Characteristic evaluation method for solid electrolyte, manufacturing method for solid electrolyte and characteristic evaluation system for solid electrolyte

Abstract

To provide a characteristic evaluation method for a solid electrolyte, a manufacturing method for a solid electrolyte, and a characteristic evaluation system for a solid electrolyte that enable the efficient determination of solid electrolytes capable of fabricating all-solid-state lithium-ion batteries with excellent battery characteristics through simulation.SOLUTION: A characteristic evaluation method for a solid electrolyte that extracts multiple physical property and crystal structure data sets for compositions containing specified elements listed in Materials Project, makes a crystal lattice along their abc axes into supercell at equal scales to ensure the total number of atoms exceeds a predetermined value for each crystal structure of composition, performs structure optimization calculation with a machine learned potential based on the crystal lattice made into supercell to calculate the Li diffusion coefficient, ranks the compositions either by the Li diffusion coefficient or by the activation energy calculated from the Li diffusion coefficient, then evaluates the property of the solid electrolyte corresponding to the ranked composition on the basis of the ranked composition.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a method for evaluating the properties of a solid electrolyte, a method for producing a solid electrolyte, and a system for evaluating the properties of a solid electrolyte. [Background technology]

[0002] With the recent rapid spread of information-related devices and communication devices such as personal computers, video cameras, and mobile phones, the development of batteries to be used as their power sources has become increasingly important. Among these batteries, lithium-ion batteries have attracted attention due to their high energy density. Furthermore, high energy density and improved battery characteristics are also required for lithium secondary batteries for large-scale applications such as vehicle-mounted power sources and load leveling.

[0003] However, in the case of lithium-ion batteries, the electrolyte is mostly organic compounds, and even if a flame-retardant compound is used, it cannot be said that the risk of fire is completely eliminated. All-solid-state lithium-ion batteries, which use a solid electrolyte, have been attracting attention in recent years as a potential alternative to liquid-based lithium-ion batteries (Patent Document 1). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2006-244734 Summary of the Invention [Problem to be solved by the invention]

[0005] Due to the recent demand for all-solid-state lithium-ion batteries, there is a demand for further research and development of solid electrolyte compounds having novel compositions that have good battery characteristics.

[0006] A good solid electrolyte is expected to be a material that has a low energy barrier (activation energy) required for lithium ions to move through the crystal lattice. A low activation energy contributes to high lithium ion conductivity. Therefore, from the perspective of improving battery performance, a solid electrolyte with a low activation energy is preferable.

[0007] However, although it is reliable to search for promising compositions by producing new compositions one by one, fabricating all-solid-state lithium-ion batteries, and evaluating their battery properties, it has a major disadvantage in terms of efficiency. For this reason, there is a need for a new method for evaluating the properties of solid electrolytes that allows battery properties to be evaluated by simulation without actually manufacturing them.

[0008] The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to provide a method for evaluating the properties of a solid electrolyte, a method for producing a solid electrolyte, and a system for evaluating the properties of a solid electrolyte, which are capable of efficiently determining, through simulation, a solid electrolyte that can be used to fabricate an all-solid-state lithium-ion battery with good battery properties. [Means for solving the problem]

[0009] The present invention, which was completed based on the above findings, is defined below. 1. A method for evaluating the properties of a solid electrolyte, comprising: extracting multiple pieces of physical property and crystal structure data for compositions containing specified elements listed in the Materials Project; supercelling the abc axis of the crystal lattice at the same magnification so that the total number of atoms in each crystal structure of the composition exceeds a specified number; calculating the diffusion coefficient of Li by performing a structural optimization calculation using a machine learning potential based on the supercelled crystal lattice; ranking the compositions by the diffusion coefficient of Li or by the activation energy calculated from the diffusion coefficient of Li; and evaluating the properties of the solid electrolyte corresponding to the composition based on the ranked compositions. 2. Extracting a plurality of pieces of physical property and crystal structure data for compositions containing predetermined elements listed in the Materials Project; a step of excluding data relating to compositions containing protons, rare gases, and actinide elements from the physical property and crystal structure data, and further excluding compositions that satisfy the following formula 1 from the remaining compositions to create composition list data; E_above_hull > 0.05 ··· (Formula 1) a step of forming a supercell of each crystal structure of a composition remaining after further excluding compositions that satisfy formula 1, with the abc axis of the crystal lattice being magnified to the same size so that the total number of atoms exceeds a predetermined number; A step of performing a structural optimization calculation using a machine learning potential based on the supercelled crystal lattice, and performing an enthalpy calculation at absolute zero using the optimized data; a step of performing an NVT-MD calculation at a predetermined temperature for the most stable structure at absolute zero obtained by the structural optimization calculation, and predicting the diffusion coefficient of Li; calculating a mean square displacement (MSD) at a predetermined temperature based on the predicted Li diffusion coefficient, and calculating the Li diffusion coefficient based on the MSD; adding the calculated Li diffusion coefficient to the composition list data; a step of ranking the compositions listed in the composition list data by the diffusion coefficient of Li or the activation energy calculated from the diffusion coefficient of Li; Evaluating the properties of the solid electrolytes corresponding to the compositions based on the ranked compositions; 2. The method for evaluating the properties of a solid electrolyte according to claim 1, comprising: 3. The method for evaluating the properties of a solid electrolyte according to 2 above, wherein in the step of performing a structural optimization calculation using a machine learning potential based on the supercelled crystal lattice and performing an enthalpy calculation at absolute zero using the optimized data, the structural optimization calculation is performed using Matlantis (registered trademark). 4. The method for evaluating the properties of a solid electrolyte according to 2 above, wherein in the step of performing an NVT-MD calculation at a predetermined temperature for the most stable structure at absolute zero obtained by the structural optimization calculation and predicting the diffusion coefficient of Li, the NVT-MD calculation is performed using Matlantis (registered trademark). 5. A step of determining a predetermined composition from a solid electrolyte whose characteristics have been evaluated by the method for evaluating characteristics of a solid electrolyte according to any one of 1 to 4 above; a step of weighing and mixing raw materials based on the determined composition to obtain a mixed powder; sintering the mixed powder; A method for producing a solid electrolyte, comprising: 6. An extraction unit that extracts multiple pieces of physical property and crystal structure data for compositions containing predetermined elements listed in the Materials Project; a simulation calculation unit for supercelling the abc axis of the crystal lattice at the same magnification so that the total number of atoms in each crystal structure of the composition exceeds a predetermined number; a structural optimization calculation unit that performs structural optimization calculations using machine learning potentials based on the supercelled crystal lattice; a Li diffusion coefficient calculation unit that performs an NVT-MD calculation on the most stable structure at absolute zero obtained by the structural optimization calculation, predicts a Li diffusion coefficient, and calculates an MSD (mean square displacement) based on the predicted Li diffusion coefficient, thereby calculating the Li diffusion coefficient; a ranking unit that ranks the compositions by the diffusion coefficient of Li or by activation energy calculated from the diffusion coefficient of Li; a characteristic evaluation unit that evaluates the characteristics of a solid electrolyte corresponding to the composition based on the ranked compositions; and A solid electrolyte characterization system equipped with 7. The extraction unit extracts a plurality of pieces of physical property and crystal structure data of compositions containing predetermined elements described in the Materials Project, excludes data relating to compositions containing protons, rare gases, and actinide elements from the physical property and crystal structure data, and further excludes compositions that satisfy the following formula 1 from the remaining compositions to create composition list data, E_above_hull > 0.05 ··· (Formula 1) the simulation calculation unit further excludes compositions that satisfy Formula 1, and in each crystal structure of the remaining compositions, supercells the abc axis of the crystal lattice at the same magnification so that the total number of atoms exceeds a predetermined number; the structural optimization calculation unit performs a structural optimization calculation using a machine learning potential based on the supercelled crystal lattice, and performs an enthalpy calculation at absolute zero using the optimized data; the Li diffusion coefficient calculation unit performs an NVT-MD calculation at a predetermined temperature for the most stable structure at absolute zero obtained by the structural optimization calculation, predicts a Li diffusion coefficient, calculates an MSD (mean square displacement) at the predetermined temperature based on the predicted Li diffusion coefficient, and calculates the Li diffusion coefficient based on the MSD; 7. The solid electrolyte characteristic evaluation system according to claim 6, wherein the ranking unit adds the calculated Li diffusion coefficient to the composition list data, and ranks the compositions listed in the composition list data by the Li diffusion coefficient or by the activation energy calculated from the Li diffusion coefficient. 8. The solid electrolyte characteristic evaluation system according to 7, wherein the structural optimization calculation unit performs the structural optimization calculation using Matlantis (registered trademark). 9. The solid electrolyte characteristic evaluation system according to 7, wherein the Li diffusion coefficient calculation unit performs the NVT-MD calculation using Matlantis (registered trademark). [Effects of the Invention]

