Gas Characterization-based Detection of Thermal Runaway Fusion in Lithium-ion Batteries

Abstract

A novel approach for real-time detection of lithium-ion battery thermal runaway has been proposed to enable the monitoring of thermal runaway states during storage, transportation, and use, and to prevent safety hazards such as fire and explosion. This approach uses the fusion of high-precision, high-sensitivity photoelectric and electrochemical detection techniques based on the dual-wavelength principle. To analyze the thermal runaway mechanism of lithium-ion batteries, four important gas parameters — CO, EX, H₂, and CO₂ — were obtained to indicate the thermal runaway state, and the characterization of these parameters under different thermal runaway states of lithium-ion batteries was studied. A real-time detection system is designed and validated through experiments involving overcharging, over-discharging, and puncturing of lithium-ion batteries. This method is suitable for the real-time detection of thermal runaway in lithium-ion battery products and can also provide a basis for evaluating the life and reliability of lithium-ion batteries.

1. Introduction

Electric energy has emerged as one of the most extensively employed renewable energy sources in the new century, offering a promising solution to the escalating energy demands and environmental challenges. Lithium-ion batteries have gained prominence in the energy storage domain for aviation, transportation, and other sectors, due to their extended cycle life, high power density, and cost-effectiveness.^1,2 Thermal runaway is a hazardous chain reaction phenomenon resulting from various factors that can cause Lithium-ion batteries to ignite and explode, owing to the emission of significant heat and harmful gases. The transition from fossil fuels to electric energy has led to a surge in frequent thermal runaway incidents involving Lithium-ion batteries, thereby posing a threat to both human life and industrial productivity.^3–5 Hence, exploring the thermal runaway characteristics of Lithium-ion batteries is of paramount significance to ensure safety measures.⁶

Scholars both domestically and abroad have conducted extensive research and discovered that the mechanism and manifestation of thermal runaway exhibit similarities. This finding provides significant inspiration for the study of thermal runaway detection methods. Xu et al. developed a method to detect battery short circuit faults using a calculated theoretical model.^7–9 Wang et al. accurately measured the thermal changes of different lithium-ion batteries during overcharging using an accelerated rate calorimeter, and studied overcharge thermal runaway by detecting the internal/external temperature and voltage of the batteries.^10–12 Liu et al. designed and built an experimental platform for lithium-ion battery fires and explosions to investigate the effect of different initial pressures on the thermal runaway characteristics of a lithium-ion battery pack in a confined space.¹³ Early warning research on lithium-ion battery thermal runaway has revealed that existing electrochemical principle sensors and semiconductor sensors are susceptible to low detection accuracy and cross-interference.¹⁴ To enhance the thermal runaway detection capability, it is necessary to design detection technologies that are more reliable, accurate, and possess faster thermal runaway recognition speed.

In this study, we propose a detection method that integrates the detection of thermal runaway in lithium-ion batteries using the Mie scattering theory and a dual-wavelength photoelectric detector in combination with an electrochemical gas sensor.¹⁵ The method aims to ensure high sensitivity of photoelectric detection while reducing the high false positive rate¹⁶ of photoelectric detection. We conducted experiments to validate the feasibility of the proposed method.

2. Theoretical Foundation

2.1 Dual-wavelength theory

Utilizing the Mie scattering model¹⁷ for aerosol optics and the photoelectric effect for statistical measurements, the power of scattered light P_n detected by an acceptable aperture at a certain scattering angle can be expressed as follows:¹⁸   

\begin{equation} \text{P}_{n} = C_{n}\int\limits_{0}^{\infty}f(d)\text{P}_{\lambda}(d,\lambda,m)\text{d}d \end{equation} (1)

Where P_λ (d,λ,m) is the single-particle Mie scattering intensity, d is the particle diameter, λ is the wavelength of incident light, m is the refractive index, f(d) is the particle size distribution function, and C_n is the mass concentration of the aerosol.

Equation 1 shows that the aerosol scattering power is proportional to its concentration. The power of scattering when using a short wavelength is denoted as P_S, while that when using a long wavelength is denoted as P_L.

