Resin design system, resin design device, and resin design method

Abstract

To provide a resin design system that can design a resin with desired performance in a short time.SOLUTION: A design system 100 includes: an input device 10 into which input information including at least information that affects a higher-order structure of a designed resin to be designed is input; an arithmetic device 20 that includes a structure DB 224 containing information on atomic positions in higher-order structures of known resins, and a structure determination unit 212 configured to determine a higher-order structure of a determined resin based on the information on atomic positions in higher-order structures of the known resins in the structure DB 224 and the input information; and an output device 30 that outputs output information regarding the determined resin.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a resin design system, a resin design device, and a resin design method. [Background technology]

[0002] The performance of a resin varies significantly depending on the type and blending ratio of raw materials used. With regard to the types of raw materials, slight differences in structural isomers and functional groups can affect the performance of the resulting resin. Traditionally, when formulating a resin composition, resins were actually produced using raw materials predicted to meet the desired performance, and performance evaluation was performed. Performance evaluation was carried out through trial and error, repeatedly changing the raw materials and blending ratios until the resin achieved the desired performance. However, this experimental method requires a large amount of labor and costs associated with material acquisition. Furthermore, resin performance can vary depending on factors that are difficult to precisely control, such as the manufacturing environment and manufacturing process. Therefore, it is difficult to obtain experimental results that maximize the performance of the selected raw materials and to ensure experimental reproducibility.

[0003] Meanwhile, in recent years, digital technologies such as computer simulations and machine learning using databases have been increasingly used in materials design. The use of digital technologies has reduced the labor hours required by workers and the costs of obtaining materials, and has made it possible to predict the performance of resins without worrying about performance changes due to the environment and process. The technologies described in Patent Document 1 and Non-Patent Document 1 are known as materials design technologies that utilize digital technologies. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2002-363199 [Non-patent literature]

[0005] [Non-Patent Document 1] Daolei Lin, et al., “Macromolecular structural evolution of polyimide chains during large-ratio uniaxial fiber orientation process revealed by molecular dynamics simulation”, ChemicalPhysicsLetters, Vol.756, 137847(2020) Summary of the Invention [Problem to be solved by the invention]

[0006] The details will be explained with reference to a comparative example, but the technique described in Non-Patent Document 1 requires a large number of calculations and takes a long time to perform. Furthermore, the technique described in Patent Document 1 is a technique related to predicting the structure of proteins, not a technique related to resins. The problem to be solved by the present disclosure is to provide a resin design system, a resin design device, and a resin design method that are capable of designing a resin having desired performance in a short period of time. [Means for solving the problem]

[0007] The resin design system of the present disclosure includes an input device to which input information including at least information that affects the higher-order structure of a design resin to be designed is input, a structural database including information about atomic positions in the higher-order structures of known resins, a computing device including a structure determination unit that determines the higher-order structure of a determined resin based on the information about atomic positions in the higher-order structures of the known resins in the structural database and the input information, and an output device that outputs output information about the determined resin. Other solutions will be described later in the description of the invention. [Effects of the Invention]

[0008] According to the present disclosure, it is possible to provide a resin design system, a resin design device, and a resin design method that are capable of designing a resin having desired performance in a short period of time. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a block diagram of a resin design system and a resin design device according to an embodiment. [Figure 2] FIG. 2 is a schematic diagram of an input screen according to an embodiment. [Figure 3] FIG. 1 is a block diagram of a computing device according to an embodiment. [Figure 4] FIG. 1 is a schematic diagram of a force field file according to one embodiment. [Figure 5] FIG. 2 is a schematic diagram of one of the structure data included in the structure database according to the embodiment. [Figure 6] FIG. 1 is a schematic diagram illustrating structure data according to an embodiment expressed as vectors. [Figure 7] FIG. 10 is a schematic diagram illustrating structure data according to an embodiment expressed as a function. [Figure 8] FIG. 10 is a schematic diagram of an output screen according to an embodiment. [Figure 9] 1 is a flowchart illustrating a resin design method according to an embodiment. [Figure 10] 1 is a diagram showing the structure of the repeating unit of Kapton in Example 1. [Figure 11] 1 shows the structure of the repeating unit (L-lactic acid) of poly-L-lactic acid in Example 2. [Figure 12] FIG. 2 is a diagram illustrating the distinction between the main chain and the side chain in the structure of the repeating unit (L-lactic acid) of poly-L-lactic acid in Example 2. [Figure 13] 1 shows the structure of the repeating unit of polycarbonate in Example 3. [Figure 14A] FIG. 10 is a diagram illustrating a method for coarse-graining repeating units of polycarbonate in Example 3. [Figure 14B] FIG. 14B is a schematic diagram of a molecular chain containing particles after coarse-graining the repeating unit shown in FIG. 14A. [Figure 15] FIG. 10 is a schematic diagram showing how a higher-order structure is created by superimposing particle chains on a higher-order structure included in a structure DB in Example 3. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, modes for carrying out the present disclosure (hereinafter referred to as "embodiments") will be described with reference to the drawings. In the following description of one embodiment, other embodiments applicable to the first embodiment will also be described as appropriate. The present disclosure is not limited to the following embodiments, and different embodiments may be combined or modified as desired without significantly impairing the effects of the present disclosure. Furthermore, the same components will be given the same reference numerals, and redundant descriptions will be omitted. Furthermore, components having the same functions will be given the same names. The contents shown are merely schematic, and for convenience of illustration, changes may be made from the actual configuration within the scope of not significantly impairing the effects of the present disclosure, and some components may be omitted or modified between drawings.