[0010] The present invention can provide a method for evaluating the properties of a solid electrolyte, a method for producing a solid electrolyte, and a system for evaluating the properties of a solid electrolyte, which can efficiently determine, through simulation, a solid electrolyte that can be used to fabricate an all-solid-state lithium-ion battery with good battery properties. [Brief explanation of the drawings]

[0011] [Figure 1]1 is a flowchart showing a procedure for determining the composition of a chloride-based solid electrolyte according to an embodiment of the present invention. [Figure 2] FIG. 1 is a schematic diagram of an all-solid-state lithium-ion battery according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0012] Next, embodiments of the present invention will be described in detail with reference to the drawings. It should be understood that the present invention is not limited to the following embodiments, and that appropriate design changes and improvements may be made based on the ordinary knowledge of those skilled in the art without departing from the spirit of the present invention.

[0013] (solid electrolyte) The solid electrolyte to be evaluated in the embodiments of the present invention is not particularly limited, and may be a sulfide-based solid electrolyte, a chloride-based solid electrolyte, or other solid electrolytes. In the embodiments of the present invention, a chloride-based solid electrolyte will be described as an example to be evaluated.

[0014] In recent years, advances in computers and computational techniques have enabled highly accurate calculations of H (enthalpy) using first-principles calculations based on quantum mechanics. First-principles calculations are a computational method for determining the electronic state of a material system without reference to experimental results. Based on quantum mechanics, they use atomic number and system structure as input parameters to ab initio elucidate physical mechanisms and predict physical properties. In first-principles calculations, crystal structure information is input as initial information, and enthalpy information can be obtained by using a generalized gradient approximation. Here, G (free energy) = H (enthalpy) - TS (temperature × entropy), so free energy and enthalpy are equal at absolute zero (0 K). While it is difficult to experimentally measure thermodynamic parameters such as enthalpy at absolute zero, it is known that experimental and calculated values are equal near room temperature for solids. The enthalpy at absolute zero for chloride-based solid electrolytes according to embodiments of the present invention can also be calculated using first-principles calculations. Software used for first-principles calculations includes Vienna Ab initio Simulation Package: VASP, ABINIT, and QUANTUM ESPRESSO.

[0015] The enthalpy at absolute zero of a chloride-based solid electrolyte may also be calculated using machine learning potential (neural network potential). Machine learning potential is an AI model trained using the results of large-scale quantum chemical calculations performed by a supercomputer as training data, and has the advantage of being able to perform calculations with similar accuracy thousands of times faster than the above-mentioned first-principles calculations. Examples of software used for machine learning potential include Matlantis (registered trademark). A method for calculating the enthalpy at absolute zero of a chloride-based solid electrolyte using Matlantis will be described in detail below in the section on characterization methods for solid electrolytes.

[0016] The activation energy of the chloride-based solid electrolyte is preferably 28.0 [kJ / mol] or less. When the activation energy is 28.0 [kJ / mol] or less, the energy barrier required for lithium ions to migrate through the crystal lattice is low, thereby improving ionic conductivity and battery characteristics. The activation energy of a chloride-based solid electrolyte evaluated as having a promising composition by the method for evaluating the properties of a solid electrolyte according to an embodiment of the present invention is more preferably 20.0 [kJ / mol] or less, and even more preferably 15.0 [kJ / mol] or less.

[0017] The ionic conductivity of a chloride-based solid electrolyte evaluated as having a promising composition by the method for evaluating the properties of a solid electrolyte according to an embodiment of the present invention is preferably 0.5 mS / cm or more, more preferably 8.0 mS / cm or more, and even more preferably 1.0 mS / cm or more at 30°C.

