Experimental Study on Cycle Aging Life of 21700 Cylindrical Batteries Under Different Heat Exchange Conditions

Abstract

Lithium-ion batteries are widely used in energy storage systems, and temperature is an important factor that affects the battery aging performance. Battery aging tests have been conducted in environmental chambers in numerous studies. The ambient temperature is usually regarded as an indicator that affects the battery aging performance. However, with the same ambient temperature but different heat exchange conditions, the battery cycle aging life can still vary. In this study, a side face temperature control device and end faces temperature control device for a cylindrical battery were designed and made. Together with the environmental chamber, three types of heat exchange conditions were used to conduct cycle aging tests for the 21700 cylindrical battery. Based on the aging results of batteries under different heat exchange conditions, the battery aging mechanisms were analyzed. At the end of the battery’s life, the maximum loss rate of the active anode material is close to 20%, and the loss rate of the lithium inventory of most test groups is approximately 10%. The internal resistance growth rate of the aged battery can exceed 50%. During the battery aging process, battery temperature data were monitored, and the cumulative time-averaged surface temperature (CTAT) was proposed as a new metric to assess the temperature level for the long-term operating battery. The aging results of the 21700 cylindrical batteries show that within the temperature range of this study, the lower the CTAT, the faster the battery capacity degrades. The correlation between the battery temperature level and aging performance was also analyzed, which can be used to predict the battery cycle life. The analysis of battery aging mechanisms and the proposed temperature metric in this study provide guidance for research on battery life sustainability, as well as the thermal management strategy design of the battery.

1. Introduction

The advantages of lithium-ion batteries enable their widespread use in various industries, including energy storage systems. Owing to the electrochemical system involving ion transport and reactions in the battery [1], the temperature is a critical factor affecting the performance of lithium-ion batteries [2]. Lithium-ion batteries generate heat during operation, and thermal management systems are used to control the temperature of the batteries in practical applications [3]. The battery cooling solutions mainly include air cooling, liquid cooling, phase change material cooling, etc. [4]. In addition, lithium-ion batteries inevitably age during long-term operation, and the battery aging mechanism is influenced by many factors [5]. Due to the temperature sensitivity of the electrochemical system of the battery, heat exchange conditions can also affect the battery cycle aging performance. The studies of battery aging typically encounter the difficulty of the long duration of experiments. Accurately assessing the thermal behavior and temperature state of batteries under long-term operating conditions is crucial for enhancing the sustainability of battery applications and improving holistic battery life management.

Song et al. carried out aging experiments on prismatic lithium iron phosphate (LFP) batteries [6]. The batteries were cycled in a constant-temperature environmental chamber with different temperature settings. It was found that when the temperature of the environmental chamber increased from 28 °C to 32 °C, the available cycles of the battery decreased by approximately 17%. Laakso et al. chose 18650 cylindrical batteries with a nickel cobalt manganese (NCM) cathode as the research subject and conducted cycle aging tests on the batteries under constant temperature conditions of 5 °C, 23 °C and 40 °C [7]. The results show that the battery aged the fastest at the ambient temperature of 5 °C. Badami et al. conducted aging research on the 26650 LFP battery [8]. The aging characteristics and mechanisms of the batteries in environments of 25 °C and 50 °C were compared and analyzed. Kirkaldy et al. conducted plenty of aging experiments on the 21700 NCM battery and studied the influences of various factors, including temperature, on the battery aging performance [9]. The temperature factor was considered through three different ambient temperatures of 10 °C, 15 °C and 30 °C. Kucinskis et al. studied the difference in the aging rate of the 21700 cylindrical batteries over a wide range of ambient temperatures from −15 °C to 60 °C [10]. The findings showed that battery aging was rapid at a low ambient temperature but also accelerated when the ambient temperature was too high, indicating that a suitable temperature range exists to achieve the optimal battery life.

Numerous studies have investigated the influence of temperature on the aging performance of lithium-ion batteries [11,12]. As described previously, many experimental studies on battery aging predominantly considered a single heat exchange condition, typically using an environmental chamber. The ambient temperature of battery testing is usually regarded as the indicator that affects the battery aging performance [13,14]. However, under the same ambient temperature, differences in heat exchange conditions can still lead to variations in the battery aging performance. In practical industrial applications, batteries employ different thermal management methods, with variations in the heat transfer modes and intensities. Thus, the temperature level of the battery itself, instead of the ambient temperature, should be one of the crucial factors that affects battery aging performance, meriting further investigation. Compared with square and pouch batteries, cylindrical batteries have certain advantages, such as stable consistency and easy procurement [15]. Cylindrical batteries are suitable for aging experimental studies [16]. This study focused on the 21700 high-energy-density cylindrical battery and investigated the battery aging performance under different heat exchange conditions. Throughout the battery aging tests, both the electrical data and the temperature data of the batteries were monitored.

