JP6838511B2 - Analysis method for polycrystalline materials - Google Patents

Description

The present invention relates to a technique for analyzing a grain boundary phase structure from an electron microscope image of a polycrystalline material.

In a functional material having a polycrystalline structure, not only the main phase (the phase having the highest phase ratio, which often has the greatest influence on the characteristics of the material) but also the grains existing between the main phase particles. The boundary phase often affects the characteristics. It is important to quantify the grain boundary phase structure in order to derive the characteristic improvement guideline. Conventionally, there are the following documents as an analysis method capable of quantifying the characteristics of a large number of grain boundary phases in a field of view by photographing a polycrystalline structure with an electron microscope.

Patent Document 1 describes an analytical means for obtaining analytical data having spectral information of each sample point, a morphological observation means for obtaining morphological observation data for each sample point, a threshold setting means for setting a threshold value for the morphological observation data, and a set threshold value. A binarizing means that compares the values of the morphological observation data and performs binarization, and a sample region is selected and analyzed by logical calculation processing of the binarizing data obtained by the binarizing means and the analysis data. A sample analyzer is described that includes a region extraction means for extracting data and a data analysis processing means for analyzing and processing the extracted analysis data.

Patent Document 2 includes _{a two-particle grain boundary phase between two R 2} T ₁₄ B main phase crystal particles and two adjacent R ₂ T ₁₄ B main phase crystal particles, and the thickness of the two-particle grain boundary phase is A rare earth magnet having a phase of 5 nm or more and 500 nm or less and having a magnetic phase different from that of a ferromagnet is described. Further, when the thickness of the two-grain boundary phase is measured in an arbitrary cross section, the thickness of the two-particle boundary phase is 5 nm or more and has a magnetism different from that of the ferromagnet with respect to all the measurement points. Rare earth magnets in which the ratio of the measurement points is 20% or more are also described. It is described that the two-particle grain boundary phase width (grain boundary phase thickness) in Patent Document 2 is an average value of 60 measured values.

The grain boundary phase includes a so-called two-particle grain boundary phase formed between two adjacent main phase crystal particles and a so-called grain boundary triple point surrounded by three or more main phase crystal particles. FIG. 2 of Patent Document 2 is a schematic diagram specifically showing a method for measuring the two-particle boundary phase width. In this figure, two grain boundary phases and grain boundary triple points are formed between _{adjacent R 2} T _{14 B main phase crystal grains.} Focusing on the two-particle boundary phase to be measured, the boundary between the two-particle boundary phase and the triple point of the grain boundary connected to the two-particle boundary phase is determined. After determining the boundary, divide this into four equal parts and draw three bisectors. The positions of these three equal division lines are set as the measurement points of the two-particle boundary phase width, and three measurement values are obtained. This measurement is performed on 20 arbitrarily selected two-particle boundary phases of interest, and the average of the measured values at a total of 60 measurement points is taken as the thickness (width) of the two-particle boundary phase.

Patent Document 3 describes between a main phase composed of a ferrite having a hexagonal M-type magnetoplumbite structure, a first grain boundary phase existing between two main phases, and three or more main phases. In any cross section of the ferrite sintered magnet having the existing second grain boundary phase, the second grain boundary phase is scattered, and the average area of the second grain boundary phase is 0.2 μm ^2. ferrite sintered magnet and less than is described and it is possible to improve the H _cJ while maintaining high B _r and a high H _{k /} H _cJ. The average area of the second grain boundary phase is obtained by binarizing the reflected electron image (BSE image) of the cross section of the ferrite sintered magnet by FE-SEM (field emission scanning electron microscope). The area and number of phases are obtained, and the total area is divided by the number.

Patent Document 4 describes a main crystal grain composed of a perovskite-type composite oxide containing Ba, Ti, a rare earth element, Mg and Mn as a metal element, and a two-sided grain boundary phase formed by the main crystal grain. A dielectric porcelain comprising a three-priority grain boundary phase, wherein the triple-point grain boundary phase is at least M ₄ R ₆ O (SiO ₄ ) type 6 crystal phase (M is selected from alkaline earth elements). A dielectric porcelain is described in which at least one kind, R is at least one kind selected from rare earth elements) is precipitated. It is described that the width da of the two-sided grain boundary phase is 0.4 nm or less. It is stated that the width of the two-sided grain boundary phase was evaluated by transmission electron microscopy and micro-region electron diffraction.

Japanese Unexamined Patent Publication No. 5-325862 Japanese Unexamined Patent Publication No. 2014-209546 WO2014 / 021149A1 Japanese Unexamined Patent Publication No. 2004-107200

The grain boundary phases of functional materials having a polycrystalline structure include (1) width and composition, (2) angle (which crystal plane of the main phase particles is in contact with), and (3) two-grain boundary phase or triple point grains. There are characteristics such as the boundary. In quantifying the grain boundary phase structure, it is necessary to measure the above characteristics in a large number of grain boundary phases and perform statistical analysis. However, the two-particle boundary phase relays the triple-point grain boundary phase and connects to each other, and the change from the two-particle grain boundary phase to the triple-point grain boundary phase is continuous. No guideline was given to classify the two, and the operator classified them in a timely manner.