When the scattering angle, receiving aperture, and short and long wavelengths of light are determined, and when the mass concentration C_n and refractive index are constant, the ratio of short-wavelength light scattering power P_S to long-wavelength light scattering power P_L is denoted as R.   

\begin{equation} \text{R} = \frac{\text{P}_{\text{S}}}{\text{P}_{\text{L}}} = \frac{\displaystyle\int\limits_{0}^{\infty}f(d)\text{P}_{\lambda}(d,\lambda_{S},m)\text{d}d}{\displaystyle\int\limits_{0}^{\infty}f(d)\text{P}_{\lambda}(d,\lambda_{L},m)\text{d}d} \end{equation} (2)

Equation 2 is a function of the scattering power ratio R and the median particle size d, where λ_S and λ_L represent the short and long wavelengths of light, respectively. By correlating the value of R with the corresponding particle size, the median particle size of the aerosol can be determined. Employing the dual-wavelength theory in photoelectric sensors, the median particle size can be inferred from the ratio of blue and red light power. This allows for differentiation of different types of aerosol particles and outperforms the accuracy of single-wavelength light source photoelectric sensors. It can also reduce false alarms in warning against lithium-ion battery thermal runaway.¹⁹

The experiment utilized a dual-wavelength optoelectronic sensor consisting of blue light LED (470 nm) and infrared LED (850 nm). The particle size was characterized by the ratio of the scattered light power of the two wavelengths. By means of analysis, the thermal runaway particles of lithium-ion batteries and interference particles could be effectively identified.

2.2 Mechanism of gas production

The gas generation mechanism of lithium-ion batteries under different thermal runaway conditions varies depending on their chemical systems. In this experiment, we used a new type of lithium-ion battery, the ternary material soft pack lithium-ion battery, which has a positive electrode composed of a ternary composite material (NCA), a negative electrode made of graphite, and an electrolyte consisting of a three-phase mixture of organic solvents and additives. This battery exhibits characteristics such as good flexibility, light weight, and bendability.²⁰

During thermal runaway of the ternary polymer lithium-ion battery, the temperature increase causes oxygen to decompose from the NCA positive electrode material, as expressed in Eq. 3:   

\begin{align} &\text{Li$_{x}$(Ni$_{0.80}$Co$_{0.15}$Al$_{0.05}$)O$_{2}$} \notag\\ &\quad\to \text{Li$_{x}$(Ni$_{0.80}$Co$_{0.15}$Al$_{0.05}$)O$_{1 + x}$} + \frac{1}{2}\text{($1 - x$)O$_{2}$} \end{align} (3)

The oxygen molecules produced can initiate various reactions. Carbon-based materials, SEI film (consisting mainly of various inorganic components such as Li₂CO₃, LiF, Li₂O, LiOH, and various organic components such as CH₂OCO₂Li, CH₂OLi, (CH₂OCO₂Li)₂, etc.), and organic electrolytes undergo varying degrees of oxidation or combustion reactions under sufficient or insufficient oxygen supply, producing various carbon-oxygen compounds, hydrogen, and combustible gases, as shown in Eqs. 4, 5, 6, 7, and 8:^21,22   

\begin{equation} \text{C} + \frac{1}{2}\text{O$_{2}$} \to \text{CO} \end{equation} (4)

  

\begin{equation} \text{CO} + \frac{1}{2}\text{O$_{2}$} \to \text{CO$_{2}$} \end{equation} (5)

  

\begin{equation} \text{(CH$_{2}$OCO$_{2}$Li)$_{2}$} + \text{2Li} \to \text{2Li$_{2}$CO$_{3}$} + \text{C$_{2}$H$_{4}$} \end{equation} (6)

  

\begin{equation} \text{Li$_{2}$CO$_{3}$} + \text{2HF} \to \text{2LiF} + \text{CO$_{2}$} + \text{H$_{2}$O} \end{equation} (7)

  

\begin{equation} \text{CO} + \text{H$_{2}$O} \to \text{CO$_{2}$} + \text{H$_{2}$} \end{equation} (8)

The gas generation mechanism during the thermal runaway process of lithium-ion batteries is complex. Based on the above analysis, thermal runaway of lithium-ion batteries inevitably results in the production of CO, CO₂, H₂, and combustible gases. These four gas parameters can be used to characterize the thermal runaway state of lithium-ion batteries under different thermal runaway conditions, and when combined with the overall thermal runaway characteristics, a new thermal runaway warning system can be established by integrating photoelectric detection technology.

