Catalyst evaluation system, catalyst evaluation program and catalyst evaluation method

Abstract

To search for a catalyst indicating a high evaluation index for a reactant more efficiently.SOLUTION: A catalyst evaluation system of the present invention comprises an output unit that generates a pair of replacement models including an initial state replacement model in which a structure of an initial state basic model has been replaced and a final state replacement model in which a structure of a final state basic model has been replaced on the basis of a pair of basic models including the initial state basic model corresponding to a reactant of a catalytic reaction and the final state basic model corresponding to the product of the catalytic reaction, and replacement information related to the replacement of the structures of the pair of basic models, and matches physical property values related to the pair of replacement models with the pair of replacement models to output a catalyst of the catalytic reaction related to the reactant and the product.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to a catalyst evaluation system, a catalyst evaluation program, and a catalyst evaluation method. [Background technology]

[0002] Catalysts are used in many chemical reactions, and methods are used to search for catalysts that promote target reactions using computational chemistry. Since the use of computational chemistry reduces the labor, cost, and time required to search for catalysts compared to actual experiments, various methods for searching for catalysts using computational chemistry are being considered.

[0003] As a method for searching for catalysts using computational chemistry, for example, an automatic compound structure generation device has been disclosed that performs machine learning to automatically generate a compound structure derived from a compound structure that serves as the basis for automatic generation, generates an automatic compound structure generation model, and performs a process of automatically generating a new compound structure using the automatic compound structure generation model (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent Publication No. 2021-81769 Summary of the Invention [Problem to be solved by the invention]

[0005] However, in devices that use conventional computational chemistry methods to search for catalysts, such as the device in Patent Document 1, when searching for a catalyst with a high evaluation index for a reactant that will further promote a target reaction from among the countless catalysts that exist, depending on the system being handled and the physical property values ​​being calculated, there is a problem that the calculation costs can increase and it can take a long time (e.g., several days) to calculate the physical property values.

[0006] For example, there are countless combinations of catalysts, but when searching for chemical reactions of hundreds to tens of thousands of types of catalysts using atomic systems of several hundred atoms, the calculation load becomes extremely large compared to when using atomic systems of a dozen atoms, making it difficult to perform calculations in a realistic time frame. Furthermore, when calculating additional physical properties such as activation energy, which require high calculation costs, the calculation load becomes even larger and the calculation time increases.

[0007] An object of the present invention is to search for a catalyst that exhibits a high evaluation index for a reactant more efficiently. [Means for solving the problem]

[0008] One aspect of the present invention is A catalyst evaluation system that generates a pair of replacement models including an initial state replacement model in which the structure of the initial state basic model has been replaced and a final state replacement model in which the structure of the final state basic model has been replaced based on a pair of basic models including an initial state basic model corresponding to reactants of a catalytic reaction and a final state basic model corresponding to products of the catalytic reaction, and replacement information related to replacement of structures of the pair of basic models, and has an output unit that matches physical property values ​​related to the pair of replacement models with the pair of replacement models and outputs catalysts of the catalytic reaction related to the reactants and the products.

[0009] Another aspect of the present invention is On the computer, A catalyst evaluation program that executes a process of generating a pair of replacement models including an initial state replacement model in which the structure of the initial state basic model has been replaced and a final state replacement model in which the structure of the final state basic model has been replaced based on a pair of basic models including an initial state basic model corresponding to reactants of a catalytic reaction and a final state basic model corresponding to products of the catalytic reaction, and replacement information related to replacement of structures of the pair of basic models, and matching physical property values ​​related to the pair of replacement models with the pair of replacement models, and outputting catalysts of the catalytic reaction related to the reactants and the products.

[0010] Another aspect of the present invention is The computer A catalyst evaluation method that generates a pair of replacement models including an initial state replacement model in which the structure of the initial state basic model has been replaced and a final state replacement model in which the structure of the final state basic model has been replaced based on a pair of basic models including an initial state basic model corresponding to reactants of a catalytic reaction and a final state basic model corresponding to products of the catalytic reaction, and replacement information related to replacement of structures of the pair of basic models, and performs a process of matching physical property values ​​related to the pair of replacement models with the pair of replacement models, and outputting catalysts of the catalytic reaction related to the reactants and the products. Effect of the Invention

[0011] The present invention makes it possible to more efficiently search for catalysts that exhibit high evaluation indexes for reactants. [Brief description of the drawings]

[0012] [Figure 1] 1 is a diagram showing a configuration of a catalyst evaluation system according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a functional block diagram showing a configuration of a catalyst evaluation device. [Diagram 3] FIG. 1 is a diagram illustrating an example of a pair of base models. [Figure 4] 4 is a diagram showing an example of a pair of substituted models obtained by substituting the pair of basic models shown in FIG. 3. FIG. [Diagram 5] FIG. 11 is an explanatory diagram showing an example of a method for generating a substitution model. [Figure 6] FIG. 1 is an explanatory diagram showing the relationship between the transition state in a catalytic reaction with an alkene coordinated to the surface of a pair of substitution models and the activation energy when passing through the transition state. [Figure 7] FIG. 2 is an explanatory diagram of activation energy and adsorption energy. [Figure 8] FIG. 1 is a diagram showing an example of a plot of the relationship between adsorption energy and activation energy. [Figure 9]2 is a block diagram showing a hardware configuration of the catalyst evaluation device. FIG. [Figure 10] 1 is a flowchart showing an example of a catalyst evaluation method according to an embodiment of the present invention. [Figure 11] FIG. 4 is a functional block diagram showing another configuration of the catalyst evaluation device. [Figure 12] FIG. 2 is a diagram showing the relationship between adsorption energy and activation energy in the catalytic reaction of the catalyst of Example 1. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] Hereinafter, embodiments of the present invention will be described in detail. In this specification, unless otherwise specified, the numerical range indicated by "to" means that the numerical range includes the numerical range before and after the range as the lower and upper limits.

[0014] <Catalyst evaluation system> A catalyst evaluation system according to an embodiment of the present invention will be described. The catalyst evaluation system according to the present embodiment evaluates catalysts that undergo a catalytic reaction with a reactant to produce a predetermined target product from the reactant, and is used to search for catalyst candidates.

[0015] FIG. 1 is a diagram showing the configuration of a catalyst evaluation system according to this embodiment. As shown in FIG. 1, the catalyst evaluation system 1 includes a catalyst evaluation device 10, a storage unit 20, and a machine learning potential 30. In the catalyst evaluation system 1, the catalyst evaluation device 10, the storage unit 20, and the machine learning potential 30 are connected via a communication network 40, and input values ​​to the catalyst evaluation device 10, the storage unit 20, and the machine learning potential 30 and output values ​​of the catalyst evaluation device 10, the storage unit 20, and the machine learning potential 30 may be transmitted via the communication network 40. At least one of the storage unit 20 and the machine learning potential 30 may be stored on the cloud.

[0016] In this embodiment, the catalyst evaluation device 10, the storage unit 20, and the machine learning potential 30 are connected via a communication network 40, but may be connected by wire. Also, the catalyst evaluation system 1 may be a standalone device such as a PC (Personal Computer) that includes each component within the device.

[0017] The catalyst evaluation device 10 evaluates candidates for catalysts that catalytically react with reactants by using the machine learning potential 30. The catalyst evaluation device 10 will be described in detail later.

