RU2808661C1 - Method for predicting specific capacity of graphite anodic material of lithium-ion battery - Google Patents

Abstract

FIELD: electrochemical energy.

SUBSTANCE: invention can be used in the field of material control in the production of lithium-ion batteries. A method for predicting the specific capacity of a graphite anode material of a lithium-ion battery includes preparing and carrying out physical and chemical testing of samples, taking measurements on the surface of the samples with placing a pressing electrochemical cell with an acidic electrolyte on their surface, supplying an input signal and receiving a measured signal, assembling an electrochemical cell with obtaining cyclic voltammetric dependences and hodographs of electrochemical impedance, processing the obtained data based on a system of linear regression equations constructed from training samples of samples with known values of the predicted parameters. Physicochemical control is carried out, which consists of measuring the pH of the aqueous extract of powdered graphite anode material. Electrodes are used as samples, which are prepared by preparing a mixture of powdered graphite anode material with a binder, applying it to the electrode collectors, drying and rolling the electrodes. Measurements are carried out on the surface of the resulting electrodes in a pressure electrochemical cell, in which a stepwise changing potential with a step duration of 20-180 s and a step value of 15-20 mV is used as an input signal, and the dependence of the amounts of electricity on the step potentials is used as the measured signal. Two electrochemical cells are prepared in the form of disk elements of type 2016, each of which contains a lithium electrode, a separator, electrode samples, an electrolyte. An electrochemical impedance hodograph is obtained in one electrochemical cell, and in the other, cyclic voltammetric dependences are obtained in the potential range: 5-1500 mV at potential scan rates of 4-100 mV/s. Processing the obtained data based on a system of linear regression equations includes finding the total amount of electricity depending on the amount of electricity on the potential in the potential region minus 700 - 0 mV, finding the angular coefficient of the dependence of the maximum current density on the square root of the potential sweep rate, determining the parameters Rs and CPE-T of the equivalent circuit of the Rs-CPE electrochemical cell in the frequency range 0.1-100 Hz, determining whether the data belongs to one of four clusters on the pH-amount of electricity plane in the potential range minus 700 - 0 mV, the boundaries of which are determined based on the results of data clustering for samples with known specific capacitance values, selecting a multiple regression design equation to calculate the predicted capacitance, each of which is based on data from samples for which the specific capacitance is known.

EFFECT: reducing the time to obtain the capacity prediction result.

1 cl, 4 dwg

Description

The invention relates to electrochemical energy and can be used in the field of materials control in the production of lithium-ion batteries.

There is a known method for predicting the capacity of electrode material, based on the use of the generalized Peukert equation (I.Yu. Bogush, N.K. Plugotarenko, T.N. Myasoedova Study of the functional characteristics of mesoporous electrodes of supercapacitors based on silicon-carbon films // Journal of Technical Physics, 2022 , volume 92, issue 12, pp. 1833-1843). The method is based on determining the Peukert equation constant for a given type of electrode material based on experimental capacitances at several discharge current densities, after which the predicted capacitance at any current density can be calculated from this equation.

The disadvantage of this solution is its limited nature, associated with prediction only depending on the discharge current density, as well as the large amount of testing required to find the constant of the Peukert equation.

The closest to the claimed method is the method for predicting the required characteristics, set out in patent RU2653775. The forecasting method described in the patent relates to the corrosion rate (ours is the predicted parameter) and consists of preparing surfaces (ours is preparing samples), visual measuring, flaw detection, acoustic emission and other non-destructive testing methods, determining the parameters of the actual technical condition of the pipeline , determining the corrosion indicator, configuration and size of existing defects, characterized in that, based on the results of visual measuring, flaw detection, acoustic emission and other non-destructive testing methods (in our case - physical and chemical testing), control cuttings are performed, and used as surface preparation dividing the control cutting into samples; on each of the obtained samples of the control cutting, a pressure electrochemical cell with an acidic electrolyte is installed, a stepwise changing current is passed (in our case, the input signal), including 30 pulses with a uniformly increasing amplitude of 0.004 mA on each pulse, pulse duration 2000 -2500 ms, the duration of the pause between pulses is 300-400 ms, the dependence of the potential on time is recorded (in our case it is the measured signal), the values of the dependence of the no-current potentials on time are selected from the obtained dependence, the obtained dependence of the no-current potentials on time is differentiated according to the obtained dependence of the derivative a local minimum is determined from time to time, in the areas between the inflection points the points with the maximum value of the derivative are found, by the potentials of which the phase of the corrosion products is identified, and the amount of this phase is found from the length of the sections between the inflection points, then each of the control cutting samples is placed in a three-electrode electrolytic cell ( in our case - an electrochemical cell) with a free volume of electrolyte, set the linear potential scan mode at a speed of 4-6 mV/s and take the curves of the dependence of the current density (in our case - cyclic voltammetry) on the potential in a 3% sodium chloride solution on a surface without corrosion products and on the surface with corrosion products, for the obtained curves of current density versus potential, calculate the derivatives of the current density with respect to the potential, find the potential value at which the derivatives differ by no more than 0.01-0.015, for this potential value find the current density on the surface without corrosion products and on the surface with corrosion products, calculate the fraction of the free surface, then each of the control cutting samples is placed in an electrolytic cell with a free volume of electrolyte, the impedance hodographs of the corrosion products themselves and the impedance hodographs of the corrosion products in the alkaline electrolyte are obtained, a one-dimensional linear regression equation is constructed according to to which the active component of the impedance is calculated, then, based on the obtained phase composition of corrosion products and their quantity, the value of the fraction of the free surface, the active component of the impedance, the corrosion indicator is calculated using a system of linear regression equations constructed from training samples of samples obtained for certain parameters of the corrosive environment, according to the distribution of values RF determines the type of corrosion lesions based on the area of the analyzed sample and the phase composition of corrosion products.