In this study, two types of novel temperature control devices for the cylindrical battery were designed and made. The aging mechanisms of the 21700 battery and the correlation between battery aging and temperature were analyzed. A new temperature metric capable of assessing the temperature level for the long-term-operating battery was proposed. The rest of the paper is arranged as follows: Section 2 states the experimental section, including the experimental equipment and self-made temperature control devices for the cylindrical battery. The test methods are also introduced. Section 3 presents the results and discussion of the battery aging tests. Section 4 concludes the paper.

2. Experimental Section

The 21700 cylindrical battery with a high energy density from Samsung (Yongin, Republic of Korea) was selected as the research subject. Its cathode material is nickel cobalt aluminum (NCA), and the anode material is graphite mixed with a small amount of silicon. The basic specifications of the battery are shown in

Table 1. Basic specifications of the battery [17].

Cell Model	Nominal Capacity (Ah)	Charging Cut-off Voltage (V)	Discharging Cut-off Voltage (V)	Height (mm)	Diameter (mm)
SAMSUNG INR21700-N50E (Samsung SDI Co., Ltd., Yongin, Republic of Korea)	4.9	4.2	2.5	70	21

. The 21700 batteries were tested for consistency, and batteries with an initial nominal capacity difference within 1% were selected for the experiments.

2.1. Experimental Equipment

This research utilized the following main equipment:

• NEWARE CT4008-5V12A battery charging (NEWARE Technology LLC, Shenzhen, China) and discharging equipment was used to control the battery operation according to the test process mentioned below. Battery charging and discharging equipment can record the terminal voltage, current and other electrical data of the operating battery.
• An ESPEC GPU-3 (ESPEC Corp., Osaka, Japan) environmental chamber with an internal integrated heating and cooling system was used to create an adjustable constant-temperature environment for battery testing.
• An OMEGA high-precision T-type thermocouple (OMEGA Engineering, Norwalk, CT, USA) was connected to a National Instruments 9214 board and cRIO-9037 board (National Instruments, Austin, TX, USA) base to form a temperature acquisition system. The thermocouples were calibrated in an oil bath using a first-class standard mercury thermometer.

The specifications of the test equipment are shown in

Table 2. Specifications of test equipment.

Equipment	Model	Specification
Battery charging and discharging equipment	Neware CT4008-5V12A	Voltage range: 25 mV–5 V; current range: 5 mA–12 A; accuracy: ±0.05%
Environmental chamber	ESPEC GPU-3	Temperature range: −40 °C–150 °C; temperature fluctuation ≤ ±0.3 °C
Thermocouple	OMEGA T-type	Diameter: 0.6 mm; range: −267 °C~260 °C
Temperature acquisition board	NI 9214	Number of channels: 16; errors: 0.33 °C
Temperature board base	NI cRIO-9037	1.33 GHz dual-core Intel Atom processor

.

2.2. Self-Made Temperature Control Devices for Cylindrical Battery

The cylindrical lithium-ion battery generates heat when operating, and the heat is dissipated through the outer faces of the battery. The heat exchange conditions directly affect the temperature state of the cylindrical battery. In this study, cycle aging performance tests of the cylindrical battery were conducted under three types of heat exchange conditions. The first heat exchange condition was the environmental chamber condition. The environmental chamber is the most widely used heat exchange condition for battery experiments. When the cylindrical battery is charging or discharging in the environmental chamber, the side face and two end faces of the cylindrical battery can all dissipate heat, which is equivalent to the convection air cooling condition with a constant ambient temperature. The second heat exchange condition was designed to maintain a constant temperature on the side face of the operating cylindrical battery. There is a strong heat exchange effect on the side face of the cylindrical battery, which is similar to the situation in engineering where the side face of the cylindrical battery is cooled by attaching serpentine cooling tubes around it. The third heat exchange condition was designed to maintain a constant temperature at the two end faces of the operating cylindrical battery. The cylindrical battery mainly dissipates heat through the end faces, which is similar to the situation in engineering where vertically arranged cylindrical batteries are grouped and cooled on the top and bottom surfaces of the battery module. The second and third heat exchange conditions were both achieved by self-made temperature control devices for the cylindrical battery. The specific setups are described below.

2.2.1. Side Face Temperature Control Device for the Cylindrical Battery

A thermoelectric cooler (TEC) was used to control the battery temperature. When direct current passes through the TEC, due to the Peltier effect, heat can be absorbed on one face of the TEC, and heat can be released on the other face of the TEC. If the current is reversed, the cooling face and heating face will be swapped, which makes it suitable for temperature control. The primary factor limiting the widespread application of the TEC in battery thermal management is the heat dissipation requirements at the hot side of the TEC.