[0018] The memory unit 20 stores information on the catalyst and reactants, a pair of basic models including an initial state basic model having a basic structure in the initial state representing the catalyst as a model and a final state basic model having a basic structure in the final state, types of hydrocarbons used to replace parts of the pair of basic models, types of energy, etc.

[0019] A catalyst is a substance that accelerates the reaction rate of a specific chemical reaction, and does not change itself before or after the reaction. The catalyst includes homogeneous catalysts that can be dissolved in a solution such as a non-aqueous organic solvent, heterogeneous catalysts that do not dissolve in a solution and exist as a solid, and homogeneous catalysts fixed to a support or heterogeneous catalyst. Examples of homogeneous catalysts include compounds containing organic alkali metals and organic alkaline earth metals. Heterogeneous catalysts are alloys or metal oxides that contain multiple metal element species. Specific examples of heterogeneous catalysts include metal catalysts, metal compound catalysts such as metal oxides, zeolites, catalysts in which metals or metal compounds are supported or ion-exchanged on a support, metal complexes, activated carbon, etc. Examples of alloys include γ-Al 2 O 3 / MoOx. The metal oxide is, for example, Al 2 O 3 , MoOx (x is a natural number, a value corresponding to the ratio of Mo to O), etc.

[0020] The catalyst information includes information on the components that make up the catalyst.

[0021] When the component constituting the catalyst is, for example, a metal, an alloy, or a metal oxide, information on the crystal structure of the component constituting the catalyst may be included.

[0022] The crystal structure of the catalyst component is, for example, Al. 2 O 3 If α-Al 2 O 3 , γ-Al 2 O 3 etc.

[0023] Reactants are components used in producing a desired product.

[0024] The information on the reactants includes, for example, the type of the reactant, such as 3-Ethyl-1-pentene.

[0025] A pair of basic models, including the initial state basic model and the final state basic model, are surface models that show a structure in which molecules present on the surface of the catalyst and in its vicinity are stacked and arranged in several layers (e.g., one layer), and show a state in which several surface molecules (e.g., one) involved in the reaction with the reactant are arranged on the surface.

[0026] The initial state refers to a state in which the molecules (surface molecules) involved in the reaction with the reactant and the reactant are arranged in the same surface model and are coordinated without any bond recombination. The final state refers to a state in which, among the surface molecules present on the surface in the surface model, the surface molecules involved in the reaction with the reactant have reacted with the reactant and bond recombination has occurred.

[0027] The initial state structure (basic structure) of the initial state basic model refers to the state in which the surface molecules involved in the reaction are arranged on the surface before the reaction with the reactant, and the final state structure (basic structure) of the final state basic model refers to the state after the surface molecules involved in the reaction on the surface have reacted with the reactant.

[0028] The types of energy include activation energy, adsorption energy, and the like.

[0029] The machine learning potential 30 is an interatomic potential using a machine learning technique that outputs energy from information on the structure of atoms. Examples of the machine learning potential include a neural network potential (NNP), a Gaussian approximation potential (GAP), a spectral neighbor analysis potential (SNAP), and a moment tensor potential (MTP). Among these, the machine learning potential is preferably an NNP in terms of the high flexibility of a neural network. As the NNP, an atomic simulator that learns the relationship between atomic coordinates and energy using quantum chemical calculations as teaching data can be used. As the NNP, a preferred potential (PFP) may be used.

[0030] [Catalyst evaluation equipment] The catalyst evaluation device 10 will be described. In this example, the catalyst is aluminum oxide (γ-Al 2 O 3 ) and metal composite oxide with MoOx (γ-Al 2 O 3 / MoOx alloy) and the reactant is ethylene, a type of alkene. 2 O 3 The structure is that the reactant alkene is coordinated to the Mo of the surface molecule of the / MoOx alloy. The substitution model is γ-Al 2 O 3 The hydrogen atoms of alkenes on the surface of the / MoOx alloy are replaced by methyl groups, which are the substituents, to form γ-Al 2 O 3 The Mo in the surface molecules of the MoOx alloy is substituted with propylene (CH 3 CHCH 2 The substituent may be any hydrocarbon.

[0031] Fig. 2 is a functional block diagram showing the configuration of the catalyst evaluation device 10. As shown in Fig. 2, the catalyst evaluation device 10 has an acquisition unit 11, an output unit 12, and a selection result output unit 13, and evaluates catalysts that undergo catalytic reactions with reactants to search for catalysts that exhibit high evaluation indexes.

[0032] The evaluation index may be any index related to a catalytic reaction with a reactant, such as reactivity during a catalytic reaction with a reactant. When the evaluation index is, for example, reactivity, the catalyst evaluation device 10 searches for a catalyst having high reactivity with the reactant.

[0033] The acquisition unit 11 acquires a pair of basic models including an initial state basic model corresponding to a reactant of a catalytic reaction and a final state basic model corresponding to a product of the catalytic reaction. That is, the acquisition unit 11 acquires an initial state basic model having a base initial state structure as a basic structure, and a final state basic model having a base final state structure as a basic structure, as a pair of basic models. Note that the number of atoms constituting the initial state basic model and the final state basic model of the pair of basic models is not particularly limited and may be set arbitrarily as appropriate.

[0034] An example of a pair of basic models is shown in Figure 3. As shown in Figure 3, for example, γ-Al 2 O 3 / Ethylene (C) is added to Mo contained in the surface molecules of the MoOx alloy. 2 H 4 A pair of basic models including an initial state basic model and a final state basic model in which the atoms (A, B, C, D, E ...

[0035] The acquisition unit 11 may allow a user to acquire the basic model from the storage unit 20, an external database, or the like.

[0036] 2 corresponds the physical property values ​​related to the pair of substituted models substituted based on the substitution information related to the substitution of the structures of the pair of basic models to the pair of substituted models, and outputs a catalyst of a catalytic reaction related to a reactant and a product. That is, the output unit 12 evaluates the catalyst and searches for a model having a high evaluation index for the reactant, such as being effective for the catalytic reaction.

[0037] The substitution information refers to reactants, catalysts, and substitution structures of surface molecules that substitute for the substituted structures of the surface molecules included in the pair of basic models.

[0038] The physical property value related to the pair of substitution models refers to the energy of the pair of substitution models, etc.

[0039] The output unit 12 includes a structure generation unit 121 and a structure selection unit 122 .

[0040] The output unit 12 preferably executes calculations using the machine learning potential 30 in at least one of the structure generation unit 121 and the structure selection unit 122. It is more preferable that the output unit 12 executes calculations using the machine learning potential 30 in at least one of the structure optimization unit 1214 of the structure generation unit 121 and the verification unit 1221, the physical property calculation unit 1222, the judgment unit 1223, and the selection unit 1224 of the structure selection unit 122. It is further preferable that the catalyst evaluation device 10 executes calculations using the machine learning potential 30 in at least one of the structure optimization unit 1214, the verification unit 1221, and the physical property calculation unit 1222, which have a large calculation load.

[0041] The structure generating unit 121 generates a pair of replacement models including an initial state replacement model in which the structure of the initial state basic model is replaced and a final state replacement model in which the structure of the final state basic model is replaced, based on the pair of basic models acquired by the acquiring unit 11 and replacement information related to replacement of the structures of the pair of basic models. That is, the structure generating unit 121 replaces a part of the basic structures of the pair of basic models with a substituent, and generates a pair of replacement models including an initial state replacement model having the replaced initial state structure as a replacement structure and a final state replacement model having the final state structure as a replacement structure.

