Crystal analysis method, crystal analysis device, and crystal analysis program

By reducing the number of anionic atom nodes in ion crystal analysis and appropriately expressing the crystal structure, the calculation accuracy problem caused by the large number of unit cell atoms in the prior art is solved, and the efficiency and accuracy of similarity calculation are improved.

CN114743611BActive Publication Date: 2024-11-15FUJITSU LTD
View PDF 8 Cites 0 Cited by

Patent Information

Application Number
CN202111292863.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2020-12-23
Filing Date
2021-11-03
Publication Date
2024-11-15
Estimated Expiration
2041-11-03

AI Technical Summary

Technical Problem

When analyzing ionic crystals, the prior art has difficulty in obtaining accurate results when calculating similarity due to the large number of atoms in the unit cell, and the annealer has a limit on the number of positions, which is difficult to deal with.

Method used

By creating a chart, expand the repeating unit cell to adjacent repeating unit cell, and treat the anionic atoms bonded to it as a node in the ion crystal, reduce the number of anionic atom nodes, properly express the repeating structure of the crystal, use the extended node and cyclic structures to process unit cells.

Benefits of technology

It effectively reduces the number of bits required to analyze ionic crystals, improves the accuracy and efficiency of calculating similarity, and avoids the similarity decrease caused by differences in the number of atoms.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114743611B_ABST
    Figure CN114743611B_ABST
Patent Text Reader

Abstract

The present invention relates to a crystal analysis method, a crystal analysis device and a crystal analysis program. A crystal analysis method for executing the following process on a computer is disclosed, comprising: creating a chart indicating data of a repeating unit cell in an ionic crystal and data of adjacent repeating unit cells adjacent to the repeating unit cell; analyzing the ionic crystal according to the chart; when the number of nodes in the first unit cell indicating data of an anion atom bonded to a cation atom in the repeating unit cell is n, the number of nodes in the second unit cell indicating data of anion atoms in the repeating unit cell is set to be less than n-1, wherein the data of the repeating unit cell includes multiple nodes in the unit cell indicating data of atoms in the repeating unit cell, and the multiple nodes in the unit cell include a first node in the unit cell and a second node in the unit cell.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to a crystal analysis method, a crystal analysis device and a crystal analysis program. Background Art

[0002] In recent years, the use of computers to analyze materials has been used to design new materials, predict the properties of existing materials, and find new uses for existing materials.

[0003] More specifically, for example, when calculating the similarity between molecules and the similarity between crystals having a repetitive structure, a "conflict graph" showing the combination of atoms in the structure for which the similarity is calculated can be used. As a technique using a conflict graph, a technique for specifying similarity by searching for a substructure common to each structure by solving the maximum independent set problem of the conflict graph with an annealer or the like has been proposed.

[0004] Here, when determining the similarity between crystals having a repetitive structure, the number of atoms contained in each unit cell used for structural comparison may increase (the number of bits required to calculate the similarity may increase). In particular, for example, when the crystal is an ionic crystal, the unit cell contains a large number of anions (e.g., oxygen atoms), and the number of atoms contained in the unit cell increases.

[0005] Furthermore, when calculating similarity using an annealer, when the number of atoms included in a unit cell is large (a larger number of bits required for calculation), it may be difficult to obtain accurate similarity because the annealer has a limit on the number of bits available for calculation.

[0006] As a prior art, Maritza Hernandez, Arman Zaribafiyan, MalihehAramon, Mohammad Naghibi "A Novel Graph-based Approach for Determining Molecular Similarity". arXiv:1601.06693 (https: / / arxiv.org / pdf / 1601.06693.pdf) is disclosed. Summary of the invention

[0007] Technical issues

[0008] An object of the present application is to provide a crystal analysis method, a crystal analysis apparatus, and a crystal analysis program capable of reducing (suppressing) the number of bits required for analyzing ionic crystals.

[0009] Solution

[0010] According to one aspect of the embodiment, a crystal analysis method for a computer to perform the following process includes: creating a graph indicating data of a repeating unit cell in an ionic crystal and data of adjacent repeating units adjacent to the repeating unit cell; analyzing the ionic crystal based on the graph; and when the number of first unit cell intra-nodes indicating data of anion atoms bonded to cation atoms in a repeating unit cell is n, setting the number of second unit cell intra-nodes indicating data of anion atoms in a repeating unit cell to less than n-1, wherein the data of the repeating unit cell includes multiple unit cell intra-nodes indicating data of atoms in the repeating unit cell, and the multiple unit cell intra-nodes include the first unit cell intra-node and the second unit cell intra-node.

[0011] Beneficial Effects of the Invention

[0012] In one aspect, the present application may provide a crystal analysis method, a crystal analysis apparatus, and a crystal analysis program capable of reducing (suppressing) the number of bits required to analyze ionic crystals. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 is a diagram illustrating an example of graphically representing the states of acetic acid and methyl acetate;

[0014] Figure 2 is a diagram illustrating an example of combination in the case where the same elements in molecules A and B are combined and made into nodes of a conflict graph;

[0015] Figure 3 is a graph illustrating an instance of a rule for creating edges in a conflict graph;

[0016] Figure 4 is a diagram illustrating an example of a conflict graph for molecules A and B;

[0017] Figure 5 is a diagram illustrating an instance of a maximal independent set in a graph;

[0018] Figure 6 is a diagram illustrating an example of a process in the case of finding the largest common substructure between molecules A and B by finding the largest independent set of a conflict graph (solving the maximum independent set problem);

[0019] Figure 7 is an explanatory diagram for explaining an example of a technique for searching for a maximum independent set in a graph having 6 nodes;

[0020] Figure 8 is an explanatory diagram for explaining an example of a technique for searching for a maximum independent set in a graph having 6 nodes;

[0021] Fig. 9 is a diagram illustrating an instance of a maximum independent set in a conflict graph;

[0022] Fig.10 is a schematic diagram illustrating an example of a periodic structure of a crystal (crystalline material) having a repeating structure;

[0023] Fig.11 is a diagram illustrating an example of a unit cell in a technique for expanding and processing the unit cell using an expansion node;

[0024] Fig.12 is a diagram illustrating an example of a unit cell in a technique for treating a repeated structure of a crystal as a cyclic structure;

[0025] Fig.13A is a diagram illustrating an example of a diagram of a unit cell of a crystal;

[0026] Fig. 13B is a diagram illustrating another example of a diagram of a unit cell of a crystal;

[0027] Fig.14A is a diagram for explaining an example of a technique for multiplying unit cell A and unit cell B by integers;

[0028] Fig. 14B is a diagram for explaining an example of a technique for multiplying unit cell A and unit cell B by integers;

[0029] Fig. 14C is a diagram for explaining an example of a technique for multiplying unit cell A and unit cell B by integers;

[0030] Fig.15A is a diagram illustrating an example of a unit cell in an ionic crystal;

[0031] Fig. 15B is a diagram illustrating another example of a unit cell in an ionic crystal;

[0032] Fig.16 is a diagram illustrating a comparison between an example of the prior art and an embodiment of the present application;

[0033] Fig.17 is a diagram illustrating an example of the relationship between a combination of nodes in the first graph and the second graph and c'ij representing a weight with respect to a bond order between cation atoms;

[0034] Fig.18 is a diagram illustrating an example of a hardware configuration of a crystal analysis device disclosed in the present application;

[0035] Fig.19 is a diagram illustrating another hardware configuration example of the crystal analysis device disclosed in the present application;

[0036] Fig. 20 is a diagram illustrating an example of a functional configuration of a crystal analysis device disclosed in the present application;

[0037] Fig.21 is a diagram illustrating an example of a flow chart for analyzing an ionic crystal by using an example of the technology disclosed in the present application;

[0038] Fig. 22 is a diagram illustrating a portion of an example of a flowchart for creating a graph by a technique for treating a repetitive structure of a crystal as a cyclic structure by using an example of the technique disclosed in the present application;

[0039] Fig.23 is a diagram illustrating a portion of an example of a flowchart for creating a graph by a technique for treating a repetitive structure of a crystal as a cyclic structure by using an example of the technique disclosed in the present application;

[0040] Fig.24 is a diagram illustrating a portion of an example of a flowchart for creating a graph by a technique for treating a repetitive structure of a crystal as a cyclic structure by using an example of the technique disclosed in the present application;

[0041] Fig.25 is a diagram illustrating a portion of an example of a flowchart for creating a graph by a technique for treating a repetitive structure of a crystal as a cyclic structure by using an example of the technique disclosed in the present application;

[0042] Fig.26 is a diagram illustrating a portion of an example of a flowchart for creating a graph by a technique for treating a repetitive structure of a crystal as a cyclic structure by using an example of the technique disclosed in the present application;

[0043] Fig. 27 is a diagram of a portion of an example of a flowchart for creating a graph by using an example of the technology disclosed in the present application, by using a technique for expanding and processing a unit cell by using an expansion node;

[0044] Fig.28 is a diagram of a portion of an example of a flowchart for creating a graph by using an example of the technology disclosed in the present application, by using a technique for expanding and processing a unit cell by using an expansion node;

[0045] Fig.29 is a diagram of a portion of an example of a flowchart for creating a graph by using an example of the technology disclosed in the present application, by using a technique for expanding and processing a unit cell by using an expansion node;

[0046] Fig.30 is a diagram illustrating an example of a functional configuration of an annealing machine that performs an annealing method;

[0047] Fig.31 is a diagram illustrating an example of the operation flow of the conversion control unit 114;

[0048] Fig.32is a diagram illustrating exemplary summary results of Example 1 and Comparative Example 1;

[0049] Fig.33 is a diagram illustrating the results of Example 2;

[0050] Fig.34 is a diagram illustrating an example of the relationship between the type of ionic crystal and the calculated similarity in Example 3 and Comparative Example 2;

[0051] Fig.35 is a diagram illustrating an example of the relationship between the type of ionic crystal and the number of bits required to calculate the similarity in Example 3 and Comparative Example 2;

[0052] Fig.36 This is a diagram illustrating an example of the relationship between the type of ionic crystal and the calculated similarity in Example 4. DETAILED DESCRIPTION

[0053] (Crystal Analysis Method)

[0054] The technology disclosed in this application is based on the discovery of the inventors that, in the prior art, when analyzing an ionic crystal, when the number of atoms contained in the unit cell of the ionic crystal is large, the number of bits required for calculation increases, and it may be difficult to accurately analyze. Therefore, before describing the details of the technology disclosed in this application, the problems of the prior art, etc. will be described.

[0055] As described above, when performing an analysis of calculating the similarity between molecules and the similarity between crystals having a repetitive structure, for example, a technique using a "conflict graph" that shows the combination of atoms in the structure for which the similarity is calculated can be used. In this technique, for example, similarity can be obtained by searching for the maximum independent set of the conflict graph (solving the maximum independent set problem) to specify a substructure common to each structure.

[0056] Here, when calculating the structural similarity between compounds by solving the maximum independent set problem in the conflict graph, the compounds are represented as processed graphs. Here, representing the compounds as graphs means representing the structure of the compounds by using, for example, information about the types of atoms (elements) in the compounds and information about the bonding states between individual atoms.

[0057] The method of creating a conflict graph is described below by taking the case of first creating a conflict graph of acetic acid (CH3COOH) and methyl acetate (CH3COOCH3) as an example of obtaining the similarity between molecules.

[0058] First, acetic acid (hereinafter sometimes referred to as "molecule A") and methyl acetate (hereinafter sometimes referred to as "molecule B") are represented as a graph and given as follows. Figure 1 As shown. Figure 1In the formula (A), the atoms forming acetic acid are represented by A1, A2, A3 and A5, and the atoms forming methyl acetate are represented by B1 to B5. Figure 1 , A1, A2, B1, B2, B4 represent carbon, A3, A5, B3, B5 represent oxygen, a thin solid line represents a single bond, and a thick solid line represents a double bond.

[0059] Next, the vertices (atoms) in molecules A and B represented as graphs are combined to create vertices (nodes) of the conflict graph. At this point, for example, Figure 2 As shown, it is advantageous to combine the same elements in molecules A and B to create nodes of the conflict graph. Figure 2 In the example shown, a combination of A1, A2, B1, B2, and B4 representing carbon and a combination of A3, A5, B3, and B5 representing oxygen are used as nodes of the conflict graph.

[0060] Subsequently, edges (branches or sides) in the conflict graph are created. When two nodes are compared, an edge is created between the two nodes when the nodes are composed of atoms in different situations (e.g., atomic number, presence or absence of bonds, bond order, etc.). When two nodes are compared and the nodes are composed of atoms in the same situation, no edge is created between the two nodes.

[0061] Here, we will refer to Figure 3 Describes the rules for creating edges in the conflict graph.

[0062] First, in Figure 3 In the example shown, it will be explained whether to create an edge between node [A1B1] and node [A2B2]. Figure 3 From the structure of molecule A represented as a graph in FIG, it can be seen that carbon A1 of molecule A contained in node [A1B1] and carbon A2 of molecule A contained in node [A2B2] are bonded to each other (single bond). Similarly, carbon B1 of molecule B contained in node [A1B1] and carbon B2 of molecule B contained in node [A2B2] are bonded to each other (single bond). In other words, for example, the bonding condition between carbon A1 and A2 and the bonding condition between carbon B1 and B2 are the same as each other.

[0063] In this way, in Figure 3 In the example, the situation of carbon A1 and A2 in molecule A is the same as the situation of carbon B1 and B2 in molecule B, and nodes [A1B1] and [A2B2] are considered to be nodes composed of atoms in the same situation. Figure 3 In the example shown, no edge is created between nodes [A1B1] and [A2B2].

