Prediction method and manufacturing method of all-solid-state battery

Abstract

To provide a predictive method for evaluating the performance of an all-solid-state battery before assembly.SOLUTION: A prediction method for predicting the performance of an all-solid-state battery includes a preparation step of preparing a negative electrode active material, an analysis step of performing XRD analysis on the negative electrode active material to obtain physical property information of the negative electrode active material, and a prediction step of predicting a charging resistance value of an all-solid-state battery using the negative electrode active material based on the physical property information.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present disclosure relates to prediction methods and methods for manufacturing all-solid-state batteries. [Background technology]

[0002] Information-related devices and communication devices such as personal computers, video cameras, and mobile phones are becoming widespread. Furthermore, from the perspective of reducing the burden on the environment, electric vehicles and other motor-driven vehicles are becoming more common. Accordingly, various studies are being conducted on the batteries used as power sources for these devices.

[0003] For example, Patent Document 1 discloses a state-of-charge measuring device that includes a voltage detection means for detecting the voltage of a secondary battery, a current detection means for detecting the current of the secondary battery, and a measurement means for measuring the state of charge of the secondary battery, and that calculates the resistance of the secondary battery using the voltage detected by the voltage detection means and the current detected by the current detection means, and measures the state of charge using the resistance. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2013-142649 Summary of the Invention [Problem to be solved by the invention]

[0005] Here, it is assumed that the inspection (battery performance evaluation) to determine whether a battery (all-solid-state battery) is a good product or not will be performed on the finished battery (battery after assembly). However, if the inspection is performed after the battery is assembled, the materials and parts used in the battery that is determined to be a bad product will be lost, which may reduce battery productivity.

[0006] The present disclosure has been made in view of the above circumstances, and a main object of the present disclosure is to provide a prediction method that can evaluate the performance of an all-solid-state battery before assembly. [Means for solving the problem]

[0007] [1] A method for predicting the performance of an all-solid-state battery, the method comprising: a preparation step of preparing a negative electrode active material; an analysis step of performing XRD analysis on the negative electrode active material to obtain physical property information of the negative electrode active material; and a prediction step of predicting a charging resistance value of an all-solid-state battery using the negative electrode active material based on the physical property information.

[0008] [2] The negative electrode active material is Li4Ti5O 12 The prediction method according to [1], wherein the negative electrode active material has a crystalline phase, and in the XRD step, the degree of crystallinity of the crystalline phase in the negative electrode active material is obtained as the physical property information.

[0009] [3] The prediction method according to [1] or [2], wherein in the prediction step, the prediction is made based on a trained model that has previously been machine-learned to describe the relationship between the physical property information of the negative electrode active material and the charging resistance value of the all-solid-state battery, and the physical property information acquired in the analysis step.

[0010] [4] A method for manufacturing an all-solid-state battery having a positive electrode active material layer, a negative electrode active material layer, and a solid electrolyte layer disposed between the positive electrode active material layer and the negative electrode active material layer, the method comprising: a first step of carrying out the prediction method according to any one of [1] to [3]; and a second step of forming the negative electrode active material layer using the negative electrode active material predicted in the first step to have the charging resistance value equal to or less than a threshold value.

[0011] [5] The negative electrode active material for which the charging resistance value was predicted to be equal to or less than the threshold value was Li4Ti5O 12 The method for producing an all-solid-state battery according to [4], wherein the all-solid-state battery has a crystalline phase, and the degree of crystallinity of the crystalline phase is 98.4% or more. [Effects of the Invention]

[0012] The present disclosure has an effect of enabling the performance of an all-solid-state battery to be evaluated before assembly. [Brief explanation of the drawings]

[0013] [Figure 1] FIG. 1 is a schematic cross-sectional view illustrating an example of an all-solid-state battery according to the present disclosure. [Figure 2] 10 is a graph showing the results of machine learning in the example. DETAILED DESCRIPTION OF THE INVENTION

[0014] The prediction method and the manufacturing method of the all-solid-state battery according to the present disclosure will be described in detail below.

