Metal Particle Analysis Method

Abstract

To provide a metal particle analysis method capable of determining a threshold when it is difficult to determine a threshold, by a mode method, in a histogram of a gray value indicating a correlation between brightness and a frequency of the brightness, the gray value being acquired when imaging a sample including multiple components, in an X-ray CT device.SOLUTION: A metal particle analysis method comprises: a first step for determining a volume which a metal substance occupies in a sample, and analyzing respective existence amounts of the metal substance and non-metal substance in a sample; a second step for determining a voxel number which metal particles occupy, in a histogram of a gray value on the basis of the volume which the metal substance occupies, and voxel volume in CT imaging; and a third step for integrating the voxel number in the CT imaging, in an order of having the higher brightness, in the histogram of the gray value, and setting a value of the brightness when the voxel number becomes equal to the voxel number occupied by the metal particles determined in the second step, as a threshold between the metal particles and bulk component.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to a method for analyzing metal particles using an X-ray CT scanner. [Background technology]

[0002] When analyzing a sample containing multiple components (for example, a sample with a bulk component that is a non-metallic substance and metal particles made of a metallic substance within that bulk component), various analytical methods exist depending on the purpose of the analysis. In this invention, "metallic substance" refers to a single metal or an alloy composed of multiple single metals; "metallic particles" refers to single metal particles or alloy particles; "non-metallic substance" is a general term for slag components and other elements other than metallic substances; and "bulk component" refers to a matrix component composed of non-metallic substances.

[0003] To analyze the weight percentage of metallic substances in a sample, there are chemical analysis methods such as the bromine-methanol method (see Non-Patent Document 1). To analyze metallic particles in a sample, one method is to polish the cross-section of the sample and observe the cross-section using an optical microscope or electron microscope. Furthermore, to analyze shape parameters such as particle volume, equivalent spherical diameter, and sphericity of the particles, there is also a method using an X-ray CT scanner (see Patent Documents 1 and 2). [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2020-34372 [Patent Document 2] Japanese Patent Publication No. 2020-134504 [Non-patent literature]

[0005] [Non-Patent Document 1] Determination of metallic nickel in nickel oxide by bromine-methanol method (Bulletin of the Central Laboratory of Customs (13), 1973-03, p7) [Non-Patent Document 2] Study on quantitative evaluation of asphalt pavement materials using X-ray CT (Journal of Japan Society of Civil Engineers, Series E1 (Pavement Engineering), Vol. 67, No. 3, (Journal of Pavement Engineering, Vol. 16)) [Non-Patent Document 3] Development of a quality evaluation method for the interior of asphalt mixtures using an X-ray CT scanner (Civil Engineering Technical Data 55-2 (2013)) [Overview of the Initiative] [Problems that the invention aims to solve]

[0006] As mentioned above, there is a method of analyzing a sample using an X-ray CT scanner to obtain particle information of metal particles, but according to the inventors' research, there were the following problems.

[0007] X-rays penetrate a sample, but are also absorbed depending on the material of the sample. As a result, there are differences in the attenuation of X-ray transmission in different parts of the sample. The higher the density and atomic number of the sample material, the more easily X-rays are absorbed, and the higher the brightness of the CT image. This gradation of brightness and darkness in the CT image is called the gray value.

[0008] On the other hand, when a sample is imaged using an X-ray CT scanner, a gray value histogram is obtained that shows the correlation between the brightness of each part of the sample and its frequency (number of voxels). A histogram is a graph in which brightness is plotted on the horizontal axis and frequency (number of voxels) on the vertical axis, with a curve drawn to show the relationship between the two. An example of such a histogram is shown in Figure 1. In this invention, a voxel is a cube consisting of the smallest unit of data used to represent a three-dimensional object in a computer.

[0009] In the histogram pattern shown in Figure 1, for example, if the sample is a mixture of metal particles and bulk components, then, based on the relationship between the density and brightness of the sample described above, we can define that "the high-brightness peak A represents metal particles, and the low-brightness peak B represents bulk components."