[0064] Next, in Figure 3 In the example shown, it will be explained whether to create an edge between node [A1B4] and node [A2B2]. Figure 3 From the structure of molecule A represented as a graph in , it can be seen that carbon A1 of molecule A contained in node [A1B4] and carbon A2 of molecule A contained in node [A2B2] are bonded to each other (single bond). However, from the structure of molecule B represented as a graph, it can be seen that carbon B4 of molecule B contained in node [A1B4] and carbon B2 of molecule B contained in node [A2B2] have oxygen B3 sandwiched between carbon B4 and B2, rather than being directly bonded. In other words, for example, the bonding situation between carbon A1 and A2 and the bonding situation between carbon B4 and B2 are different from each other.

[0065] That is, for example, in Figure 3 In the example, the situations of carbon A1 and A2 in molecule A and carbon B4 and B2 in molecule B are different from each other, and nodes [A1B4] and [A2B2] are considered to be nodes composed of atoms in different situations. Figure 3 In the example shown, an edge is created between nodes [A1B4] and [A2B2].

[0066] In this way, a conflict graph may be created based on the following rule: edges are created between nodes when they are composed of atoms in different cases, and edges are not created between nodes when they are composed of atoms in the same case.

[0067] Figure 4 is a diagram showing an example of a conflict graph for molecules A and B. Figure 4 As shown in , for example, in the nodes [A2B2] and [A5B5], the bonding situation between carbon A2 and oxygen A5 in molecule A and the bonding situation between carbon B2 and B5 in molecule B are the same as each other. Therefore, the nodes [A2B2] and [A5B5] are considered to be nodes composed of atoms in the same situation as each other, and therefore no edge is created between the nodes [A2B2] and [A5B5].

[0068] Next, an example of a method for solving the maximum independent set problem of the created conflict graph will be described.

[0069] A maximum independent set (MIS) in a conflict graph refers to a set that includes the largest number of nodes without edges between them among the set of nodes constituting the conflict graph. In other words, for example, a maximum independent set in a conflict graph refers to a set with the largest size (number of nodes) among the set of nodes that have no edges between them.

[0070] Figure 5 is a graph illustrating an example of a maximal independent set in a graph. Figure 5In the example, nodes included in the set are marked with reference numeral "1", and nodes not included in any set are marked with reference numeral "0"; if there is an edge between nodes, the nodes are connected by a solid line, and if there is no edge, the nodes are connected by a dotted line. Note that here, Figure 5 As shown, for simplicity of explanation, a graph with 6 nodes is taken as an example.

[0071] exist Figure 5 In the example shown, among the sets of nodes with no edges between them, there are three sets with the largest number of nodes, each of which has 3 nodes. In other words, for example, Figure 5 In the example shown, the three sets enclosed by a single dot-dashed line serve as the largest independent sets in the diagram.

[0072] Here, as described above, the conflict graph is created based on the following rule: when nodes are composed of atoms in different situations, edges are created between these nodes, and when nodes are composed of atoms in the same situation, no edges are created between these nodes. Therefore, in the conflict graph, calculating the maximum independent set as a set with the maximum number of nodes in a set composed of nodes with no edges between them is synonymous with calculating the maximum substructure in a substructure shared by two molecules. In other words, for example, the maximum common substructure of two molecules can be specified by calculating the maximum independent set in the conflict graph.

[0073] Figure 6 An example of a flow in the case of calculating the maximum common substructure of molecules A (acetic acid) and molecule B (methyl acetate) by calculating the maximum independent set in the conflict graph (solving the maximum independent set problem) is shown. Figure 6 As shown in FIG. 1 , a conflict graph is created in the following manner: molecules A and B are each represented as a graph, the same elements are combined and used as nodes, and edges are formed according to the atoms that constitute the nodes. Then, by calculating the maximum independent set in the created conflict graph, the maximum common substructure of molecules A and B can be calculated.

[0074] Here, an example of a specific method for calculating (searching for) a maximum independent set in a conflict graph will be described.

[0075] The maximum independent set in the conflict graph can be searched by, for example, using a Hamiltonian function, where minimization means searching for the maximum independent set. More specifically, for example, the search can be performed by using a Hamiltonian function (H) represented by the following calculation formula.

[0076] [Formula 1]

[0077]

[0078] Here, in the above calculation formula, n refers to the number of nodes in the conflict graph, and bi Refers to the numerical value representing the deviation of the i-th node.

[0079] In addition, when there is an edge between the i-th node and the j-th node, w ij is a non-zero positive number. When there is no edge between the i-th node and the j-th node, w ij is 0.

[0080] In addition, x i represents a binary variable, indicating that the i-th node is 0 or 1, x j represents a binary variable, indicating that the j-th node is 0 or 1.

[0081] Note that α and β are positive numbers.

[0082] The relationship between the Hamiltonian function represented by the above calculation formula and the search for the maximum independent set will be described in more detail. The above calculation formula is a Hamiltonian function that represents the Ising model equation in a quadratic unconstrained binary optimization (QUBO) format.

[0083] In the above formula, when x i When x is 1, it indicates that the i-th node is included in the set of candidates for the maximum independent set. i When x is 0, it means that the i-th node is not included in the set of candidates for the maximum independent set. j When x is 1, it means that the jth node is included in the set of candidates for the maximum independent set. j When it is 0, it means that the j-th node is not included in the set of candidates for the maximum independent set.

[0084] Therefore, in the above formula, by searching for a combination in which as many nodes as possible have state 1 under the constraint that there is no edge between nodes whose states are designated as 1 (bits are designated as 1), the maximum independent set can be searched.

[0085] Here, each term in the above calculation formula will be described.

[0086] The first term on the right side of the above equation (the term with a coefficient of -α) is the term whose value changes with x. i The term that becomes smaller as the number of i's that are 1 increases (the number of nodes included in the set that is a candidate for the maximum independent set increases). Note that the value of the first term on the right side of the above calculation formula becomes smaller, which means that a larger negative number is given. That is, for example, in the above calculation formula, due to the effect of the first term on the right side, when many nodes have bits that are 1, the value of the Hamiltonian function (H) becomes smaller.

[0087] The second term on the right side of the above calculation formula (the term with coefficient β) is the penalty term, and its value is when there is an edge between nodes with bit 1 (when w ijIn other words, for example, when there is no edge between nodes with bits set to 1, the second term on the right side of the above calculation formula is 0, and is a positive number in other cases. That is, for example, in the above calculation formula, due to the effect of the second term on the right side, when there is an edge between nodes with bits set to 1, the value of the Hamiltonian function (H) becomes larger.

[0088] As described above, when many nodes have bits set to 1, the value of the above calculation formula is smaller, and when there is an edge between nodes with bits set to 1, the value of the above calculation formula is larger; therefore, it can be said that minimizing the above calculation formula means searching for the maximum independent set.

[0089] Here, the relationship between the Hamiltonian function represented by the above calculation formula and the search for the maximum independent set will be explained using examples with reference to the drawings.

[0090] The following situation will be considered: In a graph with 6 nodes, Figure 7 Set the bits in each node as shown in the example. Figure 7 In the example in Figure 5 Similarly, for the case where there is an edge between the nodes, the nodes are connected with a solid line, and for the case where there is no edge, the nodes are connected with a dashed line.

[0091] for Figure 7 In the example, assume that in the above calculation formula, when there is an edge between the i-th node and the j-th node, b i is 1, w ij is 1, then the above calculation formula is as follows.

[0092] [Formula 2]

[0093] H=-α(x0+x1+x2+x3+x4+x5)

[0094] +β(λ 01 x0x1+λ 02 x0x2+λ 03 x0x3+λ 04 x0x4+λ 05 x0x5+…)

[0095] =-α(1+0+1+0+1+0)

[0096] +β(1*1*0+0*1*1+0*1*0+0*1*1+0*1*0+…)

[0097] =-3α

[0098] In this way, in Figure 7 In the example of , when there is no edge between nodes with bits set to 1 (when there is no contradiction as an independent set), the second term on the right side is 0, and the value of the first term is given as is as the value of the Hamiltonian function.

[0099] Next, we will consider Figure 8 The example shown in Figure 1 shows the case where bits are set in each node. Figure 7 Similar to the example in , assume that in the above calculation formula, when there is an edge between the i-th node and the j-th node, b i is 1, w ij is 1, then the above calculation formula is as follows.

[0100] [Formula 3]

[0101] H=-α(x0+x1+x2+x3+x4+x5)

[0102] +β(λ 01 x0x1+λ 02 x0x2+λ 03 x0x3+λ 04 x0x4+λ 05 x0x5+…)

[0103] =-α(1+ 1 +1+0+1+0)

[0104] +β(1*1* 1 +0*1*1+0*1*0+0*1*1+0*1*0+…)

[0105] =-4α+5β

[0106] In this way, in Figure 8 In the example of , since there are edges between nodes with bits set to 1, the second term on the right side is not 0, and the value of the Hamiltonian function is given as the sum of the two terms on the right side. Here, Figure 7 and Figure 8 In the example shown, for example, when it is assumed that α>5β, −3α<−4α+5β is satisfied, and therefore, Figure 7 The value of the Hamiltonian function in the example is Figure 8 The Hamiltonian function in the example is small. Figure 7 In the example, the set of non-contradictory nodes is obtained as the maximum independent set, and it can be seen that the maximum independent set can be obtained by searching for the combination of nodes whose Hamiltonian function value in the above calculation formula (1) becomes smaller.

[0107] Next, an example of a method for calculating the structural similarity between molecules based on the searched maximum independent set will be described.

[0108] The structural similarity between molecules can be calculated, for example, using the following calculation formula.

[0109] [Formula 4]

[0110]

[0111] Here, in the above similarity calculation formula, S(G A ,G B ) represents the similarity between a first molecule represented by a graph (e.g., molecule A) and a second molecule represented by a graph (e.g., molecule B), expressed as 0 to 1, meaning that the closer it is to 1, the higher the similarity.

[0112] In addition, V A represents the total number of node atoms of the first molecule represented in the graph, V c A The number of node atoms included in the maximum independent set of the conflict graph among the node atoms of the first molecule represented by the graph. Note that the node atoms refer to the atoms at the vertices of the molecule represented by the graph.

[0113] In addition, V B represents the total number of node atoms of the second molecule represented by the graph, V c B Indicates the number of node atoms of the second molecule represented by the graph that are included in the maximum independent set of the conflict graph. δ represents a number between 0 and 1.

[0114] In addition, in the above similarity calculation formula, max{A, B} means selecting a larger value from A and B, and min{A, B} means selecting a smaller value from A and B.

[0115] Here, similar to Figures 1 to 8 The example shown in FIG. 1 takes acetic acid (molecule A) and methyl acetate (molecule B) as examples to illustrate the method for calculating the similarity.

[0116] exist Fig. 9 In the conflict graph shown, the maximum independent set consists of four nodes: node [A1B1], node [A2B2], node [A3B3], and node [A5B5]. That is, for example, Fig. 9 In the example, |V A | is given as 4, |V c A | is given as 4, |V B | is given as 5, |V c B | is given as 4. In addition, in this example, when δ is assumed to be 0.5 and the average value of the first molecule and the second molecule is taken (equal treatment), the above calculation formula of the similarity is as follows.

[0117] [Formula 5]

[0118]

[0119] In this way, in Fig. 9In the example, based on the above similarity calculation formula, the similarity of the molecular structure can be calculated to be 0.9.

[0120] Although the technique for calculating the similarity between molecules has been described in detail above, a similar technique can also be appropriately used in the case of calculating the similarity between crystals having a repetitive structure.

[0121] Here, in the case of calculating the similarity between crystals, since the crystal has a repetitive structure (periodic structure, crystal structure) in which unit cells are repeated three-dimensionally, it is necessary to appropriately express this repetitive structure to calculate the similarity with high accuracy.

[0122] Fig.10 Schematic diagram illustrating an example of a periodic structure of a crystal (crystalline material) having a repeating structure. Fig.10 In the example shown, the crystal 10X has a structure in which a unit cell 10 is repeated, comprising nine distinguishable atoms (atoms A to I). Note that for ease of illustration, Fig.10 shows a structure in which atoms and unit cells are arranged two-dimensionally, but an actual crystal has a structure in which atoms and unit cells are arranged three-dimensionally. Fig.10 In the figure, the dashed lines at the top, bottom, left, and right represent the unit cell 10 repeats.

[0123] like Fig.10 As shown, the crystal 10X is a collection of a large number of unit cells 10. Therefore, when analyzing the crystal 10X (calculating similarity, etc.), it is necessary to consider not only the atoms included in one unit cell 10 but also the atoms included in the unit cells 10 existing around it.

[0124] As an example of a technique for obtaining crystal similarity in consideration of a repeated structure in a crystal, regarding repeated unit cells in a crystal, there can be cited a technique for expanding and processing one unit cell to include atoms of other unit cells adjacent to the one unit cell (expanded nodes).

[0125] Will refer to Fig.11 Describe the technique. Fig.11 As shown, in the technique of expanding and processing a unit cell by using an expansion node, for example, nodes representing atoms A to I are respectively designated as nodes A to I. Then, in the technique, for example, each node of the atom (node ​​in the unit cell) is connected by an edge to correspond to the chemical bonds between atoms in the crystal.

[0126] In addition, in the technique of extending the unit cell by using an extended node, for example, in another unit cell adjacent to one unit cell containing atoms A to I, a node corresponding to an atom adjacent to and bonded to the atom of this unit cell is prepared as an extended node. Fig.11In the example shown, for example, for node A, an extended node G1 corresponding to atom G in a unit cell adjacent to the upper side of the unit cell containing node A, and an extended node C4 corresponding to atom C in a unit cell adjacent to the left side of the unit cell containing node A are prepared. Similarly, Fig.11 As shown, for nodes A to I, extended nodes corresponding to atoms in adjacent unit cells are prepared.

[0127] Next, if Fig.11 As shown, for nodes A to I, edges between prepared extended nodes (extended edges) are created to correspond to chemical bonds between atoms in the crystal.