[0011] 1 is a block diagram of a resin design system 100 and a computing device 20 (an example of a resin design device) according to one embodiment. The design system 100 and the computing device 20 are capable of designing a resin having desired performance by predicting the higher-order structure of the resin based on input information such as raw materials. The resin may be any resin having a variety of structures, such as a linear resin, a resin having a branched structure, or a resin having a network structure.

[0012] The design system 100 includes an input device 10, a structure DB 224, a calculation device 20, and an output device 30. In the illustrated example, the structure DB 224 is provided in the calculation device 20. The input device 10, the calculation device 20, and the output device 30 are connected to each other via electric signal lines (not shown). The calculation device 20 is connected to a server 50 via a network 40. The calculation device 20 (resin design device) includes a parameter DB 223, a structure DB 224, a parameter setting unit 211, a structure determination unit 212, a performance determination unit 213, and an output unit 23.

[0013] The input device 10 is used, for example, by a user, to input input information including at least information that affects the higher-order structure of a designed resin to be designed. The designed resin is a resin whose higher-order structure, performance, etc. are to be designed. The influencing information includes at least one of the repeating units of the designed resin or the raw materials that form the repeating units. The repeating units and raw materials determine the bonding positions of covalent bonds, hydrogen bonds, etc., and the atoms to which they are bonded in the higher-order structure of the designed resin, and both are information that governs (determines) the higher-order structure of the designed resin. Therefore, since this information affects the higher-order structure of the resin, the higher-order structure of the designed resin can be determined by including at least one of these pieces of information.

[0014] The raw material is, for example, a monomer that forms a designed resin by polymerization, and the repeating unit is usually a repeating unit derived from the monomer. Hereinafter, the repeating unit of the designed resin will be referred to as the "target molecular chain," "target chain," etc., as appropriate.

[0015] In the example of the present disclosure, the input information further includes target values ​​of the performance of the designed resin, which allows the target values ​​of the performance to be used in designing the higher-order structure as well as the higher-order structure.

[0016] The input device 10 may include, for example, at least one of a mouse, a keyboard, a touch panel, a display, and the like, none of which are shown.

[0017] 2 is a schematic diagram of an input screen 300 according to one embodiment. The input screen 300 is displayed on a display device (not shown) such as a display or monitor. The input screen 300 is provided with an input field 301 for inputting input information such as raw materials, performance, target performance values, etc.

[0018] The raw material can be specified by inputting, for example, the repeating unit. In addition, by inputting a longer molecular chain, calculations can be made that take into account the stereoregularity of the sequence. In other words, when inputting raw materials, it is best to input a structure that can express as accurately as possible the higher-order structure (steric structure) of the molecular chain of the designed resin.

[0019] The raw material can be input in any format that can generate a higher-order structure. In the illustrated example, the raw material is input in SMILES format in input field 301. The raw material may also be input as a file in which the higher-order structure is represented by coordinates.

[0020] The performance may be, for example, an index that can be normally calculated by molecular dynamics calculation. Specific examples of the performance include mechanical properties such as bulk modulus and tensile strength, glass transition point, density, thermal expansion coefficient, cohesive energy density, and specific heat. In the illustrated example, the performance is the elastic modulus. As the target value of the performance, a required performance value for the performance is input. In the illustrated example, multiple performances and target values ​​of the performance are input, but only one of each may be input.

[0021] 3 is a block diagram of a calculation device 20 according to an embodiment. The calculation device 20 includes a calculation unit 21, a storage unit 22, and an output unit 23. The calculation unit 21 includes a CPU 2100, a GPU 2101, a RAM 2102, etc. The storage unit 22 includes an HDD 2200 (which may be an SSD), a ROM 2201, etc. The output unit 23 includes an I / F 2300, etc. The calculation device 20 is realized by loading a predetermined design program stored in the HDD 2200, the ROM 2201, etc. into the RAM 2102 and executing the program by the CPU 2100.

[0022] Returning to FIG. 1, the storage unit 22 includes a parameter DB 223 and a structure DB 224. In addition, the storage unit 22 also stores a calculation execution program, which is not shown in the figure. DB stands for database. The parameter DB 223 includes parameter information used in a molecular simulation (described below) when determining energy. By including the parameter DB 223, it is possible to execute a molecular simulation (described below) using parameters set for the chain to be designed.

[0023] The parameter information is information about parameters used in calculations in molecular simulations when determining energy. In the example of the present disclosure, the parameter information includes at least one of parameters used in a force field or a pseudopotential method. As parameters used in a pseudopotential method, for example, parameters used in a Projector Augmented Wave method can be used. The molecular simulation is performed by, for example, at least one of molecular dynamics calculations, first-principles calculations, first-principles molecular dynamics calculations, first-principles electronic structure calculations, molecular orbital calculations, and molecular mechanics calculations. This allows the molecular simulation to be performed.

[0024] The parameter information used in molecular simulations varies depending on the calculation method used. For example, in molecular dynamics and molecular mechanics calculations, a force field expressed by an empirical potential function is used. Furthermore, a force field that describes the interactions between coarse-grained particles rather than between atoms can be used.

[0025] Specific examples of force fields that describe the interactions between these atoms (particles) include AMBER, CHARMM, CVFF, GROMOS, OPLS, UFF, COMPASS, MMFF, PCFF, ReaxFF, DREIDING, MM2, MM3, and MM4. Force fields created by machine learning can also be used. In the memory unit 22, files consisting of files describing these force fields are stored in the parameter DB 223.