[0042] The structure generation unit 121 may randomly replace any hydrogen atom of the alkene coordinated to the surface molecule of the initial state basic model of the pair of basic models with a substituent made of a hydrocarbon such as a methyl group. When the substituent is a hydrocarbon, any site of the hydrocarbon may be randomly replaced with a hydrogen atom of a reactant present on the surface of the catalyst of the pair of basic models.

[0043] An example of a pair of substitution models in which the pair of basic models shown in FIG. 3 are substituted is shown in FIG. 4. As shown in FIG. 4, for example, the γ-Al 2 O 3 One hydrogen atom of ethylene coordinated to Mo contained in the surface molecule present as a reaction site on the surface of the MoOx alloy is replaced with a methyl group, a type of hydrocarbon. This replacement causes the methyl group to bond with the ethylene coordinated to Mo, and propylene, an alkene, is generated as another reactant. This results in the formation of γ-Al 2 O 3 A pair of substitution models is generated, including an initial state substitution model having an initial state substitution structure and a final state substitution model having a final state substitution structure, in the case where a catalyst made of a .OMEGA. / MoOx alloy reacts with propylene as a reactant. In Fig. 4, since the hydrogen atoms of the ethylene on the surface molecules of the pair of basic models in Fig. 3 are replaced with methyl groups, the number of atoms constituting the initial state substitution model and the final state substitution model in the pair of substitution models is 493.

[0044] In addition, if the structure generation unit 121 has prepared in advance a list of sites where reactants are substituted that are coordinated to surface molecules present on the surface of a pair of basic model catalysts, the structure generation unit 121 may generate a pair of substitution models according to the prepared list.

[0045] As described below, when the judgment unit 1223 of the structure selection unit 122 judges that the process has not ended, the structure generation unit 121 replaces the hydrogen atoms of the reactants coordinated to the surface molecules present on the surface of the catalyst in the initial state substitution model and the final state substitution model of the pair of substitution models with other substituents, and generates a new pair of substitution models including the initial state substitution model and the final state substitution model.

[0046] The structure generation unit 121 can generate a substitution model by replacing the alkenes coordinated to the surface molecules of the catalyst of a pair of basic models, the initial state basic model and the final state basic model, with a substituent that is a type of hydrocarbon, using vector calculations.

[0047] The structure generation unit 121 can include a vector acquisition unit 1211 , an adjustment unit 1212 , a generation unit 1213 , and a structure optimization unit 1214 .

[0048] The vector acquisition unit 1211 acquires a first vector V related to a replaced structure related to the replacement information of a pair of base models, and a second vector v related to a replaced structure related to the replacement information.

[0049] The replaced structure associated with the replacement information of the first vector V is, for example, a replaced structure of a molecule included in the pair of basic models, such as a surface molecule or a reactant included in the pair of basic models.

[0050] The replacement structure associated with the replacement information of the second vector v is, for example, a replacement structure of a molecule included in the pair of basic models, such as a surface molecule or a reactant included in the pair of basic models.

[0051] As the first vector V, a vector of the reactant that exists on the surface of a pair of basic model catalysts and corresponds to the bond axis direction (direction of the first bond axis) connecting a reactant to which a substituent is bonded and an atom bonded to the reactant to which the substituent is bonded.

[0052] As the second vector v, a vector corresponding to the bond axis direction (direction of the second bond axis) connecting a bonding atom that is included in the substituted structure and is bonded to the bonded atom of the reactant and an atom that is bonded to the bonding atom in the substituted molecule can be used.

[0053] The first vector V is a vector of the reactant that exists on the surface of the catalyst of the pair of basic models and corresponds to the bond axis direction connecting the bonded molecule of the reactant to which the substituent is bonded and the atom bonded to the bonded molecule, and the second vector v is a vector that is included in the replacement structure and corresponds to the bond axis direction connecting the bonded atom of the reactant and the atom bonded to the bonded atom in the replacement molecule. In this case, as shown in Figure 5 (a), the first vector V of the reactant coordinated to the surface molecule of the basic model in the initial state of the pair of basic models and the second vector v of the replacement molecule are obtained. Note that Figure 5 shows a case where the reactant is ethylene and the replacement molecule is methane.

[0054] The presence or absence of a bond between atoms may be determined based on a common method, for example, the presence or absence of a bond between atoms is determined based on the radius of the covalent bond of the atoms.

[0055] The first vector V of the reactants and the second vector v of the displacement molecules are calculated by the following formulas (1) and (2). React H are the coordinates of the hydrogen atoms of the reactants, and r React C is the coordinate of the reactant's bonded atom. Replace H are the coordinates of the hydrogen atoms of the substituted molecule, and r Replace C are the coordinates of the bond atoms of the substituted molecule. First vector of reactants V=r React H +r React C (1) The second vector of the displacement molecule, v=r Replace H +r Replace C (2)

[0056] 2 changes the arrangement of the replacement structure of the replacement molecule based on the acquired first vector V of the reactant and the second vector v of the replacement molecule. That is, the adjustment unit 1212 rotates the second vector v of the replacement molecule so that it is parallel or approximately parallel to the first vector V of the reactant coordinated to the surface molecule present on the surface of the catalyst. In this way, a parallelized second vector v' that is parallel or approximately parallel to the first vector V is acquired.

[0057] For example, as shown in Fig. 5(b), 2 O 3 By rotating the second vector v of the displacement molecule together with the displacement molecule so that it becomes parallel or approximately parallel to the first vector V of the reactant coordinated to the surface molecule of the / MoOx alloy, a second parallelized vector v' that is parallel or approximately parallel to the first vector V is obtained.

[0058] The rectified second vector v' is calculated by the following formula (3): In formula (3), Θ is the angle between the first vector V and the second vector v', R(Θ) is a three-dimensional rotation matrix, and α is a constant. Parallelized second vector v' = R(Θ)V = α × first vector V (3)

[0059] 2 translates the replacement molecule so that the coordinates of the bond position of the replacement molecule become the origin. The coordinates of the bond position are the position of any of the atoms constituting the replacement molecule.

[0060] For example, as shown in FIG. 5(c), the position of the bond atom of the replacement molecule is taken as the coordinate of the bond position of the replacement molecule, and the replacement molecule is translated so that the bond atom of the replacement molecule becomes the origin.

[0061] Coordinates r' of the atom of the substituted molecule after translation Replace can be calculated from the following formula (4). Note that r Replace are the coordinates of the atoms of the displaced molecule after translation. r' Replace =r Replace -r Replace C (4)

[0062] The adjustment unit 1212 removes hydrogen atoms at the bond sites between the reactant and the replacement molecule, and obtains a first adjustment vector V' by adjusting the magnitude of the first vector V. The first adjustment vector V' can be obtained by the following formula (5). In formula (5), the distance is the bond distance between the bonded atom of the reactant and the bonding atom of the replacement molecule. The bond distance can be arbitrarily specified. First adjustment vector V' = 1 / ||V|| × V × distance (5)

[0063] For example, as shown in Figure 5(d), the hydrogen atoms of the reactant and the substitution molecule coordinated to the surface molecule of the basic model catalyst are removed to obtain radicals and substitution molecules, and the value of the first vector V is applied to the above equation (5) to obtain the first adjustment vector V'.