The side face of the cylindrical battery is curved, while both faces of the TEC are flat. To achieve good heat exchange between the TEC and the side face of the cylindrical battery, two grooved aluminum blocks were designed to be wrapped around the cylindrical battery, and the outer side of each aluminum block was in contact with a TEC. In addition, a power supply, a TEC controller, two fans and other equipment were required to form the side face temperature control device for the cylindrical battery, as shown in Figure 1 and Figure 2. The main equipment parameters are listed in

Table 3. Main equipment parameters of battery temperature control device.

Component	Parameters
12706 TEC (Tianqi Star Electronics Co., Shenzhen, China)	Voltage: 12 V; size: 40 mm × 40 mm × 4 mm
TCM-M207 TEC controller (TE Technology, Traverse City, MI, USA)	Voltage: 24 V; maximum current: 7 A
Thin-film NTC thermistor	Probe thickness: 0.5 mm; measurement error: 0.35 °C
Fan	Voltage: 24 V; size: 60 mm × 60 mm × 25 mm
Copper fin	Size: 60 mm × 60 mm × 22 mm
Thermal conductive silicone pad	Thickness: 1 mm; thermal conductivity: 6.5 W/(m·K)
Power supply	Voltage: 24 V; maximum current: 15 A

. Between the 21700 cylindrical battery and the aluminum block was the thermal conductive silicone pad. The two aluminum blocks were clamped by screws and nuts to reduce the contact thermal resistance. The length of the tightened screws was marked to ensure that the preload force was consistent before and after each disassembly and assembly. The two aluminum blocks were connected to the TEC, fins, fans, etc., forming an integrated unit, which made disassembly and assembly of the battery easy and convenient. There were thermal conductive silicone pads on both sides of the TEC. The fins with fans were used to ensure the heat dissipation of TECs. The TEC controller was based on proportional–integral–derivative (PID) control to reach the target temperature. Since the controller only supports thermistors as the temperature sensor, the customized thin-film negative temperature coefficient (NTC) thermistors were calibrated and arranged between the side face of the cylindrical battery and the groove of the aluminum block located at the contact face center. This means that the temperature at the center of the side face of the cylindrical battery was the temperature input for the controller. Easyhost 6.1 software was installed on the computer to interact with the controller. The controller is dual-channel and can simultaneously control the TECs on the outside of the two aluminum blocks. The two target temperatures were consistently set to be identical to maintain the temperature on the side face of the cylindrical battery. As the side face of the battery is a steel shell with high thermal conductivity, the temperature uniformity effect of the aluminum block is even stronger. This is conducive to achieving a uniform temperature on the side face of the cylindrical battery under temperature control conditions. Since thermocouples were used in all temperature measurement scenarios except for the TEC controller in this study, the thermocouple was also arranged at the center of the side face of the cylindrical battery to monitor the temperature simultaneously. To connect the cylindrical battery clamped between aluminum blocks to the battery charging and discharging equipment, nickel sheets were spot-welded at both end faces of the battery. Meanwhile, to prevent battery short circuits, parts of the nickel sheets were wrapped with insulating tape. In addition, to reduce the heat dissipation from the end faces of the battery and decrease the heat exchange between the aluminum blocks and the environment, the assembled device was wrapped with 5 mm thick Armacell insulation tape (Armacell, Münster, Germany).

Compared with the convection air-cooling condition in the environmental chamber, in the side face temperature control device, the side face of the cylindrical battery and aluminum blocks directly exchange heat through heat conduction, which greatly enhances the heat dissipation of the operating battery. In engineering, battery thermal management systems can effectively suppress the temperature rise of batteries. For instance, in electric vehicles, the side face of the cylindrical battery is closely attached to liquid cooling tubes, or in the energy storage field, batteries can be wrapped in phase change materials. If the cooling capacity of the thermal management system is strong enough, the temperature of the battery cooling face will hardly rise when the battery is operating. The side face temperature control device achieves this effect because the maximum cooling power of the 12706 TEC exceeds 50 W, while the maximum heat generation power of the 2 C discharging battery is only about 5 W. If the target temperature of the control device is set at 25 °C, the side face of the operating cylindrical battery is maintained at 25 °C.