[0128] In this way, in the technology for expanding and processing unit cells by using an extension node, by using an extension node corresponding to an atom in another unit cell adjacent to the unit cell, the unit cell is expanded in consideration of the bonding state of atoms in the crystal. By doing so, in the technology for expanding and processing unit cells by using an extension node, since the repetitive structure of the unit cells in the crystal can be appropriately expressed, analysis (calculation of similarity, etc.) targeting the crystal can be performed with high accuracy.

[0129] Furthermore, as another example of a technique that considers a repetitive structure of a crystal, the following technique can be cited regarding unit cells in a crystal: atoms in one unit cell corresponding to atoms in another unit cell adjacent to the one unit cell are specified and represented as a cyclic structure.

[0130] Will refer to Fig.12 Explain the technology. Fig.12 As shown, in the technique of treating the repeating structure of a crystal as a cyclic structure, for example, similar to Fig.11 , the nodes representing atoms A to I are designated as nodes A to I, respectively. Then, in this technique, similar to Fig.11 In this example, each node of an atom (the node in the unit cell) is connected by edges to correspond to the chemical bonds between atoms in the crystal.

[0131] In addition, if Fig.12 As shown, in the technique of processing a repeating structure of a crystal as a cyclic structure, for example, a cyclic edge is created to correspond to chemical bonds of atoms in another unit cell adjacent to the unit cell containing nodes A to I.

[0132] Here, if Fig.12 As shown, the loop edge extending on the upper side of node A represents the same loop edge as the loop edge extending downward from node G. More specifically, for example, the loop edge connecting node A and node G corresponds to a virtual chemical bond between node A in one unit cell and node G in another unit cell adjacent to the upper side of the one unit cell.

[0133] Similarly, Fig.12 The other loop edges shown in each represent a loop edge corresponding to a chemical bond between atoms contained in adjacent unit cells.

[0134] In this way, in the technology of treating the repeating structure of a crystal as a cyclic structure, by representing the chemical bonds between atoms in adjacent unit cells with cyclic edges, information about the repeating structure of the unit cells (information about how the unit cells are arranged) can be expressed without increasing the number of nodes representing the unit cells.

[0135] That is, for example, in a technique for treating a repetitive structure of a crystal as a cyclic structure, for a unit cell in a crystal, atoms in one unit cell corresponding to atoms in another unit cell adjacent to the one unit cell are designated and expressed as a cyclic structure. By doing so, in a technique for treating a repetitive structure of a crystal as a cyclic structure, analysis (calculation of similarity, etc.) targeting the crystal can be performed with high accuracy because the repetitive structure of the unit cell in the crystal can be appropriately expressed.

[0136] Here, when determining the similarity between crystals having a repetitive structure, it is desirable to obtain the similarity by assuming a plurality of unit cells (multiplying the unit cell by an integer) as necessary to reduce the difference between the number of atoms (number of nodes) included in the structures for obtaining the similarity.

[0137] The following describes the relationship between the number of atoms (number of nodes) included in the structure for which similarity is obtained and the calculated similarity value.

[0138] First, the case where the number of atoms (number of nodes) contained in the structures for which similarity is obtained is significantly different is described. Fig.13A and 13B As shown, it is assumed that the unit cell 200A has nine distinguishable atoms (atoms A to I), and the unit cell 200B has 24 distinguishable atoms (atoms A to X).

[0139] exist Fig.13A All atoms in the unit cell 200A are connected to Fig. 13B The atoms in the unit cell 200B have a common substructure ( Fig.13A The unit cell 200A is contained in Fig. 13B In the case of the unit cell 200B), it is generally believed that the similarity of these structures is very large. However, when the similarity is calculated based on the above similarity calculation formula, the numerical value of the similarity of these structures is "0.675", which is not a large value.

[0140] From this we can see that due to Fig.13A The unit cell 200A and Fig. 13B The difference in the number of nodes (number of atoms) included in the unit cell 200B is large, so the calculated similarity is low and the accuracy of the analysis is low.

[0141] The similarity S between unit cell A and unit cell B can be calculated by the following formula by simplifying the above similarity calculation formula.

[0142] S = {N(A∩B) / N(A) + N(A∩B) / N(B)} × 0.5

[0143] Here, in the above formula, "N(A∩B) ≤ N(A)" and "N(A∩B) ≤ N(B)" are satisfied.

[0144] In the above formula, when "N(A) ≠ N(B)" is satisfied, the similarity S is not "1". Specifically, for example, when N(A) = λN(B) (0 < λ ≤ 1) is satisfied, the maximum value that the similarity S can take is "(1 + λ) / 2". In addition, since a crystal can be considered a periodic structure in which unit cells repeat in a semi-infinite manner, the fact that the similarity S does not take the value "1" due to the difference in the number of atoms (number of nodes) in each unit cell cannot be said to be a necessary reason for the crystal similarity S.

[0145] Here, as a technique for representing the values that the similarity S can take from 0 to 1, for example, the following two techniques can be cited.

[0146] Normalization of similarity S

[0147] When normalizing the similarity S, for example, the values that the similarity S (0 ≤ S ≤ (1 + λ) / 2) can take are set to 0 to 1, and are normalized to similarity S' using the following formula.

[0148] S' = {N(A∩B) / N(A) + N(A∩B) / N(B)} / (1 + λ)

[0149] Multiplying unit cell A and unit cell B by an integer

[0150] By appropriately multiplying unit cell A and unit cell B by an integer and treating them as multiple unit cells, the difference between N(A) and N(B) is made as small as possible, and the values that the similarity S can take are set to 0 to 1.

[0151] When normalizing the similarity S, when N(A) < N(B), even when S' = 1, there is a sub-structure B\A (the difference set obtained by subtracting the elements of unit cell A from unit cell B) that is not included in "A∩B".

[0152] In other words, for example, regardless of the structure of the sub-structure B\A and the number of atoms constituting each unit cell, there may be a case where the similarity S' = 1. As described above, the technique for normalizing the similarity S is not suitable as a technique for evaluating similarity.

[0153] However, in the technology of multiplying unit cell A and unit cell B by integers, N(A")=N(B") is set, where, for example, A" is the structure of multiple unit cells of unit cell A, B" is the structure of multiple unit cells of unit cell B, and the similarity S'" is expressed by the following calculation formula.

[0154] S"'={N(A"∩B") / N(A")+N(A"∩B") / N(B")} / (1+λ)

[0155] However, in order to set N(A")=N(B"), it may be necessary to increase the number of target unit cells (for example, using the least common multiple of unit cell A and unit cell B, etc.). In this case, it is ideal to multiply at least one of unit cell A or unit cell B by an integer so that N(A")≈N(B") is satisfied, in other words, for example, thereby reducing the difference in the number of nodes of the structure for which the similarity is calculated. As a specific criterion, in the case of setting N(A")=λN(B"), "9 / 10≤λ≤10 / 9" can be set. By defining in this way, the difference in the number of nodes contained in the target structure can be reduced while suppressing the number of unit cells required as a target for calculating the similarity, and the accuracy of the calculated similarity can be improved.

[0156] Here, we will refer to Figures 14A to 14C Describe a technique for multiplying cell A and cell B by an integer.

[0157] exist Figures 14A to 14C In the example of , a unit cell A represented as a graph having nodes and edges is referred to as graph A, and a unit cell B represented as a graph having nodes and edges is referred to as graph B. In the technique of multiplying unit cells A and unit cells B by an integer, for example, in order to reduce the difference between the number of nodes in graph A and the number of nodes in graph B, at least one of graph A or graph B is replicated by multiplying by an integer. At this time, it is desirable to replicate in such a manner that the absolute value of the difference between the total number of nodes in the replicated graph A and the total number of nodes in the replicated graph B is minimized, for example.

[0158] Then, in this example, in the technique of multiplying the unit cell A and the unit cell B by an integer, for example, the graph A is copied into four and the graph B is copied into nine.

[0159] For a graph obtained by integrating the replicated graphs, for example, graphs A can be combined with each other, and graphs B can be combined with each other, as Fig. 14B As shown. The combination here means providing edges between the respective graphs (between graphs A or between graphs B). In addition, in the graph obtained by integrating the replicated graphs, for example, Fig. 14C As shown, there is no need to perform joins between graphs A and between graphs B.

[0160] In this way, in the technique of multiplying unit cell A and unit cell B by an integer, for example, at least one of the diagrams of the unit cells of the two crystals is copied. By copying the diagram to reduce the difference in the number of atoms (number of nodes) between the two unit cells being compared, a decrease in similarity due to the difference in the number of atoms (number of nodes) can be avoided. As a result, the accuracy of the analysis (e.g., similarity) can be improved.

[0161] Here, when obtaining the similarity between crystals, as described above, the number of atoms contained in each unit cell of the comparison structure may increase (the number of bits required to calculate the similarity may increase). In particular, for example, when the crystal is an ionic crystal, the unit cell contains a large number of anions (for example, oxygen atoms), and the number of atoms contained in the unit cell increases.

[0162] Fig.15A Illustrate examples of unit cells in ionic crystals. Fig.15A The ionic crystal A shown has 8 lithium (Li), 4 manganese (Mn), 16 phosphorus (P), and 48 oxygen (O) atoms in the unit cell, for a total of 76 atoms. Fig. 15B Illustrate another example of the unit cell in an ionic crystal. Fig. 15B The ionic crystal B shown has 16 lithium (Li), 8 manganese (Mn), 16 phosphorus (P), and 56 oxygen (O) in the unit cell, respectively, for a total of 96 atoms.

[0163] exist Fig.15A and 15B In the example shown, in the case of calculating the similarity between ionic crystal A and ionic crystal B, for example, a node of the above-mentioned conflict graph is created for each individual atom. At this time, the number of nodes of the conflict graph (the number of bits required for calculation) is "8x16=128" for lithium, "4x8=32" for manganese, "16x16=256" for phosphorus, and "48x56=2688" for oxygen. Therefore, in order to calculate the similarity between ionic crystal A and ionic crystal B, for example, the number of nodes in the conflict graph (the number of bits required for calculation) is "128+32+256+2688=3104".

[0164] In this way, when the number of atoms (number of nodes) contained in a unit cell increases, assuming that there are a plurality of unit cells, it is difficult to calculate the similarity due to the limitation of the number of bits that can be processed by the annealing machine for calculating the similarity. For example, the number of bits that can be used by the annealing machine is up to about 8,000 bits, and even in the case where 3,000 bits or more are required when comparing unit cells as in the example described above, it is difficult to process the unit cells by multiplying integers due to the lack of available bits.

[0165] As described above, in the related art, when an ionic crystal is analyzed and the number of atoms included in the unit cell of the ionic crystal is large, there is a problem that the number of bits required for calculation increases and accurate analysis may be difficult.

[0166] Therefore, the present inventors have intensively studied methods and the like capable of reducing (suppressing) the number of bits required for ion crystal analysis, and have obtained the following findings.

[0167] In other words, for example, the present inventors have found that the number of bits required for analyzing an ionic crystal can be reduced (suppressed) by the following crystal analysis method or the like.

[0168] A crystal analysis method as an example of the technology disclosed in the present application includes: a graph creation process for creating a graph, wherein in a crystal, a repeating unit cell is extended to an adjacent repeating unit cell adjacent to the repeating unit cell, wherein the repeating unit cell includes intra-cell nodes indicating data of atoms in the repeating unit cell, and intra-cell edges indicating data of chemical bonds between two atoms in the repeating unit cell; and an analysis process for analyzing the crystal using the graph, wherein in the graph creation process, in the case where the crystal is an ionic crystal, in the ionic crystal, when the number of intra-cell nodes that are data of anion atoms bonded to cation atoms in the repeating unit cell is n, the number of intra-cell nodes that are data of anion atoms is set to less than n-1.

[0169] Here, in the example of the technology disclosed in the present application, in the process of graph creation, for example, one repeating unit cell is prepared, and the repeating unit cell includes intra-unit cell nodes indicating data of atoms in the one repeating unit cell and intra-unit cell edges indicating data of chemical bonds between two atoms in the one repeating unit cell. That is, for example, in the example of the technology disclosed in the present application, for example, in order to express chemical bonds between atoms in the unit cell in the crystal, the above-mentioned one repeating unit cell (graph of the unit cell) including intra-unit cell nodes and intra-unit cell edges is prepared.

[0170] Then, in the example of the technology disclosed in the present application, in the diagram creation process, for example, a diagram is created, wherein, in the diagram creation process, this one repeating unit cell is extended to the adjacent repeating unit cell adjacent to this one repeating unit cell. That is, for example, in the example of the technology disclosed in the present application, for example, in order to express the repeating structure (periodic structure, crystal structure) of an ionic crystal, a diagram is created in which this one repeating unit cell is extended in consideration of the relationship with the atoms of the surrounding repeating unit cells. By doing so, in the example of the technology disclosed in the present application, the repeating structure in the crystal can be appropriately expressed, and the crystal can be analyzed with high accuracy.

[0171] Furthermore, in an example of the technology disclosed in the present application, in the graph creation process, for example, in the case where the crystal is an ionic crystal, in the ionic crystal, when the number of nodes within the unit cell that are data of an anion atom bonded to a cation atom in this one repeated unit cell is n, the number of nodes within the unit cell that are data of anion atoms is set to be less than n-1. In other words, in the graph creation process, for example, by not considering (by deleting) at least one anion atom included in this one repeated unit cell as a node, the cation atom and the anion atom are considered as one node as a whole.

[0172] By doing so, in the example of the technology disclosed in the present application, since the number of nodes within the unit cell in this one repeated unit cell can be suppressed (reduced), the number of bits required for calculation can be reduced (suppressed). That is, for example, in the example of the technology disclosed in the present application, by creating a graph in which at least some nodes of anion atoms in an ionic crystal are deleted, the number of nodes included in the graph (the number of bits required for calculation) can be suppressed.

[0173] Subsequently, in the example of the technology disclosed in the present application, in the analysis process, for example, the crystal is analyzed using the graph created in the graph creation process. That is, for example, in the example of the technology disclosed in the present application, in the analysis process, for example, the ionic crystal is analyzed based on the graph in which at least some nodes of the anion atoms in the ionic crystal are deleted.