[0064] 2 obtains a pair of substitution models by replacing sites included in the reactants coordinated to the surface molecules of the catalysts of the pair of basic models, which are structures to be replaced, with the substituents, which are substitution structures, based on the magnitude of the acquired first adjustment vector V'. That is, the generation unit 1213 obtains a pair of substitution models having a new structure as a substitution structure by bonding the designated atom of the substituent to the reactant along the direction of the first adjustment vector V'.

[0065] For example, as shown in Figure 5(e), a pair of substitution models having a substitution structure is obtained by combining a substituent obtained by removing a hydrogen atom from a replacement molecule with a radical obtained by removing a hydrogen atom from a reactant coordinated to a surface molecule of a basic model catalyst in the direction of the first adjustment vector V'.

[0066] 2 has a vector acquisition unit 1211, an adjustment unit 1212, and a generation unit 1213, and thus can generate a pair of substitution models having new structures as substitution structures by respectively replacing hydrogen atoms of reactants coordinated to surface molecules of catalysts of a pair of basic models and hydrogen atoms at positions where substituents are bonded. Therefore, the structure generation unit 121 can generate a pair of substitution models having two initial state substitution models and final state substitution models having substitution structures, regardless of the type of reaction or the size of the system, simply by preparing two initial state basic models and final state basic models having the basic structures of the initial state and final state of a pair of basic models of a base reaction.

[0067] The structural optimization unit 1214 may perform a process of optimizing the replacement structures of the pair of replacement models generated by the generation unit 1213. By performing structural optimization of both the initial state structure and the final state structure of the pair of replacement models, it is possible to obtain both an initial state replacement model having a stable initial state replacement structure and a final state replacement model having a stabilized final state replacement structure.

[0068] The optimization method is not particularly limited, and for example, a method generally used for structural optimization, such as the L-BFGS method, may be used.

[0069] The structural optimization unit 1214 preferably uses NNP to execute the process of structural optimization of a pair of substitution models. By executing the structural optimization unit 1214 using NNP, it is possible to perform structural optimization of the substitution models at a higher speed.

[0070] The structure selection unit 122 includes a verification unit 1221, a physical property calculation unit 1222, a determination unit 1223, a selection unit 1224, and a structure discarding unit 1225. The structure selection unit 122 calculates physical property values ​​every time a pair of substitution models is generated until a predetermined termination condition is satisfied, that is, when repeating N times (N is an integer equal to or greater than 1) until the predetermined termination condition is satisfied.

[0071] The verification unit 1221 calculates a value related to the validity of the pair of substitution models based on information related to the structures of the pair of substitution models. That is, the verification unit 1221 verifies the validity of the optimized substitution structures of the pair of substitution models acquired by the structure optimization unit 1214.

[0072] The information related to the structure of the pair of displacement models is, for example, whether a particular bond is formed, whether the transition state in question is a transition state resulting from a reactant and a product, and the like.

[0073] The value relating to the validity of the paired substitution model is a value relating to the structure of the paired substitution model, such as bond length, whether or not NEB calculations converge, etc.

[0074] The verification unit 1221 may use a method used for verifying the validity of a general structure. The verification unit 1221 may verify the validity of the substitution structure of the generated pair of substitution models using a Nudged Elastic Band (NEB) method, or may calculate the root mean square deviation (RMSD) to verify the validity of the substitution structure of the generated pair of substitution models.

[0075] The NEB method is a method for determining the most stable reaction path with the minimum activation energy for the reaction from the initial state to the final state of a pair of models. In the NEB method, the initial state substitution structure and the final state substitution structure of a pair of substitution models are processed as input data, and n images of intermediate structures that connect these two structures are generated. Each intermediate structure is connected to another adjacent intermediate structure by a spring along the reaction path. The intermediate structures that are connected (elastic band) are optimized toward the potential energy surface that is the reaction path. When optimizing according to the force acting on each intermediate structure image, the image is moved in the vertical direction perpendicular to the elastic band (so-called spring force is applied) so that adjacent images do not get close to each other, so that they are evenly distributed along the reaction path and approach the target reaction path. The image with the highest energy maximizes energy along the elastic band and tries to minimize it in all other directions. When the images converge, they become a structure close to the transition state. The NEB calculation ends when the force of each intermediate structure image falls below a set threshold. Therefore, the NEB method determines the forces acting on each intermediate structure, and performs structural optimization while taking into account the perpendicular component to the reaction path and the restoring force of the spring, thereby making it possible to determine the reaction path with the smallest activation energy as the most stable path.

[0076] The verification unit 1221 uses the NEB method to calculate, as the most stable path, a reaction path with the minimum activation energy when a catalyst having the substitution structure of the pair of substitution models optimized by the structure optimization unit 1214 and a reactant undergo a catalytic reaction. Specifically, the NEB method calculates a reaction path with the minimum activation energy when a catalyst undergoes a catalytic reaction with a reactant coordinated to the surface of the pair of substitution models. From the most stable path obtained by the NEB method, it is possible to calculate a transition state (TS) in the catalytic reaction of the reactant coordinated to the surface of the substitution structure and the activation energy when passing through the transition state, as shown in FIG. 6.

[0077] It is preferable that the verification unit 1221 uses NNP to verify the validity of the substitution structure of the generated substitution model. When the NEB method is used to verify the validity of the substitution structure of the generated substitution model, the NEB method can search for a reaction path with the lowest activation energy, but the search takes time. By executing using NNP, the verification unit 1221 can more quickly calculate the most stable path from the substitution structure in the initial state of the substitution model to the substitution structure in the final state, and the activation energy when passing through the transition state.

[0078] When using the NEB method, if the NEB calculations converge after a predetermined number of times (e.g., several hundred times on average), the verification unit 1221 determines that the replacement structure of the generated pair of replacement models is valid, and if the NEB calculations do not converge after a predetermined number of times, it determines that the replacement structure of the generated pair of replacement models is not valid.

[0079] When the verification unit 1221 determines that the substitution structures of the generated pair of substitution models are valid, the physical property calculation unit 1222 calculates physical property values ​​of the generated pair of substitution models.

[0080] The physical property values ​​include activation energy, adsorption energy, etc. The activation energy can be calculated by the NEB method performed by the verification unit 1221. The adsorption energy can be calculated from the energy difference between the structure in the stabilized initial state and the structure optimized by separating the reactants from the stabilized initial state structure, as shown in FIG.

[0081] It is preferable that the physical property calculation unit 1222 uses NNP to calculate the physical property values ​​of the replacement model. By using NNP, the physical property calculation unit 1222 can calculate the physical property values ​​of the replacement model at a higher speed.

[0082] The determination unit 1223 determines whether to end the generation of the pair of substitution models.

[0083] The determination of whether the generation of a pair of substitution models has been completed may be made based on, for example, when a predetermined number of calculations have been completed, when a predetermined combination of calculations has been completed, when a predetermined time has elapsed, or when an end signal has been received from the user.

[0084] When the determination unit 1223 determines that the generation of the pair of substitution models should not be terminated, it causes the structure generation unit 121 to generate a substitution model again.