3. Experimental Plan

3.1 Experimental instruments and equipment

In this experiment, the IT6872A power supply was used to charge the batteries, while a 220 V to 24 V switch power supply was used to power the air pump. The PT100 thermistor attached to the surface of one side of the Li-ion battery was utilized to detect changes in surface temperature during the thermal runaway process of the Lithium-ion battery. For infrared monitoring, an H10 series thermal imager was employed. Additionally, an ADPD188BI photovoltaic detection sensor, JX series CO₂ sensor, and JXM series CO, EX, H₂ sensors were utilized to detect changes in gas type and mass concentration during the thermal runaway process. The data was recorded, saved, and uploaded through a serial port.

3.2 Experimental platform

To investigate the genuine thermal runaway process of lithium-ion batteries under conditions of being overcharged, being over-discharged and being externally punctured, an experimental platform was constructed in this study, enabling simultaneous photovoltaic and gas detection. The thermal runaway process of lithium-ion batteries releases a significant amount of heat. To minimize the impact of temperature on the sensors, a gas chamber and a detection chamber were devised. The gas chamber, measuring 10 cm in length, 6 cm in width, and 4 cm in height, and the detection chamber, measuring 12.5 cm in length, 12.5 cm in width, and 9.5 cm in height, were assembled using glass to facilitate observation and eliminate interference from other gases during the experiment. The theoretical effect is presented in Fig. 1a, and the practical platform is illustrated in Fig. 1b.

Figure 1.

Schematic diagram of the experimental platform.

3.3 Experimental process and content

The three experiments were conducted at a room temperature of 25 °C, and the experimental procedures are shown in Fig. 2. The information of the lithium-ion batteries used in the experiment is shown in Table 1.

Figure 2.

Experimental flow chart.

Table 1. The selected Lithium-ion battery information.

Rated capacity (mA h)	300 mA h	1500 mA h
Model	602030	703450
Rated voltage (V)	3.7	3.7 V
Charging termination voltage (V)	4.2	4.2 V
Discharge cut-off voltage (V)	2.7	2.7 V
Thickness/width/total length (mm)	6/20/31	8/34/51

For the lithium-ion battery being over-charged experiment, a 1500 mA h battery was chosen and a PT100 thermistor was affixed to its surface. The battery was initially at 100 % state of charge with a voltage of 4.2 V. A constant current and voltage of 6 A and 10 V, respectively, were applied to the battery using an IT6872A power supply.

For the being over-discharged experiment, a 300 mA h lithium-ion battery was selected, and a PT100 thermistor was attached to its surface. The battery was initially at 100 % SOC state with a voltage of 4.2 V. Continuous discharge was carried out using a 50 W, 0.4 Ω metal aluminum shell resistor as the discharge load.

For the external puncture experiment of a Li-ion battery, a PT100 thermistor was affixed to the surface of a 1500 mA h battery, initially at 50 % SOC state. A steel cone with a diameter of 5 mm was used to apply force to the surface of the battery. The loading was stopped and the steel cone was withdrawn when it penetrated one-third of the battery’s thickness.

4. Thermal Runaway Results and Data Analysis

4.1 Overcharge experiment

Overcharging is a common safety hazard that can cause thermal runaway in lithium-ion batteries. When lithium-ion batteries are charged with a current and voltage that exceeds their rating, it can damage the internal structure of the battery. If the power is not cut off in time, it can even lead to accidents such as fires. In this overcharging experiment, the lithium-ion battery eventually caught fire and exploded, resulting in complete failure, as shown in Fig. 3.

Figure 3.

Visible light monitoring of thermal runaway caused by overcharging of the lithium batteries.

At the beginning of the overcharging experiment, all equipment began recording, and the relative time was recorded as t = 0 s. The thermal and visible imaging cameras were used to detect the thermal runaway phenomenon in the lithium-ion battery during being overcharged, as illustrated in Fig. 4.

Figure 4.

Infrared monitoring diagram of the soft pack lithium batteries being overcharged leading to thermal runaway.

