Method for manufacturing a negative electrode active material layer, and method for manufacturing a secondary battery
By measuring strain variance and using machine learning to predict coating film quality, the method addresses the inadequacy of viscometer measurements, ensuring efficient and appropriate coating film production.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- TOYOTA JIDOSHA KK
- Filing Date
- 2024-12-10
- Publication Date
- 2026-06-22
AI Technical Summary
The thixotropy index value measured by a viscometer alone is insufficient to determine the quality of the coating film after application, leading to wasted coating work as issues are discovered only after application.
A method involving the use of a dynamic viscoelastic device to measure parameters such as strain variance, specifically storage modulus and loss modulus, to predict the quality of the coating film before application, combined with machine learning to select appropriate parameters.
Enables prediction and determination of coating film quality in advance, reducing unnecessary coating and efficiently obtaining suitable films.
Smart Images
Figure 2026100944000001_ABST
Abstract
Description
[Technical Field]
[0001] This disclosure relates to a method for manufacturing a negative electrode active material layer contained in a secondary battery. [Background technology]
[0002] In the manufacture of electrodes for secondary batteries, the electrode mixture for forming the active material layer is prepared using an active material (powdered solid), a binder (paste), and a solvent (liquid) as raw materials. Patent Document 1 discloses a manufacturing method for coating electrodes while maintaining a constant thixotropy index value of the paste measured by a viscometer. [Prior art documents] [Patent Documents]
[0003] [Patent Document 1] Patent No. 7031259 [Overview of the project] [Problems that the invention aims to solve]
[0004] However, the thixotropy index value measured by a viscometer alone is insufficient to determine the quality of the coating film after application. In other words, problems with the coating film may only be discovered after it has been applied, resulting in wasted coating work.
[0005] Therefore, the present disclosure aims to provide a method for manufacturing a negative electrode active material layer that can efficiently obtain a suitable coating film. [Means for solving the problem]
[0006] This application discloses a method for manufacturing a negative electrode active material layer, comprising the steps of: preparing a negative electrode slurry by mixing a negative electrode active material, a solid electrolyte, a conductive additive, a binder, and a solvent; acquiring parameters of the negative electrode slurry using a dynamic viscoelastic device; determining the quality of the coating film from the acquired parameters; and coating the negative electrode slurry that was determined to be good in the determination step.
[0007] The parameters may also be obtained from strain variance measurements.
[0008] The strain dispersion measurement may consist of the storage modulus and loss modulus measured in the strain range of 0.01 to 1000 (%).
[0009] Furthermore, this application discloses a method for manufacturing a secondary battery, including each of the steps for manufacturing the negative electrode active material layer as described above. [Effects of the Invention]
[0010] According to this disclosure, by measuring the physical property parameters obtained with a dynamic viscoelasticity measuring device while the negative electrode slurry is in its state, it is possible to predict and determine in advance the quality of the coating film after coating with the negative electrode slurry, thereby suppressing unnecessary coating and efficiently obtaining an appropriate coating film. [Brief explanation of the drawing]
[0011] [Figure 1] This diagram illustrates the layer structure of the all-solid-state battery 10. [Modes for carrying out the invention]
[0012] 1. Secondary battery Here, we will describe an all-solid-state battery as one form of a secondary battery, but it is sufficient that the negative electrode active material layer contains a solid electrolyte, and it may also be a secondary battery that contains an electrolyte solution.
[0013] Fig. 1 shows a schematic cross-sectional view showing an example of an all-solid-state battery. As shown in Fig. 1, the all-solid-state battery 10 includes a positive electrode active material layer 11 containing a positive electrode active material, a negative electrode active material layer 12 containing a negative electrode active material, a solid electrolyte layer 13 formed between the positive electrode active material layer 11 and the negative electrode active material layer 12, a positive electrode current collector layer 14 that collects current from the positive electrode active material layer 11, and a negative electrode current collector layer 15 that collects current from the negative electrode active material layer 12. Note that the positive electrode active material layer 11 and the positive electrode current collector layer 14 may be collectively referred to as the positive electrode layer, and the negative electrode active material layer 12 and the negative electrode current collector layer 15 may be collectively referred to as the negative electrode layer. Hereinafter, each component of the all-solid-state battery 10 will be described.
[0014] 1.1. Positive electrode active material layer The positive electrode active material layer 11 is a layer containing a positive electrode active material, and further contains a solid electrolyte, a conductive auxiliary material, and a binder. Known active materials may be used as the positive electrode active material. For example, cobalt-based (such as LiCoO2), nickel-based (such as LiNiO2), manganese-based (such as LiMn2O4, Li2Mn2O3, etc.), iron phosphate-based (such as LiFePO4, Li2FeP2O7, etc.), NCA-based (compounds of nickel, cobalt, and aluminum), NMC-based (compounds of nickel, manganese, and cobalt), etc. More specifically, there is LiNi 1 / 3 Co 1 / 3 Mn 1 / 3 O2 and so on. The surface of the positive electrode active material may be coated with an oxide layer such as a lithium niobate layer, a lithium titanate layer, or a lithium phosphate layer.