[0026] 4 is a schematic diagram of a force field file according to one embodiment. In the illustrated example, the force field includes definitions of atom types, functions of energy change when bond lengths, bond angles, and bond dihedral angles are changed according to the atom combination, and coefficients of the functions. For example, a molecular simulation can be performed by using the force field file shown in FIG. 4.

[0027] Returning to Fig. 1, in the case of first-principles calculations and first-principles molecular dynamics calculations, the parameter information included in the parameter DB 223 is, for example, pseudopotential. In particular, when calculations are performed using a plane wave basis, the plane wave basis can be used in combination with the pseudopotential method. At this time, a file describing the parameter information of the pseudopotential is stored in the parameter DB 223 in the storage unit 22 (Fig. 1).

[0028] The structure DB 224 (FIG. 1) includes higher-order structure information, which is information about the positions of atoms in the higher-order structure of a known resin. A known resin is a resin for which information such as constituent atoms and higher-order structure is already known. The structure DB 224 stores a higher-order structure corresponding to each known resin. Therefore, the structure DB 224 stores multiple known resins and their corresponding higher-order structures. Note that in molecular simulations, one or more "atoms" constituting a known resin may be coarsely grained and referred to as "particles." However, for convenience, the following will refer to them collectively as "atoms." The explanation for "atoms" can also be applied to "particles." Note that, in the illustrated example, the structure DB 224 is provided in the computing device 20, but it may also be provided in, for example, a server 50, a system other than the computing device 20, a network (none of which are shown), or the like.

[0029] The higher-order structure information can be created in the same way as a model of the higher-order structure of a resin created by ordinary molecular simulation. Each file constituting the structure DB 224 is created for each known resin, and the higher-order structure of that known resin is described in that file. The file format (extension) can be any format used to create a simulation model, such as .cif, .mol, .pdb, .xyz, .cube, etc.

[0030] In the examples of the present disclosure, each file containing higher-order structural information corresponding to a specific known resin is referred to as structural data, and a collection of these files is referred to as structural DB 224. Therefore, for example, if higher-order structural information regarding a known polyethylene resin is described in structural data 1, higher-order structural information regarding a known polystyrene resin is described in structural data 2, and structural information regarding a known polyimide resin is described in structural data 3, then structural DB 224 includes structural data 1, 2, and 3.

[0031] 5 is a schematic diagram of one of the structure data included in the structure DB 224 according to one embodiment. In the illustrated example, the atom positions included in the structure DB 224 include atomic coordinate data, which allows the positions of the atoms to be identified.

[0032] In the illustrated example, the structural data includes the size of the lattice in which the molecular chains of the known resin are arranged, as well as the x, y, and z coordinates of the atoms. However, the coordinates may be expressed not only in the Cartesian coordinate system represented by x, y, and z, but also in an internal coordinate representation (Z-matrix). In the illustrated example, the coordinates are coordinates with the element labeled C1 as the origin. The space group and crystal system of the lattice are also listed, but these may be omitted. The lattice size is composed of the length of each side and the angle between lattices, but unnecessary information can be omitted. For example, if the coordinates of the atoms are expressed as relative positions (relative coordinates) to the lattice size, the lattice size can be omitted.

[0033] In another embodiment, the atomic positions included in the structure DB 224 include at least one of a vector representing the relationship between two atoms or a function representing the shape of a molecular chain of a known resin. This makes it easier to apply (e.g., fit) the target chain to a higher-order structure than when coordinates are used, and allows for faster calculations. The vector or function representing the atomic positions can be determined, for example, by accumulating the repeating units and higher-order structures of known resins as higher-order structure information.

[0034] 6 is a schematic diagram of structural data according to one embodiment expressed as a vector. In the illustrated example, the vector indicated by vector number 1 has the length of a line segment connecting a start point having the coordinates shown in the figure and an end point having the coordinates shown in the figure. The start point atom and the end point atom are carbon atoms, and these carbon atoms are bonded by a single bond.

[0035] FIG. 7 is a schematic diagram showing structural data according to one embodiment expressed as a function. In the illustrated example, multiple functions are used to represent the molecular chain shape of a known resin. For example, the shape of a molecular chain with molecular chain number 1 is represented by the functions with function numbers 1 and 2. The functions with function numbers 1 and 2 are connected continuously. However, the shape of a molecular chain may be represented by only one function. For example, the function with function number 1 is represented by f1(x, y, z), and the applicable range of function number 1 is from the minimum value (0.0, 0.0, 0.0) to the maximum value (5.0, 5.0, 5.0). In the example of the present disclosure, the function represents only the shape of the molecular chain, including, for example, atomic positions, and does not represent information on individual atoms (element type, bond order, etc.).

[0036] Returning to Figure 5, the molecular chain of a known resin can be described using either an all-atom model or a coarse-grained model. When using an all-atom model, in addition to the coordinates of the atoms, the type of element and bonding information (type of bonding atom and bond order) can be described as shown in Figure 5. When using a coarse-grained model, various levels of coarse-graining can be used, such as the united atom model and the bead-spring model. In the case of a coarse-grained model, the coordinates of particles, which are coarse-grained atoms or groups of atoms, and bonding information between the particles are described.

[0037] In the example of the present disclosure, the coordinate data included in the structure DB 224 are the coordinates of atoms constituting the main chain of the molecular chain of the known resin. As a result, as will be described in detail later, the number of atoms to be applied can be reduced, thereby further reducing the calculation time and calculation cost when creating a higher-order structure. However, the coordinate data may also include the coordinates of atoms constituting the side chain.