The parameters of the 1500 mA soft pack lithium-ion battery in being overcharged with thermal runaway are shown in Fig. 5.

Figure 5.

Detection data of thermal runaway caused by overcharged the Lithium-ion battery.

The experiment involved subjecting a lithium-ion battery to an overcharging test using a 10 V, 6 A power supply, while monitoring the battery’s surface temperature with an infrared camera, as shown in Fig. 4. The temperature increased steadily during the overcharge, with a noticeable acceleration at t = 353 s, when a large amount of smoke was released and the temperature reached 80.86 °C. At t = 372 s, the battery system was completely destroyed, with the current dropping to 0.1 A, voltage rising to 9.9 V, and the surface temperature exceeding 100 °C, which resulted in flames. The lithium-ion battery was overcharged process current, voltage and temperature change as shown in Fig. 5a.

Gas concentration measurements were taken during the experiment, with four gases detected, and the mass concentration changes were shown in Fig. 5b. Prior to relative time 330 s, the CO, EX, H₂, and CO₂ sensors recorded no significant change in gas concentration. At t = 333 s, 334 s, and 335 s, the EX, H₂, and CO sensors, respectively, indicated changes in gas concentration. The H₂ sensor reached a maximum of 100 mg/L 16 s later, followed by the CO sensor reaching its maximum of 1000 mg/L at t = 366 s, which began to decrease after 28 s. The EX-concentration also reached its maximum of 100 mg/L at t = 371 s, before sharply dropping. The researchers analyzed the experimental data and Fig. 5a and concluded that the combustible gas peak concentration corresponded to a battery surface temperature of 147 °C, at which point the lithium-ion battery caught fire and rapidly consumed the combustible gas, leading to a linear decrease in gas concentration. The CO₂ concentration began to increase significantly after relative time t = 389 s, reaching its maximum at t = 431 s.

Analysis of the photodetector signal in Fig. 5c showed that from the beginning of the overcharging until t = 300 s, the compound at the positive electrode released lithium ions that were embedded in the negative electrode. As metallic lithium accumulated in the negative electrode, the battery temperature gradually rose and a small amount of gas was released. However, due to the shell barrier and the low gas production, neither the photodetector nor the gas sensor detected any signals, but the lithium-ion battery slightly expanded. With continuous overcharging, the surface temperature of the battery continued to rise. Almost all the metallic lithium at the positive electrode was deposited between t = 300–326 s, causing a significant increase in the negative electrode resistance, electrolyte decomposition, and temperature rise, resulting in a noticeable expansion of the battery. The photodetector detected the first significant peak signal at t = 326 s, with a high infrared light scattering signal of 121.48 nW/mW and a high blue light scattering signal of 104.37 nW/mW. After gas sensor signal verification, an early warning was issued 36 s in advance. At t = 378 s, the reaction chamber emitted a substantial amount of smoke, which was accompanied by open fire and violent variations in the photoelectric sensor signal.

The photoelectric sensor employs a 470 nm blue light-emitting diode (LED) and an 850 nm infrared LED to distinguish between various sizes of aerosol particles and mitigate false alarms, thereby enhancing detection accuracy. Examination of the ratio curve depicting the response of the blue light and infrared radiation, as shown in Fig. 5d, reveals a range of ratios from 0.825 to 0.895. Analysis of these results indicates that the detected smoke particle diameters are closely clustered, excluding the possibility of false detection and corroborating the effectiveness of our approach.

4.2 Over-discharge experiment

Thermal runaway caused by over-discharge is a form of thermal runaway in lithium-ion batteries that has received less attention in research. The degree of discharge of lithium ions and the reaction and decay mechanisms within the battery are known to affect the occurrence and severity of this phenomenon. The lithium-ion battery was detected by visible light in the phenomenon of thermal runaway caused by over-discharge, as demonstrated in Fig. 6.

Figure 6.

Visible light monitoring of thermal runaway caused by over-discharge in the lithium batteries.

The parameters of 300 mA soft pack lithium-ion battery in being over discharged leading to thermal runaway are shown in Fig. 7.

Figure 7.

Detection data of thermal runaway caused by over-discharged the lithium batteries.