[0015] An inorganic solid electrolyte is preferable as the solid electrolyte. This is because it has a higher ionic conductivity and better heat resistance compared to organic polymer electrolytes. Examples of inorganic solid electrolytes include sulfide solid electrolytes and oxide solid electrolytes. Examples of sulfide solid electrolyte materials having Li ion conductivity include, for example, Li2S-P2S5, Li2S-P2S5-LiI, Li2S-P2S5-Li2O, Li2S-P2S5-Li2O-LiI, Li2S-SiS2, Li2S-SiS2-LiI, Li2S-SiS2-LiBr, Li2S-SiS2-LiCl, Li2S-SiS2-B2S3-LiI, Li2S-SiS2-P2S5-LiI, Li2S-B2S3, Li2S-P2S5-ZmSn (where m and n are positive numbers. Z is either Ge, Zn, or Ga), Li2S-GeS2, Li2S-SiS2-Li3PO4, Li2S-SiS2-Li x MO y (where x and y are positive numbers. M is either P, Si, Ge, B, Al, Ga, or In), etc. Note that the description "Li2S-P2S5" means a sulfide solid electrolyte material formed using a raw material composition containing Li2S and P2S5, and the same applies to other descriptions.
[0016] On the other hand, examples of oxide solid electrolyte materials having Li ion conductivity include, for example, compounds having a NASICON-type structure, etc. As an example of a compound having a NASICON-type structure, a compound represented by the general formula Li 1+x Al x Ge 2-x (PO4)3 (0 ≦ x ≦ 2) (LAGP), a compound represented by the general formula Li 1+x Al x Ti 2-x (PO4)3 (0 ≦ x ≦ 2) (LATP), etc. can be mentioned. Also, as other examples of oxide solid electrolyte materials, LiLaTiO (for example, Li 0.34 La 0.51 TiO3), LiPON (for example, Li 2.9 PO 3.3 N 0.46 ), LiLaZrO (for example, Li7La3Zr2O 12 ), etc. can be mentioned.
[0017] The binder is not particularly limited as long as it is chemically and electrically stable, but examples include fluorine-based binders such as polyvinylidene fluoride (PVDF) and polytetrafluoroethylene (PTFE), rubber-based binders such as styrene-butadiene rubber (SBR), olefin-based binders such as polypropylene (PP) and polyethylene (PE), and cellulose-based binders such as carboxymethylcellulose (CMC).
[0018] As conductive additives, carbon materials such as carbon fiber, acetylene black, and Ketjenblack, as well as metallic materials such as nickel, aluminum, and stainless steel, can be used.
[0019] The content of each component in the positive electrode active material layer 11 and the shape of the positive electrode active material layer 11 may be the same as in the conventional design. In particular, a sheet-shaped positive electrode active material layer 11 is preferred from the viewpoint of easily constructing an all-solid-state battery 10. In this case, the thickness of the positive electrode active material layer 11 is preferably, for example, 0.1 μm or more and 1 mm or less, and more preferably 1 μm or more and 150 μm or less.
[0020] 1.2.Negative electrode active material layer The negative electrode active material layer 12 is a layer containing at least a negative electrode active material, and may optionally contain at least one of a solid electrolyte, a conductive additive, and a binder. The solid electrolyte, conductive additive, and binder can be considered in the same way as in the positive electrode active material layer 11.
[0021] While there are no particular limitations on the negative electrode active material, when constructing a lithium-ion battery, examples of negative electrode active materials include carbon materials such as graphite and hard carbon, various oxides such as lithium titanate (LTO), Si and Si alloys, or metallic lithium and lithium alloys.
[0022] 1.3.Solid electrolyte layer The solid electrolyte layer 13 is a solid electrolyte layer positioned between the positive electrode active material layer 11 and the negative electrode active material layer 12 in this embodiment. The solid electrolyte layer 13 contains at least a solid electrolyte. The solid electrolyte can be considered in the same way as the solid electrolyte described for the positive electrode active material layer 11.
[0023] 1.4. Current collector layer The current collector layers consist of a positive electrode current collector layer 14 that collects current from the positive electrode active material layer 11, and a negative electrode current collector layer 15 that collects current from the negative electrode active material layer 12. Examples of materials that make up the positive electrode current collector layer 14 include stainless steel, aluminum, nickel, iron, titanium, and carbon. On the other hand, examples of materials that make up the negative electrode current collector layer 15 include stainless steel, copper, nickel, and carbon.