[0038] In another embodiment, the coordinate data included in the structure DB 224 are coordinates of particles constituting the main chain of a particle chain obtained by coarsely graining the molecular chain, which will be described in detail later, and thus the number of particles to be superimposed can be reduced, thereby further reducing the calculation time and cost when creating a higher-order structure.

[0039] In the example shown in Figure 5, each atom is marked and labeled. In the atom label column, a label is defined in a format that combines the type of atom and a number. For example, the bond length between atom label C1 and atom label C2 is 1.549 angstroms, and the bond order is D (double bond). Also, for example, the bond amount between atom label O1 and atom label H5 is 1.796 angstroms, and the bond order is a hydrogen bond. Note that each atom does not have to be labeled.

[0040] The higher-order structures included in the structure DB224 preferably include as many types of higher-order structure information (files) as possible that can constitute known resins. As described above, the higher-order structure information can be created in the same manner as a higher-order structure model of a known resin created by ordinary molecular simulation. Therefore, the structure DB224 can be easily constructed by utilizing an existing higher-order structure model. However, from the viewpoint of easily creating the structure DB224, for example, higher-order structure information may be obtained only for representative known resins (e.g., approximately 50 to 100 types) to create the structure DB224. Here, it is preferable that the representative known resins include a variety of higher-order structures. By including a variety of higher-order structures, the higher-order structure of the designed resin can be predicted based on the structure DB224.

[0041] The structure DB224 preferably includes higher-order structure information (files) for known resins with different degrees of crystallinity, from completely amorphous structures to crystalline polymers. It is also preferable to provide multiple higher-order structure information files even for resins with the same degree of crystallinity. The structure DB224 preferably includes higher-order structure information for various main chains, such as hydrocarbon-based, aromatic, and other main chains with polar groups. Higher-order structures vary in the spacing between molecular chains, the period at which the molecular chains fold in the crystalline portion, the size of the helix in the case of a helical structure, or the angle at which the molecular chain bends. Therefore, storing as much higher-order structure information as possible in the structure DB224 can improve the accuracy of higher-order structure prediction.

[0042] The structure DB 224 has an index 2241 for each piece of information about a known resin, and known resins having the same information are associated with the same index 2241. This allows, when determining a higher-order structure, to select only pre-estimated higher-order structure information from all higher-order structures using the index 2241 as a clue, and determine the higher-order structure using only the selected higher-order structure information. This reduces the number of types of higher-order structure information to be referenced, further reducing calculation time.

[0043] In the illustrated example, the heading 2241 indicates the type of molecular chain that constitutes the known resin. Therefore, each piece of higher-order structure information includes the type of molecular chain. If the types of molecular chains are classified as, for example, "straight-chain polymer," "branched polymer," "network polymer," etc., and the raw material entered in the input field 301 (FIG. 2) is a branched polymer, it is sufficient to refer to only the higher-order structure information for "branched polymer" among the higher-order structure information.

[0044] Returning to FIG. 1, the calculation unit 21 includes a parameter setting unit 211, a structure determination unit 212, and a performance determination unit 213.

[0045] The parameter setting unit 211 determines parameter information to be used in molecular simulation based on the parameter DB 223. The determination is performed by executing a program that can be used in molecular simulation, such as molecular dynamics calculation, for example.

[0046] For example, when a molecular simulation is performed using molecular dynamics calculations, the parameter setting unit 211 determines a force field as parameter information based on the parameter DB 223. By determining the parameter information, the most stable bond length, bond angle, and bond dihedral angle between atoms in the chain to be designed are defined. This makes it possible to apply the chain to be designed to higher-order structure information contained in the structure DB 224, calculate the energy of the higher-order structure, and perform calculations for performance evaluation, as will be described in detail later.

[0047] The structure determination unit 212 determines the higher-order structure of the determined resin based on information on atomic positions in the higher-order structure of the known resin in the structure DB 224 and input information, such as raw materials. The determined resin is a resin having a higher-order structure determined by the structure determination unit 212. In the example of the present disclosure, the structure determination unit 212 determines the energy of the higher-order structure having the structure when the structure based on the input information is applied (substituted) to the atomic positions (e.g., coordinate data) included in the structure DB 224. This makes it possible to determine the energy of the higher-order structure of the determined resin from a molecular model created based on the structure DB 224.

[0048] In another embodiment, the structure determination unit 212 applies the atoms constituting the main chain of the repeating unit of the designed resin to the atom positions constituting the main chain among the atom positions contained in the structure DB 224. In higher-order structures, the main chain in particular tends to contribute significantly to the higher-order structure. Therefore, by ignoring the side chains and considering only the main chain, the calculation speed can be increased.

[0049] The structure determination unit 212 determines the energy of the higher-order structure using molecular simulation. By using molecular simulation, the higher-order structure can be predicted without actually producing the designed resin. As described above, the molecular simulation can be performed by at least one of molecular dynamics calculation, first-principles calculation, first-principles molecular dynamics calculation, first-principles electronic structure calculation, molecular orbital calculation, and molecular mechanics calculation. In this case, it is preferable to use a model of a material with the smallest RMSD, which is the energy calculated by the formula (1) described below, as the initial structure, and search for a structure with even lower energy. This allows the energy of a more stable structure to be calculated, allowing the higher-order structure to be predicted with higher accuracy. A structure with lower energy can be searched for, for example, by the conjugate gradient method.

[0050] In the example of the present disclosure, the structure determination unit 212 determines the energy of a higher-order structure having a structure obtained by superimposing atoms contained in a repeating unit (design target chain) of a designed resin onto the atomic positions contained in the structure DB 224. By superimposing atoms in the design target chain one by one onto the atomic positions contained in the structure DB 224, a molecular chain based on raw materials, etc. can be virtually created. Then, the energy of the higher-order structure can be calculated for the determined resin having the virtually created molecular chain.