[0018] (Method for evaluating the properties of solid electrolytes) A method for evaluating the properties of a solid electrolyte according to an embodiment of the present invention involves extracting multiple pieces of physical property and crystal structure data for compositions containing specified elements described in the Materials Project, supercelling the abc axis of the crystal lattice at the same magnification so that the total number of atoms in each crystal structure of the composition exceeds a specified number, calculating the diffusion coefficient of Li by performing a structural optimization calculation using a machine learning potential based on the supercelled crystal lattice, ranking the compositions by the diffusion coefficient of Li or by the activation energy calculated from the diffusion coefficient of Li, and evaluating the properties of the solid electrolyte corresponding to the composition based on the ranked compositions.

[0019] Hereinafter, a procedure for determining the composition of a chloride-based solid electrolyte having an enthalpy at absolute zero of -1.8132 to -0.7429 eV / atom as a chloride-based solid electrolyte evaluated as a promising composition will be described. That is, the composition of a chloride-based solid electrolyte having this enthalpy (evaluated as a promising composition) can be determined by the procedure described in the flowchart shown in FIG.

[0020] <Step 1> The physical properties and crystal structure data of compositions containing Li and Cl (or Li, O, Cl) listed in the Materials Project are extracted using an API (Application Programming Interface). The Materials Project is a database of first-principles calculation results, and in addition to crystal structures, it also contains various data on materials, such as band structures, thermodynamic quantities, phase diagrams, and magnetic moments. Examples of compositions containing Li and Cl include LiGaCl4 and LiFeCl4. Examples of compositions containing Li, O, and Cl include LiAlS3(Cl2O3)2 and LiFeMoClO4.

[0021] <Step 2> Next, similarly, the physical properties and crystal structure data of compositions containing Li and halogen species: M (or Li, O, M) listed in the Materials Project are extracted, and all M is replaced with Cl and added to the data in Step 1. M is a halogen species other than Cl (F, Br, I), and examples of compositions containing Li and M include Li2VF6, K2LiSbBr6, and Cs2LiYI6. Examples of compositions containing Li, O, and M include LiVOF3, Li4Al3Ge3BrO 12 , Li4Ga3Si3IO 12 By substituting halogen species other than Cl (F, Br, I) for Cl, Li2VCl6, K2LiSbCl6, Cs2LiYCl6, LiVOCl3, Li4Al3Ge3ClO 12 , Li4Ga3Si3ClO 12 Compositions obtained by element substitution in this way include compositions containing Li and Cl (or Li, O, Cl) that are not listed in the Materials Project, so candidates for chloride-based solid electrolytes according to embodiments of the present invention (for example, compositions that provide favorable battery characteristics for all-solid-state lithium-ion batteries) can be selected from a wide range of compositions other than those listed in the Materials Project. Note that step 2 is not essential, and this step may be omitted.

[0022] <Step 3> Next, from the data obtained in step 2, data relating to compositions containing protons, rare gases, and actinide elements are excluded. Specifically, compositions containing H, He, Ne, Ar, Kr, Xe, Ac, Th, Pa, U, Np, Pu, etc. are excluded. Examples of compositions to be excluded include LiThF5, LiHF2, LiH2ClO5, and Li6Zr6HCl. 18 , Li4H3ClO3, Li2UC l6 etc. Next, from the remaining compositions, compositions that satisfy the following formula 1 are further excluded. E_above_hull > 0.05 ··· (Formula 1) Here, "E_above_hull" is a thermodynamic index that indicates how energetically unfavorable it is from the thermodynamic convex hull, and is data listed in the Materials Project. Stable phases indicate 0. A small "E_above_hull" indicates that the phase is easy to produce experimentally (stable). Therefore, excluding compositions that satisfy Equation 1 means that only compositions with an "E_above_hull" of 0.05 or less remain, indicating that the extracted data will only include very stable compositions.

[0023] <Step 4> Next, for each crystal structure of the composition obtained in step 3, the abc axis of the crystal lattice is made into a supercell at the same magnification so that the total number of atoms is more than 60. The crystal structure is data listed in the Materials Project. Supercelling is one of the simulation calculation methods, and is a method in which the cell (periodically repeated unit) when simulating with periodic boundary conditions is made into a supercell larger than the theoretically smallest unit cell (unit lattice). Under periodic boundary conditions, the macroscopic system (10 23 Microscopic systems (10 order) that can be handled by simulation 9Boundaries are set so that the total number of atoms can be expressed in units of order (order of magnitude). However, if the cell size for MD calculations is small due to the boundary conditions, the target atom will be affected by overlapping contributions from each atom, making it impossible to obtain the correct interactions (energy and force). For this reason, the cell size is typically set to at least twice the potential cutoff distance (in Matlantis®, described below, this is up to r = 6 Å between neighboring atoms). The above issue can be resolved by using the Supercell method to expand the cell in each abc axis direction (expanding by a factor of α as shown in the formula below). The above "total number of atoms exceeding 60" is an arbitrary threshold; the unit cells of each composition material extracted using API are transformed to the same size using the formula below so that the number of atoms exceeds 60. α (integer) = Function that returns the smallest integer greater than or equal to the given number ((60 ÷ number of atoms in the unit cell) (1 / 3) )

[0024] <Step 5> Next, based on the crystal lattice supercelled in step 4, a structural relaxation (structural optimization) calculation is performed using a machine learning potential. More specifically, the structural relaxation calculation is performed using Matlantis (registered trademark), a neural network potential. The structural relaxation calculation calculates the locally stable coordinates of the system and their energy, and optimizes the force applied to each atom. The structural relaxation calculation also optimizes the model to perform the NVT-MD calculation described below with the optimal lattice length. In the structural relaxation, the initial coordinates of the system are determined, and a locally stable point is found from the initial coordinates to find the stable structure. The structural relaxation algorithm uses the Matlantis Atomic Simulation Environment (ASE), a local structural relaxation algorithm. The structural relaxation calculation is performed using the optimization algorithm implemented in the ASE. Specifically, the local structural relaxation algorithm, ASE, uses the BFGS method, which approximates the Hessian from the optimization trajectory and executes the Newton method algorithm using the approximate Hessian. Calculate the enthalpy at absolute zero using the data optimized by the BFGS method (iterator: iterator = 10, maximum argument value: fmax = 5.0 × 10-3 ) will be implemented.