[0010] Next, in order to extract the portion defined as metal particles in the CT image, the boundary value (sometimes referred to as the "threshold" in this invention) at the boundary between peak A and peak B in the histogram is determined. Then, by using the brightness value at this threshold to binarize the gray level of the CT image into metal particles and bulk components, the metal particles in the CT image can be extracted and analyzed.

[0011] Here, if the histogram pattern yields two peaks corresponding to the peak portion of A and the peak portion of B, as shown in Figure 1, then the so-called mode method can be used, where the value of the valley between these two peaks is used as the threshold.

[0012] However, our investigations have shown that even if the sample is a mixture of metal particles and bulk components, due to reasons such as a low proportion of metal particles, only one peak may be obtained in the histogram pattern, as shown in Figure 2. In such cases, it is impossible to determine the threshold using the modal method described above, and therefore binarization processing cannot be performed on the CT image. As a result, it is not possible to extract the portion defined as metal particles from the CT image, and thus it is not possible to analyze the volume of the metal particles, the equivalent spherical diameter of the metal particles, the sphericity of the metal particles, etc.

[0013] The present invention was made under the circumstances described above, and the problem it aims to solve is to provide a method for determining a threshold in a gray value histogram, which shows the correlation between brightness and the frequency of brightness obtained when imaging a sample having multiple components with an X-ray CT scanner, in cases where it is difficult to determine a threshold using the mode method. [Means for solving the problem]

[0014] In order to solve the above problem, the present inventors conducted research, and arrived at a configuration for determining a threshold value by comparing a quantitative analysis result obtained by an analysis method such as the bromine-methanol method described above with a gray value histogram obtained by imaging a sample with an X-ray CT apparatus, and completed the present invention.

[0015] That is, a first invention for solving the above problem is: A method for determining a threshold value between the metal particles and the bulk component at the luminance in a gray value histogram that indicates the correlation between luminance and the frequency of the luminance, obtained when imaging a sample, which is a mixture of metal particles that are a metallic substance and a bulk component that is a non-metallic substance, with an X-ray CT apparatus, the method comprising: a first step of analyzing the respective abundances of the metallic substance and the non-metallic substance in the sample, and determining the volume occupied by the metallic substance in the sample; a second step of determining the number of voxels occupied by the metal particles in the gray value histogram from the volume occupied by the metallic substance and the voxel volume in CT imaging; a third step of accumulating the number of voxels in CT imaging from the side of higher luminance in the gray value histogram, and setting the luminance value obtained when the accumulated number equals the number of voxels occupied by the metal particles determined in the second step as the threshold value between the metal particles and the bulk component, which is a metal particle analysis method. A second invention is: the metal particle analysis method according to the first invention, wherein the non-metallic substance is a metal oxide. A third invention is: the metal particle analysis method according to the first invention, wherein the abundance ratio of fine metal particles having a particle diameter less than the detection lower limit in CT imaging of the sample is measured in advance, and when determining the threshold value between the metal particles and the bulk component in the third step, correction is performed by subtracting the number of voxels occupied by the fine metal particles from the number of voxels occupied by the metal particles. A fourth invention is: This is a method for analyzing metal particles according to any one of the first to third inventions, wherein the amount of the metal substance in the sample is analyzed using the bromine-methanol method. The fifth invention is, This is a metal particle analysis method that performs binarization processing of a CT image using a threshold value described in any of the first to third inventions. [Effects of the Invention]

[0016] According to the present invention, when a sample is a mixture of metallic particles and non-metallic bulk components, and is imaged with an X-ray CT scanner, a threshold can be determined in the gray value histogram, which shows the correlation between brightness and the frequency of that brightness, even when it is difficult to determine the threshold between metallic particles and bulk components using the mode method. [Brief explanation of the drawing]

[0017] [Figure 1] This is an example of a schematic histogram of gray value, showing the correlation between brightness and frequency (number of voxels). [Figure 2] This is another example of a schematic histogram of gray value, showing the correlation between brightness and frequency (number of voxels). [Figure 3] This is a schematic histogram of gray value, which shows the correlation between brightness and frequency (number of voxels), and represents the threshold (brightness value) at which the cumulative number of voxels equals the number of voxels occupied by metal particles. [Figure 4] In the gray value histogram for Example 1, the threshold at which the cumulative value of the number of voxels equals the number of voxels occupied by the metal particles is shown. [Figure 5] In the CT image according to Example 1, the region of interest identified as containing metal particles is surrounded by a white line. [Modes for carrying out the invention]

[0018] X-ray CT (Computed Tomography) equipment irradiates a sample with X-rays while rotating it 360°, and uses the differences in X-ray absorption by different materials in each part of the sample to non-destructively evaluate the internal structure of a material in three dimensions.