[0174] By doing so, in an example of the technology disclosed in the present application, a graph suppressing the number of nodes within a unit cell is used, thereby making it possible to perform a highly accurate analysis, such as calculating similarity after multiplying the unit cell by an integer to make the number of nodes consistent.

[0175] Here, in the graph created in the graph creation process, the number of nodes in the unit cell as anion atom data is not particularly limited as long as it is n-1 or less as described above, and can be appropriately selected according to the purpose, but is ideally set to, for example, "0 (zero)". In other words, for example, in the example of the technology disclosed in the present application, it is ideal not to regard all anion atoms contained in one repeated unit cell as nodes (remove all anions).

[0176] Will refer to Fig.16 A description will be given of a form in which the number of nodes in the unit cell that are data of anion atoms in the graph created in the graph creation process is set to “0”.

[0177] Fig.16 This example shows that a repeating unit cell in an ionic crystal contains one lithium (Li), one manganese (Mn), one phosphorus (P) and 14 oxygen (O). Fig.16As shown in the example of the prior art in , in an ionic crystal, the number of oxygens corresponding to the valence is bonded to each cation atom (lithium, manganese and phosphorus) with a single bond. In this way, in an ionic crystal, cation atoms and anion atoms are usually bonded to each other. In the example of the prior art, a graph is created with each of these atoms considered as a node and the interatomic bonds considered as edges of single bonds.

[0178] In addition, Fig.16 In the example shown, for example, there are two bonds between phosphorus (P) and manganese (Mn) located at the lower left of the unit cell through oxygen (O), and phosphorus (P) and manganese (Mn) located at the upper part of the unit cell are bonded to each other through one single bond oxygen (O). Fig.16 In the example shown, for example, manganese (Mn) and lithium (Li) are bonded to each other via two single-bonded oxygens (O).

[0179] Then, if Fig.16 As shown in one embodiment of the present application in , in an example of the technology disclosed in the present application, a cation atom (+) and an anion atom (-) surrounding the cation atom are collectively treated as one node. That is, for example, Fig.16 As shown in one embodiment of the present application in , in an example of the technology disclosed in the present application, for example, the number of nodes included in the unit cell can be reduced by considering only cation atoms as nodes without considering anion atoms as nodes.

[0180] In addition, in the examples of the technology disclosed in the present application, the bond order between the cation atoms can be set to the number of chemical bonds via the anions that are considered not to exist between the cation atoms (the number of anions deleted between the cations). In other words, for example, in the examples of the technology disclosed in the present application, the bond order between the cation atoms can be determined by the number of bonds between the cation atoms via the anion atoms.

[0181] For example, in Fig.16 In the example shown, for example, the bond order between phosphorus (P) and manganese (Mn) located at the lower left of the unit cell can be set to "2", and the bond order between phosphorus (P) and manganese (Mn) located at the upper side of the unit cell can be set to "1". Fig.16 In the illustrated example, for example, the bond order between manganese (Mn) and lithium (Li) may be set to "2".

[0182] In the examples of the technology disclosed in this application, Fig.16 As shown in one embodiment of the present application, for example, by considering only cation atoms as nodes and not anion atoms as nodes, the number of nodes included in the unit cell can be reduced. Fig.16In the example shown, in the example of the prior art, the number of nodes included in the unit cell is 17, but in one embodiment of the present application, the number of nodes included in the unit cell can be reduced to 4.

[0183] Therefore, in the example of the technology disclosed in the present application, the number of nodes included in the unit cell can be suppressed by considering the cation atoms and the anion atoms together as one node and by deleting at least some of the anion atoms included in the unit cell being processed.

[0184] Therefore, in the example of the technology disclosed in the present application, it is possible to reduce (suppress) the number of nodes included in a graph for analyzing an ionic crystal, that is, for example, the number of bits required for calculation.

[0185] Each process in the crystal analysis method disclosed in the present application is described in detail below.

[0186] The crystal analysis method disclosed in the present application includes, for example, a graph creation process and an analysis process, and also includes other processes as necessary.

[0187] <Chart creation process>

[0188] In the graph creation process in the example of the technology disclosed in the present application, as described above, a repeating unit cell is prepared, which includes a unit cell node indicating data of atoms in the repeating unit cell and a unit cell edge indicating data of a chemical bond between two atoms in the repeating unit cell. Then, in the graph creation process, a graph is created in which the repeating unit cell is extended to adjacent repeating units adjacent to the repeating unit cell.

[0189] <<Chart>>

[0190] In the example of the technology disclosed in the present application, a graph has, for example, nodes which are data on atoms and edges which are data on chemical bonds between two atoms.

[0191] Here, in an example of the technology disclosed in the present application, a graph can be, for example, an abstract data type, including a set of nodes (vertices) and a set of edges (branches) representing connection relationships between the nodes. The graph is represented by G=(V, E), for example, where V is a set of nodes and E is a set of edges. V is a finite set and E is a set containing two elements selected from V.

[0192] It can be assumed that a node indicates an atom in one unit cell of a crystal. Hereinafter, a node indicating an atom in one unit cell of a crystal may be referred to as an "intra-unit cell node".

[0193] In addition, it can be assumed that the edge represents the chemical bond between atoms in the unit cell. Hereinafter, the edge indicating the chemical bond between atoms in the unit cell may be referred to as an "intra-unit cell edge".

[0194] The graph may have nodes (hereinafter sometimes referred to as "expanded nodes") indicating atoms in neighboring cells adjacent to the repeating cell, where the atoms have chemical bonds with atoms in the repeating cell.

[0195] In this case, the graph may have an edge indicating a chemical bond between an atom corresponding to a node in the unit cell and an atom corresponding to an extended node (hereinafter sometimes referred to as an "extended edge"). In addition, in this case, the graph may have an edge indicating a chemical bond between two atoms corresponding to an extended node (hereinafter sometimes referred to as an "edge between extended nodes").

[0196] That is, for example, in an example of the technology disclosed in the present application, the graph created in the graph creation process ideally includes intra-cell nodes that are data indicating atoms in a repeating cell, and intra-cell edges that are data indicating a chemical bond between two atoms in the cell; and extended nodes that are data of an atom that has a chemical bond with an atom in the repeating cell and is located in an adjacent repeating cell adjacent to the repeating cell; and extended edges that are data of a chemical bond between an atom corresponding to the intra-cell node in the repeating cell and an atom corresponding to the extended node.

[0197] In this way, in the examples of the technology disclosed in this application, reference can be appropriately used. Fig.11 A technique is described in which a unit cell is expanded and processed by using an expansion node.

[0198] In addition, the diagram may have an edge indicating a virtual chemical bond between an atom X in a unit cell and an atom Y in the unit cell and corresponding to an atom Y', the atom Y' having a chemical bond with the atom X in the unit cell and in an adjacent unit cell adjacent to the unit cell. Here, the atom Y in the unit cell and corresponding to an atom Y' in the adjacent unit cell refers to an atom in the unit cell that overlaps with the atom Y' in the adjacent unit cell when the adjacent unit cell and the unit cell overlap so that the constituent atoms overlap.

[0199] Note that, hereinafter, an edge indicating a virtual chemical bond between atom Y in the above-mentioned unit cell and atom X in the unit cell may be referred to as a “loop edge”.

[0200] That is, for example, in an example of the technology disclosed in the present application, the graph created in the graph creation process is ideally a cyclic graph, which includes: intra-cell nodes indicating data of atoms in one repeating cell; and intra-cell edges indicating data of chemical bonds between two atoms in the cell; and cyclic edges indicating data of virtual chemical bonds between atom X and atom Y in the one repeating cell and corresponding to atom Y', wherein atom Y' has a chemical bond with atom X in the one repeating cell and in an adjacent repeating cell adjacent to the one repeating cell.

[0201] In this way, in the examples of the technology disclosed in this application, reference can be appropriately used. Fig.12 Describes a technique for treating the repetitive structure of a crystal as a cyclic structure.

[0202] Nodes within the unit cell and expanded nodes have data such as atom type, atomic valence, and atomic charge.

[0203] For example, edges within a unit cell, loop edges, and extended edges have data such as bond type, bond angle, bond length, and bond order. Bond angles and bond lengths can be represented using, for example, coordinate data of atoms or chemical bonds.

[0204] Furthermore, it is desirable to create intra-cell edges, loop edges, and extension edges by, for example, von Lonoy partitioning between nodes.

[0205] Von Launoy segmentation is a technique that divides the nearest neighbor region of each node by drawing perpendicular bisectors on the lines connecting adjacent nodes.

[0206] Here, the von Lonoy diagram and the von Lonoy segmentation will be described.

[0207] The von Lonoy graph is a graph in which multiple points (generators) are arranged at arbitrary positions in a certain metric space, and regions are divided based on which generator other points in the same metric space are closest to. Moreover, this division of the region is called a von Lonoy partition. In the von Lonoy graph, the partitioning pattern is determined only by the positions of the generators.

[0208] In the case where a unit cell of a crystal is schematically represented as including chemical bonds, the chemical bonds in the unit cell of the crystal are generally not uniquely determined. However, by creating edges (intra-unit cell edges, loop edges, and extended edges) by von Launeui segmentation, the edges (intra-unit cell edges, loop edges, and extended edges) can be uniquely determined. As a result, the edges of a plurality of crystals can be uniquely determined, which can improve the analysis accuracy of the crystals.

[0209] In addition, intra-cell edges, loop edges, and extended edges can be created, for example, based on the standard ionic radius of atoms. For example, the following rule can be applied: if "d≤a(r1+r2)+b" is satisfied, then an edge is created between nodes, where d is the distance between any two nodes, and r1 and r2 are the ionic radii of a single node. Here, a is a constant equal to or greater than 0, and b is a constant. In addition, 1.0≤a≤1.2 and is expected.

[0210] As a method of creating the unit cell inner edges, loop edges, and extended edges, one or both of the above two techniques, such as the von Launeuy segmentation technique and the technique using the standard ionic radius, can be freely selected. In the case where both are selected, for example, among the unit cell inner edges, loop edges, and extended edges that can be determined by von Launeuy segmentation, those that are regarded as unit cell inner edges, loop edges, and extended edges in the technique using the standard ionic radius of atoms can be created as unit cell inner edges, loop edges, and extended edges.

[0211] In addition, in an example of the technology disclosed in the present application, as described above, in the process of graph creation, in the case where the crystal is an ionic crystal, in the ionic crystal, when the number of intra-cell nodes that are data of anion atoms bonded to cation atoms in a repeated unit cell is n, the number of intra-cell nodes that are data of anion atoms is set to be less than n-1.

[0212] Here, ionic crystal refers to a crystal formed by ionic bonding of cations (positive ions, cation atoms) and anions (negative ions, anion atoms), for example. In the examples of the technology disclosed in the present application, the ionic crystal analyzed is not particularly limited and can be appropriately selected according to the purpose.

[0213] In addition, in the example of the technology disclosed in the present application, as described above, it is desirable not to regard all anionic atoms contained in the repeating unit cell as nodes (remove all anions). In other words, for example, in the example of the technology disclosed in the present application, it is desirable to set the nodes of anionic atoms to "0" and consider that all cationic atoms are directly bonded to each other.

[0214] More specifically, for example, in the example of the technology disclosed in the present application, it is desirable to set the number of nodes in the unit cell that are data of anion atoms to 0 during the graph creation process. By doing so, the number of nodes included in the repeated unit cell can be further reduced, and the number of bits required for analyzing the ionic crystal can be further suppressed.

[0215] Note that during graph creation, when the number of intra-cell nodes as anion atom data is not set to 0 (when some of all anion atoms contained in a repeated cell remain), the anion atoms to be regarded as intra-cell nodes can be appropriately selected according to the purpose.

[0216] In the example of the technology disclosed in the present application, whether it is considered to be a node in the unit cell (whether to delete) can be determined, for example, from the bonding relationship between each anion atom in the repeating unit cell. In this case, for example, anion atoms bonded to more than two cationic atoms can be deleted, leaving anion atoms bonded to only one cationic atom.

[0217] In the examples of the technology disclosed in the present application, the anion atom not regarded as a node in the unit cell (anion atom to be deleted) is not particularly limited and can be appropriately selected according to the purpose. Examples thereof include oxygen atoms, chlorine atoms, fluorine atoms, etc.

[0218] The anion atoms not considered as nodes in the unit cell can be appropriately selected according to the ionic crystal being analyzed. Then, for example, by selecting anion atoms having a large number of atoms in the ionic crystal, the number of bits required for calculation can be further suppressed.

[0219] In the example of the technology disclosed in the present application, the bond order between the cation atoms when the anion atoms are deleted can be set to, for example, the number of chemical bonds of the anions that are considered not to exist between the cation atoms. In other words, for example, in the example of the technology disclosed in the present application, as described above, the bond order between the cation atoms can be determined using the number of bonds between the cation atoms via the anion atoms.

[0220] That is, for example, in an example of the technology disclosed in the present application, during the graph creation process, when the number of nodes within the unit cell that are data of anion atoms is set to 0, and it is considered that the cation atoms in a repeated unit cell are directly chemically bonded to each other, the bond order of the chemical bonds between the cation atoms is ideally set to the number of chemical bonds between the cation atoms via the anion atoms that are considered not to exist.

[0221] By doing so, in the examples of the technology disclosed in this application, such as Fig.16 As shown, the bonding state of each atom in the ionic crystal can be more appropriately considered, and a higher accuracy analysis can be performed.

[0222] <Analysis Process>

[0223] In the analysis process in the example of the technology disclosed in the present application, crystal analysis is performed by using the graph created in the graph creation process as described above.

[0224] The crystal analysis performed in the analysis process is not particularly limited and can be appropriately selected according to the purpose. For example, examples thereof include analysis of structural similarity of a plurality of crystals, prediction of crystal properties, and the like.