[0025] Specifically, the enthalpy calculation is carried out in the following procedure. (1) Calculate the energy of a single atom (E_atom) for each element and compile it in a CSV table (a data sheet for enthalpy calculations). (2) Next, for the stable structure (energy E0) that was created as a Supercell in step 5 above and subjected to structural relaxation calculations, use the data sheet in (1) to calculate the enthalpy (H_0) at absolute zero using the following formula: H_0 = E0-Σ(E_atom) Here, E0 is the ground state energy, and Σ(E_atom) is the sum of the energies of each element in the composition.

[0026] <Step 6> Next, Matlantis performs NVT-MD (100 ps) calculations at a specific temperature (800 K) on the most stable structure (ground state) at absolute zero obtained in the structural relaxation calculation in Step 5. MD simulation is a computer simulation of the physical motion of atoms and molecules. It solves Newton's equations of motion in classical mechanics for each atom that makes up a substance, tracking the time evolution of atomic positions and energy to predict the physical properties of the system. NVT-MD is a canonical ensemble (canonical state distribution) MD simulation, which assumes constant temperature and volume and can be used when volume and temperature changes are to be ignored. The NVT-MD (100 ps) calculation predicts the diffusion coefficient D of Li. When measuring the diffusivity (e.g., diffusion coefficient) of Li at room temperature, the time scale is on the order of several hours. However, such time scales cannot be reproduced in atomic simulations, so the simulation is performed on an extremely short calculation scale of 100 ps. The jump frequency of Li self-diffusion is very small at room temperature, so it is unrealistic to observe Li hopping to neighboring sites due to vibrations around a fixed point at around 100 ps. For this reason, we performed simulations at a high temperature of 800 K to observe Li hopping due to thermal vibrations and evaluate the diffusivity of Li in the material.

[0027] <Step 7> Next, based on the diffusion coefficient D of Li obtained in step 6, the MSD (mean square displacement) at a specific temperature (800K) is calculated. MSD is a statistical processing index that indicates the magnitude of movement, and is the average value of the square of the distance between the start point and end point of movement within a certain time span. The formula for calculating MSD is shown in Equation (2) below.

[0028] [ka]

[0029] In equation (2), N is the number of atoms in the grain boundary, r α (t+t0) is the position of the first atom at time t[s], rα (t0) indicates the position of the first atom in the initial configuration.

[0030] Next, the diffusion coefficient of Li (D) [m 2 / s] is calculated. MSD = 2dDt (Equation 3) In equation (3), d represents the dimension of diffusion, and t represents time [s]. The diffusion coefficient D [m 2 / s] is calculated from the slope of the MSD, where the horizontal axis is time [ps] and the vertical axis is MSD [Å], and the units are converted as follows: ps, Å system → m 2 / sec conversion D = (slope of MSD) × 1e -16 (Å 2 )×1e 12 (ps) / 6

[0031] <Step 8> Next, add the diffusion coefficients obtained in step 7 to the table listing the compositions (compositions obtained in step 3) listed in the Materials Project. This will create a list that combines the physical properties of each composition with its diffusion coefficient.

[0032] <Step 9> Next, the compositions listed in the table obtained in step 8 are ranked by lithium diffusion coefficient based on materials reported in publicly known experimental papers (previously reported materials).

[0033] <Step 10> Next, from the compositions ranked in Step 9, those with Li diffusion coefficients above a predetermined value are selected and labeled with their structural features. Structural features include whether the compound is a layered compound or not, and whether the cation polyhedra in the crystal are joined by point contacts. Layered compounds are acceptable, but because the Li diffusion path in layered compounds is a layered structure (2D plane), they may have inferior diffusivity compared to materials with 3D Li diffusion. Furthermore, structures in which the cation polyhedra in the crystal are joined by point contacts are expected to be particularly good ionic conductors. The large number of interstitial spaces is expected to form paths for Li atoms, increasing the 3D diffusivity of Li within the material. For example, in the crystal structure of LiTaCl6, TaCl octahedra and LiCl octahedra are joined by point contacts. In this way, compositions with a Li diffusion coefficient equal to or greater than a predetermined value and favorable structural characteristics are selected as promising candidate compositions for chloride-based solid electrolytes (e.g., compositions that provide favorable battery characteristics for all-solid-state lithium-ion batteries). 12 , or Li3Zr4Cl 19 , LiVOCl3, LiVOCl4, Li6V2OCl 11 , Li2V2OC l6 , Li2VOCl4, Li2VOCl5, Li3VOCl5, etc. are selected as promising candidate compositions for chloride-based solid electrolytes.

[0034] In the embodiment of the present invention, the diffusion coefficient (D) [m 2 / s] to the activation energy (E A ) [J / mol] can be calculated based on the Arrhenius equation shown in the following formula (4).

[0035] [ka]

[0036] In equation (4), D0 is the frequency factor, R is the gas constant [J / mol·K], and T is the absolute temperature [K].

[0037] Furthermore, the obtained diffusion coefficient of Li (D) [m 2 / s] to (D) [cm 2 / s], the ionic conductivity (σ) [S / cm] can be calculated based on the Nernst-Einstein equation shown in the following formula (5).

[0038] [ka]

[0039] In equation (5), z is the valence of the ion, F is the Faraday constant [C / mol], and c is the concentration of the ion [mol / cm 3 ], R is the gas constant [J / mol K], and T is the absolute temperature [K].

[0040] The composition of the chloride-based solid electrolyte may be determined from the promising candidate compositions obtained as described above. Specifically, the activation energy (E A ) is 28.0 [kJ / mol] or less can be considered to be a composition that provides favorable battery characteristics for an all-solid-state lithium ion battery. In this way, in the method for evaluating the properties of a solid electrolyte according to the embodiment of the present invention, it is possible to efficiently determine a solid electrolyte that can be used to fabricate an all-solid-state lithium ion battery with good battery properties through simulation, without actually manufacturing the battery.

