Method for evaluating adhesive property between rubber and metal
A method using X-ray absorption fine structure analysis and machine learning creates scatter diagram data to quantify adhesion changes in vulcanized rubber due to moist heat aging, addressing the challenge of assessing adhesion changes and improving product quality.
Patent Information
- Application Number
- JP2023084681
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-23
- Publication Date
- 2024-12-05
AI Technical Summary
Existing methods struggle to clearly understand the changes in adhesion between metal and vulcanized rubber due to moist heat aging, as the intricate relationship between various parameters makes it difficult to assess the impact of metal component diffusion.
A method involving three-dimensional spatially resolved X-ray absorption fine structure analysis is used to obtain three-dimensional image data of granules in a vulcanized rubber sample, treating metal compounds with different valences as separate components, and using machine learning to create scatter diagram data that plots feature points before and after moist heat aging, allowing for a clearer evaluation of adhesion changes.
This method enables a more accurate visual understanding of adhesion changes by quantifying metal component ionization and sulfurization, facilitating improvements in adhesion between vulcanized rubber and metal in products like tires.
Smart Images

Figure 2024168203000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to a method for evaluating adhesion between rubber and metal, and more particularly to a method for evaluating adhesion between rubber and metal that can more clearly grasp the degree of change in adhesion between metal and vulcanized rubber due to moist heat aging. [Background technology]
[0002] Rubber products such as tires use metal materials such as steel cords. The adhesion between metal materials and vulcanized rubber is significantly affected by moist heat aging. For example, the brass components of brass-plated steel cords and the copper components and zinc compounds of the metal sulfide layer ionize and diffuse into the vulcanized rubber, reducing adhesion.
[0003] The present inventors have proposed a method for three-dimensionally quantifying and evaluating the increase and decrease of metal components diffusing into rubber by computer tomography-X-ray absorption fine structure analysis (see Patent Document 1). In the invention proposed in Patent Document 1, the increase and decrease of metal components in rubber is grasped microscopically using a three-dimensional image of the metal component particles obtained by computer tomography-X-ray absorption fine structure analysis. From the obtained three-dimensional image, various parameters such as particle composition, particle diameter (equivalent sphere diameter), particle shape, and particle density can be obtained. However, it is difficult to grasp the adhesiveness and the degree of change due to moist heat aging because various parameters are intricately related. Therefore, further ingenuity is required to more clearly grasp the degree of change in the adhesiveness between vulcanized rubber and metal due to moist heat aging. [Prior art documents] [Patent documents]
[0004] [Patent Document 1] JP 2022-188741 A Summary of the Invention [Problem to be solved by the invention]
[0005] An object of the present invention is to provide a method for evaluating adhesion between rubber and metal, which enables a clear understanding of the degree of change in adhesion between metal and vulcanized rubber due to moist heat aging. [Means for solving the problem]
[0006] The method for evaluating adhesion between rubber and metal of the present invention, which achieves the above object, uses a vulcanized rubber obtained by vulcanizing unvulcanized rubber containing a metal as a sample, obtains three-dimensional image data of a plurality of granules having a metal component derived from the metal present in the sample by three-dimensional spatially resolved X-ray absorption fine structure analysis, and utilizes the obtained three-dimensional image data. In the method for evaluating adhesion between rubber and metal, metal compounds having different metal components and metal compounds of the same type of metal component but with different valences are treated as particles of different components for each of the granules, and an initial measurement step using the sample and a post-treatment measurement step using the sample that has been subjected to a moist heat aging treatment after the initial measurement step are carried out, In each of the measurement steps, the respective three-dimensional image data is acquired, and the acquired three-dimensional image data is used to obtain the detection intensity and detection frequency of each of the particles of the different components present in the same granular body as measurement results.The measurement results for each of the granular bodies in each of the measurement steps are data-processed by a computing device using machine learning to create scatter plot data in which characteristic points showing the characteristics of each of the particles of the different components in each of the same granular bodies before and after the moist heat aging treatment are plotted, and the adhesion between the vulcanized rubber and the metal is evaluated using the created scatter plot data. Effect of the Invention
[0007] According to the present invention, each characteristic point plotted in the scatter diagram data represents the volume or mass of each different component of the metal components in each granule before and after the moist heat aging treatment. That is, by comparing each characteristic point of each different component of the metal components in the scatter diagram data before and after the moist heat aging treatment, the progress of ionization of each metal component in the sample and the progress of the sulfurization reaction with the rubber sulfur can be understood. Therefore, by creating the scatter diagram data, the change in the adhesion between the vulcanized rubber and the metal due to moist heat aging can be visually and more clearly understood. [Brief description of the drawings]
[0008] [Figure 1] FIG. 1 is an explanatory diagram illustrating an evaluation system. [Diagram 2] FIG. 2 is an explanatory diagram illustrating three-dimensional image data of a granular body before a moist heat aging treatment. [Diagram 3] FIG. 11 is an explanatory diagram illustrating three-dimensional image data of a granular body after moist heat aging treatment. [Figure 4] FIG. 1 is a flow diagram illustrating the steps of an embodiment of a method for evaluating rubber-to-metal adhesion. [Diagram 5] FIG. 11 is an explanatory diagram illustrating measurement result data in an initial measurement process. [Figure 6] FIG. 11 is an explanatory diagram illustrating measurement result data in a post-processing measurement step. [Figure 7] FIG. 11 is an explanatory diagram illustrating feature data in an initial measurement process. [Figure 8] FIG. 11 is an explanatory diagram illustrating an example of scatter diagram data. [Figure 9] FIG. 11 is an explanatory diagram illustrating another example of scatter diagram data. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0009] Hereinafter, the method for evaluating adhesion between rubber and metal according to the present invention will be described with reference to the embodiment shown in the drawings.