2.2.2. End Faces Temperature Control Device for the Cylindrical Battery

The basic principle of the end faces temperature control device for the cylindrical battery is the same as that of the side face temperature control device introduced in the previous subsection. The schematic diagram and the physical image are shown in Figure 3 and Figure 4. TECs in combination with a controller were also utilized to achieve temperature control. The equipment presented in Table 3 was also employed in the end faces temperature control device. The achieved effect was maintenance of the temperature of the two end faces of the cylindrical battery. Two square aluminum plates with a thickness of 5 mm were used to clamp the two end faces of the cylindrical battery. The side face of the 21700 cylindrical battery was wrapped in Armacell insulation tubes with an inner diameter of 22 mm and a thickness of 30 mm to reduce the heat dissipation from the side face. The cylindrical battery was placed horizontally to ensure unobstructed airflow from the fans on both sides of the device, achieving efficient heat exchange. To avoid the risk of a battery short circuit, the two aluminum plates were connected with plastic screws. An aluminum plate was attached to a TEC (with thermal conductive silicone pads on both sides of the TEC), a fin and a fan to form a unit. The device also provided easy accessibility for battery disassembly and assembly. The two aluminum plates clamped the battery. Thermal conductive silicone pads were inserted between the two end faces of the cylindrical battery and the aluminum plate to reduce the contact thermal resistance. The thin-film NTC thermistor connected to the TEC controller was clamped between the end face of the cylindrical battery and the thermal conductive silicone pad, directly measuring the temperature at the center of the battery end face as the temperature input of the TEC controller. The target temperatures of the two end faces of the cylindrical battery were always set to be the same. The operating battery mainly dissipated heat through the end faces. Furthermore, a thermocouple was placed at the center of the side face of the cylindrical battery to monitor the temperature.

Two sets of side face temperature control devices and end faces temperature control devices were made to facilitate the experimental efficiency. Moreover, the temperature control devices were all operated in a 25 °C environmental chamber to ensure consistent environmental conditions during the operation.

2.3. Test Methods

2.3.1. Temperature Measurement

The center of the side face of the cylindrical battery is usually selected as the temperature monitoring point, as shown by the red dot in Figure 5; the measured temperature is generally called the surface temperature of the cylindrical battery [18]. The plastic wrapping film on the battery side face was stripped off to avoid any influence on the temperature measurement. Only a small part of the plastic film near the positive terminal remained for insulation protection [19]. All batteries in this study were uniformly treated in this way. The surface temperature of the cylindrical battery was measured by the thermocouple under different heat exchange conditions.

2.3.2. Charge and Discharge Tests

Referring to the specification of the 21700 cylindrical battery [17] and the Chinese national standard GB/T 36276-2023 [20], in this study, a 25 °C constant-temperature environmental chamber was selected as the condition for the reference charge and discharge test. The steps of the reference charge test involved the battery being shelved for 30 min to reach thermal equilibrium and then charged with a constant current of 0.5 C (1 C = 4.9 A). When the terminal voltage reached the charging cut-off voltage of 4.2 V, the test switched to constant voltage charging until the current dropped to 0.02 C. The reference discharge test of the battery was discharged with a constant current of 1 C in a 25 °C environmental chamber after thermal equilibrium. The discharging process ended when the terminal voltage reached the discharging cut-off voltage of 2.5 V. The charging process that includes the constant current stage first and then the constant voltage stage is simply called CC-CV charging, and the discharging process with a constant current is simply called CC discharging.

2.3.3. Nominal Capacity Test

The nominal capacity is used to evaluate the electricity storage capacity of the battery. During the battery cycle aging process, it is generally assumed that the battery reaches the end of life when the nominal capacity reduces to 80% of the initial nominal capacity. The nominal capacity of the battery can be obtained through the reference charge and discharge tests introduced above. Three consecutive cycles of reference charge and discharge tests were conducted. The discharging capacity of the last cycle was taken as the nominal capacity of the battery.

2.3.4. Direct Current (DC) Internal Resistance Test

The DC internal resistance of the battery is related to the heat generation of the battery and the energy conversion efficiency of charging and discharging. During the cycle aging process, the internal resistance of the battery changes. According to the FreedomCAR Battery Test Manual [21], the hybrid pulse power characterization (HPPC) tests were used to obtain the DC internal resistance of the battery. Charging and discharging current pulses were applied to the battery, and the internal resistance could be calculated by monitoring the voltage change. As the electrochemical system of a battery is greatly affected by the temperature, the internal resistance of the battery varies at different temperatures. A battery’s internal resistance also changes at different states of charge (SOCs). In this study, the battery’s internal resistance at 50% SOC was tested in a 25 °C environmental chamber. The fully charged battery was discharged to 50% SOC with a 1 C rate and shelved for 1 h to achieve thermal and electrochemical equilibrium. Then, the internal resistance of the battery was tested according to

Table 4. Steps of HPPC test.

Step	Time (s)	Rate (C)	Current (A)
Constant-current discharge	10	1	4.9
Shelve	40	0	0
Constant-current charge	10	0.75	3.675

[19].