[0015] A. Forecasting Method The prediction method of the present disclosure is a method for predicting the performance of an all-solid-state battery. In the prediction method of the present disclosure, first, a negative electrode active material is prepared (preparation step). Next, XRD analysis is performed on the negative electrode active material to obtain physical property information of the negative electrode active material (analysis step). Then, based on the physical property information, a charging resistance value of an all-solid-state battery using the negative electrode active material is predicted (prediction step).

[0016] According to the present disclosure, the performance of an all-solid-state battery can be predicted based on the physical property information of the negative electrode active material obtained by XRD analysis, and therefore the performance of the all-solid-state battery can be evaluated before assembly. Note that, in this specification, "battery performance" means "charging resistance" unless otherwise specified.

[0017] As described above, it is assumed that performance evaluation is performed on assembled batteries. If the battery is determined to be non-defective, the materials and parts used in the battery will be lost in production. In contrast, the prediction method disclosed herein predicts the performance of an all-solid-state battery from the physical properties of the all-solid-state battery's material (negative electrode active material). In other words, it is possible to determine whether a manufactured battery is non-defective before the battery is assembled. Therefore, by using a negative electrode active material predicted to provide good performance for the all-solid-state battery, it is possible to manufacture all-solid-state batteries with reduced production loss and improve productivity.

[0018] 1.Preparation process The preparation step in the present disclosure is a step of preparing a negative electrode active material.

[0019] The negative electrode active material is preferably crystalline. "Crystalline" means having at least one crystalline phase. The negative electrode active material may be a single-phase material having one crystalline phase, or a multi-phase material having two or more crystalline phases. In particular, the negative electrode active material is Li4Ti5O 12 It is preferred that it has a crystalline phase.

[0020] Examples of the negative electrode active material include Li2TiO3 and Li4Ti5O 12 and lithium titanates (LTO-based negative electrode active materials) such as Li2Ti2O5; and TiNb2O7 and Ti2Nb 10 O 29 Examples of niobium titanate composite oxides (NTO-based negative electrode active materials) include:

[0021] The shape of the negative electrode active material is, for example, particulate. The negative electrode active material may be primary particles or secondary particles formed by agglomeration of a plurality of primary particles. The average particle size (D 50 ) is, for example, 10 nm or more and 50 μm or less. 50 ) refers to the cumulative 50% particle size in the volume-based particle size distribution measured by a laser diffraction particle size distribution analyzer. 50 ) may be the average particle size of primary particles or the average particle size of secondary particles. 50 ) refers to the cumulative 50% particle size in the volume-based particle size distribution measured by a laser diffraction particle size distribution analyzer.

[0022] 2.Analysis process The analyzing step in the present disclosure is a step of performing an XRD analysis on the negative electrode active material to obtain physical property information of the negative electrode active material.

[0023] The XRD analysis can be performed by a conventionally known method. Examples of physical property information obtained from the XRD analysis include peak information such as peak positions, peak intensities at predetermined positions, and peak intensity ratios between multiple peaks, as well as crystallinity information obtained from the peak information. The crystallinity can be calculated using the Li4Ti5O 12 The crystallinity of (LTO) can be mentioned.

[0024] Here, the crystallinity (crystallinity of LTO) can be calculated, for example, by the following formula. Crystallinity (%) = (LTO peak area / total peak area) × 100

[0025] 3. Prediction process The prediction step in the present disclosure is a step of predicting the charge resistance value of an all-solid-state battery using the negative electrode active material, based on the physical property information.

[0026] The prediction is preferably performed, for example, based on a trained model that has previously been machine-learned to describe the relationship between the physical property information of the negative electrode active material and the charging resistance value of the all-solid-state battery, and the physical property information acquired in the analysis step. Machine learning is usually performed before the preparation step.

[0027] Although the machine learning algorithm is not particularly limited, it is preferably supervised learning. For example, LightGBM (Light Gradient Boosting Machine) can be used as the supervised learning algorithm. Note that the supervised learning method in the present disclosure is regression analysis such as multiple regression analysis and short regression analysis.