[0085] The selection unit 1224 selects a pair of substitution models whose physical property values ​​related to the pair of substitution models are equal to or less than a target threshold value from the multiple pair of substitution models as a model of the catalyst for the catalytic reaction. That is, when the determination unit 1223 determines that the generation of the pair of substitution models is to be terminated, the selection unit 1224 selects a pair of substitution models whose calculated physical property values ​​satisfy a predetermined condition from the multiple calculated physical property values ​​as a valid model.

[0086] The target threshold value is not particularly limited, and may be set appropriately depending on the physical property values ​​related to the pair of substitution models.

[0087] Valid models may be selected based on, for example, selecting a predetermined number in order of decreasing energy, such as activation energy or adsorption energy, having an energy equal to or less than a predetermined value, or having the lowest energy.

[0088] Furthermore, when the activation energy and the adsorption energy are calculated as the physical property values ​​in the physical property calculation unit 1222, the selection unit 1224 may create a relationship diagram between the adsorption energy and the activation energy as shown in FIG. 8, and select an effective model based on this relationship diagram.

[0089] When the verification unit 1221 determines that the generated substitution structure of the pair of substitution models is invalid, the structure discarding unit 1225 discards the pair of substitution models.

[0090] The selection result output unit 13 outputs the selected pair of substitution models as valid candidate models by display or the like.

[0091] (Hardware configuration of catalyst evaluation device 10) Next, an example of the hardware configuration of the catalyst evaluation device 10 will be described. FIG. 9 is a block diagram showing the hardware configuration of the catalyst evaluation device 10. As shown in FIG. 9, the catalyst evaluation device 10 is configured as an information processing device (computer), and can be physically configured as a computer system including a CPU (Central Processing Unit: processor) 101 which is an arithmetic processing unit, a RAM (Random Access Memory) 102 and a ROM (Read Only Memory) 103 which are main storage devices, an input device 104 which is an input device, an output device 105, a communication module 106, and an auxiliary storage device 107 such as a hard disk. These are connected to each other by a bus 108. The output device 105 and the auxiliary storage device 107 may be provided externally.

[0092] The CPU 101 controls the overall operation of the catalyst evaluation device 10 and performs various information processing. The CPU 101 can execute, for example, a catalyst evaluation method or a catalyst evaluation program stored in the ROM 103 or the auxiliary storage device 107, which will be described later, to search for a catalyst.

[0093] The RAM 102 is used as a work area for the CPU 101 and may include a non-volatile RAM for storing main control parameters and information.

[0094] The ROM 103 stores a basic input / output program, etc. The catalyst evaluation program may be stored in the ROM 103.

[0095] The input device 104 is an input device such as a keyboard, a mouse, operation buttons, a touch panel, a display screen, etc., and receives information input by a user as an instruction signal and outputs the instruction signal to the CPU 101 .

[0096] The output device 105 is a display device such as a monitor display, an audio device such as a speaker, a printing device such as a printer, etc. In the output device 105, for example, information such as the catalyst selection result is displayed on a display device such as a monitor display, and the displayed screen is updated in response to an input operation via the input device 104 or the communication module 106.

[0097] The communication module 106 is a data transmission / reception device such as a network card, and functions as a communication interface that receives information from an external data recording server or the like, and outputs analysis information to other electronic devices.

[0098] The auxiliary storage device 107 is a storage device such as an SSD (Solid State Drive) or an HDD (Hard Disk Drive), and stores, for example, various data, files, and the like required for the operation of the catalyst evaluation device 10.

[0099] Each function of the catalyst evaluation device 10 is realized by loading a specified computer software (including a catalyst evaluation program) from a main memory device such as RAM 102 or an auxiliary memory device 107 and executing it by the CPU 101, thereby reading and writing data in a main memory device such as RAM 102 or an auxiliary memory device 107, etc., and operating the input device 104, the output device 105, and the communication module 106.

[0100] Therefore, each part of the catalyst evaluation device 10 shown in FIG. 2 is realized by software and hardware working together in a computer equipped with the catalyst evaluation device 10, by a processor executing predetermined computer software (including a catalyst evaluation program) that is pre-stored.

[0101] The catalyst evaluation program can be stored, for example, in a main storage device or auxiliary storage device 107 of a computer. The catalyst evaluation program may be stored on a computer connected to a communication line such as the Internet, and a part or all of the catalyst evaluation program may be provided by being downloaded via the communication line. Furthermore, the catalyst evaluation program may be configured to be provided or distributed via the communication line.

[0102] The catalyst evaluation program may be recorded (including installed) into a computer from a state in which a part or the whole of the program is stored in a portable storage medium, such as an optical disk such as a CD-ROM or a DVD-ROM, or a semiconductor memory such as a flash memory.

[0103] <Catalyst evaluation method> The catalyst evaluation method according to this embodiment will be described below. The catalyst evaluation method according to this embodiment can be performed using the above-mentioned catalyst evaluation system 1. Therefore, some of the contents that have already been described will be omitted.

[0104] Fig. 10 is a flowchart showing the catalyst evaluation method according to the present embodiment. As shown in Fig. 10, the catalyst evaluation method according to the present embodiment is a catalyst evaluation method for evaluating catalysts that undergo a catalytic reaction with a reactant to produce a predetermined target product from the reactant, and searching for a catalyst that exhibits a high evaluation index for the reactant.

[0105] In the catalyst evaluation method according to this embodiment, the acquisition unit 11 acquires a pair of base models (acquisition step: step S11).

[0106] Next, the structure generation unit 121 of the output unit 12 generates a pair of substitution models (substitution model generation process: step S12).

[0107] Next, the structure selection unit 122 of the output unit 12 verifies the validity of the replacement structures of the generated pair of replacement models (verification process: step S13), and if the replacement structures of the pair of replacement models are valid (S13: Yes), calculates physical property values ​​related to the pair of replacement models (physical property calculation process: step S14).

[0108] Next, the structure selection unit 122 determines whether to terminate the generation of the pair of substitution models (determination process: step S15), and if it determines that a predetermined termination criterion is met (step S15: Yes), it selects a pair of substitution models whose physical property values ​​associated with the pair of substitution models are below a target threshold value as a model of the catalytic reaction (selection process: S16).

[0109] In this embodiment, the predetermined end condition is a condition in which the structure selection unit 122 judges that the generation of the pair of substitution models generated by the structure generation unit 121 is to be ended.

[0110] Next, the selection result output unit 13 outputs at least one of the selected pair of base models and the selected pair of replacement models as a valid candidate model (selection result output step: S17).

[0111] On the other hand, if the structure of the pair of substitution models is not valid (S13: No), the structure selection unit 122 discards the pair of substitution models (discarding step: step S18).

[0112] After discarding the pair of substitution models, the structure selection unit 122 sequentially executes the processes from the determination step (step S15) onwards.

[0113] Moreover, if the structure selection unit 122 determines that the above-mentioned predetermined termination criterion is not satisfied (step S15: No), the output unit 12 sequentially executes the processes subsequent to the generation step of the substitution model (step S11).

