Method for evaluating solid electrolyte

Abstract

To provide a method for evaluating a solid electrolyte which is low in calculation cost and can evaluate ion conductivity with high accuracy.SOLUTION: A method for evaluating a solid electrolyte causes a computer to execute processing including the following steps: (a) acquiring information on a basic crystal structure of a target solid electrolyte from a structure database; (b) generating information on a mobile-ion-removed crystal structure obtained by removing mobile ions moving in the solid electrolyte from the basic crystal structure; (c) calculating a route where the mobile ions can move from one side of a pair of surfaces of the mobile-ion-removed crystal structure to the other side thereof, on the basis of a geometric relation between a dimension of a gap of the mobile-ion-removal crystal structure and a dimension of the mobile ion; and (d) calculating, in a machine learning force field, potential energy of the mobile ions on the route.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present disclosure relates to a method for evaluating a solid electrolyte.

Background Art

[0002] In recent years, solid-state batteries having a solid electrolyte as an electrolyte have attracted attention. In order to improve the performance of a solid-state battery, it is effective to improve the ionic conductivity in the solid electrolyte. Therefore, methods for evaluating the ionic conductivity in a solid electrolyte have been developed.

[0003] For example, Patent Document 1 discloses a simulation of ionic conductivity in which, when performing molecular dynamics calculations to understand the behavior of ions in a solid electrolyte layer, specific atoms other than the ions contained in the solid electrolyte layer are constrained and an electric field is applied.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] An object of the present disclosure is to provide a method for evaluating a solid electrolyte that can evaluate ionic conductivity with low computational cost and high accuracy.

Means for Solving the Problems

[0006] The present inventors have found that the above problems can be solved by the following means. <Aspect 1> A method for evaluating a solid electrolyte, comprising causing a computer to execute a process including the following steps: (a) obtaining information on the basic crystal structure of a target solid electrolyte from a structure database, (b) Generating information on a mobile ion removal crystal structure obtained by removing mobile ions moving in the solid electrolyte from the basic crystal structure; (c) Calculating, based on the geometric relationship between the dimensions of the voids in the mobile ion removal crystal structure and the dimensions of the mobile ions, a path through which the mobile ions can move from one side to the other side of a pair of surfaces of the mobile ion removal crystal structure; and (d) Calculating the potential energy of the mobile ions on the path using a machine learning force field. <Aspect 2> The method according to Aspect 1, wherein the machine learning force field is one obtained by training the results of first-principles calculations. [Advantages of the Invention]

[0007] According to the present disclosure, it is possible to provide a method for evaluating a solid electrolyte that can evaluate ion conductivity with low computational cost and high accuracy. [Brief Description of the Drawings]

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

[0009] Hereinafter, embodiments of the present disclosure will be described in detail. Note that the present disclosure is not limited to the following embodiments and can be variously modified and implemented within the scope of the gist of the disclosure.

[0010] <Method for Evaluating Solid Electrolyte> The method of the present disclosure for evaluating a solid electrolyte causes a computer to execute a process including the following steps: (a) obtaining information on the basic crystal structure of a target solid electrolyte from a structure database; (b) generating information on a mobile-ion-removed crystal structure obtained by removing mobile ions moving in the solid electrolyte from the basic crystal structure; (c) calculating a path through which mobile ions can move from one side to the other side of a pair of surfaces of the mobile-ion-removed crystal structure based on the geometric relationship between the dimensions of the voids of the mobile-ion-removed crystal structure and the dimensions of the mobile ions; and (d) calculating the potential energy of the mobile ions on the path using a machine learning force field.

[0011] As a method for evaluating a solid electrolyte, a method of calculating using a force field that represents necessary parameters by an empirical formula, that is, a classical force field, can be considered.

[0012] Examples of the method of calculating using a classical force field include a method of causing a computer to execute the following process as illustrated in FIG. 4. That is, a process of obtaining information on the basic crystal structure of a target solid electrolyte from a structure database (see FIG. 4(a)), a process of generating information on a mobile-ion-removed crystal structure obtained by removing mobile ions moving in the solid electrolyte from the basic crystal structure (see FIG. 4(b)), an optional process of dividing the mobile-ion-removed crystal structure into microgrids (meshes) (see FIG. 4(c)), and a process of arranging diffusing ions in the mobile-ion-removed crystal structure or the optionally formed microgrids and calculating the potential energy of the mobile ions by classical force field calculation (see FIG. 4(d)).

[0013] According to the study by the present inventors, when evaluating a solid electrolyte by classical force field calculation, although the calculation cost was low, the accuracy of the calculation results was poor. That is, the calculation results deviated significantly from the experimental results evaluating the ionic conductivity of the mobile ions.

[0014] In contrast, a method of evaluating a solid electrolyte using a more accurate force field, that is, a machine learning force field having a huge number of non-empirical parameters, can be considered.