The sensor was activated at t = 0 s, with the battery in a normal condition and the circuit open, resulting in a voltage of 4 V and a current of 0 A. At t = 77 s, the circuit was closed, allowing the continuous discharge of the lithium-ion battery to a 50 w, 0.4 Ω resistor. The status of the lithium-ion battery at this time is depicted in Fig. 6a, while the voltage and current are shown in Fig. 7a. The current exhibited a surge of 3.54 A within 6 seconds, followed by a quick drop to 2.07 A, with a slight recovery before beginning to slowly decrease. At t = 110 s, the voltage hit a minimum of 1.6 V before rising to 2.2 V and continuing to gradually decline, in a trend that was consistent with the current during discharge. After t = 100 s, the surface temperature of the lithium-ion battery began to rise rapidly. By t = 545 s, severe Li⁺ depletion led to copper dissolution, electrode material collapse, and severe deformation of the lithium-ion battery, as well as the production of a significant amount of gas, with a corresponding high temperature, as illustrated in Fig. 6b. The highest temperature recorded was 111.79 °C at t = 552 s, after which it began to decrease.

The gas concentration changes are presented in Fig. 7b. Before t = 530 s, no significant change in gas concentration was detected by the CO, EX, and H₂ and CO₂ sensors. At t = 535 s, the CO concentration was first detected, followed by a rise in the EX-concentration 3 seconds later. The CO concentration peaked at 733.9 mg/L at t = 549 s, then rapidly decreased. The EX-concentration peaked at 580.8 mg/L at t = 552 s and then gradually stabilized. After t = 558 s, the CO₂ concentration began to rise significantly. The H₂ concentration was detected at t = 586 s and reached its maximum range 5 seconds later. The CO₂ concentration also reached its maximum value at t = 615 s.

In Fig. 7c, the photoelectric signal values of the lithium-ion battery during thermal runaway were compared to those of other experimental groups, revealing weak signal values. From 77 s to 530 s, the lithium-ion battery was discharged, causing the negative electrode to lose Li⁺ and the positive electrode potential to decrease. As the over-discharge continues, excessive lithium removal from the negative electrode led to the decomposition of the SEI film and gas production, causing the lithium-ion battery to deform and an internal short circuit to form. At a relative time of t = 533.24 s, a signal value was measured, with the highest scattered blue light source reception signal value being 117.9 nW/mW and the lowest being 113.4 nW/mW, with a difference not exceeding 4.5. The highest scattered infrared light source reception signal value was 117 nW/mW, the lowest was 110 nW/mW, and the difference between the two values did not exceed 7. The resulting gas concentration was approximately 1/5 of that observed in the being overcharged experiment and 1/4 of that in the being punctured experiment. Setting the alarm threshold too low may cause false alarms, while setting it too high may result in missed alarms. The fusion detection method sets a reasonable warning value and detects the gas concentration signal. Even if the photoelectric sensor detects a low signal value and no alarm is sounded, an alarm will still be triggered if there is an increase in the various gases shown in the figure to prevent the possibility of missed alarms.

Figure 7d displays the ratio of blue light response to infrared response. The ratio curve ranges from a minimum of 0.995 to a maximum of 1.065, indicating that the median particle size of the smoke particles detected did not differ significantly and the data was valid.

4.3 Puncture experiment

Puncturing lithium-ion batteries can rapidly trigger thermal runaway, resulting in the release of copious amounts of smoke, and even fire or explosion. The lithium-ion battery was detected by visible light in the phenomenon of thermal runaway caused by puncture, as demonstrated in Fig. 8.

Figure 8.

Visible light detection of thermal runaway due to punctured the lithium batteries.

The experiment was observed at t = 0 s. At t = 20 s, the battery was punctured, which was accompanied by the generation of tiny electric sparks. At t = 90 s, the battery showed a slight bulge with blue smoke and electrolyte leakage at the puncture site, as shown in Fig. 8a. At t = 200 s, thick smoke began to spread, as shown in Fig. 8b.

The parameters of 1500 mA soft pack lithium-ion battery in being punctured leading to thermal runaway are shown in Fig. 9.

Figure 9.

Detection data of thermal runaway heat caused by punctured the lithium batteries.