[0019] X-rays have the property of penetrating samples, but the amount of X-ray absorption differs depending on the material of each part of the sample. This phenomenon is called absorption contrast. When an X-ray transmission image of a sample is obtained using this absorption contrast, the transmission X-ray intensity decreases as the density, thickness, and atomic number of the sample increase, and the transmission X-ray intensity increases as the density, thickness, and atomic number of the sample decrease.

[0020] For example, if a sample is a mixture of metallic particles and non-metallic bulk components, the metallic particles will appear brighter in the CT image because they are denser than the bulk components. By using this brightness difference to analyze the CT image, it is possible to separate and observe the metallic particles in the sample from the bulk components and evaluate their volume, distribution, shape, etc.

[0021] In order to perform the image analysis described above, in order to separate metal particles from bulk components and extract the metal particles, a threshold value between the metal particles and bulk components at the given brightness level should be determined, and the CT image should be binarized using this threshold value.

[0022] As described above, if the sample is a mixture of metal particles and bulk components, and two peaks corresponding to peak A and peak B are obtained, as schematically shown in the histogram pattern in Figure 1, then from the relationship between density and brightness, it can be defined that "peak A represents metal particles, and peak B represents bulk components." Then, the threshold can be determined using the so-called mode method, where the boundary value of the valley between these two peaks is used as the threshold.

[0023] However, even if the sample is a mixture of metal particles and bulk components, there are cases where only one peak is obtained, as shown in Figure 2, due to reasons such as a low proportion of metal particles. When only one peak is obtained, it is difficult to determine the threshold using the modal method described above. For this reason, it was not possible to binarize the CT image.

[0024] On the other hand, chemical analysis methods such as the bromine-methanol method allow for the separate and quantitative analysis of the amounts of metallic substances and non-metallic substances (mainly metal oxides) in a sample. However, such chemical analysis methods cannot evaluate the distribution or shape of metal particles in a sample. Here, the inventors have come up with a configuration to determine a threshold between metal particles and bulk components by comparing the quantitative analysis results obtained by chemical analysis methods such as the bromine-methanol method with a gray value histogram.

[0025] Specifically, first, an X-ray CT scanner is used to obtain a CT image of the sample and a histogram of gray values. On the other hand, the same sample is subjected to quantitative analysis by separating the amount of metallic substances from the amount of non-metallic substances using methods such as the bromine-methanol method. Then, using the results of this quantitative analysis, the so-called P-tile method (Percentile Method) is applied to determine the luminance threshold at the boundary between metallic particles and bulk components in the gray value histogram, as shown in Figure 3. Furthermore, the P-tile method is a technique that estimates a threshold using the fraction of a specific layer or structure, and is explained in Non-Patent Documents 2 and 3.

[0026] Specifically, if the weight percentage of metallic material in the sample is known by a chemical analysis method, the total volume of the metallic material can be calculated based on the total sample mass used to create the gray value histogram and the density of the metallic material. Next, in the gray value histogram shown in Figure 3, the number of voxels is accumulated from the highest brightness, and the brightness value when the accumulated number of voxels equals the number of voxels occupied by the metallic particles (the gray area in Figure 3) can be used as the threshold.

[0027] Furthermore, if a metallic substance consists of multiple metallic elements, the weight percentage of each metallic element can be measured using an appropriate quantitative analysis method, the volume of each metallic element can be determined based on the density of each metallic element, and the total volume of the metallic substance can be calculated by summing them up.

