Structural analysis device

Abstract

To provide a structural analysis device capable of rapidly generating highly reliable crystal structure estimation models. [Solution] The structural analysis apparatus according to this disclosure comprises: an acquisition unit that acquires measurement data which is the result of XRD analysis of the crystal structure to be analyzed; a simulation unit that performs an XRD analysis simulation on a crystal structure estimation model and generates simulation data corresponding to the measurement data; an evaluation unit that calculates an R value which is an index representing the degree of agreement between the measurement data and the simulation data, and also calculates the energy value of the crystal structure estimation model; a model estimation unit that performs optimization of the crystal structure estimation model based on the R value and energy value fed back from the evaluation unit and supplies the optimized crystal structure estimation model to the simulation unit; and an output unit that outputs one or more crystal structure estimation models.

Description

【Technical Field】 , , 【0005】 【0001】 The present disclosure relates to a structure analysis device. 【Background Art】 【0002】 For a Rietveld analysis device, it is required to rapidly generate an estimation model of the crystal structure of an analysis target with high reliability. For example, Non-Patent Document 1 discloses a technique for automating Rietveld analysis. 【Prior Art Documents】 【Non-Patent Documents】 【0003】 【Non-Patent Document 1】 Yoshihiko Ozaki et al., "Automated crystal structure analysis based on blackbox optimization", npj Computational Materials (2020) 6:75 【Summary of the Invention】 【Problems to be Solved by the Invention】 【0004】 In Non-Patent Document 1, the optimization of the estimation model is performed so that the R value, which is an index representing the degree of coincidence between the measurement data, which is the XRD analysis result for the crystal structure of the analysis target, and the simulation data, which is the result of performing XRD analysis simulation on the crystal structure estimation model, shows the minimum value. However, there is a problem that only minimizing the R value may generate an estimation model of a structure that is energetically unstable (a structure with high energy), and a highly reliable estimation model still cannot be generated. 【0005】 The present disclosure has been made in view of the above background, and an object thereof is to provide a structure analysis device capable of rapidly generating a highly reliable crystal structure estimation model. 【Means for Solving the Problems】 【0006】 The structural analysis apparatus according to this disclosure includes: an acquisition unit that acquires measurement data which is the result of X-ray Diffraction (XRD) analysis of the crystal structure to be analyzed; a simulation unit that performs an XRD analysis simulation on the estimated model of the crystal structure and generates simulation data corresponding to the measurement data; an evaluation unit that calculates an R value, which is an index representing the degree of agreement between the measurement data and the simulation data, and also calculates the energy value of the estimated model of the crystal structure; a model estimation unit that optimizes the estimated model of the crystal structure based on the R value and the energy value fed back from the evaluation unit and supplies the optimized estimated model of the crystal structure to the simulation unit; and an output unit that outputs one or more estimated models of the crystal structure. In this way, the structural analysis apparatus according to this disclosure can quickly generate a highly reliable estimated model of the crystal structure by automatically optimizing the estimated model not only so that the R value becomes small, but also so that the energy of the crystal structure becomes small. [Effects of the Invention] 【0007】 This disclosure provides a structural analysis device capable of rapidly generating highly reliable crystal structure estimation models. [Brief explanation of the drawing] 【0008】 [Figure 1] This figure shows an example of the configuration of the structural analysis device described herein. [Figure 2] This is a conceptual diagram illustrating the processing flow of the structural analysis device related to this disclosure. [Figure 3] This is a flowchart showing the operation of the structural analysis device described in this disclosure. [Figure 4] This figure shows an example of the analysis results of the structural analysis device related to this disclosure. [Modes for carrying out the invention] 【0009】 The following describes specific embodiments to which the present invention is applied, with reference to the drawings. However, the present invention is not limited to the following embodiments. Also, for clarity of explanation, the following description and drawings have been simplified as appropriate. 【0010】 <Configuration of structural analysis device 100> Figure 1 shows an example of the configuration of the structural analysis apparatus 100 according to this disclosure. Figure 2 is a conceptual diagram illustrating the processing flow of the structural analysis apparatus 100 according to this disclosure. The structural analysis apparatus 100 according to this disclosure is a so-called Rietveld analyzer, which analyzes an analysis target T, such as a material, and generates an estimated model of the crystal structure of the analysis target T. Here, the structural analysis apparatus 100 according to this disclosure can quickly generate a highly reliable estimated model of the crystal structure by automatically optimizing the estimated model not only so that the R value, an index indicating the degree of agreement between the crystal structure of the analysis target T and the estimated model, becomes small, but also so that the energy of the crystal structure becomes small. A detailed explanation follows below. 【0011】 The structural analysis device 100 shown in Figure 1 comprises an acquisition unit 101, a simulation unit 102, an evaluation unit 103, a model estimation unit 104, and an output unit 105. 【0012】 The acquisition unit 101 acquires measurement data 201, which is the result of X-ray Diffraction (XRD) analysis of the crystal structure of the object T being analyzed. 【0013】 The simulation unit 102 performs an XRD analysis simulation on the initially set crystal structure estimation model 202 to generate simulation data 203 corresponding to the measurement data 201. 