Moreover, the grain boundary phase is not always a straight line, and the width and composition are not always constant. When measuring the width, composition, and angle of the grain boundary phase, there is no guarantee that the characteristics of the grain boundary phase structure can be quantified based on the results of the measurement points selected by the operator. In addition, there is a problem that it is difficult to increase the number of measurement points by the method in which the operator selects or classifies the measurement points, and statistical information that can be used for deriving the guideline for improving the characteristics of the polycrystalline material can be easily obtained. It was difficult.

One aspect of the present invention is a method of using an information processing device including a processing device, a storage device, an input device, and an output device. This method is an image input step of inputting an image of a polycrystalline sample containing a grain boundary phase from an input device, an image conversion step of converting an image into a binary image and converting a binary image into a thin line image consisting of thin lines, and a thin line. The analysis process for analyzing the grain boundary phase specified at each point of the above, and the first data for converting the analysis result of the grain boundary phase specified at each point of the thin line into data in the first format associated with each point. It is an analysis method of a polycrystalline material including a production step.

Another aspect of the present invention is a method using an information processing device including a processing device, a storage device, an input device, and an output device. In this method, the binary image obtained by binarizing the image of the polycrystal sample including the grain boundary phase is further converted into a fine line image consisting of fine lines, and the analysis result of the grain boundary phase specified at each point of the fine line is obtained with each point. The input process for inputting the associated first format data, the statistical analysis process for classifying the data in the first format according to the classification conditions and performing statistical analysis for each classification, and the results obtained in the statistical analysis process. This is an analysis method for a polycrystalline material including a second data generation step of converting data into data in a second format that can be collated with each classification condition, and an output step of outputting data in the second format.

Yet another aspect of the present invention is a method of using an information processing device including a processing device, a storage device, an input device, and an output device. This method is an image input step in which an image of a polycrystalline sample containing a grain boundary phase is input from an input device, and the image is converted into a binary image that discriminates between the grain boundary phase part and the other parts, and the binary image is converted into a binary image. An image conversion process that converts the grain boundary phase part into fine lines to generate a fine line image, extracts the branch points of the fine lines, and triple-point grains from the grain boundary phase part of the binary image based on the positional relationship with the branch points. It is an analysis method of a polycrystalline material including a classification step of extracting a boundary phase and a third data generation step of converting the result of the classification step into data in a third format that can be collated with the coordinates of the grain boundary phase.

Statistical information that can be used to derive guidelines for improving the properties of polycrystalline materials can be easily obtained.

The flowchart which shows the whole process which analyzes the grain boundary phase of a polycrystalline material. Explanatory drawing which shows the relationship between a sample shape and a crystal plane. Explanatory drawing of grain boundary phase analysis method. A flowchart showing a process of analyzing the width and intensity of the angle of the grain boundary phase specified at each point of the thin line from the electron microscope image, the binary image, and the thin line image. A flowchart showing a process of classifying a grain boundary phase into a triple point grain boundary phase and a two-particle grain boundary phase from a binary image and a fine line image. Explanatory drawing of a method of classifying a grain boundary phase into a triple point grain boundary phase and a two-particle grain boundary phase from a binary image and a thin line image. Explanatory drawing of an output example of the result of separating into a triple-point grain boundary phase and a two-particle grain boundary phase. The figure which shows an example of the output result of the 1st format. It is an output example of the statistical analysis result for each classification condition, and is the output example explanatory diagram when the classification condition is an angle range and the statistical analysis result is an average grain boundary width. It is an output example of the statistical analysis result for each classification condition, and is an explanatory diagram of the output example when the classification condition is a phase grain and the average image intensity range, and the statistical analysis result is a length ratio. It is an output example of the statistical analysis result for each classification condition, and is the output example explanatory diagram when the classification condition is a phase type and an image intensity range, and the statistical analysis result is an area ratio.

Hereinafter, embodiments will be described in detail with reference to the drawings. However, the present invention is not construed as being limited to the description of the embodiments shown below. It is easily understood by those skilled in the art that a specific configuration thereof can be changed without departing from the idea or gist of the present invention.

In the configuration of the invention described below, the same reference numerals may be used in common among different drawings for the same parts or parts having similar functions, and duplicate description may be omitted.

The notations such as "first", "second", and "third" in the present specification and the like are attached to identify the constituent elements, and do not necessarily limit the number, order, or contents thereof. is not it. In addition, numbers for identifying components are used for each context, and numbers used in one context do not always indicate the same composition in other contexts. Further, it does not prevent the component identified by a certain number from having the function of the component identified by another number.

The position, size, shape, range, etc. of each configuration shown in the drawings and the like may not represent the actual position, size, shape, range, etc. in order to facilitate understanding of the invention. Therefore, the present invention is not necessarily limited to the position, size, shape, range, etc. disclosed in the drawings and the like.

In one example of this embodiment, an analysis method for obtaining statistical information by quantifying the characteristics of the grain boundary phase using an electron microscope image of a polycrystalline material is provided. This analysis method is an imaging process in which an electron microscope image of a multi-crystal sample containing a grain boundary phase is taken. The electron microscope image is converted into a binary image of the grain boundary phase, and the binary image is thinned to be a fine line of the grain boundary phase. An image processing step for obtaining an image, an analysis step for analyzing the grain boundary phase specified at each point of the thin line, and a first format capable of collating the analysis result of the grain boundary phase specified at each point of the thin line with the coordinates of each point. The first output process to be output in, the input process to input the classification conditions and statistical analysis items of the grain boundary phase, and the analysis results output in the first output format are classified by the classification conditions, and the statistical analysis items are classified for each classification. The statistical analysis step to be carried out and the second output step of outputting the results obtained in the statistical analysis step in a second format capable of collating with each classification condition are included.