[0028] Furthermore, even when the densities of each metallic substance differ significantly, the volume of the high-density metallic substance and the volume of the low-density metallic substance are individually determined using appropriate quantitative analysis methods. Then, in the gray value histogram, the number of voxels is accumulated from the highest luminance, and the luminance value when the accumulated number of voxels equals the number of voxels occupied by the high-density metallic particles is set as the threshold between the high-density and low-density metallic particles. The luminance value when the accumulated number of voxels equals the number of voxels occupied by the low-density metallic particles can be set as the threshold between the low-density metallic particles and the bulk component. Metallic substances with significantly different densities include, for example, aluminum and iron, or iron and lead.

[0029] On the other hand, if the bulk component contains minute metal particles smaller than the resolution of the X-ray CT scanner, it is possible that the voxels occupied by these bulk components may be detected as voxels occupied by the metal particles. In such cases, it is preferable to correct the calculated threshold value. Specifically, the proportion of minute metal particles with a particle size smaller than the detection limit in CT imaging is measured in advance. Then, when determining the threshold between metal particles and the bulk component in the sample, a correction is made by subtracting the number of voxels occupied by the minute metal particles with a particle size smaller than the detection limit from the number of voxels occupied by the metal particles.

[0030] To pre-measure the relative abundance of minute metal particles with a particle size below the detection limit in CT imaging, for example, the relative abundance of minute metal particles can be measured by analyzing images acquired through SEM-EDS (Scanning Electron Microscope-Energy Dispersive X-ray Spectroscopy). Alternatively, the relative abundance of minute metal particles may be measured separately by performing ultra-high-resolution (minimum pixel size) X-ray CT imaging.

[0031] When continuously measuring the same type of sample, it is also preferable to accumulate data on the relative abundance of minute metal particles with a particle size below the detection limit in the CT scan described above, predetermine correction parameters according to the sample being measured and the measurement conditions, and perform the correction using a simple calculation formula.

[0032] As described above, the luminance value at the obtained threshold is read. Then, a binarization process is performed on the CT image, where the portion with a luminance value greater than the read value is treated as metal particles, and the portion with a luminance value less than the read value is treated as the bulk component. As a result of this binarization process, the boundary between metal particles and the bulk component is clarified in the CT image.

[0033] As a result, for example, in a sample where a non-metallic bulk component serves as a matrix and metal particles are dispersed within it, it becomes possible to obtain various data such as the volume of the metal particles, their equivalent spherical diameter, sphericity, and their distribution within the bulk component.

[0034] Furthermore, if the metallic material consists of multiple metallic elements, it is possible to obtain various data regarding the distribution and state of metallic particles composed of each metallic substance within the sample by performing EDX (energy-dispersive X-ray spectroscopy) or MLA (mineral particle analysis) analysis as desired. [Examples]

[0035] The present invention will be specifically explained using an example in which a threshold value between metal particles and bulk components (boundary) is determined in a gray value histogram obtained by an X-ray CT apparatus using a sample consisting of metal particles and bulk components. However, the present invention is not limited to this example.

[0036] (1) Sample to be measured: A sample consisting of metallic and nonmetallic substances was prepared. This sample is a mixture of metallic and nonmetallic substances in which ferronickel metal (iron-nickel alloy) particles, which are metallic particles obtained by reducing and melting limonite ore, are dispersed in bulk components (metal oxides). After cooling the above reduced molten material in a nitrogen gas atmosphere, it was allowed to cool in the air and then crushed with a hammer to particles of up to approximately 6 mm in size, and this was used as the sample to be measured.

[0037] (2) The amount of metallic substance P [wt%] in the sample was quantified using the bromine-methanol method as a chemical analysis method. In this case, if there were multiple metal elements n, the amount of each metal element Pi [wt%] was quantified. Specifically, the sample to be measured was divided, and the weight content of metallic iron and metallic nickel was quantified by the bromine-methanol method. As a result, values ​​of 3.0 wt% for iron and 0.3 wt% for nickel were obtained.

[0038] (3) Using an X-ray CT scanner, CT images of the entire area of ​​the sample were taken. At this time, the sample amount of 2.0 g was weighed in advance. For CT imaging, a Rigaku CTLabHX130 was used, and the imaging conditions were set to "FOV 15 Super High (pixel size: 5.3 μm, number of integrated images: 8)".