[0028] Furthermore, in the machine learning, it is preferable to fix manufacturing parameters other than the physical property information of the negative electrode active material, such as the type of positive electrode active material, the type of solid electrolyte, the composition of the positive electrode active material layer, and the composition of the solid electrolyte layer. This is because it is possible to predict the charging resistance value with higher accuracy. Furthermore, it is preferable that the physical property information of the negative electrode active material in the machine learning is one type. In other words, it is preferable that the machine learning method is short-term regression analysis.

[0029] All-solid-state batteries are described in "B. Manufacturing method of all-solid-state batteries."

[0030] B. Manufacturing method of all-solid-state batteries The method for manufacturing an all-solid-state battery according to the present disclosure includes a first step of carrying out the prediction method described above, and a second step of forming the negative electrode active material layer using the negative electrode active material predicted in the first step to have the charging resistance value equal to or less than a threshold value.

[0031] In the present disclosure, a negative electrode active material layer is formed using a negative electrode active material that is predicted in the above-described prediction method to have a charging resistance value of the all-solid-state battery equal to or less than a threshold value. That is, a negative electrode active material that is predicted in advance to have a good charging resistance value of the all-solid-state battery is used. Therefore, the rate of good products in the finished all-solid-state batteries can be increased, and the productivity of all-solid-state batteries can be improved.

[0032] 1.First step The first step in the present disclosure is to carry out the above-mentioned prediction method. The prediction method is the same as that described in "A. Prediction method," so a description thereof will be omitted here.

[0033] 2.Second process The second step in the present disclosure is a step of forming the negative electrode active material layer using the negative electrode active material predicted in the first step to have the charging resistance value equal to or less than a threshold value.

[0034] The negative electrode active material for which the charging resistance value was predicted to be below the threshold was Li4Ti5O 12 It is preferable that the crystalline phase has a crystallinity of, for example, 98.4% or more, or may be 98.6% or more, or may be 98.8% or more. On the other hand, the crystallinity of the crystalline phase is, for example, less than 100%, or may be 99.5% or less, or may be 99.0% or less.

[0035] The threshold value can be adjusted appropriately depending on the desired performance of the all-solid-state battery. The threshold value is, for example, 3.0 mΩ or less, or may be 2.5 mΩ or less. On the other hand, the threshold value is, for example, 2.0 mΩ or more.

[0036] The method for forming the negative electrode active material layer can be a conventionally known method, for example, a method in which a negative electrode slurry containing at least a negative electrode active material predicted to have a charging resistance value equal to or less than a threshold value is applied to a negative electrode current collector and then dried.

[0037] The negative electrode slurry contains at least the above-described negative electrode active material. The negative electrode slurry may further contain at least one of a binder, a conductive agent, and a solid electrolyte. Examples of binders include rubber-based binders such as butylene rubber (BR) and fluoride-based binders such as polyvinylidene fluoride (PVdF). Examples of conductive agents include carbon materials, metal particles, and conductive polymers. Examples of carbon materials include particulate carbon materials such as ketjen black (KB), and fibrous carbon materials such as carbon fiber, carbon nanotubes (CNT), and carbon nanofibers (CNF). Examples of solid electrolytes include oxide solid electrolytes and sulfide solid electrolytes. The negative electrode slurry can be prepared by adding the above-described materials to a solvent such as butyl butyrate and mixing them. The types, ratios, and other conditions of the above-described materials in the negative electrode slurry are preferably the conditions set in the above-described machine learning.

[0038] The negative electrode current collector can be made of a conventionally known material such as Cu foil, Ni foil, or SUS foil.

[0039] 3. Other processes The method for producing an all-solid-state battery according to the present disclosure typically includes a cathode active material layer forming step of forming a cathode active material layer, a solid electrolyte layer forming step of forming a solid electrolyte layer, and a lamination step of obtaining a laminate having the anode active material layer, the solid electrolyte layer, and the cathode active material layer.

[0040] The manufacturing conditions, such as the materials and compositions of the positive electrode active material layer and the solid electrolyte, are not particularly limited, but are preferably the conditions set in the above machine learning. In addition, the methods for forming the positive electrode active material and the solid electrolyte are not particularly limited, and can be conventionally known methods.