[0041] (Method of manufacturing solid electrolyte) In the method for producing a solid electrolyte according to an embodiment of the present invention, after the composition is determined by the method for evaluating the properties of a solid electrolyte according to an embodiment of the present invention as described above, a chloride-based solid electrolyte according to an embodiment of the present invention is actually produced as follows. First, raw materials are weighed to obtain a predetermined composition in a glove box filled with an inert gas atmosphere such as argon gas or nitrogen gas. Examples of the raw materials used here include LiCl, LiVO, LiTaO, LiFe, LiBi, LiGa, and LiZr. Other known compounds of Li with V, Ta, Fe, Bi, Ga, or Zr may also be used as raw materials.

[0042] Next, the mixture is mixed in a mortar or the like for 5 to 30 minutes to prepare a mixed powder. At this time, it is preferable to mix for a time such that the average particle size of the mixed powder becomes 5 to 40 μm.

[0043] Next, the mixed powder is pelletized and vacuum sealed in a quartz ampoule, and the quartz ampoule is fired at 400 to 800°C for 1 to 20 hours to produce a chloride-based solid electrolyte. a M b O c Cl d (wherein M is one or two of V, Ta, Fe, Sb, Ga, W, and Zr, and 1.0≦a≦3.0, 1.0≦b≦4.0, 0≦c≦1.0, 3.0≦d≦19.0).

[0044] (All-solid-state lithium-ion battery) A solid electrolyte layer is formed using a chloride-based solid electrolyte produced by the method for producing a solid electrolyte according to an embodiment of the present invention, and an all-solid-state lithium-ion battery can be produced that includes the solid electrolyte layer, a positive electrode layer, and a negative electrode layer. The positive electrode layer and the negative electrode layer that constitute the all-solid-state lithium-ion battery are not particularly limited and can be formed of known materials, and can have a known configuration as shown in FIG.

[0045] The positive electrode layer of the lithium ion battery may be a layer of a positive electrode mixture obtained by mixing a known positive electrode active material for lithium ion batteries with the chloride-based solid electrolyte prepared in the embodiment of the present invention or another solid electrolyte.

[0046] The positive electrode mixture may further contain a conductive additive. Examples of the conductive additive include carbon materials, metal materials, and mixtures thereof. The conductive additive may include at least one element selected from the group consisting of carbon, nickel, copper, aluminum, indium, silver, cobalt, magnesium, lithium, chromium, gold, ruthenium, platinum, beryllium, iridium, molybdenum, niobium, osmium, rhodium, tungsten, and zinc. The conductive additive is preferably a highly conductive carbon element, or a metal element, mixture, or compound containing carbon, nickel, copper, silver, cobalt, magnesium, lithium, ruthenium, gold, platinum, niobium, osmium, or rhodium. Examples of the carbon material include carbon black, such as ketjen black, acetylene black, denka black, thermal black, and channel black; graphite; carbon fiber; and activated carbon.

[0047] The average thickness of the positive electrode layer of the lithium ion battery is not particularly limited and can be appropriately designed depending on the purpose. The average thickness of the positive electrode layer of the lithium ion battery may be, for example, 1 μm to 100 μm, or 1 μm to 10 μm.

[0048] The method for forming the positive electrode layer of the lithium ion battery is not particularly limited and can be appropriately selected depending on the purpose. Examples of the method for forming the positive electrode layer of the lithium ion battery include sputtering using a target material of the positive electrode active material and compression molding of the positive electrode active material.

[0049] The negative electrode layer of the lithium ion battery may be a layer of a known negative electrode active material for lithium ion batteries, or may be a layer of a negative electrode mixture obtained by mixing a known negative electrode active material for lithium ion batteries with the chloride-based solid electrolyte according to the embodiment of the present invention or another chloride-based solid electrolyte.

[0050] The negative electrode layer, like the positive electrode layer, may contain a conductive additive. The conductive additive may be the same material as described for the positive electrode layer. Examples of the negative electrode active material include carbon materials, specifically, artificial graphite, graphite carbon fiber, resin-baked carbon, pyrolytic vapor-grown carbon, coke, mesocarbon microbeads (MCMB), furfuryl alcohol resin-baked carbon, polyacene, pitch-based carbon fiber, vapor-grown carbon fiber, natural graphite, and non-graphitizable carbon, or mixtures thereof. Examples of the negative electrode material include metals such as lithium metal, indium metal, aluminum metal, and silicon metal, as well as alloys of these metals combined with other elements or compounds.

[0051] The average thickness of the negative electrode layer of the lithium ion battery is not particularly limited and can be appropriately selected depending on the purpose. The average thickness of the negative electrode layer of the lithium ion battery may be, for example, 1 μm to 100 μm, or 1 μm to 10 μm.

[0052] The method for forming the negative electrode layer of the lithium ion battery is not particularly limited and can be appropriately selected depending on the purpose. Examples of the method for forming the negative electrode layer of the lithium ion battery include sputtering using a target material of the negative electrode active material, compression molding of the negative electrode active material, and vapor deposition of the negative electrode active material.

[0053] The average thickness of the solid electrolyte layer of the lithium ion battery formed using the chloride-based solid electrolyte according to the embodiment of the present invention is not particularly limited and can be appropriately designed depending on the purpose. The average thickness of the solid electrolyte layer of the lithium ion battery may be, for example, 50 μm to 500 μm, or 50 μm to 100 μm.

[0054] The method for forming the solid electrolyte layer of the lithium ion battery is not particularly limited and can be appropriately selected depending on the purpose. Examples of the method for forming the solid electrolyte layer of the lithium ion battery include sputtering using a target material for the solid electrolyte and compression molding of the solid electrolyte.

[0055] Other components constituting the lithium ion battery are not particularly limited and can be appropriately selected depending on the purpose. Examples include a positive electrode current collector, a negative electrode current collector, and a battery case.

[0056] The size and structure of the positive electrode current collector are not particularly limited and can be appropriately selected depending on the purpose. Examples of materials for the positive electrode current collector include die steel, stainless steel, aluminum, aluminum alloys, titanium alloys, copper, gold, and nickel. The positive electrode current collector may be in the form of, for example, a foil, a plate, or a mesh. The average thickness of the positive electrode current collector may be, for example, 10 μm to 500 μm, or 50 μm to 100 μm.