[0114] According to the catalyst evaluation method of the present embodiment, by selecting a candidate model as a catalyst exhibiting a high evaluation index with respect to a reactant, the candidate model can be evaluated as a catalyst exhibiting a high evaluation index with respect to the reactant, and thus a catalyst exhibiting a high evaluation index with respect to the reactant can be searched for. In the case where the evaluation index is, for example, reactivity with respect to a reactant, by selecting a candidate model as a catalyst effective in a catalytic reaction with the reactant, the candidate model can be evaluated as a catalyst effective in a catalytic reaction with the reactant, and a catalyst having high reactivity with the reactant can be searched for.

[0115] In this way, the catalyst evaluation system 1 has a catalyst evaluation device 10, which includes an output unit 12. The output unit 12 generates a pair of substitution models including an initial state substitution model and a final state substitution model based on a pair of basic models including an initial state basic model and a final state basic model and substitution information related to the substitution of the structure of the pair of basic models, and outputs a catalyst for a catalytic reaction by associating a physical property value related to the pair of substitution models with the pair of substitution models. The output unit 12 can evaluate whether the pair of substitution models is an effective model for a catalytic reaction based on a physical property value (e.g., energy, etc.) related to the pair of substitution models. Therefore, the catalyst evaluation device 10 outputs a specific pair of substitution models based on a physical property value related to the pair of substitution models among the pair of substitution models generated, thereby evaluating a catalyst that catalytically reacts with a reactant, and can easily search for a catalyst with a high evaluation index for the reactant while reducing the load of calculation.

[0116] Therefore, since the catalyst evaluation system 1 includes the catalyst evaluation device 10, it is possible to more efficiently search for a catalyst that exhibits a high evaluation index for a reactant.

[0117] When the evaluation index is, for example, the reactivity of a catalyst with a reactant, the catalyst evaluation system 1 can easily search for catalysts that are highly reactive with the reactant while reducing the calculation load, thereby making it possible to more efficiently search for catalysts that are effective in catalytic reactions with the reactant.

[0118] In addition, when the catalyst for which the characteristic value is to be calculated is, for example, a system of several hundred atoms, which requires a high calculation cost, it is difficult to calculate the characteristic value of the catalyst that reacts with the reactant in a realistic time using computational chemistry that has been generally used in the past. In particular, when the catalyst for which the characteristic value is to be calculated is a system of several hundred atoms, it is difficult to calculate even the characteristic value of one molecule using computational chemistry that has been generally used in the past, and it is difficult to search for a catalyst in terms of the cost required for searching for a catalyst. In addition, in order to be able to complete the calculation of the characteristic value of a catalyst effective for the reaction with the reactant without interruption from the structure of the basic model, it is necessary to prepare a basic model having an appropriate basic structure, but it is practically difficult to prepare such a basic model. In contrast, the catalyst evaluation system 1 uses a pair substitution model in the catalyst evaluation device 10, and can output only a pair of substitution models having a suitable substitution structure based on the physical property values ​​related to the pair of substitution models, so that a catalyst that shows a high evaluation index for the reactant, such as high reactivity with the reactant, can be easily and efficiently searched for while reducing the calculation load. Therefore, the catalyst evaluation system 1 can significantly reduce the time required to calculate the physical property values ​​of a catalyst, and can obtain the values ​​in a realistic amount of time, thereby reducing the cost required for searching for a catalyst.

[0119] The catalyst evaluation system 1 can calculate a value related to the validity of the pair of substitution models based on information related to the structure of the pair of substitution models in the output unit 12 of the catalyst evaluation device 10. The catalyst evaluation device 10 can verify the validity of the substitution structure of the pair of substitution models by calculating a value related to the validity of the pair of substitution models. As a result, the catalyst evaluation device 10 can select a pair of substitution models with high validity from the pair of substitution models, and can calculate the physical property values ​​of only the pair of substitution models with high validity from the pair of substitution models. Therefore, the catalyst evaluation system 1 can evaluate a catalyst that catalytically reacts with a reactant and efficiently search for a catalyst that shows a high evaluation index while easily reducing the calculation load. When the evaluation index is, for example, the reactivity of a catalyst with a reactant, the catalyst evaluation system 1 can efficiently search for a catalyst with high reactivity with a reactant while easily reducing the calculation load.

[0120] The catalyst evaluation system 1 can calculate the physical property values ​​each time a pair of substitution models is generated in the output section 12 of the catalyst evaluation device 10 until a predetermined termination condition is satisfied. The catalyst evaluation device 10 can generate a plurality of pairs of substitution models and calculate the physical property values ​​of each pair of substitution models until a predetermined termination condition is satisfied. The catalyst evaluation device 10 can easily and efficiently search for a catalyst that exhibits a high evaluation index for a reactant while further reducing the calculation load by selecting only a pair of substitution models with high validity from the plurality of pairs of substitution models generated.

[0121] Furthermore, in the catalyst evaluation device 10, the output unit 12 can calculate the physical property values ​​of only a pair of substitution models having a valid substitution structure among the pair of generated substitution models based on the verification result of the validity of the substitution structure of the pair of substitution models in the verification unit 1221. Therefore, the catalyst evaluation device 10 can easily and efficiently search for a catalyst that shows a high evaluation index for a reactant while further reducing the calculation load. Therefore, the catalyst evaluation system 1 can reduce the cost required for searching for a catalyst.

[0122] The catalyst evaluation system 1 can select, as a model of a catalytic reaction, a pair of substitution models having a physical property value associated with the pair of substitution models equal to or less than a target threshold value from among a plurality of pairs of substitution models in the output section 12 of the catalyst evaluation device 10. By selecting a pair of substitution models having a physical property value associated with the pair of substitution models equal to or less than a target threshold value, the catalyst evaluation device 10 can select, as an effective model, a pair of substitution models having a calculated physical property value that satisfies a condition from among the plurality of pairs of substitution models generated. Thus, the catalyst evaluation system 1 can easily and efficiently search for a catalyst having a high evaluation index for a reactant while further reducing the load of calculation by outputting only catalysts having a high evaluation index for a reactant in the catalyst evaluation device 10.

[0123] The catalyst evaluation system 1 can, for example, replace a site included in a substituted structure of a surface molecule with a substituted structure based on a first vector related to a substituted structure related to the substitution information of a pair of basic models and a second vector related to a substituted structure related to the substitution information in the output section 12 of the catalyst evaluation device 10. As a result, the catalyst evaluation device 10 can appropriately generate a pair of substitution models regardless of the type of reaction or the size of the system of the basic model by simply acquiring two structures, an initial state basic model and a final state basic model, which constitute a pair of basic models of a base reaction. Therefore, the catalyst evaluation system 1 can search for a catalyst that shows a high evaluation index for a reactant with higher accuracy by using a pair of appropriately generated substitution models.

[0124] The catalyst evaluation system 1 can change the arrangement of the substituted structure based on a parallelized second vector obtained by rotating the second vector so that the second vector is parallel to the first vector in the output section 12 of the catalyst evaluation device 10, and can replace a site included in the substituted structure of the surface molecule with the substituted structure based on the magnitude of the first vector. This allows the catalyst evaluation device 10 to easily generate a pair of substitution models. Therefore, the catalyst evaluation system 1 can more easily search for a catalyst that shows a high evaluation index for a reactant by using the generated pair of substitution models.