[0041] 4.All-solid-state battery Fig. 1 is a schematic cross-sectional view illustrating an example of an all-solid-state battery according to the present disclosure. The all-solid-state battery 10 shown in Fig. 1 has a positive electrode active material layer 1, a negative electrode active material layer 2, and a solid electrolyte layer 3 disposed between the positive electrode active material layer 1 and the negative electrode active material layer 2, and typically has a positive electrode current collector 4 that collects electrons from the positive electrode active material layer 1 and a negative electrode current collector 5 that collects electrons from the negative electrode active material layer 2.

[0042] The all-solid-state battery in the present disclosure is preferably a Li-ion secondary battery. Applications of the all-solid-state battery include, for example, power sources for vehicles such as hybrid electric vehicles (HEVs), plug-in hybrid electric vehicles (PHEVs), electric vehicles (BEVs), gasoline-powered vehicles, and diesel-powered vehicles. It is particularly preferred that the battery be used as a driving power source for hybrid electric vehicles (HEVs), plug-in hybrid electric vehicles (PHEVs), or electric vehicles (BEVs). The battery in the present disclosure may also be used as a power source for mobile objects other than vehicles (for example, trains, ships, and aircraft), or as a power source for electrical appliances such as information processing devices.

[0043] The present disclosure is not limited to the above-described embodiments. The above-described embodiments are merely examples, and any configuration that is substantially identical to the technical idea described in the claims of the present disclosure and that provides similar effects is included within the technical scope of the present disclosure. [Example]

[0044] [Experimental Example] We fabricated multiple all-solid-state batteries with different manufacturing parameters (described below). We also obtained charging resistance values ​​(evaluation data) for the multiple all-solid-state batteries we fabricated. Based on the obtained data, we used the machine learning method (described below) to determine the correlation between each manufacturing parameter and charging resistance value.

[0045] <Obtaining manufacturing parameters> As a manufacturing parameter, the negative electrode active material (Li4Ti5O 12 For the negative electrode active material having a crystalline phase (LTO negative electrode active material), the Ti content (%) was obtained by ICP-MS inductively coupled plasma mass spectrometry, and Li4Ti5O was obtained by XRD analysis. 12 The crystallinity was also obtained. The temperature at which the ionic conductivity of the sulfide solid electrolyte used as the negative electrode material was measured was also obtained.

[0046] <Fabrication of all-solid-state batteries> The negative electrode active material (LTO particles), sulfide solid electrolyte (Li2S-P2S5-based sulfide solid electrolyte), conductive agent (VGCF), and binder (BR-based binder) were added to a solvent (butyl butyrate) and mixed using an ultrasonic disperser. This produced a negative electrode slurry. The negative electrode slurry was applied to a negative electrode current collector foil (Ni foil) using a blade method. This was then dried on a hot plate at 100°C for 30 minutes. This produced a negative electrode having a negative electrode current collector and a negative electrode active material layer.

[0047] A positive electrode active material (NCA-based active material), a sulfide solid electrolyte (Li2S-P2S5-based sulfide solid electrolyte), a conductive agent (VGCF), and a binder (PVdF-based binder) were added to a solvent (butyl butyrate) and mixed using an ultrasonic disperser. This resulted in a positive electrode slurry. The positive electrode slurry was applied to a positive electrode current collector foil (Al foil) using a blade method. This was then dried on a hot plate at 100°C for 30 minutes. This resulted in a positive electrode having a positive electrode current collector and a positive electrode active material layer.

[0048] A sulfide-based solid electrolyte (Li2S-P2S5-based sulfide solid electrolyte) and a binder (PVdF-based binder) were added to a solvent (butyl butyrate) and mixed using an ultrasonic disperser. This produced a solid electrolyte slurry. The solid electrolyte slurry was then applied to a release foil (Al foil) using a blade method and dried on a hot plate at 100°C for 30 minutes. This resulted in a transfer member having a release foil and a solid electrolyte layer.