[0010] The evaluation system 1 illustrated in FIG. 1 is used to carry out an evaluation method for the adhesion between rubber and metal. In this evaluation method, a measurement is performed using three-dimensional image data D1, D2 (FIGS. 2 and 3 described later) of granular material contained in a sample 10 acquired using a three-dimensional spatially resolved X-ray absorption fine structure imaging analyzer 2 (hereinafter, XAFS imaging analyzer 2). In this measurement, an initial measurement step (initial measurement step S110 illustrated in FIG. 4 described later) and a post-treatment measurement step (post-treatment measurement step S130 illustrated in FIG. 4 described later) are performed. Next, the measurement results obtained in each measurement are processed by a calculation device 3 using machine learning to create scatter diagram data D3 illustrated in FIG. 8 described later. Then, the created scatter diagram data D3 is used to evaluate the adhesion between vulcanized rubber and metal.
[0011] First, the evaluation system 1 will be described.
[0012] The evaluation system 1 includes an XAFS imaging analysis device 2 and a calculation device 3. Various data are input and stored in the calculation device 3, and the calculation device 3 performs data processing using the input data. Various known computers can be used for the calculation device 3. The calculation device 3 includes a central processing unit (CPU), a main storage unit (memory), an auxiliary storage unit (e.g., HDD), an input unit (keyboard, mouse), and an output unit (display).
[0013] The XAFS imaging analysis device 2 may be a known XAFS imaging analysis device that measures and analyzes the X-ray absorption spectrum of the sample 10 by three-dimensional spatially resolved X-ray absorption fine structure analysis. Various known techniques such as computer tomography-X-ray absorption fine structure analysis may be used for the three-dimensional spatially resolved X-ray absorption fine structure analysis.
[0014] For sample 10, a metal-rubber composite material manufactured by vulcanizing unvulcanized rubber containing metal is used. In the vulcanized rubber of sample 10, there are many granules having metal components derived from metal. Each granule contains particles of one type of metal component (particles of zero-valent metal component) and particles of a metal compound of the metal component and sulfur of the rubber (particles of monovalent, divalent, etc. metal components). For example, a metal-rubber composite material such as sample 10 roughly reproduces the state of vulcanized rubber in contact with a brass-plated steel cord in a rubber product such as a tire. Therefore, by grasping the degree of change in the adhesiveness between the rubber and metal in sample 10 due to moist heat aging and evaluating the adhesiveness, it is possible to find measures to improve the adhesiveness against moist heat aging.
[0015] Sample 10 is, for example, vulcanized rubber containing 0.5% by mass to 10% by mass of metal powder having a particle diameter (major axis) of 1 μm to 100 μm. The metal powder preferably contains at least one of copper and zinc, and may contain both copper and zinc. The metal powder may be, for example, brass powder containing 50% by mass to 90% by mass of copper and 50% by mass to 10% by mass of zinc. By evaluating sample 10 containing such brass powder, the adhesion between a brass-plated steel cord and vulcanized rubber in a rubber product such as a tire can be evaluated more faithfully in accordance with the actual state of the rubber product.
[0016] Unvulcanized rubber may be any rubber component (polymer) that can be crosslinked by a crosslinking agent containing sulfur, such as natural rubber (NR) or synthetic rubbers such as styrene-butadiene rubber (SBR), butadiene rubber (BR), and isoprene rubber (IR). The crosslinking agent may contain at least sulfur, and may be used in combination with an organic peroxide such as peroxide. Unvulcanized rubber may contain additives other than the crosslinking agent, such as crosslinking accelerators (vulcanization accelerators), fillers (silica, carbon black, etc.), antioxidants, flame retardants, reinforcing materials, and colorants.