The disadvantage of the prototype is the limitation of the functional purpose of the method related to corrosion prediction.

The problem of predicting the specific capacity of the graphite anode material LIB is the long time to obtain the result, which, taking into account molding and cycling, takes several months, while production requirements are reduced to 1-2 days.

The technical result of the invention is to reduce the time to obtain the result of capacity prediction through the use of independent express methods.

This problem is solved by the proposed method for predicting the specific capacity of the graphite anode material of a lithium-ion battery. A method for predicting the specific capacity of a graphite anode material of a lithium-ion battery includes preparing and carrying out physical and chemical testing of samples, taking measurements on the surface of the samples with placing a pressing electrochemical cell with an acidic electrolyte on their surface with supplying an input signal and receiving a measured signal, assembling an electrochemical cell with obtaining cyclic voltammetric dependences and electrochemical impedance hodographs, processing the obtained data based on a system of linear regression equations constructed from training samples of samples with known values of the predicted parameters, and carrying out physicochemical control, which consists of measuring the pH of the aqueous extract of powdered graphite anode material, as samples use electrodes that are prepared by preparing a mixture of powdered graphite anode material with a binder, applying it to electrode collectors, drying and rolling the electrodes; measurements are taken on the surface of the resulting electrodes in a pressure electrochemical cell, in which a stepwise changing potential with a step duration is used as an input signal 20-180 s and a step size of 15-20 mV, the dependence of the amounts of electricity on the step potentials is used as the measured signal, two electrochemical cells are prepared in the form of disk elements of type 2016, each of which contains a lithium electrode, a separator, electrode samples, an electrolyte, One electrochemical cell receives a hodograph of electrochemical impedance, and the other receives cyclic voltammetric dependences in the potential range: 5-1500 mV at potential sweep rates of 4-100 mV/s, processing the obtained data based on a system of linear regression equations includes finding the total amount of electricity depending on the amount of electricity on potential in the potential region minus 700 - 0 mV, finding the angular coefficient of the dependence of the maximum current density on the square root of the potential sweep rate, determining the parameters Rs and CPE-T of the equivalent circuit of the Rs-CPE electrochemical cell in the frequency range 0.1-100 Hz, determining the affiliation data to one of four clusters on the pH-amount of electricity plane in the potential range minus 700 - 0 mV, the boundaries of which are determined based on the results of data clustering for samples with known specific capacitance values, the choice of a multiple regression design equation for calculating the predicted capacity, each of which is constructed according to samples for which the specific capacitance is known.

The specific capacity of the graphite anode material of a lithium-ion battery ( _qa ) is determined by the rate of diffusion of intercalated lithium ions in the volume of the structure and the rate of penetration of intercalated lithium ions through the structure of the solid electrolyte film, which is formed in the process of cathodic and anodic polarization during cycling. It is known that the structure and ionic conductivity of a solid electrolyte film depends on the type and number of surface functional groups (SFGs) on the graphite surface. Identification of PFGs is carried out by their acid-base and redox properties, therefore the first stage of predicting the specific capacity is to measure the pH of the aqueous extract of graphite and carry out the reduction of PFGs in the mode of stepwise potentiostatic polarization in a clamping structure sensor. The result of measurements in this mode is the dependence of the amount of electricity - potential (q _i (E _i )), in which the restoration of each type of PFG corresponds to an inflection point or maximum. Studies have established that the most critical is the presence of quinone-type PFGs, which are restored at potentials in the range from 0 to minus 700 mV, and therefore the total amount of electricity in this potential range (Q _s ) in conjunction with the pH of the aqueous extract was put in basis for data clustering. On the pH-quantity of electricity plane, four regions can be distinguished, the boundaries of which were established on the basis of an array of data obtained for samples with a certain specific capacity:

1) pH<7, Q _s >-0.1 C, acidic region with a small amount of quinone-type PFG;

2) pH<7, Q _s <-0.1 C, acidic region with a large amount of quinone-type PFG;

3) pH>7, Q _s <-0.1 C alkaline region with a small amount of quinone-type PFG;

4) pH>7, Q _s >-0.1 C alkaline region with a large amount of quinone-type PFG.

Whether the sample for which the specific capacitance is predicted belongs to one or another of these four clusters determines the type of solid electrolyte film formed on its surface, which makes it possible to consider samples located in the same cluster as of the same type and to construct for them a single multiple linear regression equation with respect to other parameters . The process of formation of a solid electrolyte film on the surface of graphite begins with its chemical interaction with the electrolyte, which makes it possible to determine the method of electrochemical impedance spectroscopy. On the hodographs of the electrochemical impedance of electrochemical cells with the electrode under study, a separator impregnated with electrolyte and a lithium anode, areas corresponding to the frequency ranges of the lithium electrode, transfer in the separator and transfer in the pores of the graphite anode and directly on its surface are identified. The studies have established that the latter processes occur in the frequency range of 0.1-100 Hz, this region is described by a simple resistance-constant phase element (Rs-CPE) equivalent circuit, and the parameters that determine the specific capacitance are the resistance values Rs and the CPE parameter -T, obtained as a result of processing the specified section of the hodograph. The kinetics of lithium transfer through a solid electrolyte film can be taken into account by cyclic voltammetry (CVA). In the anodic region of the CV, as the studies have shown, for all samples there is a maximum, the value of which increases with increasing speed of potential sweep. If the limiting stage of cycling is solid-state diffusion of lithium, this dependence will have a low slope, kv. If this slope coefficient has a high value, this means that the solid electrolyte film has low permeability and the transfer in the electrolyte associated with the separator is limiting. Thus, to predict specific capacity, a multiple linear regression model of the form can be used:

q _a =b ₀ +b ₁ Rs+b ₂ CPET+b ₃ kv

As a result of the research, the coefficients of the multiple linear regression equations were calculated for each of the four regions (1-4) on the pH-Q _s plane. The regression equations had the following form:

For area 1:

q _a =246+0.387⋅Rs+2285⋅CPET-325%5⋅kv

For area 2:

q _a =306-5.95⋅Rs+64909⋅CRETE - 27119⋅kv

For area 3:

q _a =211+1.387⋅Rs+5212⋅CPET-24512⋅kv

For area 4:

q _a =207+5.426⋅Rs+5123⋅SRET-3041⋅kv

The essence of the invention is illustrated by figures.

In fig. Figure 1 shows the dependence of the amount of electricity on the potential for a sample of the electrode being tested.

In fig. Figure 2 shows the cyclic voltammetric dependences of the test sample at different scan rates. Sweep rates in mV/s are shown on the curves.

In fig. Figure 3 shows the dependence of the current of the maximum of the anode part of the cyclic voltammetric dependences shown in Fig. 2 from the square root of the sweep speed.

In fig. Figure 4 shows the result of processing the electrochemical impedance hodograph for the test sample.

An example of the method implementation.

To predict the specific capacity of the graphite anode material, the pH of the aqueous extract was determined, which was 7.12. After this, electrodes were made from the material under study using the following method. 0.912 g of CMC (carboxymethylcellulose VTT-770) and 50 g of distilled water were placed in a mixer container (volume 250 ml). The components were mixed using a laboratory mixer AX-2000 at a speed of 1000 rpm. After complete dissolution of CMC in water, 0.428 g of carbon black (SP) and 57.84 g of carbon material (SG-NT) were added. Stirring was carried out for 2 hours (1-2 minutes without vacuum, then using vacuum, controlling the achievement of vacuum pressure to 0.1 MPa). After complete mixing of the mass (control was carried out visually), 1.71 g of a polymer anionic emulsion of styrene-butadiene rubber (SBR) and 0.875 N-methyl-2-pyrrolidone (NMP) was added and stirred for 1-2 minutes. After this, the resulting mass was applied using an applicator onto copper foil, dried and rolled on rollers. As a result, an electrode sheet was obtained.