[0049] The negative electrode and the transfer member were laminated so that the negative electrode active material layer and the solid electrolyte layer were in contact with each other, and a coating density of 1 ton / cm 2 After pressing, the Al foil of the transfer member was removed to obtain a negative electrode with a solid electrolyte layer. Next, the positive electrode and the negative electrode were laminated so that the positive electrode active material layer and the solid electrolyte layer were in contact with each other, and the pressure was 3 ton / cm. 2 The laminate was pressed at 100°C. This resulted in a laminate having a positive electrode active material layer, a solid electrolyte layer, and a negative electrode active material layer. A positive electrode current collector (Al foil), a negative electrode current collector (Cu foil), and a tab with welding tape were ultrasonically bonded to this laminate, and the whole was sealed with an aluminum laminate film. This resulted in the production of a sulfide all-solid-state battery.

[0050] <Acquisition of evaluation data> The sulfide all-solid-state battery was subjected to one charge-discharge cycle, then charged again to an SOC of 80%. It was then charged for 10 seconds at a current equivalent to 3C, and the charging resistance value was obtained as evaluation data.

[0051] <Machine Learning> Using the acquired manufacturing parameters and evaluation data (charging resistance value), we used machine learning techniques to derive explanatory variables (manufacturing parameters that contribute highly to charging resistance value) with the manufacturing parameters as explanatory variables and the evaluation data as the objective variable. The LightGBM method, a supervised learning method, was used as the machine learning technique. A simple regression model was used for supervised learning. The SHAP (SHapley Additive exPlanations) method was used to calculate the degree of contribution. The results are shown in Table 1 and Figure 2.

[0052] [Table 1]

[0053] As shown in Table 1 and Figures 2(a) to (c), the crystallinity (Li4Ti5O 12 The degree of crystallinity of the negative electrode active material had the highest correlation with the charging resistance value, while other physical properties of the negative electrode active material (Ti content) and information about the sulfide solid electrolyte (temperature at which ionic conductivity was measured) did not show any correlation with the charging resistance value. In particular, as shown in Figure 2(a), the Li4Ti5O 12 If the crystallinity is 98.4% or higher, it is predicted that the resulting all-solid-state battery will exhibit a charging resistance value of 3.0 mΩ or less, and it is predicted that an all-solid-state battery with good performance can be manufactured.

[0054] This indicates that the charging resistance value of a manufactured all-solid-state battery can be predicted by performing XRD analysis of the negative electrode active material in an upstream process in the manufacturing process of the all-solid-state battery. This suggests that the prediction method of the present disclosure makes it possible to predict the performance of an all-solid-state battery without assembling the all-solid-state battery, and also makes it possible to evaluate whether an all-solid-state battery to be manufactured is a non-defective product before assembling the all-solid-state battery. [Explanation of symbols]

[0055] 1...Cathode active material layer 2...Negative electrode active material layer 3...Solid electrolyte layer 4...Positive electrode current collector 5...Negative electrode current collector 10...All-solid-state battery

Claims

[Claim 1] A method for manufacturing an all-solid-state battery having a positive electrode active material layer, a negative electrode active material layer, and a solid electrolyte layer disposed between the positive electrode active material layer and the negative electrode active material layer, the method comprising: A first step of performing a predetermined prediction method; a second step of forming the negative electrode active material layer using the negative electrode active material predicted in the first step to have the charging resistance value equal to or less than a threshold value; The predetermined prediction method includes: a preparation step of preparing a negative electrode active material having a Li4Ti5O12 crystalline phase; an analysis step of performing XRD analysis on the negative electrode active material to obtain the crystallinity of the Li4Ti5O12 crystalline phase as physical property information of the negative electrode active material; and a prediction step of predicting a charging resistance value of an all-solid-state battery using the negative electrode active material based on the physical property information, In the prediction step, the prediction is performed based on a trained model that has previously been machine-learned to describe a relationship between the physical property information of the negative electrode active material and a charging resistance value of the all-solid-state battery, and the physical property information acquired in the analysis step; The method for producing an all-solid-state battery, wherein the negative electrode active material whose charging resistance value is predicted to be equal to or less than the threshold value has a crystallinity of the Li 4 Ti 5 O 12 crystalline phase of 98.4% or more.

Images