【0014】 The evaluation unit 103 compares the measurement data 201 with the simulation data 203 and calculates the R value Rwp, which is an index representing the degree of agreement between the measurement data 201 and the simulation data 203. The evaluation unit 103 also calculates the energy value Ef of the estimated crystal structure model 202 from the simulation data 203, for example using Density Functional Theory (DFT). 【0015】 Typically, the degree of agreement between measured data 201 and simulation data 203 increases as the R value Rwp decreases and decreases as the R value Rwp increases. Furthermore, crystal structures become more stable as the energy value Ef decreases and more unstable as the energy value Ef increases. In other words, crystal structures become more physically plausible as the energy value Ef decreases. Therefore, crystal structures tend to settle at structures with low energy values ​​Ef. Consequently, the crystal structure estimation model 202 is more likely to approach the crystal structure of the object being analyzed as both the R value Rwp and the energy value Ef decrease. In other words, the crystal structure estimation model 202 has higher reliability as both the R value Rwp and the energy value Ef decrease. 【0016】 The model estimation unit 104 optimizes the estimation model 202 used to calculate the R value Rwp and energy value Ef based on the R value Rwp and energy value Ef fed back from the evaluation unit 103, and generates an optimized crystal structure estimation model 202. 【0017】 For example, the model estimation unit 104 performs optimization on the estimation model 202 used for calculating the R value Rwp and the energy value Ef, expecting that the R value Rwp and the energy value Ef will be further reduced, and generates an estimation model 202 of the optimized crystal structure. More specifically, the model estimation unit 104 predicts model parameters such that the R value Rwp and the energy value Ef are further reduced from a plurality of combinations of the estimation model 202 of the crystal structure and the corresponding R value Rwp and energy value Ef, and generates an estimation model 202 of the crystal structure optimized according to the predicted model parameters. The estimation model 202 of the crystal structure optimized by the model estimation unit 104 is supplied to the simulation unit 102 instead of the initially set estimation model 202 of the crystal structure. 【0018】 Note that the model estimation unit 104 may be configured to perform machine learning using a plurality of combinations of the estimation model 202 of the crystal structure and the corresponding R value Rwp and energy value Ef. In this case, the model estimation unit 104 can perform optimization on the estimation model 202 used for calculating the R value Rwp and the energy value Ef, expecting that the R value Rwp and the energy value Ef will be further reduced, by using the learned model generated by machine learning. More specifically, the model estimation unit 104 can predict model parameters such that the R value Rwp and the energy value Ef are further reduced by using the learned model generated by machine learning, and generate an estimation model 202 of the crystal structure optimized according to the predicted model parameters. 【0019】 However, the model estimation unit 104 is not limited to performing machine learning using a combination of the estimation model 202 of the crystal structure and the corresponding R value Rwp and energy value Ef, and may perform machine learning using a combination of the estimation model 202 of the crystal structure and the corresponding R value Rwp, or may perform machine learning using a combination of the estimation model 202 of the crystal structure and the corresponding energy value Ef. 【0020】 The simulation unit 102 performs a simulation of XRD analysis on the estimation model 202 of the optimized crystal structure to regenerate simulation data 203 corresponding to the measurement data 201. The evaluation unit 103 compares the measurement data 201 with the regenerated simulation data 203 and calculates an R value Rwp, which is an index representing the degree of agreement between the measurement data 201 and the regenerated simulation data 203. Further, the evaluation unit 103 calculates an energy value Ef that the estimation model 202 of the crystal structure has from the regenerated simulation data 203. The model estimation unit 104 optimizes the estimation model 202 of the crystal structure based on the R value Rwp and the energy value Ef fed back from the evaluation unit 103, and generates an estimation model 202 of the optimized crystal structure. 【0021】 The optimization of the estimation model by the model estimation unit 104, the simulation of the estimation model by the simulation unit 102, and the evaluation between the measurement data 201 and the simulation data 203 by the evaluation unit 103 are repeated until the R value Rwp and the energy value Ef converge. 【0022】 The output unit 105 outputs information regarding a plurality of estimation models 202 of the optimized crystal structure. Alternatively, the output unit 105 may output information regarding an estimation model 202 in which each of the R value Rwp and the energy value Ef shows a value as low as possible among the plurality of estimation models 202 of the optimized crystal structure. The output content of the output unit 105 is, for example, displayed on a monitor. 【0023】 Note that the output unit 105 causes a monitor to display a plurality of estimation models 202 of the optimized crystal structure (or their R value Rwp and energy value Ef), and may cause the estimation model 202 (or its R value Rwp and energy value Ef) in which each of the R value Rwp and the energy value Ef shows a value as low as possible among the plurality of estimation models 202 of the optimized crystal structure to be highlighted on the monitor. 【0024】 Thus, the structural analysis device 100 according to this disclosure can quickly generate a highly reliable crystal structure estimation model by automatically optimizing the estimation model not only so that the R value, which is an index indicating the degree of agreement between the crystal structure of the object T to be analyzed and the estimation model, becomes smaller, but also so that the energy of the crystal structure becomes smaller. 【0025】 <Operation of structural analysis device 100> Figure 5 is a flowchart showing the operation of the structural analysis device 100. 【0026】 First, the structural analysis device 100 acquires measurement data 201, which is the result of XRD analysis of the crystal structure of the object to be analyzed T (step S101). 