[0051] The superposition can be performed by any method that superimposes two types of coordinates (the coordinates of the structure DB 224 and the coordinates of each atom included in the target chain). For example, a method can be used in which the coordinates of the target chain are superimposed on the coordinates of each atom in the structure DB 224, and the atomic coordinates of the target chain are optimized so that the RMSD (Root Mean Square Deviation) at that time is equal to or less than a set threshold. The threshold can be set arbitrarily. Also, depending on the structure DB 224, it may not be equal to or less than the set threshold. In that case, the coordinates of the target chain with the smallest RMSD can be used.

[0052] The RMSD is also described in Patent Document 1 and is expressed by formula (1), where N is the number of atoms to be superimposed in the chain to be designed, and δ is the distance between the atoms constituting the chain to be designed and the atoms in the structure DB 224.

[0053]

number

[0054] Even when the atomic positions included in the structure DB 224 are vectors (FIG. 6) or functions (FIG. 7), the higher-order structure can be similarly optimized by optimizing the arrangement of the chain to be designed using formula (1).

[0055] When applying a design target chain to higher-order structure information contained in the structure DB 224, it is preferable to apply the design target chain to each higher-order structure so that the number of atoms is the same. Energy varies depending on the number of atoms. Therefore, by applying the design target chain with the same number of atoms to each higher-order structure, it is possible to accurately compare energies between higher-order structures and determine an appropriate higher-order structure.

[0056] The structure determination unit 212 determines the higher-order structure having the lowest energy among the determined energies as the higher-order structure of the determined resin. Since the higher-order structure having the lowest energy exists most stably, by adopting the higher-order structure having the lowest energy, it is possible to predict the higher-order structure of the actual resin.

[0057] The performance determination unit 214 determines the performance of the determined resin based on the determined higher-order structure. The determination of the performance, i.e., the quantification of the performance, can be performed using any molecular simulation method, such as molecular dynamics calculation or first-principles calculation.

[0058] If the determined performance does not reach the target value entered in the input field 301 (FIG. 2), the performance determination unit 214 outputs to the output device 30 a message prompting the user to re-enter the input information. This allows the output information of the raw materials, etc. to be reset, and the higher-order structure can be predicted again based on the reset output information. The output can be performed, for example, by displaying a screen (not shown) on the display device (not shown) prompting the user to re-enter the information.

[0059] The output unit 23 outputs information about the determined resin as output information through the output device 30. The output unit 23 is connected to an external server 50 via a network 40. Therefore, the calculation device 20 has the function of performing input and output from the external network 40 via the server 50, so the input device 10 and the output device 30 may be omitted. In addition, information about the parameter DB 223 and the structure DB 224 stored in the memory unit 22 may be obtained from the server 50 and the information in the memory unit 22 may be updated.

[0060] As described above, the output device 30 outputs output information relating to the determined resin. The output device 30 is, for example, a display device such as a display, a printer, or the like.

[0061] FIG. 8 is a schematic diagram of an output screen 303 according to one embodiment. In the illustrated example, the output screen 303 displaying output information is displayed on a display device. The output information includes at least one of the determined higher-order structure and the predicted performance of the determined resin. By displaying the higher-order structure, the user can determine from the displayed higher-order structure what higher-order structure is necessary to achieve the desired performance. This can be used as a reference for the user in process design when actually manufacturing the resin. For example, if a structure with low crystallinity is displayed as the predicted higher-order structure, this can be reflected in the manufacturing process, such as by increasing the cooling rate of the resin. Furthermore, by displaying the predicted value, it is possible to visually determine whether the target value entered in the input field 301 ( FIG. 2 ) is exceeded. In the illustrated example, both the higher-order structure and the predicted performance are displayed, and the performance entered in the input field 301 (in the illustrated example, "elastic modulus") is also displayed.

[0062] Fig. 9 is a flowchart showing a resin design method according to one embodiment. Unless otherwise specified, the flowchart shown in Fig. 9 can be executed by the design system 100 and the computing device 20 shown in Fig. 1. Therefore, Fig. 9 will be described below with reference to Fig. 1 as needed.

[0063] When input information such as raw materials, performance, and target values ​​of performance is input to the input device 10, the parameter setting unit 211 sets the raw materials, performance, and target values ​​of performance (step S201). The parameter setting unit 211 determines parameter information to be used based on a target chain determined from the raw materials, and in the illustrated example, assigns a force field to the target chain (step S202). In the illustrated example, the structure determination unit 212 applies the target chain to which the force field has been assigned to higher-order structure information in the structure DB 224 (an example of application) to create a higher-order structure (step S203). The creation of a higher-order structure is performed for each higher-order structure included in the structure DB 224.

[0064] The structure determination unit 212 calculates the energy of the created higher-order structure (a molecular model to which the design target chain is fitted) by molecular simulation such as molecular dynamics calculation (step S204). The structure determination unit 212 compares the energies of each higher-order structure and determines the higher-order structure with the lowest energy (step S205). The above steps S202 to S205 are steps for determining the higher-order structure of the determined resin based on the structure DB 224 and the above input information input to the input device 10.