[0225] The technique for analyzing the similarity of the structures (crystal structures) of a plurality of crystals is not particularly limited and can be appropriately selected according to the purpose, as long as the technique is used to analyze the similarity between the crystal structure of a specific crystal and the crystal structure of another crystal by using the graph created in the graph creation process. For example, examples of techniques for analyzing the similarity of the structures of a plurality of crystals include techniques for calculating scores of similarity and determining higher similarity as the scores are larger, etc.

[0226] The method of calculating the similarity score is not particularly limited and may be appropriately selected according to the purpose. For example, there are the following techniques (1) and (2).

[0227] Technology (1), Fingerprinting

[0228] Technology (2) is a technology for analyzing structural similarity by expressing the maximum independent set problem of the conflict graph using the Ising model equation and solving it using an annealing machine, etc. to search for common substructures.

[0229] In the fingerprint method as the technique (1), for example, whether a substructure in a reference structure is included in a compared structure is represented by 0 or 1, and the similarity is evaluated.

[0230] In addition, as a technique using a conflict graph in technique (2), it is possible to appropriately use a reference Figures 1 to 9 Note that details of analyzing similarity by using a conflict graph in an example of the technology disclosed in the present application will be described later.

[0231] Note that the technique for calculating the similarity score is not limited to these techniques. For example, the similarity score can be calculated based on cosine similarity, correlation coefficient, correlation function, edit distance (also called Levenshtein distance), etc. related to calculating the similarity.

[0232] The technique for predicting crystal characteristics is not particularly limited and can be appropriately selected according to the purpose as long as it is a technique for predicting crystal characteristics by using a graph created by a graph creation process. For example, examples thereof include predicting characteristics by using a learning model or the like.

[0233] For example, examples of predicting characteristics by using a learning model include a technique of creating a learning model by machine learning and predicting characteristics of a crystal by using the learning model, and the like.

[0234] Examples of machine learning include, for example, supervised learning, unsupervised learning, etc. The learning technique in the case of creating a learning model by machine learning is not particularly limited and can be appropriately selected according to the purpose. For example, in the case of supervised learning, a technique for learning by using a training data set including a crystal diagram and crystal properties as labels can be used.

[0235] As training data in machine learning, for example, training data having a graph created in a graph creation process and data on characteristics of a crystal corresponding to the graph can be used.

[0236] In addition, the properties to be predicted are not particularly limited and can be appropriately selected according to the purpose, and may be chemical properties or physical properties. Specific examples of the properties include, for example, electrical conductivity, ion conductivity, discharge potential, relative dielectric constant, thermal conductivity, specific heat, and the like.

[0237] In addition, when predicting characteristics, the graph (graph structure data) created in the graph creation process can be processed and used. For example, the graph can be converted into a tensor, so that the graph structure data can be learned with high accuracy by using machine learning (such as deep learning) technology.

[0238] Here, tensors refer to data represented as multidimensional arrays, where concepts such as matrices and vectors are already common sense.

[0239] In the example of the technology disclosed in the present application, as described above, it is desirable to calculate the similarity of ionic crystals during the analysis process.

[0240] That is, for example, in an example of the technology disclosed in the present application, it is desirable to create a first graph for the first crystal and a second graph for the second crystal in a graph creation process, and calculate the similarity between the first crystal and the second crystal based on the first graph and the second graph in an analysis process.

[0241] In addition, in the example of the technology disclosed in the present application, as described above, when calculating the similarity, it is ideal to obtain the similarity by assuming multiple unit cells when necessary (multiplying the unit cell by an integer) to reduce the difference in the number of atoms (number of nodes) contained in the structure to obtain the similarity.

[0242] That is, for example, in an example of the technology disclosed in the present application, it is ideal to calculate the similarity during the analysis process after multiplying at least one of a repeated unit cell of the first graph or a repeated unit cell of the second graph by an integer so that the number of nodes within the unit cells in the first and second graphs is approximate.

[0243] By doing so, in the example of the technology disclosed in the present application, by reducing the difference in the number of atoms (number of nodes) of two repeated unit cells of the comparison structure, the reduction in similarity due to the difference in the number of atoms can be avoided, and the accuracy of the similarity calculation results can be further improved.

[0244] In addition, in the example of the technology disclosed in the present application, as described above, it is desirable to calculate the similarity in the analysis process by solving the maximum independent set problem of the conflict graph. As a technique for calculating the similarity between the first crystal and the second crystal by solving the maximum independent set problem of the conflict graph, reference can be appropriately used. Figures 1 to 9 Describe the technology.

[0245] In the conflict graph used to calculate the similarity between the first crystal and the second crystal, for example, combinations of respective node atoms constituting repeated unit cells of the first crystal represented as a graph and respective node atoms constituting repeated unit cells of the second crystal represented as a graph are defined as nodes.

[0246] Here, in the technical examples disclosed in the present application, the case where "two nodes are compared and are identical to each other" means that when the two nodes are compared, the nodes are composed of node atoms that are in the same situation (bonding situation) as each other. Similarly, in the examples of the technology disclosed in the present application, the case where "two nodes are compared and are different from each other" means that when the two nodes are compared, the nodes are composed of node atoms that are in different situations (bonding situations) from each other.

[0247] More specifically, for example, in an example of the technology disclosed in the present application, for example, the calculation of similarity in the analysis process is ideally calculated by searching for the maximum independent set of the conflict graph between the first graph and the second graph using the following calculation formula (1). In addition, the following calculation formula (1) is a calculation formula used when the number of nodes in the unit cell of the data of the anion atom is set to 0 during the graph creation process.

[0248] [Formula 6]

[0249]

[0250] Here, in the calculation formula (1), E is the Hamiltonian function, where minimizing E means searching for the maximum independent set, c' ij is a coefficient representing the weight of the key order between the i-th node and the j-th node in the conflict graph between the first graph and the second graph, x i is a binary variable indicating whether the i-th node is 0 or 1, x j is a binary variable indicating whether the j-th node is 0 or 1, and α and β are positive numbers.

[0251] Here, the calculation of c' in the above formula (1) will be described in detail. ij .

[0252] First, as described above, in the example of the technology disclosed in the present application, when an anion atom is deleted, the bond order between cation atoms can be determined by, for example, the number of bonds between cation atoms via the anion atom.

[0253] Therefore, the bond order between cation atoms is determined by the combination of cation atoms, and thus can be incorporated as a weight into the penalty term (term with coefficient β) of the Ising model equation (calculation formula (1)) to solve the maximum independent set problem of the conflict graph. That is, for example, in the example of the technology disclosed in the present application, by expressing the bond order between cation atoms as the weight (c' ij ), the bond order between cation atoms can be considered to calculate the similarity.

[0254] For example, c' represents the weight of the bond order between cation atoms. ij It can be determined by, for example, the following rules.

[0255] In the absence of a bond between the atomic combination in the i-th node and the atomic combination in the j-th node, set "c' ij =0".

[0256] If there is a bond between only one atom combination in the i-th node and the j-th node, set "c' ij =λ1”.

[0257] Between the atomic combination in the i-th node and the atomic combination in the j-th node, if there is a bond in both atomic combinations, set "c' ij =λ2(1-min{c1 / c2,c2 / c1})”.

[0258] Here, λ1 and λ2 are positive constants, and it is usually desirable to set λ1 = λ2 = 1. Note that in the following description, a case where λ1 = λ2 = 1 will be described as an example.

[0259] Note that min{A,B} means selecting the smaller value of A and B. In addition, c1 represents the bond order between the atom at the i-th node and the atom at the j-th node in the first graph, and c2 represents the bond order between the atom at the i-th node and the atom at the j-th node in the second graph.

[0260] Fig.17 An example of the relationship between the combination of nodes in the first graph and the second graph and c'ij representing the weight with respect to the bond order between cation atoms is described.

[0261] like Fig.17 As shown, according to c' which represents the weight about the bond order ij For example, if there is no bond between node A1 and node A2 in the first graph and between node B1 and node B2 in the second graph, c' ij That is, for example, in this case, since these nodes are composed of atoms in the same situation (bonding situation) as each other, that is, the bonding situation is not contradictory, c' ij is "0", so that the value of the penalty term with coefficient β is "0".

[0262] Likewise, in the case where there are keys with the same key order between nodes A1 and A2 in the first graph and between nodes B1 and B2 in the second graph (in Fig.17 In the example, when the bond order of both is 1), c' ij is "0".

[0263] In addition, if Fig.17 As shown, according to c' ijFor example, if there is a bond between node A1 and node A2 in the first graph and there is no bond between node B1 and node B2 in the second graph, c' ij That is, for example, in this case, these nodes are composed of atoms that are in completely different situations (bonding situations) from each other, that is, the bonding situations are completely contradictory, c' ij is "1", so the penalty term with coefficient β becomes a large value.

[0264] In addition, if Fig.17 As shown, according to c' ij According to the rule, for example, in the case where there are bonds with different bond orders between nodes A1 and A2 in the first graph and between nodes B1 and B2 in the second graph, c'ij will be a "value of 0 to 1" according to the bond order. That is, for example, in this case, these nodes have no contradiction in bonding, but due to the different bond orders, they will be composed of atoms in different situations (bonding situations) to a certain extent. Therefore, c'ij is a "value of 0 to 1", so that the value of the penalty term with coefficient β is a value according to the bond order.

[0265] Therefore, in the example of the technology disclosed in the present application, c'ij representing the weight about the bond order can be a multi-valued weight according to the bond order (weight according to the difference in the bond order), rather than a weight that only considers whether there is a bond between nodes (binary of 0 or 1). By doing so, in the example of the technology disclosed in the present application, the similarity can be calculated with higher accuracy considering the bond order between cation atoms.

[0266] Note that in the above description, the case where λ1=λ2=1 has been described as an example, but the technology disclosed in the present application is not limited to this, and λ1 and λ2 can be set to positive constants other than 1 when necessary.

[0267] In an example of the technology disclosed in the present application, when the maximum independent set is searched using the above calculation formula (1), the priority of creating a conflict graph is not high, and at least the above calculation formula (1) can be minimized. In other words, for example, in an example of the technology disclosed in the present application, the search for the maximum independent set in the conflict graph is replaced by a combinatorial optimization problem in the Hamiltonian function, where minimization means searching for the maximum independent set and solving it. Here, by performing an annealing method (annealing) using an annealing machine or the like, the minimization of the Hamiltonian function represented by the Ising model equation in the QUBO format as in the above calculation formula (1) can be performed in a short time.

[0268] Therefore, in the technology disclosed in the present application, on the one hand, by using the above calculation formula (1), the maximum independent set can be searched by an annealing method using an annealing machine, etc., so that the similarity of the crystal can be calculated in a shorter time. In other words, for example, in the technology disclosed in the present application, on the one hand, the maximum independent set can be searched by minimizing the Hamiltonian function (H) in the above calculation formula (1) by an annealing method, thereby calculating the similarity of the crystal in a shorter time. Note that the details of the annealing method will be described later.

[0269] Here, the annealing machine is not particularly limited and can be appropriately selected according to the purpose, as long as the annealing machine is a computer that adopts an annealing method that searches for the ground state of the energy function represented by the Ising model. Examples of annealing machines include, for example, a quantum annealing machine, a semiconductor annealing machine using semiconductor technology, a machine that performs simulated annealing performed by software using a central processing unit (CPU) or a graphics processing unit (GPU), etc. In addition, for example, a digital annealer (registered trademark) can be used as an annealing machine.

[0270] In addition, in the above example, the calculation formula has been described when the number of nodes in the unit cell as data of anion atoms is set to 0. However, when the number of nodes in the unit cell as data of anion atoms is not set to 0 (some anion atoms remain), for example, the following calculation formula can be used.

[0271] [Formula 7]

[0272]

[0273] Here, in the above calculation formula (1), E is the Hamiltonian function, where minimizing E means searching for the maximum independent set, c" i is a coefficient representing the weight of the i-th node in the conflict graph in the first graph and the second graph, c' ij is a coefficient representing the weight of the key order of the i-th node and the j-th node in the conflict graph in the first graph and the second graph, x i is a binary variable indicating whether the i-th node is 0 or 1, x j is a binary variable indicating whether the j-th node is 0 or 1, and α and β are positive numbers.

[0274] In the above calculation formula, "c" is introduced into the calculation formula (1). i ", and by introducing this "c" i "Combinations of atoms of the same anion can be added to nodes of the graph.

[0275] Here, "c" i" may be a constant related to selecting a combination of the i-th node in the conflict graph between the first graph and the second graph. More specifically, for example, when the constant of the combination between cation atoms is assumed to be "c" j " and the constant of the combination between anion atoms is assumed to be "c" k ", you can set "c" j ≥c” k ≥0".

[0276] This is because, since the number of anionic atoms is reduced in the example of the technology disclosed in the present application, the calculation result of the similarity is not affected by the selection of anionic atoms to be deleted, which is not necessary.

[0277] In addition, in the examples of the technology disclosed in this application, similar to the reference Figures 1 to 9 The described technology can obtain the similarity associated with the features of the maximum independent set searched by using the following calculation formula (2). In other words, for example, in an example of the technology disclosed in the present application, the calculation of the similarity in the analysis process is completed for the maximum independent set searched by using the following calculation formula (2).

[0278] [Formula 8]

[0279]

[0280] In the above calculation formula (2), S(G A ,G B )) represents the similarity between the first crystal represented by the first graph and the second crystal represented by the second graph, and is expressed by 0 to 1, meaning that the closer to 1, the higher the similarity, V A represents the total number of node atoms in the first crystal represented by the first graph, V c A represents the number of node atoms included in the maximum independent set of the conflict graph in the node atoms in the first crystal represented by the first graph, V B represents the total number of nodal atoms in the second crystal represented by the second graph, V c B represents the number of node atoms included in the maximum independent set of the conflict graph among the node atoms in the second crystal represented by the second graph, and δ is a number between 0 and 1.