[0017] The degree of vulcanization of sample 10 may be set to a desired level, provided that granules containing metal compounds produced by a vulcanization reaction between the metal components and the sulfur contained in the rubber are present. The vulcanized rubber of sample 10 contains a large number of granules with different metal components (e.g., copper and zinc). In addition, when granules containing the same type of metal component contain metal compounds with different valences of the metal components, each metal compound is treated as an individual particle with a different component. For example, in the case of granules containing a copper component in vulcanized rubber, Cu is contained in the granules as a zero-valent copper component (Cu), and 1-valent copper component (Cu + ) as Cu2S, divalent copper component (Cu 2+ In some cases, particles of CuS may exist as CuS particles. In that case, each is treated as an individual particle of a different component. Therefore, in a granular material containing copper, there are zero-valent copper components (Cu), monovalent copper components (Cu + ), divalent copper content (Cu 2+ ) are treated as three separate particles with different components. Similarly, in a granular material containing zinc, Zn is treated as zero-valent zinc component (Zn), divalent zinc component (Zn 2+ ), and each ZnS particle is treated as two separate particles with different components. In this way, by treating metal compounds with different valences of the metal component that exist in granules containing the same type of metal component as separate particles with different components, the change in valence of the metal component (ionization and sulfidation reaction) before and after moist heat aging of the metal-rubber composite material can be accurately understood.
[0018] In the measurement using the XAFS imaging analysis device 2, three-dimensional image data (three-dimensional image data D1, D2 exemplified in Figs. 2 and 3 described later) of each granular body present in the sample 10 is acquired, and the detection intensity and detection frequency of each particle of different components in the granular body are measured from the acquired three-dimensional image data D1, D2 to obtain the measurement results. In a method of acquiring the three-dimensional image data D1, D2, for example, a tube 11 containing the sample 10 is used as a specimen, and X-ray absorption fine structure analysis is performed while rotating the specimen. The specimen is created by creating a cylindrical sample 10 with a diameter of 100 μm to 2 mm and storing the created sample in a hollow tube 11 made of a polyimide film. The specimen is rotated at small angles of ±90 degrees with respect to the optical axis of the X-ray and irradiated with X-rays. When copper and zinc powders are used as the metal powders of sample 10, the X-ray irradiation energy includes the copper K absorption edge (K shell absorption edge) of 8857 eV to 9357 eV and the zinc K absorption edge of 9640 eV to 9700 eV, and in order to obtain more detailed chemical information, it is preferable to further include the region of +50 eV to +1500 eV from these absorption edges.
[0019] The conditions for obtaining the three-dimensional image data D1 and D2 using the XAFS imaging analysis device 2 are not particularly limited, and can be, for example, as follows. When the specimen is irradiated with X-rays and the energy of the X-rays is increased, the absorption of the X-rays is detected at the energy of the K-absorption edge of the metal component of the specimen 10. The X-rays transmitted through the specimen are projected two-dimensionally and imaged to obtain a two-dimensional image of the amount of absorption in the specimen 10. In addition, a two-dimensional image of the absorption is captured while rotating the specimen by a small angle, and an image showing the valence of the metal component and the local structure at the atomic level (atomic distance, coordination number) is obtained by analyzing the amount of absorption of X-rays and the detection intensity and shape of the absorption peak in each image. Specifically, the valence of the metal component is obtained by analyzing the X-ray absorption edge structure region. In addition, the local structure at the atomic level (atomic distance, coordination number) is obtained by analyzing the X-ray wide area fine structure region. The three-dimensional image data D1 and D2 are obtained by reconstructing each image.
[0020] The three-dimensional image data D1 shown in FIG. 2 is obtained by an XAFS imaging analysis device 2 using a sample 10 that has not been subjected to a moist heat aging treatment. The three-dimensional image data D1 shows a granular body containing a copper component, and the granular body contains copper components (Cu, Cu) with different valences (0 valence, 1 valence, 2 valence). + , Cu 2+ In detail, zero-valent copper (Cu) exists in the center of the granule, and monovalent copper (Cu + ) is present. 2+ ) is not shown in Figure 2 because its abundance ratio is small compared to particles of zero-valent and monovalent copper components, but it has been quantified. + ), divalent copper content (Cu 2+ ), with the abundance ratio of zero-valent copper (Cu) being the highest.
[0021] The three-dimensional image data D2 shown in FIG. 3 was obtained by the XAFS analysis device 2 using the sample 10 that was subjected to the moist heat aging treatment. This three-dimensional image data D2 shows the same granules as the three-dimensional image data D1 shown in FIG. 2, and the granules are present at the same positions of the sample 10. The abundance ratio of the zero-valent copper component (Cu) present in the center of the granules is smaller than that in FIG. 2. On the other hand, the monovalent copper component (Cu + ) is more abundant and more widely dispersed than in Figure 2. 2+ ) is also more abundant than in Figure 2.
[0022] Next, the evaluation method illustrated in FIG. 4 will be described.