[0057] The size and structure of the negative electrode current collector are not particularly limited and can be appropriately selected depending on the purpose. Examples of materials for the negative electrode current collector include die steel, gold, indium, nickel, copper, and stainless steel. The negative electrode current collector may be in the form of, for example, a foil, a plate, or a mesh. The average thickness of the negative electrode current collector may be, for example, 10 μm to 500 μm, or 50 μm to 100 μm.

[0058] The battery case is not particularly limited and can be appropriately selected depending on the purpose, and examples thereof include known laminate films that can be used in conventional all-solid-state batteries, such as resin laminate films and films in which metal is vapor-deposited on resin laminate films. The shape of the battery is not particularly limited and can be appropriately selected depending on the purpose. Examples include cylindrical, square, button, coin, and flat types.

[0059] (Solid Electrolyte Characterization System) The solid electrolyte property evaluation system according to an embodiment of the present invention includes an extraction unit that extracts multiple pieces of physical property and crystal structure data for compositions containing specified elements described in the Materials Project; a simulation calculation unit that supercells the abc axis of the crystal lattice at the same magnification so that the total number of atoms in each crystal structure of the composition exceeds a specified number; a structural optimization calculation unit that performs structural optimization calculations using a machine learning potential based on the supercelled crystal lattice; a Li diffusion coefficient calculation unit that performs NVT-MD calculations on the most stable structure at absolute zero obtained by the structural optimization calculation, predicts the Li diffusion coefficient, and calculates the MSD (mean square displacement) based on the predicted Li diffusion coefficient to calculate the Li diffusion coefficient; a ranking unit that ranks compositions by the Li diffusion coefficient or the activation energy calculated from the Li diffusion coefficient; and a property evaluation unit that evaluates the properties of the solid electrolyte corresponding to the composition based on the ranked compositions.

[0060] In the solid electrolyte property evaluation system according to an embodiment of the present invention, the extraction unit may extract multiple pieces of physical property and crystal structure data for compositions containing predetermined elements listed in the Materials Project, exclude data on compositions containing protons, rare gases, and actinide elements from the physical property and crystal structure data, and further exclude compositions that satisfy the following formula 1 from the remaining compositions to create composition list data. E_above_hull > 0.05 ··· (Formula 1) The extraction unit may also carry out steps 1 to 3 in the method for evaluating properties of a solid electrolyte according to the embodiment of the present invention described above.

[0061] In the solid electrolyte property evaluation system according to an embodiment of the present invention, the simulation calculation unit may further exclude compositions that satisfy Formula 1, and in each crystal structure of the remaining compositions, supercell the abc axis of the crystal lattice at the same magnification so that the total number of atoms exceeds a predetermined number. The simulation calculation unit may also perform step 4 in the method for evaluating the properties of a solid electrolyte according to the embodiment of the present invention.

[0062] In the solid electrolyte property evaluation system according to an embodiment of the present invention, the structural optimization calculation unit may perform structural optimization calculations using a machine learning potential based on a supercelled crystal lattice, and may perform enthalpy calculations at absolute zero using the optimized data. The structural optimization calculation unit may also perform step 5 in the method for evaluating the properties of a solid electrolyte according to the embodiment of the present invention.

[0063] In the solid electrolyte characteristic evaluation system according to an embodiment of the present invention, the Li diffusion coefficient calculation unit may perform an NVT-MD calculation at a predetermined temperature for the most stable structure at absolute zero obtained by the structural optimization calculation, predict the Li diffusion coefficient, calculate the MSD (mean square displacement) at the predetermined temperature based on the predicted Li diffusion coefficient, and calculate the Li diffusion coefficient based on the MSD. The Li diffusion coefficient calculation unit may also perform steps 6 and 7 in the method for evaluating the properties of a solid electrolyte according to the embodiment of the present invention.

[0064] In the solid electrolyte characteristic evaluation system according to an embodiment of the present invention, the ranking unit may add the calculated Li diffusion coefficient to the composition list data, and rank the compositions listed in the composition list data by the Li diffusion coefficient or by the activation energy calculated from the Li diffusion coefficient. The ranking unit may also perform steps 8 and 9 in the method for evaluating the properties of a solid electrolyte according to the embodiment of the present invention.

[0065] In the solid electrolyte characteristic evaluation system according to the embodiment of the present invention, the characteristic evaluation unit evaluates the characteristics of the solid electrolyte corresponding to the composition based on the ranked compositions. The characteristic evaluation unit may also perform step 10 in the method for evaluating the characteristics of a solid electrolyte according to the embodiment of the present invention described above.

[0066] The solid electrolyte characterization system according to the embodiment of the present invention can be operated in a computer or a computer network, and may include a processor, program storage memory, data storage memory, input / output devices, and support circuits that are conventional components known in the art, for executing the processes in the extraction unit, simulation calculation unit, structural optimization calculation unit, Li diffusion coefficient calculation unit, ranking unit, characterization unit, etc. [Example]

[0067] The following examples are provided to provide a better understanding of the present invention and its advantages, but the present invention is not limited to these examples.

[0068] Example 1 In Example 1, the composition of the chloride-based solid electrolyte was determined as follows, based on the flow chart shown in FIG.

[0069] <Step 1> First, the physical properties of compositions containing Li and Cl, and the physical properties of compositions containing Li, O, and Cl, as well as the respective crystal structure data, were extracted using an API (Application Programming Interface).

[0070] <Step 2> Next, similarly, the physical properties of compositions containing Li and halogen species: M, as well as the physical properties of compositions containing Li, O, and M, as well as the respective crystal structure data, were extracted, and all M was replaced with Cl and added to the data in Step 1. M was set to halogen species other than Cl (F, Br, I).

[0071] <Step 3> Next, data on compositions containing protons, rare gases, and actinide elements were excluded from the data obtained in Step 2. Specifically, compositions containing H, He, Ne, Ar, Kr, Xe, Ac, Th, Pa, U, Np, Pu, etc. were excluded. Next, from the remaining compositions, compositions that satisfy the following formula 1 were further excluded. E_above_hull > 0.05 ··· (Formula 1) This resulted in extracting data for only very stable compositions.