[0024] 1.5. Battery case The all-solid-state battery may include a battery case (not shown). The battery case is a case that houses the various components, and examples include a stainless steel battery case.
[0025] 2. Manufacturing method of secondary batteries The following describes the manufacturing method of a secondary battery, using an all-solid-state battery as an example. This also includes the manufacturing method for the negative electrode active material layer.
[0026] 2.1. Obtaining the correlation coefficient In this disclosure, in order to obtain an appropriate negative electrode active material layer, we obtain parameters that are physical properties of the negative electrode slurry, specifically parameters that have a high probability of yielding a good negative electrode active material layer, and coefficients that indicate their contribution. For this purpose, machine learning methods are used in this embodiment. Specifically, it will be done as follows:
[0027] [Preparation of negative electrode slurry] The negative electrode slurry is a composition (paste) for forming the negative electrode active material layer. Specifically, it is formed by mixing and dispersing the negative electrode active material, solid electrolyte, conductive additive, binder, and solvent, and then further mixing them by stirring.
[0028] [Acquisition of physical property data of negative electrode slurry] To evaluate the physical properties of the negative electrode slurry, a dynamic viscoelasticity measuring device (rheometer) is used, and parameters such as shear dependence (flow curve), stress-strain, strain dispersion, and frequency dispersion are obtained using various measurement methods. The flow curve is obtained in both directions: from high shear to low shear and from low shear to high shear.
[0029] [Acquisition of evaluation data for coating films] The prepared negative electrode slurry is coated onto aluminum foil using a blade method and dried to obtain a negative electrode coating film. The coating film is visually inspected to determine its quality, record the results, and obtain evaluation data. Quality can be determined by whether the desired coating film shape is achieved. Failure is not limited to specific defects, but examples include protrusions, film defects (or localized thin films), and coating streaks.
[0030] [Machine Learning] In machine learning, correlations are obtained for each parameter by using a large amount of "physical property data of negative electrode slurry" and "evaluation data of coating films," and a learning model is selected that uses the parameter with the highest correlation. While there are no specific limitations on the machine learning method used, for example, a random forest method (open source) similar to a decision tree model can be used to select appropriate explanatory variables based on the correlation coefficient values output as a learning result. The target variable is set to evaluation data of the coating film, and the explanatory variables are set to physical property data (parameters) of the negative electrode slurry measured using a dynamic viscoelasticity measurement method. Then, machine learning is performed using this data to derive parameters (explanatory variables) that show a high correlation with the target variable.
[0031] According to the inventor's analysis using a large amount of data, "strain dispersion" was found to have the highest correlation among the physical properties of the negative electrode slurry. Therefore, it is advisable to use a machine learning model derived from strain dispersion. Among these, it is preferable that the storage modulus and loss modulus be composed of those measured in the strain range of 0.01 to 1000 (%).
[0032] 2.2. Manufacturing of rechargeable batteries [Fabrication of positive and negative polarity layers] A positive electrode active material is prepared, and the necessary materials (solid electrolyte, binder, conductive additive, etc.) are mixed with this positive electrode active material to obtain a positive electrode paste. Next, the obtained positive electrode paste is coated onto the layer that will become the positive electrode current collector layer to a predetermined thickness and dried to form a positive electrode layer in which the positive electrode active material layer is laminated on the positive electrode current collector layer.
[0033] [Preparation of a solid electrolyte layer] A solid electrolyte material (e.g., a sulfide solid electrolyte) is prepared, and the necessary materials (such as a binder) are mixed with it to obtain a solid electrolyte layer paste. Then, the obtained solid electrolyte paste is applied to the positive electrode active material layer of the positive electrode layer prepared as described above to a predetermined thickness, and dried to obtain a solid electrolyte layer. This results in a laminate in which a positive electrode current collector layer, a positive electrode active material layer, and a solid electrolyte layer are laminated together.
[0034] [Fabrication of the negative electrode layer] In the fabrication of the negative electrode layer, a negative electrode slurry is first prepared as described above. Then, physical property data (parameters) of this negative electrode slurry are obtained, and using the learning model selected as described above, a quality judgment is made on the negative electrode active material layer after coating it with these parameters. Then, the negative electrode slurry that is judged as "good" is coated onto the layer that will become the negative electrode current collector layer to a predetermined thickness and dried. This produces a negative electrode active material layer laminated on the negative electrode current collector layer.