[0023] In this evaluation method, the XAFS imaging analyzer 2 is used to measure each granular body present in the sample 10 (S110 to S130). This measurement is performed in an initial measurement step S110 and a post-treatment measurement step S130. The same sample 10 is used in the initial measurement step S110 and the post-treatment measurement step S130. The sample 10 prepared in the initial measurement step S110 is used as is, and the sample 10 used in the initial measurement step S110 is used after being subjected to a moist heat aging treatment (S120) in the post-treatment measurement step S130. Next, the measurement results for each granular body obtained by the measurement are processed by the calculation device 3 using machine learning (S140). The adhesion between the vulcanized rubber and the metal in the sample 10 is evaluated using the scatter diagram data D3 created by this data processing (S150). The contents of each step (S110 to S150) are described in detail below.
[0024] In the initial measurement step S110, the XAFS imaging analysis device 2 acquires the above-mentioned three-dimensional image data D1 in Fig. 2, and the acquired three-dimensional image data D1 is used to measure the detection intensity (density) and detection frequency (proportion, abundance ratio) of each particle of different components in the granular material. The measurement results in the initial measurement step S110 are stored in the auxiliary storage unit of the calculation device 3 as measurement result data D4 (exemplified in Fig. 5 described later).
[0025] A large number of granules are present in the sample 10, and in the initial measurement step S110, a plurality of granules are selected from each of the granules. Specifically, the large number of granules present in the sample 10 are classified into groups according to the metal components contained therein, and a plurality of granules are selected from each of the classified groups. Furthermore, when metal compounds with different valences of the metal components are generated in the granules, granules containing all metal compounds that may be generated are adopted as the granules to be measured. For example, when granules containing a copper component and granules containing a zinc component are present in the sample 10, the granules containing the copper component and the granules containing the zinc component are classified into different groups, and a plurality of granules are selected from each group. Furthermore, when the granules to be measured contain a copper component, zero-valent, monovalent, and divalent copper components (Cu, Cu + , Cu2 +In addition, when the granular material to be measured contains zinc, it is advisable to use a granular material containing zero-valent and divalent zinc components (Zn, Zn 2+ It is preferable to use a granular material having particles of the above formula (1).
[0026] In the post-treatment measurement step S130 described below, granular bodies present in the sample 10 at the same positions as those measured in the initial measurement step S110 are measured, i.e., the same granular bodies are measured and compared in each step S110 and S130, so the positions of each measured granular body are made identifiable in the initial measurement step S110. The positions of the measured granular bodies are identified using a relative positional relationship with a mark attached to the tube 11 containing the sample 10. This makes it possible to use the same granular body as the measurement target in both the initial measurement step S110 and the post-treatment measurement step S130, and makes it possible to more reliably grasp changes in the same granular body due to the moist heat aging treatment.
[0027] In the three-dimensional image data D1 and D2 in Figs. 2 and 3 described above, there are particles with different detection intensities (densities) even for the same copper component contained in the same granular material. In the present invention, particles having the same (similar) detection intensities for zero-valent, monovalent, and divalent copper components are treated as particles of the same component. For example, in Fig. 2 described above, divalent copper components are scattered at multiple locations in the granular material, but particles of the same valence and having the same (similar) detection intensities scattered in a single mass of granular material are collectively regarded as one particle. Therefore, in this initial measurement step S110, for example, the detection frequency (proportion, abundance ratio) for each detection intensity (density) of zero-valent copper component (Cu) is measured, and the detection frequency (proportion, abundance ratio) for monovalent copper component (Cu + The detection frequency for each detection intensity of divalent copper (Cu 2+ ) is measured for each detection intensity.
[0028] The measurement result data D4 illustrated in Fig. 5 is created based on the measurement results using the three-dimensional image data D1 in Fig. 2 and stored in the auxiliary memory of the calculation device 3. The measurement result data D4 indicates the detection intensity (density) and detection frequency (proportion, abundance ratio) of each particle of different components in the same granular material. Therefore, this measurement result data D4 is created for each granular material that is the measurement target in the initial measurement step S110. If the number of granular materials to be measured is, for example, 100, 100 pieces of measurement result data D4 are also created.
[0029] In the measurement result data D4, the horizontal axis is the detection intensity (density) and the vertical axis is the detection frequency (ratio, abundance ratio). In FIG. 5, a, b, c, ... are attached to the horizontal axis to indicate a predetermined detection intensity range, and the detection intensity increases in the order of detection intensity range a. The measurement result data D4 indicates the total mass and number of each particle. The horizontal and vertical axes of the measurement result data D4 can be set arbitrarily as long as the measurement result data D4 can indicate the total mass and number of each particle of different components in the same granular body. For example, the detection intensity of a particle is highly correlated with the density of the particle, and the higher the detection intensity, the higher the density of the metal component, and the lower the detection intensity, the lower the density of the metal component. In addition, the detection frequency of a particle is highly correlated with the ratio of the total volume of the particle (the ratio of the total volume of the particle in the granular body) and the abundance ratio in the granular body, and the higher the detection frequency, the larger the ratio of the total volume of the particle and the larger the abundance ratio, and the lower the detection frequency, the smaller the ratio of the total volume of the particle and the smaller the abundance ratio. Therefore, these parameters can also be used for the horizontal and vertical axes of the measurement result data D4.