2.3.5. Quasi-Open Circuit Voltage (OCV) Test

The standard battery OCV test involves adjusting the SOC and being shelved under each SOC to reach electrochemical equilibrium before measuring the voltage. This method is time-consuming, and the results are discrete. Sometimes, it is desired to obtain the variation curve of the battery OCV vs. the SOC. The OCV curve contains the information of the battery electrochemical system, and after processing, it can be used to analyze the aging state of the battery [22]. Some studies charge batteries with a very small current and record the terminal voltage curve of the battery [22,23]. Since the charging current is sufficiently low, the terminal voltage curve is close to the OCV curve. This curve can also be used to obtain the aging information of the battery and is called the quasi-OCV curve [24].

The test steps for measuring the battery quasi-OCV curve are as follows: The fully charged and shelved battery was charged at a constant current of 0.05 C until the terminal voltage reached 4.2 V. The battery terminal voltage curve of this charging process was recorded, which was the quasi-OCV curve, as shown in Figure 6a. There are several voltage plateaus in the quasi-OCV curve. The quasi-OCV curve measured after the battery cycle aging shifted, and the width of the voltage plateaus also changed slightly, which could be used to analyze the battery aging mechanisms. The differential voltage (DV) curve was obtained by performing the differential processing on the quasi-OCV curve, as shown in Figure 6b.

Under different aging states, the DV curves of batteries gradually change. Meanwhile, the distance variation between the characteristic peaks on the DV curve can be used to evaluate the loss of active electrode material (LAM) and loss of lithium inventory (LLI) due to the battery aging mechanism [25,26]. Equation (1) is used to calculate the loss rate of the active anode material. Equation (2) is used to calculate the loss rate of the active cathode material. Equation (3) is used to calculate the loss rate of the lithium inventory. Q represents the distance between the characteristic peaks, as shown in Figure 6b; the second digit of the subscript, i.e., 0, indicates a new battery; and k indicates the battery after k cycles. The calculation of the LLI is different from the LAM, where it uses the capacity of the new battery Q_tot,0 instead of the characteristic peak spacing Q_3,0 as the denominator. This is because the capacity of the new battery is approximately equal to the total active lithium inventory before the battery ages.

$$G_{LAM\_AN}={\displaystyle \frac{Q_{1,0}-Q_{1,k}}{Q_{1,0}}}\times 100\%$$

(1)

$$G_{LAM\_CA}={\displaystyle \frac{Q_{2,0}-Q_{2,k}}{Q_{2,0}}}\times 100\%$$

(2)

$$G_{LLI}={\displaystyle \frac{Q_{3,0}-Q_{3,k}}{Q_{tot,0}}}\times 100\%$$

(3)

2.3.6. Cycle Aging Test

In the test of battery aging performance, it is necessary to equally evaluate the aging state of batteries under different heat exchange conditions through reference performance tests. The reference performance tests include the battery nominal capacity test, DC internal resistance test at 50% SOC and quasi-OCV test. The reference performance tests were all conducted in the 25 °C environmental chamber. The process of the battery cycle aging test is shown in Figure 7. Charge–discharge cycles were carried out under the specific heat exchange condition. CC-CV charging and CC discharging were adopted. After each instance of charging or discharging, the battery was shelved for 30 min to reach thermal equilibrium.

3. Results and Discussion

3.1. Aging Results of Batteries Under Different Heat Exchange Conditions

To explore the relationship between the heat exchange conditions and the cycle aging performance of the 21700 cylindrical battery, cycle aging tests were conducted on 21700 batteries under three types of heat exchange conditions: environmental chamber, side face temperature control device and end faces temperature control device. Both the charging rate and discharging rate were set at 1 C. In the experiment, when the cycle charging and discharging were carried out at 1 C, the battery degraded to 80% of its initial nominal capacity within several hundred cycles, reaching its end of life [27]. The corresponding number of completed cycles is generally called the number of available cycles. The number of available cycles is already limited, as this 21700 cylindrical battery is an energy-type battery with a high energy density but relatively weak power performance. However, due to the long duration of battery aging experiments, it took almost two years to obtain results. Therefore, the influence of different charging and discharging rates was not considered in this study.

The battery aging tests were carried out in the environmental chamber, side face temperature control device and end faces temperature control device conditions with constant temperatures of 15 °C, 25 °C and 35 °C. A total of nine sets of battery cycle aging test data were obtained, and each set of tests had two batteries to avoid being unable to detect abnormal data from a single sample. The discharge capacity data measured in each cycle during the battery aging process are shown in Figure 8. It can be seen that under different heat exchange conditions with various temperature settings, the discharge capacities of the two batteries measured during the cycle aging process were very close. Thus, the results indicate that the 21700 cylindrical batteries exhibit good consistency. Unless otherwise specified, the subsequent aging data represent the average values of the two batteries.