[0015] As a method for evaluating a solid electrolyte using a machine learning force field, for example, there is a method of causing a computer to execute the following processes. That is, a process of acquiring information on the basic crystal structure of a target solid electrolyte from a structure database, a process of generating information on a crystal structure with mobile ions removed by removing mobile ions moving in the solid electrolyte from the basic crystal structure, an optional process of dividing the crystal structure with mobile ions removed into unit cells, and a process of arranging diffusing ions in the crystal structure with mobile ions removed or in the optionally formed unit cells and causing a computer to calculate the potential energy of the mobile ions by machine learning force field calculation. Note that, except for the calculation method for calculating the potential energy of the mobile ions in this method, it is the same as the method for evaluating a solid electrolyte using the above classical force field. Therefore, for this method as well, FIG. 4 can be referred to.

[0016] According to the study by the present inventors, when evaluating a solid electrolyte by machine learning force field calculation for all voids in the crystal structure, the deviation from the experimental results evaluating the ionic conductivity of mobile ions is reduced, and thus the accuracy of the calculation results is improved, but the calculation cost increases.

[0017] Regarding these, the present inventors have found that, in the crystal structure, by geometrically calculating the path along which mobile ions can move and causing a computer to perform machine learning force field calculation on that path, the ionic conductivity of mobile ions can be evaluated with low calculation cost and high accuracy.

[0018] Hereinafter, the method of the present disclosure for evaluating a solid electrolyte will be described with reference to the flowchart shown in FIG. 1 and the schematic diagram shown in FIG. 2.

[0019] As shown in FIG. 1, in step S101, the computer executes a process of acquiring information on the basic crystal structure of the target solid electrolyte from the structure database (see FIG. 2(a)). Examples of the process of acquiring information include a process of downloading information stored on a network or in the cloud to a terminal such as a personal computer. As the structure database, for example, public databases such as THE Materials Project and QQMD can be used.

[0020] In step S102, the computer executes a process of generating information on the mobile-ion-removed crystal structure obtained by removing mobile ions moving in the solid electrolyte from the basic crystal structure (see FIG. 2(b)). For example, when the solid electrolyte is used in a lithium-ion battery, the mobile ions may be lithium ions.

[0021] In step S103, the computer executes a process of calculating a path through which mobile ions can move from one side to the other side of a pair of surfaces of the mobile-ion-removed crystal structure based on the geometric relationship between the dimensions of the voids in the mobile-ion-removed crystal structure and the dimensions of the mobile ions.

[0022] Specifically, for example, the crystal structure is divided into micro-lattices (see FIG. 2(c)), the positions of the voids are calculated for each micro-lattice, and a path through which the voids are connected from one side to the other side of a pair of surfaces of the mobile-ion-removed crystal structure can be calculated based on the above geometric relationship (see FIG. 2(d)).

[0023] In the method of the present disclosure, in order to calculate the potential energy of the mobile ions described later, a calculation using a machine learning force field considering electrochemical actions in the crystal structure, that is, for example, Coulomb force, is executed. On the other hand, in order to calculate a path through which mobile ions can move, as described above, the calculation is executed without considering such electrical actions. Therefore, the calculation cost is significantly reduced compared to the case where all voids in the crystal structure are calculated by the machine learning force field.

[0024] As shown in FIG. 3(b), a pair of planes of the mobile ion removal crystal structure may be, for example, the ab plane. In this case, the path along which the mobile ions can move may be a path extending in the c-axis direction from one ab plane of the crystal structure toward the other ab plane facing it.

[0025] In step S104, the computer executes a process of calculating the potential energy of the mobile ions on the path along which the mobile ions can move by using a machine learning force field (see FIG. 2(e)). That is, in step S104, instead of calculating for all the voids in the crystal structure as shown in FIG. 3(a) by using the machine learning force field, for the specific path calculated in step S103 as shown in FIG. 3(b), a calculation considering the electrochemical action in the crystal structure is executed by using the machine learning force field. The ionic conductivity of the solid electrolyte can be evaluated based on the magnitude of the potential energy of the mobile ions calculated in step S104. That is, if the above potential energy is small, it can be evaluated that the ionic conductivity of the solid electrolyte is high.

[0026] In the method of the present disclosure, the machine learning force field may be one obtained by learning the results of first-principles calculations. According to first-principles calculations, the ionic conductivity of the solid electrolyte can be evaluated with higher accuracy, but the calculation cost becomes enormous. According to the machine learning force field obtained by learning the results of first-principles calculations, the calculation cost can be significantly reduced without greatly degrading the evaluation accuracy.

Claims

1. A method for evaluating a solid electrolyte, which causes a computer to execute a process including the following steps: (a) Obtaining information on the basic crystal structure of a target solid electrolyte from a structure database; (b) Generating information on a mobile-ion-removed crystal structure in which mobile ions moving in the solid electrolyte are removed from the basic crystal structure; (c) Calculating a path along which the mobile ions can move from one side to the other side of a pair of surfaces of the mobile-ion-removed crystal structure based on a geometric relationship between the dimensions of voids in the mobile-ion-removed crystal structure and the dimensions of the mobile ions; and (d) Calculating the potential energy of the mobile ions on the path using a machine learning force field.

2. The method according to claim 1, wherein the machine learning force field is one obtained by learning the results of first-principles calculations.

Images