[0030] In step (S120), the sample 10 used in the initial measurement step S110 is subjected to a moist heat aging treatment. The conditions of the moist heat aging treatment are set, for example, at a heating temperature of 70° C., a humidity of 96% RH, and a treatment period of two weeks. The conditions of the moist heat aging treatment can be changed as appropriate depending on the type of metal component contained in the sample 10, the rubber composition and vulcanization conditions, the usage environment as a metal-rubber composite material, and the like. For example, the heating temperature is set to 50° C. to 100° C., the humidity to 50% RH to 100% RH, and the treatment period to one day to one month. The moist heat aging treatment is performed while the sample 10 is stored in the tube 11.
[0031] In the post-processing measurement step S130, the same measurement as in the initial measurement step S110 is performed on the sample 10 that has been subjected to the above-mentioned moist heat aging treatment in step (S120). In the post-processing measurement step S130, the XAFS imaging analysis device 2 acquires the above-mentioned three-dimensional image data D2 in FIG. 3, and the acquired three-dimensional image data D2 is used to measure the detection intensity (density) and detection frequency (proportion, abundance ratio) of each particle of different components in the granular material. The measurement results in the post-processing measurement step S130 are stored in the auxiliary storage unit of the calculation device 3 as measurement result data D5 illustrated in FIG. 6, which will be described later, similar to the measurement result data D4 illustrated in FIG. 5.
[0032] In the post-processing measurement step S130, a granular body existing at the same position as the position measured in the initial measurement step S110 is used as the measurement target. Specifically, in the post-processing measurement step S130, the position measured in the initial measurement step S110 is identified using a mark attached to the tube 11 containing the sample 10 and the relative positional relationship identified in the initial measurement step S110. In this way, even if an attempt is made to identify the same granular body in the initial measurement step S110 and the post-processing measurement step S130, it may not be possible to determine that they are the same granular body, so only granular bodies that can be reliably determined to be the same granular body are used.
[0033] The measurement result data D5 illustrated in FIG. 6 is created based on the measurement results using the three-dimensional image data D2 in FIG. 3 and is stored in the auxiliary memory of the calculation device 3. The measurement result data D5, like the measurement result data D4 in FIG. 5 described above, indicates the detection intensity (density) and detection frequency (proportion, abundance ratio) of particles of different components in the same granular material. Comparing this measurement result data D5 with the measurement result data D4 in FIG. 5, the measurement result data D5 shows that the detection frequency of zero-valent copper components (Cu) is generally low, and the detection intensity is also low. Also, in the measurement result data D5, the detection frequency of monovalent and divalent copper components (Cu + , Cu 2+ ) detection intensity (density) and detection frequency (ratio, abundance ratio) are high. Therefore, by comparing the measurement result data D5 with the measurement result data D4, it is clear that particles of zero-valent copper components (Cu) are ionized and diffused by the moist heat aging treatment, and are converted into monovalent and divalent copper components (Cu) by a sulfurization reaction with the sulfur in the rubber. + , Cu 2+ ) is increasing.
[0034] In step (S140), data processing is performed to create scatter plot data D3, as shown in FIG. 8, which will be described later, from the respective measurement results measured by the calculation device 3 in the initial measurement step S110 and the post-treatment measurement step S130 using machine learning. In the scatter plot data D3, feature points are plotted that indicate the characteristics of each particle of different components in each of the same granules before and after moist heat aging treatment. The features of each particle of different components in the same granules refer to the features relating to each of the different components. Specifically, the calculation device 3 performs data processing to extract feature amounts for a plurality of types of features from the measurement result data D4 and D5, using feature amount extraction by machine learning, that indicate the characteristics relating to each of the different components in each granule before and after moist heat aging treatment. For example, in a granule containing a copper component, zero-valent, monovalent, and divalent copper components (Cu, Cu + , Cu 2+) are extracted. Next, the calculation device 3 performs data processing to create scatter plot data D3 using the principal component scores obtained by performing dimensionality reduction analysis on the extracted feature data of each particle as feature points.
[0035] The feature extraction can be performed using various known feature extraction methods such as a filter method, a wrapper method, and an embedding method. Also, the feature can be extracted for a feature specified using the input unit of the arithmetic device 3. A person skilled in the art knows the features closely related to the adhesion between vulcanized rubber and metal by referring to the experience gained in the manufacturing process of rubber products, various known literature, the accumulation of a large number of experimental and test data, the accumulation of computer simulation results, and the like. Therefore, in the feature extraction, a feature closely related to the adhesion previously known may be specified, and a feature indicating the specified feature may be extracted.