In Figure 8, the initial discharge capacity of the battery differs in different heat exchange conditions. There were three types of heat exchange conditions, and the battery exhibits a lower initial discharge capacity and faster battery aging in each type of heat exchange condition at a lower temperature. For example, the number of available cycles of the battery in the 15 °C environmental chamber is much lower than that in the 35 °C environmental chamber. Furthermore, compared with the results in the environmental chamber, the battery exhibits an overall shortened lifespan in the side face temperature control device. This indicates that an enhanced heat exchange condition may accelerate the aging of this specific 21700 battery. It is noted that the capacity fade rate of the battery accelerates in the later period of the aging process. The potential causes are that the loss of active materials and lithium inventory is cumulative, and the influence of these factors becomes more pronounced in the later period of battery aging, thus accelerating the battery degradation.

The reference performance tests included testing the DC internal resistance of the battery at 50% SOC. The battery internal resistance under pulse discharge and pulse charge could be obtained, and Figure 9 shows the average internal resistance during the battery aging process under different heat exchange conditions.

Figure 10 shows the quasi-OCV curves measured during the battery aging process in different heat exchange conditions. The quasi-OCV curve was differentiated and further processed according to the method described in Section 2.3 to obtain the battery aging information, which is plotted in Figure 11. It can be seen that there are certain differences in the aging mechanisms of the batteries in different conditions. For instance, consider the test results in the environmental chamber as an example for analysis. The loss rate of the active anode material rises more rapidly in a lower-temperature environmental chamber. The loss rate of the active cathode material only shows a positive value under the high-temperature condition and in the later stage of battery aging. For example, in the 35 °C environmental chamber, there is a significant loss of the active cathode material in the later stage of the battery aging. However, throughout the entire aging process of the battery in the 15 °C environmental chamber, there is almost no loss of the active cathode material (since the active electrode material of the actual battery cannot increase, it is generally believed that a negative calculated loss rate result indicates that there is no loss of the active electrode material) [28]. The loss rate of the battery lithium inventory rises more slowly in the environmental chamber at a higher temperature.

3.2. Analysis of Battery Aging Mechanism

To further analyze the battery aging mechanisms, aging data from nine sets of batteries in different heat exchange conditions are processed and consolidated. The loss rates of active anode material, active cathode material, lithium inventory and growth rate of the internal resistance are plotted against the fade percentage of the nominal capacity, as shown in Figure 12.

Figure 12a reveals that the loss rate of the active anode material exhibits a predominantly linear relationship with the fade percentage of the nominal capacity, indicating an obvious correlation between the active anode material loss and the capacity degradation, with minimal influence from the heat exchange conditions. Thus, the data points under different heat exchange conditions cluster closely in the plot. After the battery aging tests, the maximum loss rate of the active anode material is close to 20%. Considering that the anode material of this 21700 battery is graphite–silicon, the silicon enhances the energy density due to a higher specific capacity compared with graphite [29]. However, silicon’s excessive volume expansion during battery charging can accelerate the loss of active anode material and adversely affect the cycle life [30].

Figure 12b indicates no clear correlation between the active cathode material loss rate and the fade percentage of the nominal capacity. For instance, the batteries reaching a nominal capacity fade of 20% under different heat exchange conditions show significant differences in the active cathode material loss. High cathode material loss occurs in high-temperature conditions during later aging stages. For example, the maximum loss rate of the active cathode material of the battery in the 35 °C environmental chamber is about 18%. Active electrode materials can be categorized as lithium-containing or lithium-free active materials. Active electrode material loss during battery aging results from particle fracture, structural collapse and detachment of active materials. Consequently, loss of lithium-containing active materials also contributes to loss of lithium inventory.

Figure 12c illustrates the relationship between the loss of lithium inventory and the nominal capacity fade. The loss rate of lithium inventory partially reflects the combined effect of the active cathode and anode material loss rates. However, the overall trend shows rapid initial growth followed by slower progression in later aging stages as the capacity fade increases, diverging from the loss patterns observed for active electrode materials. Therefore, lithium inventory loss may involve additional factors, such as the growth of the solid electrolyte interface (SEI) film and lithium plating on the anode [31]. At the end of the battery’s life, the loss rate of lithium inventory of most test groups is approximately 10%, which has a direct impact on the reduction in the battery nominal capacity.

Figure 12d presents the growth rate of the DC internal resistance versus the fade percentage of the nominal capacity. The growth rate of the internal resistance exceeds 50% in some groups. Even the group with the minimum internal resistance growth exhibits about a 30% increase. The increase in internal resistance also contributes to the battery capacity fade. It is generally believed that the growth of the SEI film and lithium plating can lead to an increase in the internal resistance [32]. The growth pattern of internal resistance also partially reflects the combined effect of the cathode and active anode material loss rates, which indicates that active electrode material degradation may also contribute to resistance growth. For example, particle fracture and detachment compromise the electrical contact network within the electrode, thereby reducing the overall electrical conductivity inside the battery.