[0065] The performance determination unit 213 evaluates the performance of the determined resin having the higher-order structure determined in step S205. The output unit 23 outputs the evaluation result to the output device 30 (step S207). Step S207 is a step of outputting information about the determined resin as output information through the output device 30. The performance determination unit 213 determines whether the performance exceeds the target value (step S208), and if it does, ends the calculation (Yes). On the other hand, if it does not exceed the target value (No), the performance determination unit 213 prompts the user to reset the raw materials through the output device 30 (step S209). Once the raw materials have been reset, step S202 and subsequent steps are performed again.

[0066] In the design system 100 and computing device 20 of the present disclosure, and the resin design method of the present disclosure, a design target chain based on raw materials, etc., input by a user is applied to the structure DB 224. Then, the most stable higher-order structure is predicted based on a comparison of energy values, and the performance of the determined resin is evaluated based on the predicted higher-order structure. In this way, it is possible to provide a resin design system, resin design device, and resin structure database that can speed up resin material design and design resins with desired performance in a short period of time.

[0067] Here, the present disclosure will be explained while touching upon the technology described in Patent Document 1. As described above, the technology described in Patent Document 1 is a method for predicting protein structure. When generating the three-dimensional structure of a protein to be investigated, the three-dimensional structure is superimposed on a template protein.

[0068] Like resins, proteins are a type of polymer, but they have a structure consisting of 20 types of amino acids, each composed of carbon, hydrogen, nitrogen, phosphorus, oxygen, and sulfur atoms. The difference between proteins and resins is that the types of constituent elements and the types of structures (amino acids) that make up the polymer are both limited. Resins may contain elements such as fluorine, chlorine, bromine, iodine, silicon, and other metal elements in addition to the above elements. While proteins are limited to amino, carboxyl, sulfide, and hydroxyl groups in terms of polar groups, resins possess a much wider variety of polar groups. The diversity of polar group types leads to a diversity of interactions between polar groups, such as hydrogen bonds, which ultimately affects the diversity of higher-order structures.

[0069] Furthermore, unlike amino acids, which have aliphatic main chains, resins may have aromatic rings that form the main chain. The structure of the main chain also affects the higher-order structure. Therefore, the higher-order structures that molecular chains can take are far more diverse than proteins. Thus, for example, in terms of the diversity of elemental species and structures, proteins and resins are completely different. Therefore, it is difficult to conceive of a method for determining the higher-order structure of resins simply from a method for predicting the three-dimensional structure of proteins.

[0070] The key elements in completing this disclosure are the following two points. The first is to create a diverse structure DB224 that can accommodate the diversity of resin constituent elements and molecular chain structures. This improves the predictability of higher-order structures. The second is to inductively predict higher-order structures by applying (e.g., superimposing) the target chain to each of the higher-order structures contained in the structure DB224. Although the structure DB224 contains structures that the target chain cannot stably form, such unstable structures can be excluded by energy comparison. Therefore, the presence of such unstable structures has little impact on higher-order structure prediction. Note that in protein three-dimensional structure prediction, a template structure is selected based on the amino acid sequence, which is a completely different procedure from the inductive higher-order structure determination flow of the present disclosure. [Example]

[0071] The present disclosure will be described in more detail below based on examples.

[0072] [Example 1] Figure 10 shows the structure of the repeating unit of Kapton (known resin) in Example 1. In Example 1, a polyimide (designed resin) with a small thermal expansion coefficient was searched for. In step S201 (Figure 9), a polyimide (hereinafter referred to as Kapton) produced from pyromellitic dianhydride (PMDA) and 4,4'-oxydianiline (ODA) was set as the raw material to be investigated, and the repeating unit (design target chain) shown in Figure 10 was input in SMILES format. A mouse and keyboard, which are input device 10 (Figure 1), were used for input. The performance was the thermal expansion coefficient, and the target performance value was set to 30 ppm / K or less.

[0073] To evaluate the performance by classical molecular dynamics calculation, a force field was applied to the Kapton repeating unit (design target chain shown in Figure 10) in step S202 (Figure 9). The force field used was CVFF (Consistent Valence Force Field).

[0074] Next, in step S203, the Kapton molecular chain (FIG. 10) was superimposed on each of the higher-order structures contained in the structure DB224 (an example of application). Because Kapton is a linear polymer, only higher-order structures classified as "linear" in the index 2241 were used. The Kapton molecular chain was superimposed on 500 patterns of higher-order structures related to linear polymers from the higher-order structures contained in the structure DB224.

[0075] Next, the energy of each of the 500 higher-order structures of Kapton created by superimposing them onto the higher-order structure was calculated by molecular dynamics calculation (step S204). The structure showing the lowest energy value was determined as the predicted higher-order structure (step S205). From the above, the higher-order structure could be determined by performing energy calculation once for each of the 500 higher-order structures. LAMMPS was used as the software for performing the molecular dynamics calculation.

[0076] The thermal expansion coefficient was calculated for the determined higher-order structure (the higher-order structure of the determined resin) (step S206). Using molecular dynamics calculations, a simulation was performed under constant pressure conditions with varying temperatures to calculate the change in volume with temperature. As a result, the thermal expansion coefficient was determined to be 28 ppm / K. These results were displayed on the monitor, which is the output device 30 (Figure 1) (step S207). From the above, it was found that Kapton met the target performance.

[0077] [Example 2] FIG. 11 shows the structure of the repeating unit (L-lactic acid; target chain for design) of poly-L-lactic acid in Example 2. In Example 2, the glass transition temperature (one example of performance) of poly-L-lactic acid, a type of biomass plastic, having the repeating unit shown in FIG. 11 was evaluated. The target value was set to 25°C or higher, which is room temperature. The input information was input in mol (MDL file) format, in which the structure of poly-L-lactic acid (including at least the repeating unit) was converted into coordinates. The input device 10 (FIG. 1) was the same as in Example 1.