Figure 9a illustrates the lithium-ion battery’s surface temperature change as a result of the thermal runaway process caused by a puncture, while Fig. 9b shows the variations in the mass concentrations of different gases detected during the experiment. Specifically, four types of gases were detected. The lithium-ion battery is punctured starting at t = 20 s, and after about 45 s, the mass concentrations of H₂ and CO detected by the gas sensor increased significantly. As the temperature rose, the mass concentration of EX increased substantially, surpassing that of H₂ and CO at t = 118 s and t = 126 s, respectively. At t = 174 s, the mass concentration stabilized at 17.5 mg/L and fluctuated around this value thereafter. Concurrently, the mass concentration of CO₂ started to increase rapidly, reaching the full scale at t = 216 s. The mass concentrations of CO and H₂ reached full scale at t = 219 s and t = 232 s, respectively, and then gradually stabilized.

The experimental photoelectric signal values, presented in Fig. 9c, indicate that the puncture began at t = 20 s, and the photoelectric sensor detected the first signal and the appearance of the first peak at t = 63 s. The signal subsequently started to decrease. From t = 86 s onwards, the photoelectric sensor detected a large quantity of smoke generated by an internal short circuit of the lithium-ion battery, and the signal peaked at t = 133 s. The photoelectric signal then gradually decreased, but the signal strength remained strong. The trend of the received blue light and infrared scattering signal values was generally consistent.

Figure 9d illustrates the ratio of the received blue light scattering signal value to the received infrared scattering signal value. When no smoke signal was detected, the ratio fluctuated around 1. When the photoelectric signal showed the first peak, the ratio reached as high as 1.34. When a substantial amount of smoke signal was collected by the photoelectric sensor, the ratio was relatively high, reaching a maximum of 1.56. The ratio of the dual-wavelength scattered light signal was significantly higher than that of the experimental ratios of being overcharged and being over-discharged, indicating that the smoke particles generated during the puncture had a larger size.

5. Conclusion

To investigate the different thermal runaway phenomena of lithium-ion batteries under conditions of overcharging, over-discharging, and external puncture, this article designs corresponding experimental simulations. The fusion of dual-wavelength photoelectric and electrochemical detection is adopted to determine the thermal runaway parameters of lithium-ion batteries, and the feasibility of this fusion detection technology is verified. The experimental results satisfy the expected outcomes. During the overcharging-induced thermal runaway experiment, the detection system responds, verifies, and warns 36 seconds in advance. In the over-discharging experiment, the amount of generated gas is approximately one-fifth that of the overcharging experiment and one-quarter that of the puncture experiment, with no false alarms. During the puncture experiment, the blue/infrared signal response ratio of the detection system ranges from 1.06 to 1.56, compared to 0.82 to 0.89 in the overcharging experiment and 0.99 to 1.06 in the over-discharging experiment, indicating larger smoke particle diameters and no false alarms. The research results indicate that the fusion detection technology provides a new method for thermal runaway warning of lithium-ion batteries with characteristics such as fast response, high reliability, and a wide application range. Additionally, this technology can preliminarily analyze the causes of thermal runaway based on gas concentration characteristics, signal response amplitude, and blue/infrared response ratio, providing a reference for subsequent fault troubleshooting, inspection, and maintenance, saving costs, and having high application development and value for complex environmental detection.

Nomenclature

d

Particle diameter

m

Refractive index

λ

Incident light wavelength

f(d)

Particle size distribution function

C_n

Aerosol mass concentration

P_S

Scattered power at short wave

P_L

Scattered power at long wave

R

Light scattering power ratio

t

Relative Time

Abbreviations

CO

Carbon monoxide

EX

Combustible gases

H₂

Hydrogen

CO₂

Carbon Dioxide

NCA

Ternary Composites (Ni,CO,Ai)

LED

Light emitting diode

SOC

State of Charge

CRediT Authorship Contribution Statement

Wen Li: Conceptualization (Lead), Methodology (Lead)

Hao Zhou: Validation (Lead), Writing – original draft (Lead)

XueKe Luo: Project administration (Lead)

BinBin Lyu: Methodology (Lead), Software (Lead)

SiJia Hao: Writing – review & editing (Supporting)

Conflict of Interest

The authors declare no conflict of interest in the manuscript.