3.3. Temperature Results of Batteries During Cycle Aging Process

Thermocouples were used to monitor the temperature at the side face of the 21700 batteries during the cycle aging process. Figure 13 depicts the battery temperature results measured in the environmental chamber. The monitoring temperature remained constant in the side face temperature control device for the cylindrical battery. Figure 14 presents the temperature results from the end faces temperature control device for the cylindrical battery. Figure 13 and Figure 14 only display the results of some selected cycles, including charging and discharge during the aging process. As the number of cycles increases, the aged batteries exhibit a significantly increased temperature rise rate during discharging, along with a higher maximum temperature. This phenomenon is primarily driven by the growth of internal resistance during battery aging. However, the increase in the maximum temperature during charging for aged batteries is not pronounced. Due to the increased internal resistance, the battery reaches the charging cut-off voltage and its peak temperature earlier during the CC charging period.

Since the charging and discharging rates and cut-off conditions are identical for all batteries in this study, the heat exchange condition is the sole factor responsible for the variations in the battery aging rate. When considering the influence of temperature on battery aging, some studies utilize ambient temperature as the metric, and the ambient temperature is directly considered as a variable in semi-empirical aging models. The ambient temperature does affect the battery aging rate, but the actual influencing factor should be the temperature of the battery itself. Under stable heat exchange conditions, such as an environmental chamber, raising or lowering the ambient temperature directly corresponds to higher or lower battery temperatures. However, when diverse heat exchange conditions are considered, heat transfer modes and heat transfer intensities exhibit variations. It is inadequate to regard ambient temperature as a metric. For example, in this study, under the conditions of a 25 °C environmental chamber, 25 °C side face temperature control device and 25 °C end face temperature control device, the ambient temperature for the battery can be regarded as 25 °C in all three conditions. Nevertheless, the battery aging rates exhibit significant differences.

The influence of different heat exchange conditions on the battery is reflected in the battery temperature state. The temperature measurement at the center point on the side face of the cylindrical battery is normally regarded as a standardized approach. The measured temperature can be called the surface temperature of the cylindrical battery. It is a non-destructive monitoring method and convenient for temperature recording throughout the battery’s lifecycle. Considering that the battery cycle aging process takes a long time and the surface temperature is continuously changing, the extreme value of the battery surface temperature that only appears at a particular moment is obviously not suitable as a temperature metric for the battery aging process. The influence of the battery temperature state on the cycle aging process should be a long-term cumulative effect. Therefore, we propose a new metric—the cumulative time-averaged surface temperature (CTAT), which is calculated by integrating the surface temperature value over time during the cycle aging process and dividing by the time length—as a metric to represent the temperature level of the cylindrical battery during cycle aging. The formula for calculating the CTAT is shown in Equation (4). In the equation, i denotes the i-th cycle, and n denotes the number of cycles. $\overline{T_i}$ represents the time-averaged temperature of the battery during the i-th cycle, and t_i denotes the duration of the i-th cycle. It should be noted that only temperature data recorded during the charging and discharging periods are considered in the CTAT calculation, while the data of the rest periods without current are excluded. The CTAT values under the nine heat exchange conditions in this study are calculated and presented in

Table 5. Temperature metrics during battery aging under different heat exchange conditions.

Heat Exchange Condition	CTAT (°C)	Number of Available Cycles
15 °C environmental chamber	20.5	75
25 °C environmental chamber	30.4	172
35 °C environmental chamber	40.3	495
15 °C side face controlled	15.0	40
25 °C side face controlled	25.0	97
35 °C side face controlled	35.0	243
15 °C end faces controlled	17.8	58
25 °C end faces controlled	27.7	125
35 °C end faces controlled	37.7	348

. The CTAT values range from 15 °C to 40.3 °C. The number of available cycles of the batteries under different heat exchange conditions is also listed in Table 5. Among the nine sets of battery aging data, the maximum CTAT is 40.3 °C, corresponding to 495 available cycles, and the minimum CTAT is 15.0 °C, corresponding to 40 available cycles. It can be observed that for the 21700 battery, within the temperature range of this study, the lower the CTAT, the fewer the available cycles.

$$CTAT={\displaystyle \frac{{\sum }_{i=1}^n\overline{T_i}·t_i}{{\sum }_{i=1}^n{t}_i}}$$

(4)