Further, in another example, an analysis method is provided in which the characteristics of the grain boundary phase are quantified and statistical information is obtained by using an electron microscope image of a polycrystalline material and an analysis image having the same field of view. This analysis method is an image input step of inputting an electron microscope image of a multi-crystal sample containing a grain boundary phase, an image conversion that converts an electron microscope image into a binary image of a grain boundary phase and converts a binary image into a fine line image. Extract the branch point of the thin line from the process and the thin line image, draw a circle around the branch point on the binary image of the grain boundary phase, and find the radius at which the ratio of the grain boundary phase on the circumference is within the specified range. , The grain boundary phase inside the circle is classified as a triple point grain boundary phase, the third output step that outputs the classification result in a format that can be compared with the coordinates of the grain boundary phase, and the analysis of the visual field taken by the electron microscope. Analytical image input process for inputting images, statistical analysis process for extracting triple point grain boundary phase and double grain boundary phase from analytical images and outputting each statistical analysis result, and results obtained in each statistical analysis process It includes a second output step of outputting in a second format that can be collated with the classification result.

In Example 1, an analysis method for quantifying the characteristics of the grain boundary phase and obtaining statistical information will be described using an electron microscope image of a polycrystalline material. Examples of polycrystalline materials include, for example, magnetic materials, dielectric materials, and structural materials such as steel frames.

FIG. 1 shows a flowchart of the entire process of taking an electron microscope image of a polycrystalline material, quantifying the characteristics of the grain boundary phase, and obtaining statistical information.

First, a sample is prepared and inserted, and the sample is moved to a position where an electron microscope image can be obtained (S101). Next, the misalignment, nonpoint, and focus of the detector amplifier and the electron optics system are adjusted so that the required image contrast and spatial resolution can be obtained (S102). In the case of automatic imaging, the sample stage is aligned (S103).

After making the above adjustments, a shooting recipe is input (S104), and an electron microscope image is taken with the input recipe (S105). The image of the polycrystalline sample including the grain boundary phase taken is input from the input device of the information processing device. In the information processing apparatus, the captured electron microscope image is converted into a binary image of the grain boundary phase (S106), and the binary image is thinned to obtain a fine line image of the grain boundary phase (S107 to S109). I do. Here, the line width of the thin line is 1 pixel, and as a result, the thin line can be defined by continuous points. Since the thin line can be regarded as the center line of the grain boundary phase, it is possible to express the grain boundary phase by a limited point (for example, coordinates).

Analyzing the characteristics of the grain boundary phase specified at each point of the thin line, that is, the angle, width, average image intensity, type (triple point grain boundary phase / binary grain boundary phase), etc. using an electron microscope image and a binary image. (S110 to S112). Here, the triple point grain boundary phase is a region surrounded by three or more main phase crystal grains, and the two-particle grain boundary phase is a region between two main phase crystal grains.

First data generation is performed, in which the analysis result of the grain boundary phase specified at each point of the thin line is output in the first format that can be collated with the coordinates of each point (S113). Next, the classification conditions and statistical analysis items of the grain boundary phase are input (S114), and statistical analysis is performed for each classification condition (S115). As the classification conditions, the angle range, width range, average image intensity range, type, etc. of the grain boundary phase are specified, and as the statistical analysis items, the integrated value, average value, variance, ratio, etc. are specified. The statistical analysis result is output to the second numerical table by the second data generation that is output in the second format that can be collated with each classification condition (S116). The details of each step will be described below.

<1. Preparation process>
The preparation step starts with the preparation of the sample used for the analysis (S101). Since many functional materials have anisotropy in the sample structure, it is better to process the sample so that the direction of the anisotropy can be confirmed. For example, in a general anisotropic sintered magnet, the easy magnetization axis (easy magnetization direction) is called the c-axis, the plane perpendicular to the easy magnetization direction is called the c-plane, and the plane parallel to the easy-magnetization direction is called the ab plane. There is. Depending on the purpose of analysis, a sample may be prepared by arbitrarily processing from the c-plane and / or the ab-plane. In this analysis, the sample is cut into a rectangular parallelepiped with the c-plane on the side and the ab-plane on the top.

FIG. 2 shows the relationship between the sample shape and the crystal plane. In order to efficiently analyze multiple samples, the time required for sample replacement can be reduced by processing the samples into small pieces of the same size and pasting them together with a conductive adhesive or conductive tape. can do. If it is difficult to make the sizes uniform, it is possible to combine multiple samples into one piece by filling with resin. If the analysis surface has irregularities, an image intensity distribution depending on the irregularities is observed in the electron microscope image. Therefore, it is better to process the analysis surface flat by polishing before starting imaging.