[0078] To evaluate the performance using classical molecular dynamics calculations, a force field was applied to L-lactic acid (the target chain shown in Figure 11) in step S202 (Figure 9). At this time, the United Atom method was used, which represents the methyl group of L-lactic acid as a single particle. The DREIDING force field was used.

[0079] FIG. 12 is a diagram illustrating the distinction between the main chain and the side chain in the structure of the repeating unit (L-lactic acid) of poly-L-lactic acid in Example 2. In step S203, a molecular chain (design target chain) was superimposed on each of the higher-order structures contained in the structure DB224, as in Example 1. However, in the structure DB224, not all atoms but only atoms constituting the main chain are represented in vector format. Therefore, the coordinates of the oxygen atom in box 901 and the methyl group in box 902, both of which are side chains, are not used. Note that the methyl group in box 902 is represented as a single particle at the completion of step S203. Then, only the molecular chain in box 903, which is the main chain of L-lactic acid, was applied to the vector format data.

[0080] Based on the coordinates of the main chain applied in step S203, the coordinates of the oxygen atom of the double bond in the side chain and the methyl group (represented by one particle) were determined, and the higher-order structure of poly-L-lactic acid prepared based on each higher-order structure data was obtained.

[0081] In step S204, the higher-order structure created in step S203 was used as the initial structure, and further structural optimization was performed using the conjugate gradient method, after which the energy was calculated. Of the 500 higher-order structure patterns, the structure with the lowest energy was determined as the higher-order structure of poly-L-lactic acid. Structural optimization calculations were performed approximately 100 times per higher-order structure. Since the number of higher-order structures stored in the structure DB 224 was 500, the higher-order structure was determined through a total of 100 (times per higher-order structure) x 500 (pattern higher-order structures) = 50,000 calculations.

[0082] The determined higher-order structure was a structure in which part of the helical higher-order structure had crystallized. A molecular dynamics calculation was then performed to calculate the volume while changing the temperature. The specific volume versus temperature was calculated, and the temperature at the inflection point was evaluated as the glass transition point. As a result, the glass transition point was predicted to be 55°C, achieving the target performance value. The above results were displayed on the monitor, which is the output device 30 (Figure 1), as in Example 1.

[0083] [Example 3] FIG. 13 shows the structure of the polycarbonate repeating unit in Example 3. In Example 3, polycarbonate was used as the raw material in order to search for a resin with a high tensile modulus of elasticity. The performance was determined by the tensile modulus, and the target performance value was set to 1 GPa or more. The polycarbonate repeating unit (design target chain) shown in FIG. 13 was input in SMILES format. Unlike Examples 1 and 2, a server 50 was used instead of the input device 10.

[0084] In step S202, potential parameters were applied to polycarbonate using a coarse-grained molecular dynamics method, with reference to the description in "A. Kubo, et al., "Construction of master yield stress curves for polycarbonate: A coarse-grained molecular dynamics study", Polymer, Vol. 177, pp. 84-90 (2019)".

[0085] Fig. 14A is a diagram illustrating a method for coarse-graining the repeating units of polycarbonate in Example 3. Fig. 14B is a schematic diagram of a molecular chain containing particles after coarse-graining the repeating units shown in Fig. 14A. In Fig. 14B, the phenylene group is named particle A, the isopropylidene group is named particle B, and the carbonate group is named particle C, and these are shown bonded by bond 1101. Note that hereinafter, particles A, B, and C may be collectively referred to as particles 1102.

[0086] In this method, instead of atoms, phenylene groups, isopropylidene groups, and carbonate groups are considered as particles A, B, and C, respectively, as shown in particles 1102, as shown in Figures 14A and 14B, and particles 1102 are connected to each other with bonds 1101 that resemble "springs" to represent a particle chain (equivalent to a molecular chain in the case of atoms). The parameter information describes the interactions between particles 1102 for molecular chains that include particles 1102, as shown in Figure 14B. Coarse-graining reduces the number of particles to be handled, further speeding up calculations to create higher-order structures.

[0087] In step S203, a molecular chain (design target chain) was superimposed on each of the higher-order structures contained in the structure DB224, as in Example 1. However, in the structure DB224, the coordinates of particle chains containing coarse-grained particles are expressed in the form of a function, similar to the notation method for polycarbonate. Therefore, the polycarbonate (design target chain) expressed as a particle chain created in step S202 was superimposed on the particle chain in functional form. In this way, by reducing the number of atoms to be considered through coarse-graining, the amount of calculation required for superposition can be further reduced.

[0088] Figure 15 is a schematic diagram of how a higher-order structure is created by superimposing a particle chain on a higher-order structure contained in structure DB224 in Example 3. In Figure 15, only a portion of the higher-order structure is cut out and depicted. Function 1201 is a portion of the function contained in structure DB224 before superimposition. Structure 1202 is a structure in the process of superimposing a polycarbonate particle chain (design target chain) on function 1201. Structure 1203 is the structure of the polycarbonate particle chain after superimposition is complete.

[0089] In step S204, similar to Example 2, structural optimization of the higher-order structure was further performed using the conjugate gradient method, followed by energy calculation, and the structure with the lowest energy was determined as the higher-order structure of the polycarbonate. The structural optimization calculation using the conjugate gradient method was performed approximately 100 times per higher-order structure. Since the number of higher-order structure information stored in the structure DB 224 was 500, the higher-order structure was determined by a total of 100 (times per higher-order structure) x 500 (pattern higher-order structure) = 50,000 calculations.