[0072] <Step 4> Next, for each crystal structure of the composition obtained in Step 3, the abc axis of the crystal lattice was magnified to create a supercell so that the total number of atoms was greater than 60. The crystal structure data was listed in the Materials Project. The "total number of atoms greater than 60" was determined arbitrarily using a single threshold, and the unit cell of each composition material extracted using API was transformed to the same size using the following formula so that the number of atoms was greater than 60. α (integer) = Function that returns the smallest integer greater than or equal to the given number ((60 ÷ number of atoms in the unit cell) (1 / 3) )

[0073] <Step 5> Next, based on the crystal lattice supercelled in step 4, structural relaxation (structural optimization) calculations were performed using a machine learning potential. More specifically, structural relaxation calculations were performed using Matlantis (registered trademark), a neural network potential. The structural relaxation algorithm used was the ASE (Atomic Simulation Environment) of Matlantis, a local structural relaxation algorithm, and structural relaxation calculations were performed using the optimization algorithm implemented in the ASE. Specifically, in the ASE, a local structural relaxation algorithm, the BFGS method was used, which approximates the Hessian from the optimization trajectory and executes the Newton method algorithm using the approximate Hessian. The data optimized by the BFGS method was used to calculate the enthalpy at absolute zero (iterator: iter = 10, maximum argument value: fmax = 5.0 × 10 -3 ) was carried out.

[0074] The enthalpy calculation was carried out specifically according to the following procedure. (1) The energy of a single atom (E_atom) for each element was calculated and compiled in a CSV table (a data sheet for enthalpy calculation). (2) Next, for the stable structure (energy E0) that was created as a Supercell in step 5 above and subjected to structural relaxation calculations, the enthalpy (H_0) at absolute zero was calculated using the data sheet in (1) using the following formula: H_0 = E0-Σ(E_atom) Here, E0 is the ground state energy, and Σ(E_atom) is the sum of the energies of each element in the composition.

[0075] <Step 6> Next, NVT-MD (100 ps) calculations were performed at a specific temperature (800 K) using Matlantis for the most stable structure (ground state) at absolute zero obtained in the structural relaxation calculation in Step 5. The diffusion coefficient D of Li was predicted using this NVT-MD (100 ps) calculation.

[0076] <Step 7> Next, the mean square displacement (MSD) at a specific temperature (800 K) was calculated based on the diffusion coefficient D of Li obtained in step 6. The formula for calculating the MSD is shown in the following formula (2).

[0077] [ka]

[0078] In equation (2), N is the number of atoms in the grain boundary, r α (t+t0) is the position of the first atom at time t[s], r α (t0) indicates the position of the first atom in the initial configuration.

[0079] Next, the diffusion coefficient of Li (D) [m 2 / s] was calculated. MSD = 2dDt (Equation 3) In equation (3), d represents the dimension of diffusion, and t represents time [s]. The diffusion coefficient D [m 2 / s] is calculated from the slope of the MSD, where the horizontal axis is time [ps] and the vertical axis is MSD [Å], and the units are converted as follows: ps, Å system → m 2 / sec conversion D = (slope of MSD) × 1e -16 (Å 2 )×1e 12 (ps) / 6

[0080] <Step 8> Next, the diffusion coefficients obtained in Step 7 were added to the table listing the compositions (compositions obtained in Step 3) listed in the Materials Project. This resulted in a list combining the physical properties of each composition with its diffusion coefficient.

[0081] <Step 9> Next, the compositions listed in the table obtained in step 8 were ranked by lithium diffusion coefficient based on materials reported in publicly known experimental papers (previously reported materials).

[0082] <Step 10> Next, from the compositions ranked in step 9, we selected those with a Li diffusion coefficient of 7.3428E-06 [cm / s] or higher for chloride-based solid electrolytes and 4.98053E-06 [cm / s] or higher for oxychloride-based solid electrolytes, and labeled their structural features. The structural feature was whether the cation polyhedra in the crystal were joined by point contact (corner sharing) or line contact (corner + edge sharing). In this way, compositions with a diffusion coefficient equal to or greater than a predetermined value and favorable structural characteristics were selected as promising candidate compositions for chloride-based solid electrolytes (for example, compositions that provide favorable battery characteristics for all-solid-state lithium-ion batteries). In this way, LiVCl5, LiVCl6, Li2VCl5, Li2VCl6, LiTaCl6, LiFeCl4, LiGaCl4, LiWCl6, LiSb2Cl7, Li3V2Cl 12 , or Li3Zr4Cl 19 , LiVOCl3, LiVOCl4, Li6V2OCl 11 , Li2V2OC l6 , Li2VOCl4, Li2VOCl5, and Li3VOCl5 were selected as the composition of the chloride-based solid electrolyte according to Example 1.

[0083] In addition, the diffusion coefficient of Li (D) [m 2 / s] to the activation energy (E A ) [J / mol] was calculated based on the Arrhenius equation shown in the following formula (4).

[0084] [ka]

[0085] In equation (4), D0 is the frequency factor, R is the gas constant [J / mol·K], and T is the absolute temperature [K].

[0086] From the above simulation, LiVCl5, LiVCl6, Li2VCl5, Li2VCl6, LiTaCl6, LiFeCl4, LiGaCl4, LiWCl6, LiSb2Cl7, Li3V2Cl were selected as the composition of the chloride-based solid electrolyte according to Example 1. 12 , or Li3Zr4Cl 19 , LiVOCl3, LiVOCl4, Li6V2OCl 11 , Li2V2OC l6 , Li2VOCl4, Li2VOCl5, Li3VOCl5, the diffusion coefficient of Li (D) [m 2 / s], enthalpy at absolute zero (0 K) [eV / atom], activation energy (E A ) [J / mol] and structural characteristics are shown in Table 1 below. In Table 1, there are several examples where the evaluation results differ even for the same composition, but this is because, as mentioned above, halide materials are substituted with chlorine and added as chloride data. Also, because the crystal symmetry of the original structure (e.g., orthorhombic, cubic, tetragonal) is maintained, the diffusion coefficient values may differ even for the same composition.