[0281] <Other Process>

[0282] The other processes are not particularly limited and may be appropriately selected according to purpose.

[0283] (Crystal Analysis Equipment)

[0284] The crystal analysis device disclosed in the present application may be, for example, a device that executes the crystal analysis method disclosed in the present application. In addition, an appropriate mode in the crystal analysis device disclosed in the present application may be similar to an appropriate mode in the crystal analysis method disclosed in the present application.

[0285] Here, the crystal analysis device disclosed in the present application has a chart creation unit and an analysis unit, and may also have other units (parts) when necessary. Note that the chart creation unit and the analysis unit may be implemented by a combination of a processor such as a central processing unit (CPU) and a memory such as a random access memory (RAM), or may be implemented as a part of an annealing machine.

[0286] (Crystal Analysis Program)

[0287] The crystal analysis program disclosed in the present application may be, for example, a program that causes a computer to execute the crystal analysis method disclosed in the present application. In addition, for example, an appropriate mode of the crystal analysis program disclosed in the present application may be similar to an appropriate mode of the crystal analysis method disclosed in the present application.

[0288] The crystal analysis program disclosed in the present application can be created using various known programming languages ​​according to the configuration of the computer system used, the type and version of the operating system, and the like.

[0289] The crystal analysis program disclosed in the present application can be recorded in a recording medium such as an internal hard disk or an external hard disk, or in a recording medium such as a compact disk read-only memory (CD-ROM), a digital compact disk read-only memory (DVD-ROM), a magneto-optical (MO) disk, or a universal serial bus (USB) flash drive.

[0290] In addition, in the case where the crystal analysis program disclosed in the present application is recorded in a recording medium as described above, the program can be used directly, or installed on a hard disk if necessary, and then used by a recording medium reader included in a computer system. In addition, the crystal analysis program disclosed in the present application can be recorded in an external storage area (other computers, etc.) accessible from a computer system via an information communication network. In this case, the crystal analysis program disclosed in the present application is recorded in an external storage area and can be used directly, or installed in a hard disk if necessary, and then used from an external storage area via an information communication network.

[0291] Note that the crystal analysis program disclosed in the present application may be divided for each segment of any processing and recorded in a plurality of recording media.

[0292] (Computer readable recording medium)

[0293] The computer-readable recording medium disclosed in the present application records the crystal analysis program disclosed in the present application.

[0294] The computer-readable recording medium disclosed in the present application is not limited to any specific medium and can be appropriately selected according to the purpose. For example, examples of computer-readable recording media include internal hard disks, external hard disks, CD-ROMs, DVD-ROMs, MO disks, USB memories, etc.

[0295] Furthermore, the computer-readable recording medium disclosed in the present application may be a plurality of recording media that divide and record the crystal analysis program disclosed in the present application for each segment of any processing.

[0296] Hereinafter, an example of the technology disclosed in the present application will be described in more detail using a configuration example of an apparatus, a flowchart, and the like.

[0297] Fig.18 Another hardware configuration example of the crystal analysis device disclosed in the present application will be described.

[0298] In the crystal analysis device 100 , for example, a control unit 101 , a main storage device 102 , an auxiliary storage device 103 , an input / output (I / O) interface 104 , a communication interface 105 , an input device 106 , an output device 107 , and a display device 108 are connected to one another via a system bus 109 .

[0299] The control unit 101 performs arithmetic operations (e.g., four arithmetic operations, comparison operations, arithmetic operations for annealing method, etc.), hardware and software operation control, etc. The control unit 101 may be, for example, a central processing unit (CPU), a part of an annealing machine used in the annealing method, or a combination thereof.

[0300] The control unit 101 realizes various functions by executing a program (for example, a crystal analysis program disclosed in the present application) read from the main storage device 102 or the like, for example.

[0301] The processing performed by the graph creation unit and the analysis unit in the crystal analysis device disclosed in the present application can be performed by, for example, the control unit 101 .

[0302] The main storage device 102 stores various programs and data required for executing various programs, etc. As the main storage device 102, for example, a device having at least one of a read only memory (ROM) and a random access memory (RAM) can be used.

[0303] The ROM stores various programs such as a basic input / output system (BIOS) and the like. In addition, the ROM is not particularly limited and can be appropriately selected according to the purpose. For example, examples thereof include a mask ROM, a programmable ROM (PROM), and the like.

[0304] The RAM is used, for example, as a working range expanded when various programs stored in the ROM or the auxiliary storage device 103 are executed by the control unit 101. The RAM is not particularly limited and can be appropriately selected according to the purpose. For example, examples thereof include a dynamic random access memory (DRAM), a static random access memory (SRAM), etc.

[0305] The auxiliary storage device 103 is not particularly limited as long as the device can store various information, and can be appropriately selected according to the purpose. For example, examples thereof include a solid state drive (SSD) or a hard disk drive (HDD), etc. In addition, the auxiliary storage device 103 may be a portable storage device such as a CD drive, a DVD drive, or a Blu-ray (registered trademark) disc (BD) drive.

[0306] Furthermore, the crystal analysis program disclosed in the present application is stored in, for example, the auxiliary storage device 103 , loaded into a RAM (main memory) of the main storage device 102 , and executed by the control unit 101 .

[0307] The I / O interface 104 is an interface for connecting various external devices. The I / O interface 104 can input data from, for example, a compact disk ROM (CD-ROM), a digital compact disk ROM (DVD-ROM), a magneto-optical disk (MO disk), a universal serial bus (USB) memory (USB flash drive), or output data to it.

[0308] The communication interface 105 is not particularly limited, and a known communication interface can be used as appropriate. For example, examples thereof include a communication device using wireless or wired communication or the like.

[0309] The input device 106 is not particularly limited, and any known device can be used as long as it can receive various requests and information input to the crystal analysis device 100. For example, examples thereof include a keyboard, a mouse, a touch panel, a microphone, etc. In addition, in the case where the input device 106 is a touch panel (touch display), the input device 106 can also serve as the display device 108.

[0310] The output device 107 is not particularly limited, and a known device can be used as appropriate. Examples thereof include printers and the like.

[0311] The display device 108 is not particularly limited, and a known device can be used appropriately. Examples thereof include a liquid crystal display, an organic EL display, etc.

[0312] Fig.19 Another hardware configuration example of the crystal analysis device disclosed in the present application will be described.

[0313] exist Fig.19In the example shown, the crystal analysis apparatus 100 is divided into a computer 200 that performs processing for creating a graph, analysis processing for a crystal, etc., and an annealing machine 300 that performs optimization (ground state search) in the Ising model equation. Fig.19 In the example shown, the computer 200 and the annealing machine 300 in the crystal analysis apparatus 100 are connected via a network 400 .

[0314] exist Fig.19 In the illustrated example, for example, as the control unit 101 a of the computer 200 , a CPU or the like can be used, and as the control unit 101 b of the annealing machine 300 , an annealing method (annealing) dedicated device can be used.

[0315] exist Fig.19 In the example shown, the computer 200 performs various settings for creating a graph, for crystal analysis processing, and for defining the Ising model equation. Then, regarding the weight (w ij ) and deviation (b i ) is transmitted from the computer 200 to the annealing machine 300 via the network 400.

[0316] Next, the annealer 300 optimizes (minimizes) the Ising model equation based on the received information about the values ​​of the weight (wij) and the bias (bi), and obtains the minimum value of the Ising model equation and the state (state) of the bit that gives the minimum value. Then, the obtained minimum value of the Ising model equation and the obtained state (state) of the bit that gives the minimum value are transmitted from the annealer 300 to the computer 200 via the network 400.

[0317] The computer 200 then analyzes the crystal based on the state of the bits that give the minimum value to the received Ising model equation.

[0318] Fig. 20 An example of the functional configuration of the crystal analysis device disclosed in the present application is described.

[0319] like Fig. 20 As shown, the crystal analysis device 100 includes a communication function unit 120 , an input function unit 130 , an output function unit 140 , a display function unit 150 , a storage function unit 160 , and a control function unit 170 .

[0320] The communication function unit 120 transmits and receives various data to and from an external device, for example. For example, the communication function unit 120 may receive graph data, data on bias and weight in the Ising model equation, and the like from an external device.

[0321] The input function unit 130 receives, for example, various instructions for the crystal analysis apparatus 100. In addition, the input function unit 130 may receive inputs such as graph data, data on biases and weights in the Ising model equation.

[0322] The output function unit 140 prints and outputs, for example, crystal analysis result data and the like.

[0323] The display function unit 150 displays, for example, crystal analysis result data and the like on a display.

[0324] The storage function unit 160 stores, for example, various programs, chart data, data on biases and weights in the Ising model equation, and the like.

[0325] The control function unit 170 has a graph creation unit 171 and an analysis unit 172 .

[0326] The graph creation unit 171 performs, for example, a process of creating a graph (expanding one repeating unit cell in a crystal to an adjacent repeating unit cell adjacent to the one repeating unit cell), etc. In addition, the graph creation unit 171 performs, for example, the following process: when the number of nodes in the unit cell that are data of an anion atom bonded to a cation atom in the one repeating unit cell is n, the number of nodes in the unit cell that are data of anion atoms in the ionic crystal is set to be less than n-1.

[0327] The analysis unit 172 performs processing such as analysis of crystals (ionic crystals) by using the graph created by the graph creation unit 171. More specifically, for example, the analysis unit 172 performs processing such as calculating similarity between the first crystal and the second crystal based on the first graph and the second graph.

[0328] Fig.21 An example of a flow chart for analyzing an ionic crystal using an example of the techniques disclosed in this application is shown.

[0329] First, the chart creation unit 171 acquires crystal structure data from a database (S1). Here, the database for acquiring the crystal structure data is not particularly limited and can be appropriately selected according to the purpose. Examples thereof include, for example, an inorganic crystal structure database (ICSD) and the like. Note that the number and type of crystal structure data acquired are not particularly limited and can be appropriately selected according to the purpose. For example, only crystal structure data about chemicals having a specific element can be acquired.

[0330] Next, the graph creation unit 171 creates a graph based on the acquired crystal structure data (S2). Note that the details of the flow when creating a graph in S2 will be described later.

[0331] Then, the chart creation unit 171 analyzes the crystal by using the created chart (S3). For example, an example of crystal analysis includes the calculation of the above-mentioned similarity. The calculation of the similarity is performed, for example, by solving the maximum independent set problem in the conflict graph represented by the Ising model equation using an annealing machine, etc. For example, the calculation of the similarity can be performed with reference to the following non-patent literature.

[0332] Non-patent literature: Maritza Hernandez, Arman Zaribafiyan, Maliheh Aramon, Mohammad Naghibi "A Novel Graph-based Approach for Determining MolecularSimilarity".arXiv:1601.06693(https: / / arxiv.org / pdf / 1601.06693.pdf).

[0333] Therefore, the similarity of the crystal structures can be evaluated, which is one of the analyses of crystal structures.

[0334] Figure 22 to Figure 26 By using an example of the technology disclosed in the present application, an example of a flowchart for creating a graph by a technology for processing a repetitive structure of a crystal as a cyclic structure is described.

[0335] First, if Fig. 22 As shown, the graph creation unit 171 determines whether the analyzed crystal is an ionic crystal (S100). In S100, in the case where the graph creation unit 171 determines that the analyzed crystal is an ionic crystal, the process shifts to S101. In the case where the graph creation unit 171 determines that the analyzed crystal is not an ionic crystal, the process ends.

[0336] Subsequently, the graph creation unit 171 reads the crystal structure data ( S101 ).

[0337] Next, the graph creation unit 171 calculates the relative coordinates of all atoms in a single unit cell (S102). More specifically, for example, in S102, the graph creation unit 171 calculates the relative coordinates of all atoms in a single unit cell to obtain the relative positional relationship of all atoms.

[0338] Then, the graph creation unit 171 sets all atoms in the single unit cell as nodes of the graph (S103). Further, the nodes prepared in S103 are intra-cell nodes in the repeated unit cell.

[0339] Here, S104 and S116 refer to executing (repeating) the processes S105 to S115 described later for all atoms corresponding to a certain atom A in a single unit cell. Similarly, S105 and S108 refer to executing (repeating) the processes S106 and S107 described later for all combinations of a certain atom (atom A) and another atom B.

[0340] In addition, for a combination of two atoms (A and B), that is, for example, a combination of a certain atom (atom A) among all atoms in a single unit cell and another atom (atom B) other than the atom, the graph creation unit 171 determines whether the distance between A and B is suitable as a bond length (S106). Then, if the distance between A and B is suitable as a bond length, the graph creation unit 171 combines the node A and the node B (S107). In other words, for example, in S107, the graph creation unit 171 provides an edge corresponding to the node A and the node B. In addition, the edge provided in S107 is an intra-unit cell edge in a repeated unit cell.

[0341] Furthermore, if the distance between A and B is not suitable as a bond length, the graph creation unit 171 shifts the process to S108 .

[0342] Later, in Fig.23 In the process, S109 and S115 refer to the process of executing (repeating) S110 and S114 described later for all combinations of a certain atom (atom A) and another atom B'.

[0343] Then, the graph creation unit 171 determines whether there is already a bond equivalent to the bond between A and B', between a certain atom (atom A) in a single unit cell and any atom B' in an adjacent unit cell (S110). If there is no bond equivalent to the bond between A and B' in S110, the graph creation unit 171 determines whether the distance between A and B' is suitable as the bond length (S111).

[0344] Furthermore, if the distance between A and B' is suitable as a bond length, the graph creation unit 171 determines whether a node corresponding to B' already exists (S112). Thereafter, if it is determined that there is no node corresponding to B', the graph creation unit 171 adds atom B' as an extended node (S113). Next, after adding atom B' as an extended node, the graph creation unit 171 combines node A and node B' (S114). In other words, for example, in S114, the graph creation unit 171 provides an edge corresponding to node A and node B'.