3.4. Correlation Between Battery Temperature and Nominal Capacity Fade

During the battery aging process, the reference performance tests, including the nominal capacity test, were conducted. The fade percentage of the nominal capacity is shown in Figure 15, which is the ratio of the reduction in nominal capacity to the initial nominal capacity. It can be seen that different heat exchange conditions have a significant impact on the nominal capacity fade rate. The CTAT of the battery from Table 5 is correspondingly marked on the corresponding nominal capacity fade curves in the figure (only the temperature values are labeled without the unit “°C”). The battery in the 35 °C environmental chamber exhibits the slowest aging rate, with a CTAT of 40.3 °C, and after 500 cycles, the fade percentage of its nominal capacity reaches 21.6%. It can be observed that all nine sets of data of the 21700 battery show that a lower CTAT results in faster battery aging. Based on the analyses of battery aging mechanisms in the previous section, the main reason for this phenomenon is the accelerated loss of active anode materials and lithium inventory, as well as an increase in the internal resistance at lower temperatures. Furthermore, it is observed that the shapes of the increasing internal resistance curves in Figure 9 has similarities with the fade percentage of the nominal capacity in Figure 15, indicating that there are correlations between the increases in the battery internal resistance and the capacity fade. This suggests that the battery degradation mechanisms, such as the loss of active materials and lithium inventory, may drive both the resistance rise and capacity loss in a closely synchronized manner.

Based on Figure 15, a semi-empirical battery aging model can be constructed based on experimental data. In this study, since the factors influencing battery aging are limited to different heat exchange conditions, the variables in the semi-empirical model do not include other factors such as charging/discharging rates and cut-off conditions. The Arrhenius equation is often used to describe the relationship between chemical reactions and temperature [33]. The Arrhenius equation can be used to fit the battery aging data, as shown in Equation (5). The variables are the temperature metric T and the number of cycles n, and Q indicates the fade percentage of the nominal capacity. A and B are constants. E is the apparent activation energy, and R is the gas constant, with a value of 8.314 J/(mol·K). Based on the aforementioned research, the temperature metric T can be regarded as the cumulative time-averaged surface temperature of the battery CTAT proposed in this study. When fitting the nine sets of data in Figure 15, the CTAT should be expressed in Kelvin (K). The fitting result is shown in Equation (6), with an R² value of 0.88. During the fitting process, −E/R was treated as a constant, and the obtained result is positive, indicating that the apparent activation energy E was negative. This suggests that the overall aging-related reactions for the 21700 battery are accelerated at lower temperatures within the temperature range of this study. Equation (6) can be used to predict the battery nominal capacity fade and cycle aging life with the CTAT and the number of cycles n.

$$Q=A{e}^{-\frac{E}{RT}}{n}^B$$

(5)

$$Q=5.17\times {10}^{-24}{e}^{\frac{\mathrm{13,200}}{CTAT}}{n}^{1.64}$$

(6)

4. Conclusions

4.1. Summary of Key Results

This paper presents an experimental study on the cycle aging life of the 21700 cylindrical batteries under different heat exchange conditions. The heat exchange conditions include the environmental chamber, as well as two types of independently designed battery temperature control devices, which correspond to the ideal heat dissipation effects on the side face and end faces of cylindrical batteries. Cycle aging tests were conducted on the 21700 cylindrical batteries under the three types of heat exchange conditions. Plenty of experimental data were obtained, and the aging mechanism of the 21700 battery was analyzed. Some key results are as follows:

• The maximum loss rate of active anode material is close to 20% after the battery aging tests, with an obvious correlation with the nominal capacity fade. Meanwhile the correlation between the active cathode material loss and the capacity degradation is insignificant.
• The loss rate of lithium inventory of most test groups is approximately 10% after the battery aging tests, which directly impacts the reduction in the battery’s nominal capacity. The internal resistance growth rate can exceed 50% at the end of the battery’s life.
• Experimental data of the 21700 cylindrical batteries show that the lower the CTAT, the fewer the number of available cycles within the temperature range of this study.
• A semi-empirical battery aging model was established that describes the correlation between the fade percentage of battery nominal capacity and the CTAT, as well as the number of cycles.

4.2. Limitations and Future Work

The designed temperature control devices for the long-term operating battery and the analysis of battery aging mechanisms in this study provide guidance for research on battery life sustainability. The proposed battery temperature metric CTAT provides valuable insights for the temperature control strategy in battery thermal management design, which may contribute to extending the battery’s lifespan in energy storage systems, thereby advancing sustainable development.

This study focuses on one type of 21700 battery. Different lithium-ion batteries vary in aspects such as the aging rate and the correlation between the battery aging rate and the temperature. The battery types and the temperature range of the heat exchange condition can be extended in future work. The CTAT was proposed and analyzed in this study. In further research, other metrics such as RMS temperature and entropy-based metrics can be explored. In future work, electrochemical simulation of the battery can be integrated to conduct further analysis, including determining the contribution ratio of each aging mechanism to the fade percentage of the nominal capacity.