【0027】 Subsequently, the structural analysis device 100 performs an XRD analysis simulation on the initially set crystal structure estimation model 202 to generate simulation data 203 corresponding to the measurement data 201 (step S102). 【0028】 Subsequently, the structural analysis device 100 compares the measured data 201 with the simulation data 203 and calculates the R value Rwp, which is an index representing the degree of agreement between the measured data 201 and the simulation data 203 (step S103). The structural analysis device 100 also calculates the energy value Ef of the estimated crystal structure model 202 from the simulation data 203, for example using DFT (step S103). 【0029】 Subsequently, the structural analysis device 100 outputs the estimated crystal structure model 202 used to calculate the R value Rwp and the energy value Ef (step S104). 【0030】 Subsequently, if the R-value Rwp and the energy value Ef have not converged (NO in step S105), the structural analysis device 100 optimizes the estimation model 202 used to calculate the R-value Rwp and the energy value Ef based on the feedbacked R-value Rwp and the energy value Ef, and generates an optimized crystal structure estimation model 202 (step S106). More specifically, the structural analysis device 100 predicts model parameters that further reduce the R-value Rwp and the energy value Ef from multiple combinations of the crystal structure estimation model 202 and its corresponding R-value Rwp and energy value Ef, and generates an optimized crystal structure estimation model 202 according to the predicted model parameters (step S106). 【0031】 Subsequently, the structural analysis device 100 performs an XRD analysis simulation on the optimized crystal structure estimation model 202 to regenerate simulation data 203 corresponding to the measurement data 201 (step S107). 【0032】 Subsequently, the structural analysis device 100 compares the measured data 201 with the regenerated simulation data 203 and calculates the R value Rwp, which is an index representing the degree of agreement between the measured data 201 and the regenerated simulation data 203 (step S103). The structural analysis device 100 also calculates the energy value Ef of the estimated crystal structure model 202 from the regenerated simulation data 203 (step S103). 【0033】 Subsequently, the structural analysis device 100 outputs the estimated crystal structure model 202 used to calculate the R value Rwp and the energy value Ef (step S104). 【0034】 The structural analysis device 100 repeatedly optimizes the estimation model 202 until the R value Rwp and the energy value Ef converge. When the R value Rwp and the energy value Ef converge and stop decreasing (YES in step S105), the process is terminated. 【0035】 In this disclosure, the structural analysis device 100 has described, as an example, the case in which it terminates the structural analysis when the R value Rwp and the energy value Ef converge and no longer decrease even after repeated structural analyses, but it is not limited to this. For example, the structural analysis device 100 may terminate the structural analysis when the number of estimated models generated reaches a predetermined number. 【0036】 Figure 4 shows an example of the analysis results of the structural analysis device 100. The analysis results of the structural analysis device 100, as shown in Figure 4, are displayed on a monitor, for example. In the example in Figure 4, a planar coordinate system is shown in which the horizontal axis represents the R value Rwp of the estimated model 202 and the vertical axis represents the energy value Ef of the estimated model 202. 【0037】 In the example shown in Figure 4, several plots located at the outer edge of the plot group (boundary plots) among the plots of multiple estimated models 202 are highlighted as plots of estimated model 202 with high reliability. Estimated models 202 represented by multiple boundary plots have either an extremely low R value Rwp, an extremely low energy value Ef, or both an extremely low R value Rwp and an extremely low energy value Ef overall. For example, a user of the structural analysis device 100 can appropriately select one of the estimated models 202 of the highlighted boundary plots and adopt it as a candidate for the crystal structure of the object T to be analyzed. 【0038】 Thus, the structural analysis device 100 according to this disclosure can quickly generate a highly reliable crystal structure estimation model by automatically optimizing the estimation model not only so that the R value, which is an index indicating the degree of agreement between the crystal structure of the object T to be analyzed and the estimation model, becomes smaller, but also so that the energy of the crystal structure becomes smaller. 【0039】 Furthermore, this disclosure can be realized by having a CPU (Central Processing Unit) execute a computer program to perform part or all of the processing of the structural analysis device 100. 【0040】 The program described above includes, when loaded into a computer, a set of instructions (or software code) for causing the computer to perform one or more of the functions described in the embodiments. The program may be stored in a non-temporary computer-readable medium or a physical storage medium. Examples, but not limited to, include RAM (Random-Access Memory), ROM (Read-Only Memory), flash memory, SSD (Solid-State Drive), or other memory technologies, CD-ROM, DVD (Digital Versatile Disc), Blu-ray® disc, or other optical disc storage, magnetic cassette, magnetic tape, magnetic disk storage, or other magnetic storage devices. The program may be transmitted over a temporary computer-readable medium or a communication medium. Examples, but not limited to, include temporary computer-readable medium or a communication medium that includes electrically, optically, acoustically, or otherwise propagating signals. 【0041】 Although the present disclosure has been described above with reference to embodiments, the present disclosure is not limited to the embodiments described above. Various modifications to the structure and details of the present disclosure can be made as can be understood by those skilled in the art within the scope of the present disclosure. Furthermore, each embodiment can be combined with other embodiments as appropriate. [Explanation of Symbols] 【0042】 100 Structural analysis equipment 101 Acquisition Department 102 Simulation Department 103 Evaluation Department 104 Model Estimation Unit 105 Output section 201 Measurement data to be analyzed 202 Estimation Models for Crystal Structure 203 Simulation Data