[0090] The determined higher-order structure was a completely amorphous structure. A molecular dynamics calculation was performed to calculate the tensile modulus of polycarbonate, and the result predicted a tensile modulus of 2.4 GPa, achieving the target performance value. The results were output to an external output device (not shown) via server 50.

[0091] [Comparative Example 1] The thermal expansion coefficient of Kapton calculated in Example 1 was calculated by molecular dynamics calculation with reference to the conditions described in Non-Patent Document 1 above.

[0092] A higher-order structure was created using 10 Kapton molecular chains each consisting of 20 repeating units. First, the structure of the Kapton molecular chain consisting of 20 repeating units was created using software for creating a simulation model, and a force field (using COMPASS) was applied. These 10 molecular chains were randomly placed in a simulation box, and then 3 million constant-pressure and temperature molecular dynamics calculations and 3 million constant-volume and temperature molecular dynamics calculations were performed to create the higher-order structure of Kapton. Furthermore, 3 million constant-pressure and temperature molecular dynamics calculations and 10 million constant-volume and temperature molecular dynamics calculations were performed again to obtain the final higher-order structure of Kapton. Therefore, a total of 19 million calculations were performed to create the higher-order structure of Kapton. This number of calculations is significantly greater than the number of calculations shown in Examples 1 to 3.

[0093] When the thermal expansion coefficient was calculated based on the fabricated higher-order structure using the same method as in Example 1, the thermal expansion coefficient obtained was 28 ppm / K, the same as in Example 1. Therefore, it was found that according to the present disclosure, it is possible to predict the performance of a resin with the same accuracy as conventional methods, but with a much smaller number of calculations than conventional methods. [Explanation of symbols]

[0094] 10 Input Devices 100 Design Systems 20 Arithmetic unit 21 Arithmetic section 211 Parameter setting section 212 Structure determination part 213 Performance determining section 214 Performance determining section 22 Memory section 224 Structure DB (Structure Database) 2241 Headline 23 Output section 30 Output Devices 300 Input Screen 301 Input field 303 Output Screen

Claims

1. an input device into which input information including at least information that will affect the higher-order structure of a resin to be designed is input; a structural database containing information on the positions of atoms in the higher-order structures of known resins; a computing device including a structure determination unit that determines the higher-order structure of a determined resin based on information about atomic positions in the higher-order structure of the known resin in the structure database and the input information; an output device that outputs output information regarding the determined resin; A resin design system characterized by:

2. The structure determination unit determines the energy of a higher-order structure having a structure based on the input information when the structure is applied to the atomic positions included in the structure database. The resin design system according to claim 1 .

3. The structure determination unit determines the energy of a higher-order structure having a structure obtained by superimposing atoms contained in a repeating unit of the designed resin onto the atomic positions. The resin design system according to claim 2 .

4. The structure determination unit determines the higher-order structure having the lowest energy among the determined energies as the higher-order structure of the determined resin. The resin design system according to claim 3 .

5. The structure determination unit determines the energy of the higher-order structure using a molecular simulation. The resin design system according to any one of claims 2 to 4.

6. The molecular simulation is performed by at least one of molecular dynamics calculation, first-principles calculation, first-principles molecular dynamics calculation, first-principles electronic structure calculation, molecular orbital calculation, and molecular mechanics calculation. The resin design system according to claim 5 .

7. The computing device includes a parameter database containing parameter information including at least one of parameters used in the force field or pseudopotential method used in the molecular simulation. The resin design system according to claim 5 .

8. The structure determination unit applies atoms constituting a main chain of a repeating unit of the designed resin to atomic positions constituting a main chain among the atomic positions included in the structure database. The resin design system according to any one of claims 2 to 4.

9. The atomic positions include coordinate data of the atoms. The resin design system according to any one of claims 1 to 4.

10. The coordinate data is the coordinates of atoms constituting the main chain of the molecular chain of the known resin. The resin design system according to claim 9 .

11. The coordinate data is the coordinates of particles constituting a molecular chain among particle chains obtained by coarse-graining the molecular chain. The resin design system according to claim 9 .

12. The atomic positions include at least one of a vector representing the relationship between two atoms and a function representing the shape of the molecular chain of the known resin. The resin design system according to any one of claims 1 to 4.

13. The influencing information includes at least one of the repeating unit of the designed resin and the raw material forming the repeating unit. The resin design system according to any one of claims 1 to 4.

14. The input information further includes a target value for the performance of the designed resin. The resin design system according to claim 13.

15. the computing device includes a performance determination unit that determines performance of the determined resin based on the determined higher-order structure; When the determined performance does not reach the target value, the performance determination unit outputs to the output device a prompt to re-input the input information. The resin design system according to claim 14.

16. the structural database includes an index for each piece of information about the known resin; The known resins having the same information are associated with the same heading. The resin design system according to any one of claims 1 to 4.

17. The output information includes at least one of the determined higher-order structure and the predicted performance of the determined resin. The resin design system according to any one of claims 1 to 4.

18. a structural database containing information on the positions of atoms in the higher-order structures of known resins; and a structure determination unit that determines the higher-order structure of a resin to be determined based on input information including at least information that influences the higher-order structure of a resin to be designed and information on atomic positions in the higher-order structure of the known resin in the structure database. A resin design device characterized by:

19. determining the higher-order structure of the resin to be designed based on a structure database containing information on the atomic positions in the higher-order structures of known resins and input information input to an input device and containing at least information that will affect the higher-order structure of the resin to be designed; and outputting the information about the determined resin as output information through an output device. A resin design method characterized by:

Images