The processed sample is inserted into an electron microscope, the acceleration voltage and lens mode are set, and the detector amplifier, electro-optical system misalignment, non-point, and focus are adjusted according to the adjustment procedure manual of the apparatus (S102). If the electron microscope has an automatic adjustment function for a detector or an electron optical system, the automatic adjustment function may be used. When automatically photographing a plurality of fields of view, it is better to align the sample stage (S103). The coordinate system fixed to the mirror body is XYZ, the electron beam incident direction is the Z axis, and for example, the c-plane is adjusted to be perpendicular to the Y-axis and the ab-plane is adjusted to be perpendicular to the Z-axis.

<2. Shooting process>
In the photographing step, a photographing recipe is input (S104), and an electron microscope image is photographed with the input recipe (S105). In the shooting recipe, the position of each field of view, the field of view diameter (magnification), the detection signal, the electron beam scanning mode, the number of pixels, and the shooting time are input. Further, if necessary, the automatic focus correction function, the pass / fail judgment function of the field of view, and the drift correction function during shooting can be operated before shooting.

The pass / fail judgment function of the visual field is a function of determining whether or not the visual field is unsuitable for analysis because there are voids, foreign substances, or polishing scratches at the position initially input. By selecting the energy and emission angle of the electron beam to be detected with an electron microscope, it is possible to obtain a secondary electron image that emphasizes the unevenness of the sample and a backscattered electron image that emphasizes the element species of the sample. .. If there are voids, foreign matter, or polishing scratches, the analysis field of view becomes uneven, and the secondary electron image intensity in that region increases. Therefore, before taking the backscattered electron image for analysis, the secondary electron image was taken, and the region where the secondary electron image intensity was equal to or higher than the threshold was extracted as the unsuitable region for analysis, and the ratio of the unsuitable region for analysis reached the threshold or higher. The field of view is determined to be unsuitable for analysis, and re-imaging is performed in a nearby field of view.

The captured electron microscope image is recorded as image data together with the shooting recipe and correction information. This information is stored in the storage device as an electron microscope image database, and can be used for interpreting the results obtained in the next image processing step, analysis step, and statistical analysis step.

<3. Image processing process>
After the image processing step, it shall be executed by the information processing device. As the information processing device, for example, a server including a processing device, a storage device, an input device, and an output device is used. The processing unit is, for example, a central processing unit (processor), which usually includes a control unit and an arithmetic unit. Storage devices include magnetic disk devices and semiconductor storage devices. As an input device, there is an interface for inputting data from the outside such as a keyboard, a mouse, or a network. The output device includes a display that displays an image, a printer, or an interface that outputs data to the outside such as a network.

In this embodiment, functions such as calculation and control are realized in cooperation with other hardware by executing a program stored in the storage device by a processor. A program executed by a computer or the like, its function, or a means for realizing the function may be referred to as a "function", a "means", a "part", a "unit", a "module", or the like. The above configuration may be configured by a single computer, or any part of the input device, output device, processing device, and storage device may be configured by another computer connected by a network.

Subsequent processing shall be executed by inputting the electron microscope image database acquired in the photographing process from the input device or storing it in the recording device and using the electron microscope image database.

FIG. 3 shows a schematic diagram of each image in the image processing step. FIG. 3A is a grayscale electron microscope image. FIG. 3B is a binary image obtained by binarizing an electron microscope image. FIG. 3C is a thin line image obtained by thinning a binary image. FIG. 3D shows an analysis example of the width of the grain boundary phase specified by a point with a thin line.

In the image processing step, first, the electron microscope image (FIG. 3 (A)), which is a grayscale image, is binarized from the image intensity into the main phase 10 and the grain boundary phase 20, and the region of the grain boundary phase 20 is extracted. Make an image (FIG. 3 (B)) (S106). Further, the binary image is thinned to a line width of 1 pixel and converted into a thin line image (FIG. 3 (C)) (S109).

At this time, the thin line is regarded as the center of the grain boundary phase 20. The meaning of "center" will be supplementarily explained later with reference to FIG. Since there are grain boundary phases having a narrow width and grain boundary phases having a small difference in image intensity from the main phase, there are also grain boundary phases that cannot be extracted by simple binarization. In this case, correction by image processing or manual correction is required (S107 to S108). Image processing applied to an electron microscope image includes application of a noise reduction filter, shading removal caused by a mirror body, and the like. Further, as an image processing applied to a binary image, there is a method of dividing the binary image into regions, calculating the area of each region, and removing the region whose area is equal to or less than the threshold value. There is also a method of removing small holes in the region by expanding / contracting each region. The grain boundary phase that could not be extracted even after performing the above processing is extracted by manually correcting the image intensity of the binary image.

Even if it is not possible to extract all the grain boundary phases at the stage of the binary image, it is possible to extract at the stage of the thin line image. As an image process applied to a thin line image, there is a curve complement process that complements between line segments. The grain boundary phase that could not be extracted even by applying the curve complement processing is extracted by manually correcting the image intensity of the thin line image (S110 to S111).

The binary image and the thin line image obtained in the above image processing step are stored and stored as an image database in association with the electron microscope image database obtained in the photographing step.

<4. Analysis process>
In the analysis step, the angle, width, and average image intensity of the grain boundary phase specified at each point of the fine line are analyzed from the electron microscope image, the binary image, and the fine line image. The analysis step may be executed by the information processing apparatus that has performed the image processing step, or may be executed by another information processing apparatus.