[0087] [Table 1]

[0088] (Consideration) According to the simulation described in Example 1, the composition formula: Li a M b O c Cl d (wherein M is one or two of V, Ta, Fe, Sb, Ga, W, and Zr, and 1.0≦a≦3.0, 1.0≦b≦4.0, 0≦c≦1.0, 3.0≦d≦19.0) Several chloride-based solid electrolytes were obtained with absolute zero enthalpies of -1.8132 to -0.7429 eV / atom and low activation energies of 28.0 kJ / mol or less. All of these were predicted to have good ionic conductivity when used in all-solid-state lithium-ion batteries.

Claims

1. a method for evaluating the properties of a solid electrolyte, comprising: extracting a plurality of pieces of physical property and crystal structure data for compositions containing predetermined elements described in the Materials Project; supercelling the abc axes of the crystal lattice at the same magnification in each crystal structure of the compositions so that the total number of atoms is greater than a predetermined number; calculating a diffusion coefficient of Li by performing a structural optimization calculation using a machine learning potential based on the supercelled crystal lattice; ranking the compositions by the diffusion coefficient of Li or by activation energy calculated from the diffusion coefficient of Li; and evaluating the properties of a solid electrolyte corresponding to the composition based on the ranked compositions.

2. extracting a plurality of physical property and crystal structure data of compositions containing predetermined elements described in the Materials Project; a step of excluding data relating to compositions containing protons, rare gases, and actinide elements from the physical property and crystal structure data, and further excluding compositions that satisfy the following formula 1 from the remaining compositions to create composition list data; E_above_hull > 0.05... (Formula 1) a step of supercelling the abc axes of the crystal lattice at the same magnification in each crystal structure of the remaining compositions after further excluding compositions that satisfy the formula 1 so that the total number of atoms exceeds a predetermined number; A step of performing a structural optimization calculation using a machine learning potential based on the supercelled crystal lattice, and performing an enthalpy calculation at absolute zero using the optimized data; a step of performing an NVT-MD calculation at a predetermined temperature for the most stable structure at absolute zero obtained by the structural optimization calculation, and predicting the diffusion coefficient of Li; calculating a mean square displacement (MSD) at a predetermined temperature based on the predicted Li diffusion coefficient, and calculating the Li diffusion coefficient based on the MSD; adding the calculated diffusion coefficient of Li to the composition list data; A step of ranking the compositions listed in the composition list data by the diffusion coefficient of Li or the activation energy calculated from the diffusion coefficient of Li; Evaluating the properties of the solid electrolytes corresponding to the compositions based on the ranked compositions; The method for evaluating the properties of a solid electrolyte according to claim 1 , comprising:

3. 3. The method for evaluating properties of a solid electrolyte according to claim 2, wherein in the step of performing a structural optimization calculation using a machine learning potential based on the supercelled crystal lattice and performing an enthalpy calculation at absolute zero using the optimized data, the structural optimization calculation is performed using Matlantis (registered trademark).

4. 3. The method for evaluating properties of a solid electrolyte according to claim 2, wherein in the step of performing an NVT-MD calculation at a predetermined temperature for the most stable structure at absolute zero obtained by the structural optimization calculation and predicting a diffusion coefficient of Li, the NVT-MD calculation is performed using Matlantis (registered trademark).

5. determining a predetermined composition from a solid electrolyte whose characteristics have been evaluated by the method for evaluating characteristics of a solid electrolyte according to any one of claims 1 to 4; Weighing and mixing raw materials based on the determined composition to obtain a mixed powder; sintering the mixed powder; A method for producing a solid electrolyte, comprising:

6. an extraction unit that extracts a plurality of pieces of physical property and crystal structure data of compositions containing predetermined elements described in the Materials Project; a simulation calculation unit that supercells the abc axis of the crystal lattice at the same magnification so that the total number of atoms in each crystal structure of the composition exceeds a predetermined number; a structural optimization calculation unit that performs structural optimization calculation using a machine learning potential based on the supercelled crystal lattice; a Li diffusion coefficient calculation unit that calculates the Li diffusion coefficient by performing an NVT-MD calculation on the most stable structure at absolute zero obtained by the structural optimization calculation, predicting the Li diffusion coefficient, and calculating an MSD (mean square displacement) based on the predicted Li diffusion coefficient; a ranking unit that ranks the compositions by the diffusion coefficient of Li or by activation energy calculated from the diffusion coefficient of Li; a characteristic evaluation unit that evaluates the characteristics of a solid electrolyte corresponding to the composition based on the ranked compositions; and A solid electrolyte characterization system equipped with

7. the extraction unit extracts a plurality of pieces of physical property and crystal structure data of compositions containing predetermined elements described in the Materials Project, excludes data on compositions containing protons, rare gases, and actinide elements from the physical property and crystal structure data, and further excludes compositions that satisfy the following formula 1 from the remaining compositions to create composition list data; E_above_hull > 0.05... (Formula 1) the simulation calculation unit further excludes compositions that satisfy Formula 1, and in each crystal structure of the remaining compositions, supercells the abc axes of the crystal lattice at the same magnification so that the total number of atoms exceeds a predetermined number; The structural optimization calculation unit performs a structural optimization calculation using a machine learning potential based on the supercelled crystal lattice, and performs an enthalpy calculation at absolute zero using the optimized data; the Li diffusion coefficient calculation unit performs an NVT-MD calculation at a predetermined temperature for the most stable structure at absolute zero obtained by the structural optimization calculation, predicts a Li diffusion coefficient, calculates an MSD (mean square displacement) at the predetermined temperature based on the predicted Li diffusion coefficient, and calculates a Li diffusion coefficient based on the MSD; 7. The solid electrolyte characteristic evaluation system according to claim 6, wherein the ranking unit adds the calculated Li diffusion coefficient to the composition list data, and ranks the compositions listed in the composition list data by the Li diffusion coefficient or by the activation energy calculated from the Li diffusion coefficient.

8. 8. The solid electrolyte characteristic evaluation system according to claim 7, wherein the structural optimization calculation unit performs the structural optimization calculation using Matlantis (registered trademark).

9. 8. The solid electrolyte characteristic evaluation system according to claim 7, wherein the Li diffusion coefficient calculation unit performs the NVT-MD calculation using Matlantis (registered trademark).

Images