[0039] In this invention, FOV (Field of View) refers to the diameter (mm) of the cylindrical sample container. A smaller FOV results in a smaller minimum pixel size and higher resolution. However, if the FOV is too small, the amount of sample measured decreases, leading to a larger variation error in the CT scan sample relative to the sample being measured. Therefore, an FOV of 15, which is the optimal condition for the sample being measured, was selected. The number of integrated images is the number of images taken to capture one transmission image. In other words, eight images are combined to create one transmission image in order to eliminate the effects of fluctuations due to noise.

[0040] (4) Here, the total volume V [cm³] of metallic material in the sample, which was imaged in its entirety using an X-ray CT scanner. 3 ] was calculated from (Equation 1). JPEG0007913431000001.jpg16170 where ρi: density of each metallic element Here, P obtained in (2) Fe =3.0wt%, P Ni = 0.3% by weight, the value w = 2.0 [g] obtained in (3), and ρ Fe =7.87, ρ Ni = 8.90, therefore V = 8.26 × 10 -3 cm 3 The following value was obtained.

[0041] (5) The number of voxels corresponding to metal particles in the sample imaged in the entire region in (3) above was determined as follows. <1> The size of a single voxel is determined by the imaging conditions of the X-ray CT scanner. (For example, when the length of one side of the cube is l [cm], the volume v of one voxel is v [cm]3 / voxel]=l 3 [cm 3 . ) Under the present imaging conditions, since the pixel size is 5.3 μm, v=(5.3 μm) 3 =149μm 3 =1.49×10 -10 cm 3 is obtained. <2> In the sample subjected to full-area imaging, the voxel number C corresponding to the total volume V of the metal substance calculated in (4) above was calculated by (Equation 2). JPEG0007913431000002.jpg11170Here, V=8.26×10 -3 cm 3 , v=1.49×10 -10 cm 3 from which, C=5.55×10 7 was obtained.

[0042] (6) In the histogram of gray values obtained by CT imaging of the entire region of the sample shown in FIG. 4, the number of voxels is accumulated starting from the higher luminance side, and when the value of "total" in the upper right column of FIG. 4 reaches the voxel number C=5.55×10 7 calculated in (5) <2>, the luminance value "left", which is the threshold (boundary) between the metal particles and the bulk component, was 2931.

[0043] (7) Using the obtained luminance threshold of 2931, FIG. 5 shows a CT image in which the luminance value is 2931 or higher, and the periphery of the region of interest determined to be metal particles is surrounded by a white line.

Claims

1. A method for determining a threshold value between the metal particles and the bulk component in terms of brightness, in a gray value histogram showing the correlation between brightness and the frequency of brightness obtained when a sample, which is a mixture of metallic particles and a non-metallic bulk component, is imaged with an X-ray CT scanner, A first step involves analyzing the respective amounts of the metallic substance and the non-metallic substance in the sample and determining the volume occupied by the metallic substance in the sample. A second step of determining the number of voxels occupied by the metal particles in the gray value histogram from the volume occupied by the metal substance and the voxel volume in the CT scan, A metal particle analysis method comprising: a third step in which, in the gray value histogram, the number of voxels in the CT scan is accumulated from the highest brightness, and the brightness value when it becomes equal to the number of voxels occupied by the metal particles obtained in the second step is set as a threshold between the metal particles and the bulk component.

2. The metal particle analysis method according to claim 1, wherein the nonmetallic substance is a metal oxide.

3. The metal particle analysis method according to claim 1, wherein in CT imaging of the sample, the proportion of minute metal particles having a particle size below the detection limit is measured in advance, and in the third step, when determining the threshold between the metal particles and the bulk component, a correction is made by subtracting the number of voxels occupied by the minute metal particles from the number of voxels occupied by the metal particles.

4. A method for analyzing metal particles according to any one of claims 1 to 3, wherein the amount of the metallic substance in the sample is analyzed using the bromine-methanol method.

5. A method for analyzing metal particles, comprising performing binarization processing of a CT image using a threshold value described in any one of claims 1 to 3.

Images