FIG. 4 shows a partial analysis flow of the process S112. An electron microscope image, a binary image, and a thin line image are input to the information processing apparatus (S401), and the grain boundary phase designated by each point of the thin line in the thin line image is sequentially analyzed (S402). For a certain point, several pixels before and after the point, for example, 5 pixels before and after the point are linearly approximated, the angle formed by the thin line with the x-axis is obtained, and this is defined as the angle of the grain boundary phase (S403). In the present invention, if the x-axis of the image is set parallel to the c-axis of the sample, the angle formed by the grain boundary phase with the c-axis is measured. A line profile that is orthogonal to the linearly approximated line segment and passes through the point is extracted from the electron microscope image and the binarized image (S404). For example, when the line segment AB is a line segment that is linearly approximated in FIG. 3C, the line profile at the position 31 shown in FIGS. 3A and 3B is extracted.

As shown in FIG. 3 (D), the line profile 300 extracted from the position 31 of the electron microscope image (FIG. 3 (A)) is function-fitted (S405), and the half width of the function is defined as the width of the grain boundary phase. (S406). As the fitting function, the Gaussian function is suitable for the narrow grain boundary phase, and the error function is suitable for the narrow grain boundary phase. This time, we will fit with the above two functions. The grain boundary width obtained from the line profile at position 31 of the binary image (FIG. 3B) was used as the initial value of fitting, and a criterion for adopting a result close to the initial value was used. The intensity of the grain boundary phase was defined by the average image intensity of pixels within the half width of the electron microscope image (S407).

As a result of the above processing, the result for each point of the thin line image is stored and output as the analysis result of the first format (S408).

When the S / N of the electron microscope image is low, a fitting error may occur in a narrow grain boundary phase or a grain boundary phase having a small difference in image intensity from the main phase. When the frequency of errors is high, there is a measure to increase the S / N of the line profile by adding and averaging the line profiles extracted from nearby points. Further, in order to shorten the analysis time, the thin line may be divided into several pixels for analysis. For example, a thin line is divided every 5 pixels, the grain boundary phase is analyzed with a line profile obtained by adding and averaging the line profiles extracted from each point in the divided section, and the analysis result is defined as the analysis result of each point in the section. .. Further, in the grain boundary phase added by curve complementation or manual correction of the thin line image, the width of the grain boundary phase obtained from the binarized image becomes zero. In this case, the initial value used in the function fitting of the line profile is set to an appropriate value, for example, about 5 pixels, and if the center position is greatly displaced by the fitting or if it is calculated widely, it is judged as a fitting error. Was provided.

The analysis step also includes a step of separating the grain boundary phase into a triple point grain boundary phase and a two-particle grain boundary phase.

FIG. 5 shows the flow of the separation process. First, a binary image and a thin line image are input (S501). Each point of the thin line image is classified into the following three types, that is, an end point in which one thin line exists in eight adjacent pixels, a passing point in which two exist, and a branch point in which three or more exist. A branch point is extracted from the branch points (S502), and each branch point is analyzed.

In the branch point analysis, a circle with a radius R = 1 pixel centered on the branch point is drawn on the binary image (S504). Initially, all pixels on the circumference are in grain boundary phase. The radius R is increased (S507), and the increase of the radius R is stopped when the ratio of the grain boundary phase on the circumference (that is, the point having the same value as the branch point) becomes a threshold value, for example, 90% or less. (S505 to S506). Alternatively, the increase in radius R may be stopped when the ratio of grain boundary phases in the circle becomes a threshold value, for example, 90% or less.

FIG. 6 shows an example of the classification result. For the sake of explanation, the main phase 10 is displayed in black, the grain boundary phase 20 is displayed in white, and the center 23 of the grain boundary phase is displayed. As the center 23, for example, a line (so-called equidistant intermediate line) consisting of a set of points at the same distance from the two closest opposing points at the boundary between the main phase 10 and the grain boundary phase 20 can be used. As a result, the range in which the grain boundary phase spreads around the branch point is estimated. The grain boundary phase inside the circle is classified into the triple point grain boundary phase 22 (S508). After the classification at all the bifurcation points is completed (S503), the grain boundary phase not classified as the triple point grain boundary phase is classified into the two-particle grain boundary phase 21 (S509). In the above embodiment, the treatment is performed using a perfect circle, but a similar effect can be obtained by using an ellipse or a polygon.

FIG. 7 is an output example of the result of separation into the triple point grain boundary phase and the double grain boundary phase. 7 (A) is a binary image of the grain boundary phase, FIG. 7 (B) is a fine line image of the grain boundary phase, FIG. 7 (C) is a binary image of the triple point grain boundary phase, and FIG. 7 (D) is a binary image. It is a binary image of the grain boundary phase. In the example of FIG. 7, a circle centered on the above branch point is drawn on the binary image, the grain boundary phase inside the circle is the triple point grain boundary phase, and the grain boundary phase outside is the two grain boundary phase. It is divided into. Also, draw the circle on the thin line image, and the grain boundary phase (thin line) specified by the point inside the circle is the triple point grain boundary phase, and the grain boundary phase (thin line) on the outside is the two grain boundary phase. It may be divided. 7 (A) is the same image as FIG. 3 (B), and FIG. 7 (B) is the same image as FIG. 3 (C).