[0345] Furthermore, if the graph creation unit 171 determines in S112 that a node corresponding to B′ already exists, the graph creation unit 171 shifts the process to S114 and combines the node A and the node B′.

[0346] Here, in any case when it is determined in S110 that there is no bond equivalent to the bond between A and B', when it is determined in S111 that the distance between A and B' is not suitable as the bond length, and when executing the process of S114, the graph creation unit 171 transfers the process to S115, so that the process from S110 to S114 is executed (repeated) for all combinations of a certain atom (atom A) and another atom B'.

[0347] Later, in Fig.24 In the process, S117 and S120 refer to the process of executing (repeating) S118 and S119 described later in all extended nodes.

[0348] Then, the graph creation unit 171 specifies the node A′ in the unit cell corresponding to the expanded node B′ ( S118 ).

[0349] Next, the graph creation unit 171 regards the expanded node B' as the node A' (S119). That is, for example, in S119, by regarding the expanded node B' as the node A', the graph creation unit 171 sets the edge corresponding to the node A and the node A' as a loop edge.

[0350] Then, in Fig.25 In the above, S121 and S126 refer to the process of executing (repeating) the following S122 to S125 in all individual cation nodes A'. Similarly, S122 and S125 refer to the process of executing (repeating) the following S123 and S124 in all individual cation nodes B' except the cation node A'.

[0351] The graph creation unit 171 calculates n C' , which is the number of anion nodes (cation atom nodes) C' bonded to both the cation node (cation atom node) A' and the cation node B' (S123).

[0352] Next, the graph creation unit 171 sets the bond order between the cation node A' and the cation node B' to n C' ( S124 ) That is, for example, in S124 , the graph creation unit 171 specifies the bond order between the cation node A′ and the cation node B′ based on the number of the anion node C′.

[0353] Then, the graph creation unit 171 deletes all bonds (edges) between the cation nodes and the anion nodes (S127). In addition, the graph creation unit 171 deletes all anion nodes (S128). That is, for example, in S128, the graph creation unit 171 sets the number of nodes in the unit cell that are data of anion atoms to 0.

[0354] By doing this, Figures 22 to 26 In the example shown, the number of nodes included in the repeated unit cell can be reduced, and the number of bits required for analyzing the ionic crystal can be further suppressed.

[0355] Next, in Fig.26 In S129, the graph creation unit 171 calculates the total number of nodes (m) (S129). That is, for example, in S129, the graph creation unit 171 counts the number of cation nodes in the unit cell and specifies the total number of nodes (m).

[0356] Next, the graph creation unit 171 sets the ideal number of nodes of the graph to n, calculates the value of n / m, then specifies a number obtained by rounding the value of n / m, and copies the graph as many times as the specified number by multiplying the graph by an integer (S130). That is, for example, in S130, the graph creation unit 171 multiplies the graph by an integer to approach the ideal number of nodes n of the graph. Note that the ideal node n of the graph can be appropriately selected according to the target ion crystal.

[0357] Subsequently, the graph creation unit 171 outputs graph data ( S131 ).

[0358] Then, when the output of the graph data is completed, the graph creation unit 171 ends the process of creating the graph.

[0359] Figures 27 to 29 An example of a flow chart illustrating the creation of a graph by using an example of the techniques disclosed in this application, using techniques for expanding and manipulating cells using expansion nodes.

[0360] Note that in the technique of expanding and manipulating the cell by using the expansion node, Figures 22 to 26 S100 to S128 in the flow described in perform similar processes, so their description will be omitted.

[0361] In the technique of extending the unit cell by using an extension node, such as Fig. 27 As shown, similar to Fig.26 In S129 in the graph creation unit 171, the graph creation unit 171 calculates the total number of nodes (m) (S229). Next, that is, for example, in S129, the graph creation unit 171 counts the number of cation nodes in the unit cell and specifies the total number of nodes (m).

[0362] Next, if Fig. 27 As shown, Fig.26 Similar to S130 in, the graph creation unit 171 sets n to the ideal number of nodes of the graph, calculates the value of n / m, then specifies a number obtained by rounding the value of n / m, and copies the graph as many times as the specified number by multiplying the graph by an integer (S230).

[0363] exist Fig.28 In the example, S231 and S234 mean that the processes of S232 and S233 described later are executed (repeatedly) in all individual cation nodes A'.

[0364] The graph creation unit 171 determines whether there is a node having a bonding relationship with the atom corresponding to A' among atoms in the adjacent unit cell that is not the target of the graph (S232). In S232, if the graph creation unit 171 determines that there is a node having a bonding relationship with the atom corresponding to A', the process shifts to S233. If the graph creation unit 171 determines that there is no node having a bonding relationship with the atom corresponding to A', the process shifts to S234.

[0365] Next, the graph creation unit 171 sets the atom determined in S232 as a node having a bonding relationship with the atom corresponding to A′ as a node B′, and adds to the list B″.

[0366] Subsequently, the graph creation unit 171 deletes redundant elements among the elements (nodes) in the list B″ ( S235 ). Next, the graph creation unit 171 adds the nodes included in the list B″ to the graph ( S236 ).

[0367] exist Fig.29 In the above, S237 and S239 refer to the process of executing (repeating) S238 described later in all individual nodes B' included in the list B".

[0368] The graph creation unit 171 adds a bond (edge) between the node B′ and the node C′ for each node C having a bond relationship with the node B′ in the graph ( S238 ).

[0369] Subsequently, the graph creation unit 171 outputs the graph data ( S240 ).

[0370] Then, when the output of the graph data is completed, the graph creation unit 171 ends the process of creating the graph.

[0371] In addition, Figures 21 to 29 In the examples of the technology disclosed in the present application, the processing flow in the examples of the technology disclosed in the present application has been described according to a specific order. However, in the technology disclosed in the present application, the order of each step can be appropriately switched within the technically possible range. In addition, in the technology disclosed in the present application, multiple steps can be performed together within the technically possible range.

[0372] Examples of the annealing method and the annealing machine will be described below.

[0373] Annealing is a method for solving problems in a probabilistic manner using a superposition of random values ​​and qubits. The following is an example of minimizing the value of an evaluation function to be optimized. The value of the evaluation function is called energy. In addition, in the case of maximizing the value of the evaluation function, only the sign of the evaluation function needs to be changed.

[0374] First, the process starts from an initial state, where one of the discrete values ​​is assigned to each variable. For the current state (combination of variable values), a state close to the current state is selected (for example, a state where only one variable is changed), and the state transitions therebetween are considered. The energy change relative to the state transition is calculated. Based on this value, it is probabilistically determined whether to adopt a state transition to change the state or to maintain the initial state without adopting a state transition. In the case where the adoption probability when the energy decreases is selected to be greater than the adoption probability when the energy increases, it can be expected that a state transition will occur in the direction of an average decrease in energy, and over time, the state transition will occur to a more suitable state. Therefore, there is a probability of eventually obtaining an optimal solution or an approximate solution in which the energy is close to the optimal value.

[0375] If this is adopted when the energy is deterministically decreasing but not adopted when the energy is increasing, the energy change is generally monotonically decreasing with respect to time, but no further change occurs when a local solution is reached. As mentioned above, since there are very many local solutions in discrete optimization problems, it is almost certain that you will get stuck in a state in a local solution that is not very close to the optimal value. Therefore, when solving discrete optimization problems, it is important to probabilistically determine whether to adopt this state.

[0376] In the annealing method, it has been proven that by determining the adoption (allowed) probability of state transition as follows, the state reaches the optimal solution within the limit of infinite time (number of iterations).

[0377] In the following, a method of finding an optimal solution using the annealing method will be described step by step.

[0378] (1) For the energy change (energy decrease) value (-ΔE) due to the state transition, the allowed probability p of the state transition is determined by any of the following functions f(·).

[0379] [Formula 9]

[0380] p(ΔE, T) = f(-ΔE / T) (Calculation formula 1-1)

[0381] [Formula 10]

[0382] f metro (x) = min(1, e x ) (METROPOLIS method) (Calculation formula 1-2)

[0383] [Formula 11]

[0384]

[0385] Here, T represents a parameter called a temperature value, and can be changed, for example, as follows.

[0386] (2) The temperature value T decreases logarithmically with respect to the iteration count t, as shown in the following calculation formula.

[0387] [Formula 12]

[0388]

[0389] Here, T0 is an initial temperature value, and is ideally a sufficiently large value depending on the problem.

[0390] In the case of using the allowed probability represented by the calculation formula in (1), if a steady state is reached after sufficient iterations, the occupation probability of each state follows the Boltzmann distribution for the thermal equilibrium state in thermodynamics.

[0391] Then, when the temperature is gradually reduced from high temperature, the occupation probability of the low energy state increases. Therefore, it is considered that a low energy state is obtained when the temperature is sufficiently reduced. Since this state is very similar to the state change caused by material annealing, this method is called annealing (or pseudo annealing). Note that the probability of increasing the state transition of energy corresponds to thermal excitation in physics.

[0392] Fig.30 An example of a functional configuration of an annealing machine that performs the annealing method is explained. However, in the following description, a case where a plurality of state transition candidates are generated is also described, but the basic annealing method generates one transition candidate at a time.

[0393] The annealing machine 300 includes a state holding unit 111 that holds a current state S (a plurality of state variable values). In addition, the annealing machine 300 includes an energy calculation unit 112 that calculates energy change values ​​{-ΔEi} for respective state transitions when a state transition from the current state S occurs due to a change in any one of the plurality of state variable values. In addition, the annealing machine 300 includes a temperature control unit 113 that controls a temperature value T and a transition control unit 114 that controls a state change. Note that the annealing machine 300 may be a part of the above-described crystal analysis apparatus 100.

[0394] The transition control unit 114 probabilistically determines whether to accept any one of the plurality of state transitions based on the temperature value T, the energy change value {-ΔEi} and the random value according to the relative relationship between the energy change value {-ΔEi} and the thermal excitation energy.

[0395] Here, the conversion control unit 114 includes a candidate generation unit 114a that generates state transition candidates and an availability determination unit 114b that probabilistically determines whether to allow state transition of each candidate based on the energy change value {-ΔEi} and the temperature value T. In addition, the conversion control unit 114 includes a conversion determination unit 114c that determines a candidate to be adopted from the candidates that have been allowed and a random number generation unit 114d that generates a random variable.

[0396] The operation of the annealer 300 in one iteration is as follows.

[0397] First, the candidate generation unit 114a generates one or more state transition candidates (candidate number {Ni}) from the current state S held in the state holding unit 111 to the next state. Next, the energy calculation unit 112 calculates the energy change value {-ΔEi} of each state transition listed as a candidate by using the current state S and the state transition candidate. The availability determination unit 114b uses the temperature value T generated by the temperature control unit 113 and the random variable (random value) generated by the random number generation unit 114d to allow state transitions according to the energy change value {-ΔEi} of each state transition using the allowable probability of the above calculation formula (1).

[0398] Then, the availability determination unit 114b outputs the availability {fi} of each state transition. In the case where there are multiple allowed state transitions, the transition determination unit 114c randomly selects one of the allowed state transitions using a random value. Then, the transition determination unit 114c outputs the transition number N and the transition availability f of the selected state transition. In the case where there are allowed state transitions, the state variable value stored in the state holding unit 111 is updated according to the adopted state transition.

[0399] Starting from the initial state, the above iterations are repeated while the temperature control unit 113 decreases the temperature value. When completion determination conditions such as reaching a certain number of iterations or energy falling below a certain value are met, the operation is completed. The answer output by the annealing machine 300 is the state when the operation is completed.

[0400] Fig.30 The annealing machine 300 shown can be implemented by using, for example, a semiconductor integrated circuit. For example, the conversion control unit 114 may include a random number generation circuit used as the random number generation unit 114d, a comparison circuit used as at least a part of the availability determination unit 114b, or a noise table described later.

[0401] about Fig.30 The transition control unit 114 shown in FIG. 1 will further explain the details of the mechanism for allowing state transition with the allowed probability expressed in the calculation formula (1).

[0402] A circuit that outputs 1 with a probability p and outputs 0 with a probability (1 - p) can be implemented by inputting into a comparator a uniform random number that takes a value in the interval [0, 1) for the input probability p of input A and for input B. The comparator has two inputs A and B, and outputs 1 when A > B and outputs 0 when A < B. Therefore, if the value of the probability p calculated based on the energy change value and the temperature value T using the calculation formula in (1) is input to input A of this comparator, the above function can be implemented.

[0403] In other words, for example, using a circuit that outputs 1 when f(ΔE / T) is greater than u, where f is the function used in the calculation formula in (1) and u is a uniform random number that takes a value in the interval [0, 1), the above function can be implemented.

[0404] In addition, the same function as the above can also be implemented by making the following modifications.

[0405] Applying the same monotonically increasing function to two numbers does not change the magnitude relationship. Therefore, even if the same monotonically increasing function is applied to the two inputs of the comparator, the output will not change. If the inverse function f -1 of f is used as this monotonically increasing function, it can be seen that a circuit that outputs 1 when -ΔE / T is greater than f -1 (u) can be used. Moreover, since the temperature value T is positive, it can be seen that a circuit that outputs 1 when -ΔE is greater than Tf -1 (u) can be used.

[0406] Fig.30 The conversion control unit 114 in -1 can include a noise table, which is a conversion table that implements the inverse function f

[0407] [Equation 13]

[0408]

[0409] [Equation 14]

[0410]

[0411] Fig.31 is a diagram showing an example of the operation flow of the conversion control unit 114. Fig.31The operation flow shown includes a step of selecting a state transition as a candidate (S0001), a step of determining the availability of the state transition by comparing the energy change value of the state transition with the product of the temperature value and the random value (S0002), and a step of adopting the state transition when the state transition is available and not adopting the state transition when the state transition is not available (S0003).