[0036] Examples of the extracted feature amounts include the ratio of the volume of components with different valences in a lump of granular material (the ratio of the volume of each of zero-valent, monovalent, and divalent copper components in the granular material to the volume of the granular material), the ratio of components with different valences in a predetermined detection intensity range (the ratio of zero-valent, monovalent, and divalent copper components in detection intensity range a), etc. The number of extracted feature amounts (types of features) is not particularly limited, but is, for example, about 50.
[0037] The characteristic data D6 illustrated in FIG. 7 is generated by the calculation device 3 processing the measurement result data D4 in the initial measurement step S110 using machine learning, and is stored in the auxiliary storage unit of the calculation device 3. The characteristic data D6 is the copper content (Cu, Cu) of each of the particles (A1, . . . , An) listed in the leftmost column of the table in FIG. + , Cu 2+7, a group of data (a collection of feature amounts for each particle and for each component with different valences) in which feature amounts indicating features related to a large number of types of features are accumulated. In the table of FIG. 7, some of the features and feature amounts are omitted by "···". At the same time as the feature data D6, another feature data D7 is created by data processing of the calculation device 3 using machine learning on the measurement result data D5 measured in the post-processing measurement step S130. This feature data D7 has a data structure similar to that of the feature data D6 illustrated in FIG. 7, and only the feature amounts described in the table are different, so a detailed description will be omitted.
[0038] In creating the scatter plot data D3 (illustrated in FIG. 8), the feature data D6 and D7 are integrated, and the integrated data is subjected to dimension reduction by principal component analysis, and the principal component scores for each particle and each component obtained are used as feature points. The dimension reduction analysis is not limited to principal component analysis, and various known machine learning dimension reduction methods such as non-negative matrix factorization, t-distribution type stochastic neighborhood embedding method, and autoencoder can be used. Note that the reduction of the dimension of data by machine learning is not limited to reduction from multi-dimensions to two-dimensions, and may be reduction from multi-dimensions to three-dimensions, for example.
[0039] The scatter plot data D3 illustrated in FIG. 8 is created using the principal component scores for each particle and each component as feature points, and is stored in the auxiliary storage unit of the calculation device 3. The scatter plot data D3 regards each extracted feature amount as a parameter, and has a first principal component axis PC1, one of two axes created from these parameters, as the horizontal axis, and a second principal component axis PC2 as the vertical axis. The open circles in FIG. 8 are plots of the feature points (principal component scores) of the zero-valent copper component (Cu) before the moist heat aging treatment for each particle. The black circles are plots of the feature points (principal component scores) of the zero-valent copper component (Cu) after the moist heat aging treatment for each particle. The white triangles are plots of the monovalent copper component (Cu) before the moist heat aging treatment for each particle. + The black triangles indicate the characteristic points (principal component scores) of the monovalent copper component (Cu +The white squares indicate the divalent copper component (Cu 2+ The black squares indicate the divalent copper component (Cu) after moist heat aging treatment. 2+ ) are plotted for each particle. In the actual scatter plot data D3, the same number of feature points as the number of granular objects to be measured are plotted for each particle and for each component, but in Fig. 8, 10 representative feature points are used to avoid complicating the diagram.
[0040] Each plotted feature point represents the state of each different component in each granular material before and after the moist heat aging treatment. More specifically, each feature point is the result of analyzing each feature amount reflecting the detection intensity (density) and detection frequency (proportion, abundance ratio) in the measurement result data D4 and D5, and represents the state of each component of the same metal component but with different valences in the granular material, such as the total mass, number of particles, or total volume.
[0041] The three arrows in FIG. 8 indicate the copper content (Cu, Cu) before and after the moist heat aging treatment (S120). + , Cu 2+ ) changes in state. In Fig. 8, the changes in the state of each copper component (Cu, Cu + , Cu 2+ ), three arrows are shown showing a comprehensive trend from before to after the moist heat aging treatment. However, in reality, the copper components (Cu, Cu) present in the same granules are + , Cu 2+ ) are indicated by arrows. Each arrow indicates the copper component (Cu, Cu + , Cu 2+) before and after the moist heat aging treatment are taken as the starting and ending points. The first principal component axis PC1 and the second principal component axis PC2 of the scatter plot data D3 can be used to estimate the features represented by each axis by analyzing the results of principal component analysis. From this estimation, the first principal component axis PC1 of the scatter plot data D1 generally represents the detection intensity (density) and abundance ratio of components with different valences in the granular material, and the value increases from the left side to the right side of the graph in FIG. 8. The second principal component axis PC2 generally represents the concentration (mass percentage, volume percentage, etc.) of components with different valences in the granular material, and the concentration increases toward the lower side of the graph in FIG. 8 and decreases toward the upper side. This supposition is found to be generally correct, even when taking into consideration that the measurement result data D4 and D5 in the above-mentioned Figures 5 and 6, on which the scatter plot data D3 is based, represent the detection intensity (density) and detection frequency (proportion, abundance ratio) of each particle. Therefore, it can be seen that each arrow in this scatter plot data D3 shows the increase / decrease in the total mass and particle number of each component and the diffusion state before and after the moist heat aging treatment (S120) of sample 10. Specifically, by using the scatter plot data D3, it is possible to determine the decrease in the mass of zero-valent copper component (Cu) in sample 10 due to moist heat aging and the decrease in the mass of monovalent and divalent copper components (Cu + , Cu 2+ The increase in mass of each copper component (Cu, Cu + , Cu 2+ ) is widely diffused in the sample 10.