In the example of FIG. 7, the type of the grain boundary phase is first output in a third format that can be collated with the coordinates of the grain boundary phase. From the binary image of the grain boundary phase (FIG. 7 (A)) and the thin line image (FIG. 7 (B)), the binary image of the triple point grain boundary phase (FIG. 7 (C)) and the binary value of the binary grain boundary phase An image (FIG. 7 (D)) can be obtained.

Further, the value of the pixel classified as the main phase in one image is 0, the value of the pixel classified as the triple point grain boundary phase is 1, and the value of the pixel classified as the double grain boundary phase is 2. It may be output as a value image. Whether the grain boundary phase specified at each point of the thin line is the triple point grain boundary phase or the two-particle grain boundary phase is a binary image of the triple-point grain boundary phase or the two-particle grain boundary phase at the coordinates of each point. Classify by referring to the value of the binary image of.

The analysis result of the grain boundary phase specified at each point of the thin line is output in the first format that can be collated with the coordinates of each point (S510).

FIG. 8 shows an example of the output result of the first format. The analysis results of the grain boundary phase specified by the points of the thin line are summarized in one row, and in each column, the identification number 801 of the line which is the thin line separated by the branch point and the identification number of each point of the thin line. The point number 802, the x-coordinate 803, the y-coordinate 804, the type 805 of the grain boundary phase specified by the point, the angle 806, the width 807, the average intensity 808, and the intensity dispersion 809 are recorded. By recording the initial value of the function fitting and the residual 810 in addition to the analysis result, the judgment criteria can be easily modified or added. In addition, it is convenient to number each thin line to display the analysis result and interpret the statistical analysis image. Therefore, a column for entering the line number 801 and the point number 802 indicating the number of the line belonging to the line and the number of the point from the end is also provided. Note that the line number, point number, and the like may be saved as separate data and collated with the coordinates of the points. The type of grain boundary phase is, for example, indicated as "2" for a two-particle grain boundary phase and "3" for a triple-point grain boundary phase.

As described above, in this embodiment, by processing the electron microscope image with the information processing apparatus, various analysis results corresponding to each point of the fine line (indicating the center of the grain boundary phase of the polycrystalline structure) can be obtained. Can be done. By creating a database of such analysis results and using it for subsequent statistical analysis processing and the like, it is possible to perform multifaceted analysis of the polycrystalline structure.

<5. Statistical analysis process>
In the statistical analysis step, using the output result of the first format, the classification condition and the statistical analysis item are input (S114), and the statistical analysis item is executed for each classification condition (S115). The statistical analysis step may be executed by the information processing device that performed the analysis step, or may be executed by another information processing device. As the statistical analysis items, various items generally used for statistical processing such as integrated value, mean value, variance, and ratio may be selected.

FIG. 9A shows the output results when the analysis conditions are set to the angle range and the statistical analysis conditions are set to the average grain boundary width. Points that satisfy each angle range are extracted from the output result of the first format, and the average value of the widths at the extracted points is calculated. The obtained statistical analysis result is output in a second format that can be collated with each classification condition. If a function for displaying the output result as a graph is provided as shown in FIG. 9A (b), the result can be easily interpreted.

FIG. 9B shows the output results when the classification conditions are the type of grain boundary phase and the average image intensity range, and the statistical analysis item is the length ratio. It is assumed that three phases that can be classified according to the image intensity range are included, and the image intensity range of phase 1, the image intensity range of phase 2, and the image intensity range of phase 3 are specified. The length ratio indicates the ratio occupied by the length of each phase in the material. The length may be the number of pixels constituting the thin line, but is equivalent to the number of data in the first format data (FIG. 8) extracted according to the classification conditions. Although the output data in the second format of FIG. 9B (a) shows only the result of one of the plurality of materials, as shown in FIG. 9B (b), the analysis results of a plurality of samples are integrated into one graph. If the function shown by is provided, the result can be easily interpreted.

FIG. 9C shows the output results when the type and image intensity range of the grain boundary phase are input as the classification conditions and the area ratio is input as the statistical analysis item. In the statistical analysis step, analysis using the output result of the third format may be performed. In the first output format, the analysis result of the grain boundary phase specified at each point of the thin line is output in the first format that can be collated with the coordinates of each point, but in the third output format, the classification result is the grain boundary. Output in a format that can be collated with the coordinates of the phase. It may be output as a binary image shown in FIGS. 7 (C) and 7 (D), or may be output as a numerical table in which xy coordinates and classification results are described.

When outputting as a binary image, the binary image of the triple point grain boundary phase is multiplied by the electron microscope image to create an electron microscope image from which the triple point grain boundary phase is extracted, and the image intensity of the created image is classified as a classification condition. Classify according to the image intensity range of, calculate the areas of phase 1, phase 2, and phase 3, and output the area ratio. The two-particle boundary phase is also statistically analyzed by the same procedure.

As a classification condition, it is input that the width of the two-grain boundary phase is 5 nm or more and 500 nm or less, and the image intensity range is the image intensity range of the phase having magnetism different from that of the ferromagnet, and the length ratio is used as a statistical analysis item. Can be entered. In Patent Document 2, only about 60 grain boundary phases can be analyzed, but in this method, ratio information can be obtained after analyzing the grain boundary phases of the entire visual field.