[0412] Example

[0413] Embodiments of the technology disclosed in the present application will be described. However, the technology disclosed in the present application is not limited to these embodiments.

[0414] (Example 1)

[0415] As Example 1, the similarity of ionic crystals was calculated using an example of the crystal analysis apparatus disclosed in the present application.

[0416] The similarity of the ionic crystal in Example 1 was determined by using Fig.19 The hardware configuration shown is a crystal analysis device based on Figures 21 to 26 Note that in Example 1, without executing Fig.26 A graph is created without multiplying the graph by an integer in S129 and S130 in , and the similarity is calculated.

[0417] More specifically, for example, in Example 1, a graph is expanded using a technique for treating a repetitive structure of an ionic crystal as a cyclic structure, and the following calculation formula (1) is minimized by an annealing method using Digital Annealer (registered trademark) as an annealing machine.

[0418] [Formula 15]

[0419]

[0420] Here, in the calculation formula (1), E is the Hamiltonian function, where minimizing E means searching for the maximum independent set, c' ij is a coefficient representing the weight of the key order between the i-th node and the j-th node in the conflict graph between the first graph and the second graph, x i is a binary variable indicating whether the i-th node has 0 or 1, x j is a binary variable indicating that the j-th node has 0 or 1, and α and β are positive numbers.

[0421] Furthermore, in Embodiment 1, for the obtained maximum independent set, the similarity is calculated by using the following calculation formula (2).

[0422] [Formula 16]

[0423]

[0424] In the above calculation formula (2), S(G A ,G B )) represents the similarity between the first crystal represented by the first graph and the second crystal represented by the second graph, and is expressed by 0 to 1, meaning that the closer to 1, the higher the similarity, V A represents the total number of node atoms in the first crystal represented by the first graph, V c A represents the number of node atoms included in the maximum independent set of the conflict graph in the node atoms in the first crystal represented by the first graph, V B represents the total number of nodal atoms in the second crystal represented by the second graph, V c B represents the number of node atoms included in the maximum independent set of the conflict graph among the node atoms in the second crystal represented by the second graph, and δ is a number between 0 and 1.

[0425] Here, in Example 1, the similarity is calculated for ionic crystal A whose number of atoms included in the repeating unit cell is 76 and ionic crystal B whose number of atoms included in the repeating unit cell is 96. Note that the structure of the repeating unit cell of ionic crystal A is as described above. Fig.15A The structure shown in the figure, the structure of the repeating unit cell of the ionic crystal B is as above Fig. 15B The structure shown.

[0426] Here, if Fig.15A As shown, ionic crystal A has 8 lithium (Li), 4 manganese (Mn), 16 phosphorus (P) and 48 oxygen (O) atoms in the unit cell, for a total of 76 atoms. Fig. 15B As shown, ionic crystal B has 16 lithium (Li), 8 manganese (Mn), 16 phosphorus (P) and 56 oxygen (O) atoms in the unit cell, for a total of 96 atoms.

[0427] Note that the ionic crystal A is a crystal of "ICSD#253225" whose composition formula is represented by "Li2MnP2O7", and the ionic crystal B is a crystal of "ICSD#248405" whose composition formula is "Li2Mn(PO3)4".

[0428] In Example 1, a conflict graph is created after deleting nodes of 48 oxygen atoms included in the repeated unit cell of ionic crystal A and nodes of 56 oxygen atoms included in the repeated unit cell of ionic crystal B. That is, for example, in Example 1, the number of nodes is set to 28 by deleting nodes of 48 oxygen atoms from ionic crystal A having 76 nodes, and the number of nodes is set to 40 by deleting nodes of 56 oxygen atoms from ionic crystal B having 96 nodes.

[0429] Fig.32 The following is a diagram summarizing the results of Example 1 and Comparative Example 1.

[0430] In Example 1, the number of nodes in the conflict graph (the number of bits required for calculation) is "8×16=128" for lithium nodes, "4×8=32" for manganese nodes, and "16×16=256" for phosphorus nodes. Therefore, in Example 1, the number of nodes in the conflict graph is "128+32+256=416".

[0431] However, when the similarity is calculated without deleting the oxygen atom nodes included in the unit cells of ionic crystal A and ionic crystal B (Comparative Example 1; Comparative Example of Example 1), the number of nodes required to calculate the similarity as described above is "3104", which is much larger than that of Example 1.

[0432] Furthermore, in Example 1, for the ionic crystal A and the ionic crystal B, the number of atoms as a common substructure is "18", and the similarity calculated based on the above calculation formula (2) is "0.546".

[0433] In the comparative example of Example 1 (Comparative Example 1), the number of atoms in the common substructure is "55", and the similarity calculated based on the above calculation formula (2) is "0.648".

[0434] From these results, it can be confirmed that the difference in similarity between Example 1 and Comparative Example 1 is about 0.1, which is a relatively close value.

[0435] (Example 2)

[0436] In Example 2, in addition to Fig.26 Except for S129 and S130 in , the similarity between the ionic crystal A and the ionic crystal B is calculated similarly to Example 1, and the graph is multiplied by an integer, and then the similarity is calculated.

[0437] Specifically, for example, for ionic crystal A and ionic crystal B in which all oxygen atom nodes are deleted, the difference between the number of nodes included in the unit cells of the respective crystals is reduced by doubling the unit cell of ionic crystal A so that the number of nodes is "80" and tripling the unit cell of ionic crystal B so that the number of nodes is "84".

[0438] Fig.33 This is a diagram summarizing the results of Example 2.

[0439] In Example 2, for ionic crystal A and ionic crystal B, the number of bits required for calculation is "2496", the number of atoms as a common substructure is "43", and the similarity calculated based on the above calculation formula (2) is "0.525".

[0440] Since this similarity is calculated by making the number of nodes in the unit cells in each crystal uniform, it can be considered that the accuracy is higher than the similarity in Example 1.

[0441] Furthermore, in Example 2, as described above, the number of bits required for calculation is "2496", and after unifying the number of nodes in the unit cells in the respective crystals, the similarity can be calculated with a smaller number of bits than in Comparative Example 1 in which oxygen atoms are not deleted.

[0442] (Example 3)

[0443] In Example 3, similarity is calculated in a similar manner to Example 1 except that the ionic crystal as the target of similarity calculation is a crystal (6 types) represented by the composition formula "Li2MP2O7" (M=Fe, Co, Mn, Ba) and the oxygen atom node is deleted.

[0444] Furthermore, as a comparative example (Comparative example 2) corresponding to Example 3, the similarity was calculated without deleting the node of the oxygen atom.

[0445] Fig.34 The relationship between the type of ionic crystal and the calculated similarity in Example 3 and Comparative Example 2 is described.

[0446] Here, in Fig.34 Among the six ionic crystals shown, the crystals "M=Ba" (ICSD#107724 and ICSD#41433) are crystals having a different structural type from the other crystals.

[0447] like Fig.34 As shown, in Example 3, the similarity between the crystals where M is any one of Fe, Co, or Mn is high, and the similarity between the crystals where M is Ba is also high. However, the similarity between the crystals where M is any one of Fe, Co, or Mn and the crystals where M is Ba is low, and it can be seen that the classification has been correctly performed according to the type of crystal structure.

[0448] also, Fig.35 The relationship between the type of ionic crystal and the number of bits required to calculate the similarity in Example 3 and Comparative Example 2 is described.

[0449] like Fig.35 As shown, the number of bits required to calculate the similarity is 656 or more and 832 or less in Example 3, and is 3792 or more and 3968 or less in Comparative Example 2. It can be confirmed that in Example 3, the number of bits required for calculation can be reduced, and a high-accuracy similarity that can be correctly classified according to the type of crystal structure can be calculated.

[0450] (Example 4)

[0451] In Example 4, the similarity is calculated similarly to Example 3, except that the ionic crystal as the target of calculating the similarity is a crystal represented by the respective composition formulas of CsCl, KCl, NaCl, NaF, LiF and LiCl and the halogen atom (Cl or F) as an example of the anion atom is deleted.

[0452] Note that, in the ionic crystals that are the targets of Example 4, CsCl is ICSD#257256, KCl is ICSD#28938, NaCl is ICSD#28048, NaF is ICSD#29128, LiF is ICSD#44879, and LiCl is ICSD#52235.

[0453] In Example 4, for CsCl (ICSD#257256), two unit cells in each direction of the a-axis, b-axis, and c-axis are represented as a graph, and for other crystals, two unit cells in each direction of the a-axis and b-axis and one unit cell in the c-axis direction are represented as a graph. The number of nodes in each graph is 8.

[0454] Fig.36 The relationship between the type of ionic crystal and the calculated similarity in Example 4 is described.

[0455] Here, in Fig.36 Among the six ionic crystals shown, CsCl (ICSD#257256) has a simple cubic unit cell, while the other crystals have a face-centered cubic unit cell.

[0456] like Fig.36 As shown, in Example 4, although the similarity between crystals as face-centered cubic unit cells is high, the similarity between crystals of simple cubic unit cells and crystals of face-centered cubic unit cells is low, which indicates that correct classification has been performed according to the type of crystal structure.

Claims

1. A crystal analysis method for executing the following process on a computer, the process comprising: creating a graph indicating data of a repeating unit cell in an ionic crystal and data of adjacent repeating unit cells adjacent to the repeating unit cell, the graph being a conflict graph; analyzing the ionic crystal according to the graph; and When the number of nodes in the first unit cell indicating data of anion atoms bonded to cation atoms in the repeating unit cell is n, the number of nodes in the second unit cell indicating data of anion atoms in the repeating unit cell is set to n-1 or less, wherein The data of the repeating cell includes a plurality of intra-cell nodes indicating data of atoms in the repeating cell, and The plurality of intra-unit-cell nodes include the first intra-unit-cell node and the second intra-unit-cell node, wherein the anion atom is removed by setting the number of nodes in the second unit cell to 0, Wherein, the creation includes: creating a first graph for a first crystal; and Create a second chart for the second crystal, The analysis includes calculating the similarity between the first crystal and the second crystal based on the first graph and the second graph, The process further includes: When the cation atoms in the repeating unit cell are chemically bonded to each other, the bond order of the chemical bonds between the cation atoms is determined as the number of chemical bonds between the cation atoms via the anion atoms whose number is zero.

2. The crystal analysis method according to claim 1, wherein The data of the repeating unit cell includes an inner edge of the unit cell indicating data of a chemical bond between two atoms in the repeating unit cell, The data of the adjacent repeating unit cell includes an extended node indicating data of atoms in the adjacent repeating unit cell and an extended edge indicating data of a chemical bond between an atom corresponding to an intra-cell node in the repeating unit cell and an atom corresponding to the extended node.

3. The crystal analysis method according to claim 1, wherein The graph includes loop edges of data indicating virtual chemical bonds between atom X in the repeating cell and atom Y in the repeating cell, the atom Y corresponding to atom Y′ in the adjacent repeating cell to which the atom X is chemically bonded.

4. The crystal analysis method according to claim 1, wherein The calculating includes multiplying at least one of the repeating unit cells of the first graph or the repeating unit cells of the second graph by an integer such that the number of nodes within the unit cells in the first graph and the number of nodes within the unit cells in the second graph are similar.

5. The crystal analysis method according to claim 1, wherein The calculation includes searching for a maximum independent set of conflict graphs between the first graph and the second graph by using the following calculation formula 1: [Calculation formula 1] in, In formula 1, E is a Hamiltonian function, where minimizing E means searching for the maximum independent set, c' ij is the coefficient representing the weight of the key order between the i-th node and the j-th node in the conflict graph, x i is a binary variable indicating whether the i-th node is 0 or 1, x j is a binary variable indicating whether the jth node is 0 or 1, and α and β are positive numbers.

6. The crystal analysis method according to claim 5, wherein The search includes minimizing the Hamiltonian function in the above calculation formula 1 by an annealing method.

7. The crystal analysis method according to claim 1, wherein The calculation involves searching for the maximum independent set using the following calculation formula 2: [Calculation formula 2] in, In formula 2, S(G A ,G B ) is a number between 0 and 1, indicating the similarity between the first crystal and the second crystal, V A represents the total number of atoms in the first crystal, V c A represents the number of atoms in the first crystal contained in the largest independent set, V B represents the total number of atoms in the second crystal, V c B represents the number of atoms in the second crystal contained in the largest independent set, and δ is a number between 0 and 1.

8. A crystal analysis device, comprising: a graph creating unit that creates a graph indicating data of a repeating unit cell in an ionic crystal and data of adjacent repeating unit cells adjacent to the repeating unit cell, the graph being a conflict graph; an analyzing unit that analyzes the ionic crystal according to the graph and sets the number of second unit cell nodes indicating data of anionic atoms bonded to cation atoms in the repeating unit cell to n-1 or less when the number of first unit cell nodes indicating data of anionic atoms bonded to cation atoms in the repeating unit cell is n, wherein The data of the repeating cell includes a plurality of intra-cell nodes indicating data of atoms in the repeating cell, and The plurality of intra-unit-cell nodes include the first intra-unit-cell node and the second intra-unit-cell node, wherein the anion atoms are removed by setting the number of nodes in the second unit cell to 0, and when the cation atoms in the repeating unit cell are chemically bonded to each other, the bond order of the chemical bonds between the cation atoms is determined to be the number of chemical bonds between the cation atoms via the anion atoms whose number is 0, The chart creation unit creates a first chart for the first crystal and creates a second chart for the second crystal in the chart creation. The analyzing unit calculates the similarity between the first crystal and the second crystal based on the first graph and the second graph during the analysis.

Citation Information

Patent Citations

  • Method for predicting photoelectric characteristics of Sr, Ba, La and Er doped c-ZrO2 under high pressure

    CN110046445A

  • Method for describing material crystal structure and application thereof

    CN110176280A

  • Crystallographic metrology and process control

    US20060231752A1

  • Method and system for syncing data structures

    US20160055226A1

  • Device, system, method and program for producing fragment model

    US20160078155A1