[0042] In step (S150), the adhesiveness between the rubber and the metal is evaluated using the created scatter plot data D3. From the increase or decrease in the mass of each copper component in the scatter plot data D3 of FIG. 8, the progress of the ionization and sulfurization reactions in sample 10 before and after moist heat aging can be understood. The faster the ionization and sulfurization reactions in sample 10 progress, the earlier the adhesiveness between the vulcanized rubber and the metal decreases, and the slower the ionization and sulfurization reactions progress, the less likely the adhesiveness between the vulcanized rubber and the metal decreases. The progress of the ionization and sulfurization reactions can be understood by the decrease in the mass of the zero-valent copper component (Cu) and the decrease in the mass of the monovalent and divalent copper components (Cu + , Cu 2+The increase in mass of the zero-valent copper component (Cu) is large, and the decrease in mass of the monovalent and divalent copper components (Cu + , Cu 2+ The greater the increase in mass of each copper component (Cu, Cu) the faster the sulfurization reaction proceeds. + , Cu 2+ ) is diffused, the faster the sulfurization reaction proceeds. Therefore, from the scatter plot data D3 in Fig. 8, it can be seen that in the measured sample 10, the adhesion between the vulcanized rubber and the metal decreases as the ionization and sulfurization reactions proceed due to moist heat aging.
[0043] As described above, according to this embodiment, each characteristic point plotted in the scatter diagram data D3 represents the total mass and number of particles of each component as the state of particles of components with different valences in each granular body before and after the moist heat aging treatment. That is, by comparing each characteristic point of each different component in the scatter diagram data D3 before and after the moist heat aging treatment, it is possible to grasp the progress of ionization of each component in the sample 10 and the progress of the sulfurization reaction with the rubber sulfur. Therefore, by creating the scatter diagram data D3, the change in the adhesion between the vulcanized rubber and the metal can be visually grasped more clearly, and the moist heat adhesion of the metal-rubber composite material can be accurately evaluated.
[0044] The scatter diagram data D3a shown in Fig. 9 was created by the above-mentioned evaluation method using a sample 10a with a different composition from that of the previous embodiment. Comparing the scatter diagram data D3a in Fig. 9 with the scatter diagram data D3 in Fig. 8, in the scatter diagram data D3a in Fig. 9, the mass of the zero-valent copper component (Cu) is reduced to the same extent by the moist heat aging treatment, but it is more narrowly diffused (denser), and the monovalent copper component (Cu +) is more widely diffused, but the increase in mass is significantly smaller. Therefore, it can be seen from the scatter plot data D3a in FIG. 9 that the ionization and vulcanization reactions of the measured sample 10a proceed more slowly than those of the sample 10 of the embodiment, and the decrease in adhesion between the vulcanized rubber and the metal is suppressed. Even if the composition of the sample 10 is the same as that of the previous embodiment, if the sample 10 is manufactured under different vulcanization conditions, scatter plot data different from the scatter plot data D3 of the previous embodiment can be obtained.
[0045] Therefore, many kinds of vulcanized rubbers with different unvulcanized rubber compounding and / or vulcanization conditions are used as samples 10, and scatter diagram data D3 for each sample 10 is obtained, and the relationship between compounding and / or vulcanization conditions and the degree of change in adhesiveness is understood based on each scatter diagram data D3. In this way, by accumulating evaluation results using the evaluation method of this embodiment, the degree of progress of ionization and vulcanization reaction of metal components can be used as indicators to evaluate the adhesiveness between vulcanized rubber and metal. This makes it easier to find the compounding of unvulcanized rubber and vulcanization conditions that can improve the adhesiveness between vulcanized rubber and metal under moist heat, which greatly contributes to improving the quality of rubber products such as tires, hoses, and conveyor belts that use metal materials.
[0046] The present invention is not limited to a specific embodiment, and various modifications and variations are possible within the scope of the gist of the present invention.