[0035] [Manufacturing of secondary batteries] The positive electrode current collector layer, positive electrode active material layer, and solid electrolyte layer are laminated onto the positive electrode current collector layer of the laminate fabricated up to this point. This solid electrolyte layer is then pressed onto the negative electrode current collector layer of the negative electrode layer to densify it and create a secondary battery.
[0036] 3. Effects, etc. According to the manufacturing method of this disclosure, by measuring the physical property parameters obtained with a dynamic viscoelasticity measuring device, it is possible to predict and determine in advance the quality of the coating film after coating with the negative electrode slurry, thereby suppressing unnecessary coating and efficiently obtaining an appropriate coating film.
[0037] 4. Example Test 4.1. Preparation of negative electrode slurry and acquisition of physical property data An anode slurry was prepared by using LTO-based anode active material, sulfide-based solid electrolyte, vapor-evolved carbon fiber, PVDF-based binder, and butyl butyrate as raw materials. These materials were mixed and dispersed using an ultrasonic dispersion device, and then further mixed with a stirring blade. To evaluate the physical properties of the negative electrode slurry, a dynamic viscoelasticity measuring device (rheometer) was used, and parameters such as shear dependence (flow curve), stress-strain, strain dispersion, and frequency dispersion were obtained using various measurement methods. The flow curve was obtained in both directions: from high shear to low shear and from low shear to high shear.
[0038] 4.2. Acquisition of evaluation data for coating films The negative electrode slurry was coated onto aluminum foil using a blade method, and the negative electrode coating film was obtained by drying it on a hot plate at 100°C for 30 minutes. As described above, the fabricated coating films were evaluated by visually checking their condition based on a pass / fail judgment, and the number of pass / fail judgments was recorded to obtain evaluation data.
[0039] 4.3. Machine Learning Using a random forest method (open source) similar to a decision tree model, appropriate explanatory variables were selected based on the correlation coefficient values output as learning results. The target variable was set to evaluation data of the coating film, and the explanatory variables were set to physical property data (parameters) of the negative electrode slurry measured using a dynamic viscoelasticity measurement method. Machine learning was performed using this data, and parameters (explanatory variables) that showed a high correlation with the target variable were derived and obtained from the machine learning method.
[0040] Table 1 shows the correlation coefficient results for machine learning models created using parameters obtained from dynamic viscoelasticity measurement techniques. The results are listed in descending order of correlation coefficient. From these results, the machine learning model obtained from strain dispersion measurement data had the highest correlation coefficient at 0.53, and this learning model was selected.
[0041] [Table 1]
[0042] Table 2 shows the TI value, the NG ratio (standard value) predicted by the trained model, and the NG ratio (standard value) measured from the coating film properties through visual inspection. The standard value was set with a minimum acceptable limit of 1 for the coating film NG ratio to determine the usability of the negative electrode slurry; a value of 1 or less was judged as good, and a value lower than 1 was judged as bad. From Table 2, it was found that even within the range of TI values that would be considered bad for the dispersion state of the negative electrode slurry, there were slurries that were usable with a value of 1 or less. This shows that prediction judgment based on TI values is difficult, but it is possible to make such predictions using a machine learning model. The TI value is commonly used as an evaluation index for thixotropic fluids. Here, viscosity information A (2 rpm) and B (20 rpm) were obtained using a viscometer at 2 rpm and 20 rpm, and the TI value was calculated from A / B.
[0043] [Table 2]
[0044] Therefore, it was found that, regarding the anode active material layer, a machine learning model created using strain dispersion data can predict in advance the quality of the anode coating film applied in subsequent processes from the slurry physical property data of the anode kneading process. [Explanation of Symbols]
[0045] 10...All-solid-state battery, 11...Positive electrode active material layer, 12...Negative electrode active material layer, 13...Solid electrolyte layer, 14...Positive electrode current collector layer, 15...Negative electrode current collector layer
Claims
1. A process of preparing a negative electrode slurry by mixing a negative electrode active material, a solid electrolyte, a conductive additive, a binder, and a solvent, A step of obtaining the parameters of the negative electrode slurry using a dynamic viscoelastic device, A step of determining the quality of the coating film from the acquired parameters, The process includes a step of coating the negative electrode slurry that was determined to be good in the aforementioned determination step, A method for manufacturing a negative electrode active material layer.
2. The method for manufacturing a positive electrode active material layer according to claim 1, wherein the aforementioned parameters are obtained from strain dispersion measurement.
3. The method for manufacturing a negative electrode active material layer according to claim 2, wherein the strain dispersion measurement comprises the storage modulus and loss modulus measured in the region of strain amount from 0.01 to 1000 (%).
4. A method for manufacturing a secondary battery, comprising each step of the method for manufacturing a positive electrode active material layer according to any one of claims 1 to 3.