Claims

[Claim 1] An acquisition unit that acquires measurement data, which is the result of X-ray Diffraction (XRD) analysis of the crystal structure to be analyzed, A simulation unit that performs XRD analysis simulation on the estimated crystal structure model to generate simulation data corresponding to the measurement data, An evaluation unit calculates an R value, which is an index representing the degree of agreement between the measurement data and the simulation data, and also calculates the energy value of the estimated crystal structure model. A model estimation unit optimizes the crystal structure estimation model based on the R value and energy value fed back from the evaluation unit, and supplies the optimized crystal structure estimation model to the simulation unit. An output unit that outputs one or more estimated models of the aforementioned crystal structure, A structural analysis device equipped with the following features. [Claim 2] The model estimation unit performs machine learning using a combination of the estimated crystal structure model and the corresponding R value and energy value, and uses the trained model generated by the machine learning to optimize the estimated crystal structure model so that the R value and energy value become lower. The structural analysis apparatus according to claim 1. [Claim 3] The evaluation unit uses Density Functional Theory (DFT) to calculate the energy values ​​of the estimated crystal structure model. The structural analysis apparatus according to claim 1. [Claim 4] The output unit displays information regarding one or more estimated models of the crystal structure on a monitor. The structural analysis apparatus according to claim 1. [Claim 5] The output unit highlights, on the monitor, multiple plots located at the outer edge of the plot group, among multiple plots representing each of the multiple estimated models of the crystal structure, which are shown on a planar coordinate system where the horizontal axis represents the R value and the vertical axis represents the energy value. The structural analysis apparatus according to claim 1.

Images