Further, as the classification condition, the classification condition can be set to the triple point grain boundary phase, and the statistical analysis item can be set to the grain size of the grain boundary phase. In this case, the area of all triple point grain boundaries can be automatically calculated by dividing the binary image of the triple point grain boundary phase into regions and calculating the area of each region. The area cannot be calculated automatically in the conventional binary image in which the triple point grain boundary phase and the double grain boundary phase are not separated. Further, since the operator classifies the triple point grain boundary phase and the binary image in a timely manner, the classification result depends on the operator.

Patent Document 4 describes that the dielectric porcelain is characterized in that the width of the two-particle boundary phase is 0.4 nm or less. However, since there is no description of a method for measuring the width of the two-particle boundary phase, various results are output depending on the interpretation of the width of the two-particle boundary phase. For example, the classification condition is that the grain boundary phase is a two-grain boundary phase, and the statistical analysis item is the result of statistical analysis with the average grain boundary width, and it can be interpreted that the average grain boundary width is 0.4 nm or less. it can. The classification condition is that the grain boundary phase is a two-grain grain boundary phase and the width of the grain boundary phase is 0.4 nm or less, and the statistical analysis item is the result of statistical analysis using the length ratio of the grain boundary phase as the length. It can also be interpreted that the ratio is 100%. In the analysis of grain boundary phase structure in polycrystalline samples, it is essential to specify classification conditions and statistical analysis items. Further, in order to classify the grain boundary phase according to the classification conditions, it is essential to output the characteristics of the grain boundary phase specified at each point of the thin line as information of each point. In addition, by outputting the information of each point together with the coordinates of each point, it becomes possible to perform analysis in combination with the results of other analysis methods and analysis methods.

The analysis method using the electron microscope image was described in Example 1, but in Example 2, the analysis result and the statistical analysis result obtained from the electron microscope are used to obtain an analysis image of the same field as the field image taken by the electron microscope image. The method of analysis will be described. By the method shown in Example 1, the field of view taken with an electron microscope is separated into a main phase, a triple point grain boundary phase, and a two-particle grain boundary phase, and the separation result is output in the third format.

The white region of FIG. 7 (A) is the main phase, the black region of FIG. 7 (C) is the triple point grain boundary phase, and the black region of FIG. 7 (D) is the two-particle grain boundary phase. An analytical image of the same field of view as this field of view is obtained. As the analysis image, an image obtained from various signals emitted by the incident of an electron beam, for example, an EDX (Energy Dispersive X-ray Spectroscopy) image, an Auger electron image, an EBSD (Electron Backscatter Diffraction Patterns) image, a spin SEM Can be given. Images obtained with an optical microscope or probe microscope other than an electron microscope may be used.

To obtain statistical analysis information of the triple-point grain boundary phase, create a binary image with the black area as 1 and the white area as 0 in FIG. 7 (C), and multiply the analysis image to obtain the triple-point grain boundary phase. It is possible to extract the analytical values and obtain the average value of the analytical values and the area ratio of the values in a certain range. For the two-particle boundary phase, a binary image with the black region shown in FIG. 7 (D) set to 1 and the white region set to 0, and for the main phase, the white region shown in FIG. 7A set to 1 and the black region set to 0. A value image is created, each region is extracted, and statistical analysis processing is performed.

According to the above examples, statistical analysis can be performed under the specified classification conditions for the characteristics of the grain boundary phase in the polycrystalline material. This analysis method can be automated, and a large amount of data can be statistically analyzed. It will be possible to obtain statistical information that can be used to derive characteristics improvement guidelines, and research and development of characteristic improvement by grain boundary phase control will be accelerated. It also contributes to countermeasures against product defects caused by grain boundary phases.

10 ... Main phase, 20 ... Grain boundary phase, 21 ... Two-particle grain boundary phase, 22 ... Triple point grain boundary phase, 23 ... Center of grain boundary phase, 300 ... Line profile

Claims (15)