[0125] The catalyst evaluation system 1 can calculate a value related to the validity of a pair of substitution models using NNP in the output unit 12 of the catalyst evaluation device 10. By executing the verification unit 1221 of the output unit 12 using NNP, the catalyst evaluation device 10 can predict energy without solving strict equations as in computational chemistry, and can therefore perform calculations at high speed and speed up the calculation of the validity of the structure of a pair of substitution models. Therefore, the catalyst evaluation system 1 can search for a catalyst that shows a high evaluation index for a reactant more simply and in a short time.

[0126] In the catalyst evaluation system 1, the catalyst evaluation device 10 can use the adsorption energy or the activation energy as the physical property value related to the pair of substitution models. In this way, when the physical property value related to the pair of substitution models is the adsorption energy or the activation energy, the catalyst evaluation device 10 can appropriately search for a catalyst that exhibits a high evaluation index for a reactant by using the calculated adsorption energy or the activation energy.

[0127] In this embodiment, the output unit 12 of the catalyst evaluation device 10 generates a pair of substitution models in the structure generation unit 121 from the pair of basic models first acquired by the acquisition unit 11, and then verifies the validity of the substitution structure of the pair of substitution models, but is not limited to this. As shown in Fig. 11, the output unit 12 may verify the validity of the basic structure of the pair of basic models without replacing the pair of basic models first acquired by the acquisition unit 11 with the pair of substitution models. After the determination unit 1223 determines that the generation of the pair of basic models or the pair of substitution models is not to be terminated, the structure generation unit 121 may replace a part of the pair of basic models to generate a pair of substitution models, and perform verification of the validity of the substitution structure of the generated pair of substitution models.

[0128] That is, when the output unit 12 repeats the process N times (N is an integer greater than or equal to 1) until a predetermined termination condition is satisfied, the first time, the verification unit 1221 may verify the validity of the basic structure of the pair of basic models, and if the validity is verified, the physical property calculation unit 1222 may calculate the physical property values ​​of the pair of basic models.

[0129] In this case, the catalyst evaluation method according to the present embodiment uses a pair of basic models in the verification step (step S13), the physical property calculation step (step S14), and the judgment step (step S15) which are performed first (the first time when the method is repeated N times (N is an integer equal to or greater than 1)) in the catalyst evaluation method according to the present embodiment shown in FIG. 10. Then, after it is determined in the judgment step (step S15) that the calculation is not completed, a generation step (step S12) of a pair of replacement models is performed, and a pair of replacement models is generated by replacing the pair of basic models. Other than these, the catalyst evaluation method according to the present embodiment shown in FIG. 10 is the same as that shown in FIG. Therefore, a detailed explanation of the catalyst evaluation method when using the catalyst evaluation device 10 shown in FIG. 11 will be omitted.

[0130] In addition, in this embodiment, the case where the catalyst is a heterogeneous catalyst such as a metal compound has been described, but the catalyst evaluation system 1 may be a homogeneous catalyst or the like. In this case, the catalyst evaluation system 1 uses a pair of basic models including an initial state basic model having a basic structure in which a surface molecule is arranged at a predetermined site of the homogeneous catalyst, and a final state basic model having a basic structure in which a reactant is coordinated to the surface molecule of the homogeneous catalyst. Using this pair of basic models, the catalyst evaluation system 1 may generate a pair of substitution models including an initial state substitution model having a substitution structure in which a hydrogen atom at a predetermined site of the surface molecule of the pair of basic models is replaced with a substituent, and a final state substitution model having a basic structure in which a reactant that is another type of alkene is arranged on the surface molecule of the homogeneous catalyst.

[0131] As described above, the embodiment has been described, but the above embodiment is presented as an example, and the present invention is not limited to the above embodiment. The above embodiment can be implemented in various other forms, and various combinations, omissions, substitutions, modifications, etc. can be made without departing from the gist of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are included in the scope of the invention and its equivalents described in the claims. EXAMPLES

[0132] The embodiments will be described in more detail below with reference to examples and comparative examples, but the embodiments are not limited to these examples and comparative examples.

[0133] <Example 1> A catalyst evaluation device 10 having the configuration shown in FIG. 2 was manufactured. In the structure optimization section 1214, the verification section 1221, and the calculation section 1222 of the physical property value of the catalyst evaluation device 10, the operation was performed using a preferred potential (PFP). The catalyst to be calculated by the catalyst evaluation device 10 was aluminum oxide (γ-Al 2 O 3 ) and metal composite oxide with MoOx (γ-Al 2 O 3 A mixture of alkene bound to the surface of a ZnO / MoOx alloy was used as a reactant.

[0134] [Preparation of a pair of base models] As shown in Figure 3, γ-Al 2 O 3 A pair of basic models was prepared, consisting of an initial-state basic model having a basic structure of the initial state of the γ-Al / MoOx alloy and a final-state basic model having a basic structure of the final state. 2 O 3 Ethylene was bonded to the surface molecules present on the surface of the / MoOx alloy. The number of atoms constituting the pair of basic models, the initial state basic model and the final state basic model, was set to 490.

[0135] [Generating a pairwise substitution model] The hydrogen atoms of the ethylene on the surface molecules of the pair of basic models were randomly replaced with hydrocarbons to form alkenes, and a pair of replacement models was generated consisting of an initial state replacement model having the replacement structure in the substituted initial state and a final state replacement model having the replacement structure in the final state. Ethylene was bound to multiple hydrocarbons obtained from a database in 3590 different patterns to generate a pair of replacement models with alkenes in 3590 different patterns.

[0136] [Structural optimization] The substitution structures of the pairwise substitution models were optimized using the L-BFGS method.

[0137] [Verification of structural validity] Using the NEB method, we verified whether the substitution structure of the optimized pair of substitution models was valid. If the NEB calculations converged after a certain number of times, we determined that the substitution structure of the generated pair of substitution models was valid, and if the NEB calculations did not converge after a certain number of times, we determined that the substitution structure of the generated pair of substitution models was not valid.

[0138] [Calculation of physical properties] The adsorption energy and activation energy were calculated as physical properties of a pair of substitution models whose substitution structures were deemed appropriate. (adsorption energy) The adsorption energy is the initial state basic model of γ-Al 2 O 3 The energy was calculated from the difference between the energy of the structure optimized by removing the reactants from the / MoOx alloy and the energy of the basic structure in the initial state. (Activation Energy) The activation energy used was the activation energy used when verifying the validity of the substitution structure of an optimized pair of substitution models using the NEB method.

[0139] [Destroy Structure] Pairwise substitution models in which the substitution structure was deemed implausible were discarded.

[0140] [judgement] It was determined whether the generation of the pair-based substitution models had been completed, and the calculations were repeated until the generation of the pair-based substitution models was completed, and the calculations were completed after about two months. The number of pair-based substitution models calculated was 3,590, and of the 3,590 pair-based substitution models for each catalyst, only 255 converged. The average number of atoms in the initial-state substitution model and the final-state substitution model for each pair-based substitution model was 501.3. The average number of NEB calculations performed when verifying the validity of the structure was 487.9.

[0141] [evaluation] Using the adsorption energy and activation energy calculated in the above [Calculation of physical properties], a relationship diagram between adsorption energy and activation energy was created. The results of plotting the relationship between adsorption energy and activation energy are shown in Figure 12. In Figure 12, the cross marks indicate cases where the carbon chain of the hydrocarbon used to form the alkene in the pair of substitution models is 4 or less, and A, B, and C are alkenes extracted as being particularly effective. The structures of alkenes A, B, and C are shown below.