Funding

National Natural Science Foundation of China: No. 51205005

Shanghai Engineering Technology Research Center: No. PXM2017-014212-000013

References

• 1)  K. C. Abbott, J. E. H. Buston, J. Gill, S. L. Goddard, D. Howard, G. Howard, E. Read, and R. C. E. Williams, J. Power Sources, 539, 231585 (2022).
• 2)  Y. Zhu, M. Y. Liu, L. Wang, and J. X. Wang, Sustainability, 14, 7048 (2022).
• 3)  D. S. Ren, X. N. Feng, L. S. Liu, H. J. Hsu, L. G. Lu, L. Wang, X. M. He, and M. G. Ouyang, Energy Storage Materials, 34, 563 (2021).
• 4)  W. S. Huang, X. N. Feng, X. B. Han, and W. F. Zhang, Cell Rep. Phys. Sci., 2, 100285 (2021).
• 5)  L. Z. Jia, D. Wang, T. Yin, X. C. Li, L. W. Li, Z. Q. Dai, and L. L. Zheng, ACS Omega, 7, 14562 (2022).
• 6)  X. Lai, J. Yao, C. Y. Jin, X. N. Feng, H. B. Wang, C. S. Xu, and Y. J. Zheng, Batteries, 8, 248 (2022).
• 7)  Q. Yuan, X. P. Chen, K. P. Meng, P. X. Wang, L. Tang, T. Tang, J. Cao, and Y. B. Wu, J. Electrochem. Energy Convers. Storage, 19, 030911 (2022).
• 8)  B. H. Liu, Y. K. Jia, J. Li, S. Yin, C. H. Yuan, Z. H. Hu, L. B. Wang, Y. X. Li, and J. Xu, J. Mater. Chem. A, 6, 21475 (2018).
• 9)  B. H. Liu, Y. K. Jia, C. H. Yuan, L. B. Wang, X. Gao, S. Yin, and J. Xu, Energy Storage Materials, 24, 85 (2020).
• 10)  Y. L. Zhu, C. J. Wang, F. Gao, M. X. Shan, P. L. Zhao, Q. F. Meng, and Q. B. Wu, J. Energy Storage, 38, 102571 (2021).
• 11)  Q. P. Zhang, P. Z. Li, C. H. Liu, F. L. Wei, M. Wang, J. X. Li, S. H. Zhu, G. S. Shao, and J. Mao, J. Electrochem. Soc., 169, 060502 (2022).
• 12)  L. Feng, L. H. Jiang, J. L. Liu, Z. Y. Wang, Z. S. Wei, and Q. S. Wang, J. Power Sources, 507, 230262 (2021).
• 13)  Q. Y. Liu, Q. Zhu, W. T. Zhu, X. Y. Yi, and X. Han, Electrochemistry, 90, 087004 (2022).
• 14)  Z. H. Liao and G. Q. Zhang, Advanced Technology of Electrical Engineering and Energy, 38(10), 61 (2019).
• 15)  W. Li, B. B. Lyu, M. G. Xu, and S. J. Hao, Optoelectron. Lett., 18, 502 (2022).
• 16)  C. Zhong, C. Sun, Y. C. Zhang, and H. J. Ding, Proc. SPIE, 11606, 1160611 (2020).
• 17)  H. Y. Zuo and J. G. Yang, Acta Phys. Sin., 56, 6132 (2007).
• 18)  T. Deng, S. Wang, and M. Zhu, Sens. Actuators, B, 236, 334 (2016).
• 19)  W. Shu, D. Tian, D. Zheng, and Z. Ming, Fire Saf. Sci., 25(4), 224 (2016).
• 20)  W. F. Li, H. W. Wang, Y. J. Zhang, and M. G. Ouyang, J. Energy Storage, 24, 100775 (2019).
• 21)  A. W. Golubkov, S. Scheikl, R. Planteu, G. Voitic, H. Wiltsche, C. Stangl, G. Fauler, A. Thaler, and V. Hacker, RSC Adv., 5, 57171 (2015).
• 22)  Z. Z. Jia, P. Qin, Z. Li, Z. S. Wei, K. Q. Jin, L. H. Jiang, and Q. S. Wang, J. Energy Storage, 50, 104302 (2022).