For measurements in a clamping electrochemical cell, samples were cut out of the electrode sheet, on which measurements were carried out in a clamping electrochemical cell, in which a stepwise changing potential with a step duration of 40 s and a step value of 20 mV was used as the input signal; the dependence of the amounts of electricity was used as the measured signal from the stage potentials. Based on the obtained dependence (Fig. 1), the amount of electricity in the range of minus 700-0 mV was calculated by summing the values of the amounts of electricity falling within the specified range. This value was 0.42 C.

Two electrodes were cut out of the electrode sheet for the cell in the body of the 2016 cell. From the resulting electrodes, two electrochemical cells were assembled, including the cut out electrodes, a three-layer Jimitek separator, onto which 0.3 ml of TSE 2016 electrolyte was applied, a lithium auxiliary electrode (the same electrode was the reference electrode).

Electrochemical impedance was measured in one cell. The resulting hodograph (Fig. 2) was saved as a file compatible with the Z-View processing program.

In another cell, cyclic voltammetric dependences were obtained at potential scan rates of 4, 10, 20, 50 and 100 mV/s (Fig. 2).

Data processing included:

1. Calculation of the Qs value based on the obtained dependence q _i (E _i ). Summation of qs values in the potential range minus 700-0 mV gives a value of - 0.14 C.

2. Processing the electrochemical impedance hodograph in the Z-View program. The electrochemical impedance measurement file (Figure 4) was divided into three sections corresponding to the lithium electrode, separator and graphite electrode. The last section in the frequency range 1-100 Hz was processed in accordance with the Rs-CPE equivalent circuit, for which the following values were obtained:

Rs=6.782 Ohm

CPET=0.0097 F

3. On each curve of Fig. 2, the values of the maximum current on the anode branch of the CV were selected, the obtained values were plotted in the coordinates Im - √v (Fig. 3), for these data the coefficients of the linear regression equation were calculated using the standard Excel option. The kv coefficient value was:

kv=0.00317.

4. Based on the pH value of the aqueous extract equal to 7.12 and the Qs value equal to -0.14, it was determined that it is necessary to use the regression equation for region 4. Substituting the obtained parameter values into the regression equation for region 4 gives:

q _a =207+5.426⋅6.782+5123⋅0.0097-0.00317⋅3041=284 mAh/g

When determining the capacity of the material using the galvanostatic cycling method, a value of 278 mAh/g was obtained. The tests were carried out over 5 days. The forecast error was 2.2% of the determined value. The total time for measurements and data processing was 2.5 hours.

Claims (1)

A method for predicting the specific capacity of a graphite anode material of a lithium-ion battery, including preparing and carrying out physical and chemical testing of samples, taking measurements on the surface of the samples with placing on their surface a pressing electrochemical cell with an acidic electrolyte with supplying an input signal and receiving a measured signal, assembling the electrochemical cell with obtaining cyclic voltammetric dependences and hodographs of electrochemical impedance, processing the obtained data based on a system of linear regression equations constructed from training samples of samples with known values of the predicted parameters, characterized in that they carry out physico-chemical control, which consists of measuring the pH of an aqueous extract of powdered graphite anode material, electrodes are used as samples, which are prepared by preparing a mixture of powdered graphite anode material with a binder, applying it to the electrode collectors, drying and rolling the electrodes; measurements are taken on the surface of the resulting electrodes in a pressure electrochemical cell, in which stepwise steps are used as the input signal changing potential with a step duration of 20-180 s and a step value of 15-20 mV, the dependence of the amounts of electricity on the step potentials is used as the measured signal, two electrochemical cells are prepared in the form of disk elements of type 2016, each of which contains a lithium electrode, a separator, samples electrodes, electrolyte, in one electrochemical cell a hodograph of electrochemical impedance is obtained, and in the other cyclic voltammetric dependences are obtained in the potential range: 5-1500 mV at potential scan rates of 4-100 mV/s, processing the obtained data based on a system of linear regression equations includes finding the total amount electricity dependence of the amount of electricity on the potential in the potential region minus 700 - 0 mV, finding the angular coefficient of the dependence of the maximum current density on the square root of the potential sweep rate, determining the Rs and CPE-T parameters of the equivalent circuit of the Rs-CPE electrochemical cell in the frequency range 0.1 -100 Hz, determination of whether the data belongs to one of four clusters on the pH-amount of electricity plane in the potential range minus 700 - 0 mV, the boundaries of which are determined based on the results of data clustering for samples with known specific capacitance values, selection of the multiple regression calculation equation for calculating the predicted capacities, each of which is built according to data from samples for which the specific capacitance is known.