[0142] [ka]

[0143] [ka]

[0144] [ka]

[0145] 12, it was confirmed that alkenes A, B, and C have smaller activation energy and higher reactivity when reacting with a catalyst than propylene. Therefore, it can be said that the catalyst evaluation system according to this embodiment can more efficiently search for a catalyst that shows a high evaluation index for a reactant by evaluating a pair of substitution models corresponding to the catalyst.

[0146] The aspects of the embodiment of the present invention are as follows, for example. <1> A catalyst evaluation system comprising: a pair of basic models including an initial state basic model corresponding to reactants of a catalytic reaction and a final state basic model corresponding to products of the catalytic reaction; and replacement information related to the replacement of structures of the pair of basic models, which generates a pair of replacement models including an initial state replacement model in which the structure of the initial state basic model has been replaced and a final state replacement model in which the structure of the final state basic model has been replaced; and an output unit that matches physical property values ​​related to the pair of replacement models with the pair of replacement models, and outputs catalysts of the catalytic reaction related to the reactants and the products. <2> The output unit calculates a value related to the validity of the pair of substitution models based on information related to the structure of the pair of substitution models. <1> The catalyst evaluation system according to claim 1. <3> The output unit includes a calculation unit that calculates the physical property value every time the pair of substitution models is generated until a predetermined termination condition is satisfied. <1> or <2> The catalyst evaluation system according to claim 1, <4> The output unit selects, from the plurality of pairs of substitution models, a pair of substitution models in which the physical property value associated with the pair of substitution models is equal to or less than a target threshold value, as a model of a catalyst for the catalytic reaction. <3> The catalyst evaluation system according to claim 1. <5> The output unit replaces a site included in the replaced structure with the replaced structure based on a first vector related to the replaced structure associated with the replacement information of the pair of basic models and a second vector related to the replaced structure associated with the replacement information. <1> ~ <4> 13. A catalyst evaluation system according to any one of the above. <6> The output unit changes the arrangement of the substituted structure based on a parallelized second vector obtained by rotating the second vector so that the second vector is parallel to the first vector, and replaces a site included in the replaced structure with the substituted structure based on the magnitude of the first vector. <5> The catalyst evaluation system according to claim 1, <7> The output unit calculates a value related to the validity of the pair of substitution models using NNP. <2> ~ <6> 13. A catalyst evaluation system according to any one of the above. <8> The physical property value associated with the pair of substitution models is an adsorption energy or an activation energy. <1> ~ <7> 13. A catalyst evaluation system according to any one of the above. <9> On the computer, A catalyst evaluation program that executes a step of generating a pair of replacement models including an initial state replacement model in which the structure of the initial state basic model has been replaced and a final state replacement model in which the structure of the final state basic model has been replaced based on a pair of basic models including an initial state basic model corresponding to reactants of a catalytic reaction and a final state basic model corresponding to products of the catalytic reaction, and replacement information related to the replacement of structures of the pair of basic models, and matching physical property values ​​related to the pair of replacement models with the pair of replacement models, and outputting catalysts of the catalytic reaction related to the reactants and the products. <10> The computer A catalyst evaluation method comprising the steps of: generating a pair of replacement models including an initial state replacement model in which the structure of the initial state basic model has been replaced and a final state replacement model in which the structure of the final state basic model has been replaced based on a pair of basic models including an initial state basic model corresponding to reactants of a catalytic reaction and a final state basic model corresponding to products of the catalytic reaction, and replacement information related to the replacement of structures of the pair of basic models; associating physical property values ​​related to the pair of replacement models with the pair of replacement models, and outputting catalysts of the catalytic reaction related to the reactants and the products. [Explanation of symbols]

[0147] 1. Catalyst evaluation system 10. Catalyst evaluation equipment 11 Acquisition Department 12 Output section 13 Selection result output section 20 Memory section 30 Machine Learning Potential 121 Structure generation part 122 Structure Selection Section 1211 Vector Acquisition Section 1212 Adjustment section 1213 Generation part 1214 Structural Optimization Department 1221 Verification Department 1222 Calculation of physical properties 1223 Judgment section 1224 Selection 1225 Structure Destruction Department

Claims

1. A catalyst evaluation system comprising: a pair of basic models including an initial state basic model corresponding to reactants of a catalytic reaction and a final state basic model corresponding to products of the catalytic reaction; and replacement information related to the replacement of structures of the pair of basic models, which generates a pair of replacement models including an initial state replacement model in which the structure of the initial state basic model has been replaced and a final state replacement model in which the structure of the final state basic model has been replaced; and an output unit that matches physical property values ​​related to the pair of replacement models with the pair of replacement models, and outputs catalysts of the catalytic reaction related to the reactants and the products.

2. The catalyst evaluation system according to claim 1 , wherein the output unit calculates a value relating to the validity of the pair of substitution models based on information relating to a structure of the pair of substitution models.

3. 3 . The catalyst evaluation system according to claim 1 , wherein the output unit includes a calculation unit that calculates the physical property value each time the pair of substitution models is generated until a predetermined termination condition is satisfied.

4. The catalyst evaluation system according to claim 3 , wherein the output unit selects the pair of substitution models in which the physical property value associated with the pair of substitution models is equal to or less than a target threshold value from among the plurality of pairs of substitution models as a model of the catalyst for the catalytic reaction.

5. 3. The catalyst evaluation system according to claim 1, wherein the output unit replaces a site contained in the substituted structure with the substituted structure based on a first vector relating to the substituted structure related to the substitution information of the pair of basic models and a second vector relating to the substituted structure related to the substitution information.

6. 6. The catalyst evaluation system of claim 5, wherein the output unit changes the arrangement of the substituted structure based on a parallelized second vector obtained by rotating the second vector so that the second vector is parallel to the first vector, and replaces a site contained in the substituted structure with the substituted structure based on the magnitude of the first vector.

7. The catalyst evaluation system according to claim 2 , wherein the output unit calculates a value relating to the validity of the pair of substitution models using NNP.

8. The catalyst evaluation system according to claim 1 , wherein the physical property value related to the pair of substitution models is an adsorption energy or an activation energy.

9. On the computer, A catalyst evaluation program that executes a step of generating a pair of replacement models including an initial state replacement model in which the structure of the initial state basic model has been replaced and a final state replacement model in which the structure of the final state basic model has been replaced based on a pair of basic models including an initial state basic model corresponding to reactants of a catalytic reaction and a final state basic model corresponding to products of the catalytic reaction, and replacement information related to the replacement of structures of the pair of basic models, and matching physical property values ​​related to the pair of replacement models with the pair of replacement models, and outputting catalysts of the catalytic reaction related to the reactants and the products.

10. The computer A catalyst evaluation method comprising the steps of: generating a pair of replacement models including an initial state replacement model in which the structure of the initial state basic model has been replaced and a final state replacement model in which the structure of the final state basic model has been replaced based on a pair of basic models including an initial state basic model corresponding to reactants of a catalytic reaction and a final state basic model corresponding to products of the catalytic reaction, and replacement information related to the replacement of structures of the pair of basic models; associating physical property values ​​related to the pair of replacement models with the pair of replacement models, and outputting catalysts of the catalytic reaction related to the reactants and the products.

Images