when analyzing the particles.

[0104] 21 shows the content of the calculations executed by the calculator shown in FIG. 1. Based on a particle of interest image 148, the particles of interest contained therein are counted (see reference numeral 149). For example, the number of particles of interest per unit area may be calculated. Furthermore, based on the particle of interest image 148, the length of each particle of interest may be measured (see reference numeral 150), and an average length may be calculated based on a plurality of lengths determined from a plurality of particles of interest (see reference numeral 152). Furthermore, based on the particle of interest image 148, the width of each particle of interest may be measured (see reference numeral 154), and an average width may be calculated based on a plurality of widths determined from a plurality of particles of interest (see reference numeral 156). Whether or not each particle of interest is an asbestos particle may be determined based on the length, width, area, etc. of each particle of interest.

[0105] For example, the above series of processes may be performed in a particle measurement system equipped with a laser microscope. In this case, a segmented detector that detects laser light from the sample may be used. In the above embodiment, elongated crystals may be analyzed instead of asbestos particles. Alternatively, particles having shapes other than elongated shapes may be analyzed. [Explanation of symbols]

[0106] 10 particle measurement system, 12 scanning electron microscope, 14 information processing device (particle analysis device), 16 measurement unit, 18 calculation control unit, 26 backscattered electron detector, 30 signal processing unit, 42 angle calculator, 44 divergence calculator, 46 convex region determiner, 48 normalizer, 50 angle group extractor, 52 particle of interest analyzer.

Claims

1. a calculator that calculates an angle representing the orientation of the surface at each coordinate within the beam scanning range on the sample based on an intensity distribution obtained by detecting a signal emitted from the coordinate using the detection area array; and a normalizer that calculates a plurality of normalized angles by applying normalization corresponding to a shape of interest to a plurality of angles corresponding to a plurality of coordinates within the beam scanning range; an analyzer for each candidate particle within the beam scanning range, analyzing whether the candidate particle is a particle of interest based on a group of normalized angles corresponding to the candidate particle; A particle analysis device comprising:

2. 2. The particle analysis device according to claim 1, the normalizer calculates the plurality of normalized angles by multiplying each of the plurality of angles by a coefficient corresponding to the shape of interest. A particle analysis device characterized by:

3. 3. The particle analysis device according to claim 2, the target shape is a needle shape or a string shape, The coefficient is 2. A particle analysis device characterized by:

4. 2. The particle analysis device according to claim 1, a determiner that determines a convex region within the beam scanning range as the candidate particle based on the plurality of angles corresponding to the plurality of coordinates, A particle analysis device characterized by:

5. 5. The particle analysis device according to claim 4, the determiner determines the convex region by applying a calculation for determining divergence to the plurality of angles corresponding to the plurality of coordinates. A particle analysis device characterized by:

6. 2. The particle analysis device according to claim 1, The analyzer calculating dispersion information based on the set of normalized angles corresponding to the candidate particles; analyzing whether the candidate particle is the particle of interest based on the dispersion information; A particle analysis device characterized by:

7. 2. The particle analysis device according to claim 1, The analyzer creating a histogram based on the set of normalized angles corresponding to the candidate particles; analyzing whether the candidate particle is the particle of interest based on the histogram; A particle analysis device characterized by:

8. 8. The particle analysis device according to claim 7, The analyzer determining, based on the histogram, that the shape of the candidate particle is a combination of a plurality of shapes of interest; If the shape of the candidate particle is a combination of the plurality of shapes of interest, the candidate particle is divided into a plurality of particles of interest. A particle analysis device characterized by:

9. a step of calculating an angle representing the orientation of the surface at each coordinate within the beam scanning range on the sample based on an intensity distribution obtained by detecting a signal emitted from the coordinate using the detection area array; calculating a plurality of normalized angles by applying normalization corresponding to a shape of interest to a plurality of angles corresponding to a plurality of coordinates within the beam scanning range; analyzing, for each candidate particle within the beam scanning range, whether the candidate particle is a particle of interest based on a group of normalized angles corresponding to the candidate particle; A particle analysis method comprising:

10. A program for executing a particle analysis method in an information processing device, a function of calculating an angle representing the orientation of the surface at each coordinate within the beam scanning range on the sample based on an intensity distribution obtained by detecting a signal emitted from the coordinate using a detection area array; and a function of calculating a plurality of normalized angles by applying normalization corresponding to a shape of interest to a plurality of angles corresponding to a plurality of coordinates within the beam scanning range; a function of analyzing, for each candidate particle within the beam scanning range, whether the candidate particle is a particle of interest based on a group of normalized angles corresponding to the candidate particle; A program comprising:

Images