[0047] The present disclosure includes the following inventions. Invention (1): A method for evaluating adhesion between rubber and metal, comprising the steps of: using a vulcanized rubber sample obtained by vulcanizing unvulcanized rubber containing a metal; obtaining 3D image data of a plurality of granules having metal components originating from the metal present in the sample by 3D spatially resolved X-ray absorption fine structure analysis; and utilizing the obtained 3D image data, For each of the granular bodies, metal compounds having different metal components and metal compounds having the same type of metal component but different valences are treated as particles of different components, an initial measurement step using the sample and a post-treatment measurement step using the sample that has been subjected to moist heat aging treatment after the initial measurement step are carried out, and in each of the measurement steps, the three-dimensional image data is acquired, and using the acquired three-dimensional image data, a detection intensity and a detection frequency of each of the particles of the different components present in the same granular body are obtained as measurement results; By using machine learning to process the measurement results for each of the granular bodies in each of the measurement steps with a computing device, scatter plot data is created in which characteristic points showing characteristics of each of the particles of different components in each of the same granular bodies before and after the moist heat aging treatment are plotted; A method for evaluating adhesion between rubber and metal, comprising the steps of: evaluating adhesion between the vulcanized rubber and the metal using the scatter diagram data thus created. Invention (2): A method for evaluating adhesion between rubber and metal as described in invention (1), in which a large number of types of vulcanized rubber having different formulations and / or vulcanization conditions for the unvulcanized rubber are used as the samples, the scatter plot data for each of the samples is obtained, and the relationship between the formulations and / or vulcanization conditions and the degree of change in adhesion is understood based on each of the scatter plot data. Invention (3): A method for evaluating adhesion between rubber and metal as described in invention (1) or (2), in which the scatter plot data is created by using the machine learning to process the measurement results for each of the granules in each of the measurement steps using the computing device, thereby extracting feature amounts indicating the characteristics of each of the different components for each of the granules before and after the moist heat aging treatment for a large number of types of the features, and using the principal component scores obtained by performing dimensionality reduction by principal component analysis for each of the extracted features as the feature points. Invention (4): A method for evaluating adhesion between rubber and metal according to any one of Inventions (1) to (3), in which the vulcanized rubber containing 0.5% by mass or more and 10% by mass or less of metal powder having a particle diameter of 1 μm or more and 100 μm or less is used as the sample. [Explanation of symbols]
[0048] 1. Rating System 2. Three-dimensional spatially resolved X-ray absorption fine structure imaging analyzer 3 Computing device 10 Samples 11 Tube D1, D2 3D image data D3 Scatter Plot Data D4, D5 Measurement result data D6, D7 feature data
Claims
1. A method for evaluating adhesion between rubber and metal, comprising the steps of: using a vulcanized rubber sample obtained by vulcanizing unvulcanized rubber containing a metal; obtaining three-dimensional image data of a plurality of particles having metal components originating from the metal present in the sample by three-dimensional spatially resolved X-ray absorption fine structure analysis; and utilizing the obtained three-dimensional image data, For each of the granular bodies, metal compounds having different metal components and metal compounds having the same type of metal component but different valences are treated as particles of different components, an initial measurement step using the sample and a post-treatment measurement step using the sample that has been subjected to moist heat aging treatment after the initial measurement step are carried out, and in each of the measurement steps, the three-dimensional image data is acquired, and the obtained three-dimensional image data is used to obtain, as measurement results, detection intensities and detection frequencies of the particles of the different components present in the same granular body, By using machine learning to process the measurement results for each of the granular bodies in each of the measurement steps with a computing device, scatter plot data is created in which characteristic points showing characteristics of each of the particles of different components in each of the same granular bodies before and after the moist heat aging treatment are plotted; A method for evaluating adhesion between rubber and metal, comprising the steps of: evaluating adhesion between the vulcanized rubber and the metal using the scatter diagram data thus created.
2. A method for evaluating adhesion between rubber and metal as described in claim 1, in which a number of types of vulcanized rubber having different formulations and / or vulcanization conditions of the unvulcanized rubber are used as the samples, the scatter plot data for each of the samples is obtained, and the relationship between the formulation and / or the vulcanization conditions and the degree of change in adhesion is understood based on each of the scatter plot data.
3. A method for evaluating adhesion between rubber and metal as described in claim 1 or 2, in which the scatter plot data is created by using the machine learning to process the measurement results for each of the granular materials in each of the measurement steps using the computing device, extracting feature amounts indicating the characteristics related to each of the different components for each of the granular materials before and after the moist heat aging treatment for a large number of types of the features, and using principal component scores obtained by performing dimensionality reduction by principal component analysis for each of the extracted features as the feature points.
4. 3. The method for evaluating adhesion between rubber and metal according to claim 1 or 2, wherein the vulcanized rubber contains 0.5% by mass or more and 10% by mass or less of metal powder having a particle diameter of 1 μm or more and 100 μm or less as the sample.
Citation Information
Patent Citations
Rubber composition assessment method
JP2022188741A