Using an information processing device equipped with a processing device, a storage device, an input device, and an output device,
An image input step of inputting an image of a polycrystalline sample containing a grain boundary phase from the input device.
An image conversion step of converting the image into a binary image and converting the binary image into a thin line image composed of thin lines.
An analysis step for analyzing the grain boundary phase specified at each point of the thin line,
A first data generation step of converting the analysis result of the grain boundary phase designated at each point of the thin line into data in a first format associated with each point.
Method for analyzing polycrystalline materials including. In the image conversion step,
The image, which is a grayscale image, is converted into a binary image in which a portion corresponding to the grain boundary phase is extracted using image intensity, and the portion corresponding to the grain boundary phase is thinned to a predetermined line width to form a thin line. Convert to image,
The method for analyzing a polycrystalline material according to claim 1. The predetermined line width is a line width of 1 pixel, and the thin line is defined by continuous points.
The method for analyzing a polycrystalline material according to claim 2. In the analysis step,
The direction of the thin line is obtained by linearly approximating the thin line consisting of a predetermined number of pixels before and after the point extracted from the thin line, the angle formed with the specified direction is calculated, and the angle of the grain boundary phase specified at that point. Output as
In the first data generation step,
The angle of the grain boundary phase specified at each point of the thin line is converted into data in a first format associated with each point.
The method for analyzing a polycrystalline material according to claim 1. In the analysis step,
The direction of the thin line is obtained by linearly approximating the thin line consisting of a predetermined number of pixels before and after the point extracted from the fine line, a line profile orthogonal to the obtained direction is extracted from the image, and the width of the grain boundary phase is determined. Analyze and output as the width of the grain boundary phase specified at that point,
In the first data generation step,
The width of the grain boundary phase specified at each point of the thin line is converted into data in a first format associated with each point.
The method for analyzing a polycrystalline material according to claim 1. In the analysis step,
The direction of the thin line is obtained by linearly approximating the thin line consisting of a predetermined number of pixels before and after the point extracted from the fine line, the line profile orthogonal to the obtained direction is extracted from the image, and the width of the grain boundary phase is analyzed. Then, the average value of the image intensity inside the width is calculated, and it is output as the average image intensity of the grain boundary phase specified at that point.
In the first data generation step,
The average image intensity of the grain boundary phase specified at each point of the thin line is converted into data in a first format associated with each point.
The method for analyzing a polycrystalline material according to claim 1. The analysis step is
A branch point extraction step for extracting the branch point of the thin line, and
Assuming a circle centered on the branch point on the binary image, a radius determination step for obtaining a radius in which the ratio of points having the same value as the branch point on the circumference is within a specified range.
A triple point grain boundary phase classification step of classifying a grain boundary phase designated by a point inside the circle of radius on the binary image or the thin line image into a triple point grain boundary phase.
The method for analyzing a polycrystalline material according to claim 1. Using an information processing device equipped with a processing device, a storage device, an input device, and an output device,
The binary image obtained by binarizing the image of the polycrystalline sample including the grain boundary phase is further converted into a thin line image consisting of fine lines, and the analysis result of the grain boundary phase specified at each point of the fine line is associated with each of the points. Input process for inputting data in the first format,
A statistical analysis step in which data in the first format is classified according to classification conditions and statistical analysis is performed for each classification.
A second data generation step of converting the results obtained in the statistical analysis step into data in a second format that can be collated with each classification condition.
An output process that outputs data in the second format,
Method for analyzing polycrystalline materials including. The analysis result is obtained by linearly approximating a thin line consisting of a predetermined number of pixels before and after the point extracted from the thin line to obtain the direction of the thin line and calculating the angle formed with the specified direction. The angle of the grain boundary phase specified by
The data of the first format associates the angle of the grain boundary phase specified at each point of the thin line with each point.
The classification condition is a range of the angles.
The method for analyzing a polycrystalline material according to claim 8. In the analysis result, a thin line consisting of a predetermined number of pixels before and after the point extracted from the thin line is linearly approximated to obtain the direction of the thin line, and a line profile orthogonal to the obtained direction is extracted from the image, and the grain is obtained. It is the width of the grain boundary phase specified at that point, which is obtained by analyzing the width of the boundary phase.
The data of the first format associates the width of the grain boundary phase specified at each point of the thin line with each point.
The classification condition is in the range of the width.
The method for analyzing a polycrystalline material according to claim 8. In the analysis result, a thin line consisting of a predetermined number of pixels before and after the point extracted from the thin line is linearly approximated to obtain the direction of the thin line, and a line profile orthogonal to the obtained direction is extracted from the image to obtain a grain boundary. It is the average image intensity of the grain boundary phase specified at that point, which is obtained by analyzing the width of the phase and calculating the average value of the image intensity inside the width.
The data of the first format corresponds to the average image intensity of the grain boundary phase specified at each point of the thin line with each point.
The classification condition is in the range of the average image intensity.
The method for analyzing a polycrystalline material according to claim 8. The analysis result is the type of grain boundary phase specified at that point.
The data of the first format corresponds to the type of the grain boundary phase specified at each point of the thin line with each point.
The classification condition is the type of the grain boundary phase.
The method for analyzing a polycrystalline material according to claim 8. The item of the statistical analysis is at least one of the integrated value, the average value, the variance, and the ratio.
The method for analyzing a polycrystalline material according to claim 8. Using an information processing device equipped with a processing device, a storage device, an input device, and an output device,
An image input step of inputting an image of a polycrystalline sample containing a grain boundary phase from the input device.
An image conversion step of converting the image into a binary image that discriminates between the grain boundary phase portion and the other portion, and converting the grain boundary phase portion of the binary image into fine lines to generate a fine line image.
A classification step of extracting a branch point of the fine line and extracting a triple point grain boundary phase from a portion of the grain boundary phase of the binary image based on the positional relationship with the branch point.
A third data generation step of converting the result of the classification step into data in a third format that can be collated with the coordinates of the grain boundary phase.
Method for analyzing polycrystalline materials including. The classification step is
Assuming a circle centered on the branch point on the binary image, a radius determination step for obtaining a radius in which the ratio of points having the same value as the branch point on the circumference is within a specified range.
A triple point grain boundary phase classification step of classifying a grain boundary phase designated by a point inside the circle of radius on the binary image into a triple point grain boundary phase is included.
The method for analyzing a polycrystalline material according to claim 14.

Images