Lithium-ion batteries (LIBs) have emerged as key devices for new energy storage and conversion, with large-scale applications in new energy vehicles, energy storage power plants, aerospace, and other fields due to their advantages of high power density, high energy density, long cycle life, and low self-discharge rate. However, with the widespread commercial use of LIBs in recent years, there has been a steady increase in the incidence of explosions, fires in energy storage power plants, and spontaneous combustion in new energy vehicles. Most incidents occur during the charging and charging rest periods of LIBs. At these time, the safety and dependability of LIBs during charging and discharging operations have become significant barriers to their continued development. Most technologies face difficulties in measuring in-situ battery conditions due to the unique internal environment. Micro-electro-mechanical system (MEMS) fiber-optic sensors offer advantages such as inherent safety, resistance to electromagnetic interference, electrolyte corrosion resistance, high measurement precision, and the potential for mass manufacturing. These advantages enable real-time, in-situ, and accurate monitoring of battery state parameters.

MEMS microcavities are produced utilizing Au-Au bonding technology to minimize residual pressure inside the microcavities and to mitigate the influence of mismatches in coefficients of thermal expansion across various material surfaces. A Fabry-Perot (F-P) interferometer for high-sensitivity pressure measurement is created by fusing a MEMS diaphragm to the end of an optical fiber. When external pressure is applied to the MEMS diaphragm, the F-P cavity length of the fiber-optic sensor changes, which results in a shift in the optical range difference. Pressure can be measured by demodulating the sensor’s optical range difference with a polarized low-coherence interferometric demodulation device.

The MEMS fiber-optic sensors are placed in a closed pressure tank inside a temperature chamber. Sensor performance tests are conducted under constant temperature and variable pressure conditions, ranging from 40 to 280 kPa at 40 kPa intervals. Subsequently, the temperature environment varies from -40 to 60 ℃ at 20 ℃ intervals, and the full-scale pressure is measured at each constant temperature. The absolute phase values of each cosine Gaussian signal are obtained using a monochromatic frequency domain demodulation algorithm to determine the MEMS fiber-optic sensors’ temperature and pressure response characteristics. The F-P cavity length of the MEMS fiber-optic sensors shows a highly linear relationship with the external pressure, and the pressure-temperature cross-sensitivity is as low as 0.091 kPa/℃. The sensor has a pressure measurement error of only 0.019 rad at 20 ℃, with an accuracy of 1.7×10^-4f_FS, f_FS is the full scale of the sensor. Although the sensor’s accuracy decreases as the temperature deviates from room temperature, it still maintains a pressure accuracy of 8.6×10^-4f_FS at 60 ℃, which provides a solid foundation for capturing the detailed state characteristics of LIBs under actual operating conditions. The battery in-situ monitoring experimental system is shown in Fig. 7, where the battery test system provides the corresponding current and voltage to the battery, and the optical information from the MEMS fiber-optic sensor is collected in real-time by the optical signal demodulator and demodulated by the host computer. In this experiment, the LIBs are charged and discharged for 40 cycles at a rate of 1 C, during which their state characteristics change, as shown in Fig. 8(a). The currents and voltages of the LIBs, along with their internal pressures, exhibit a stable and reproducible correlation, and the battery’s state of charge determines the cyclic pressure inside the battery. As the number of charge/discharge cycles increases, the battery pressure baseline gradually rises, and the pressure rate increases steadily. Similarly, the peak battery cycle value grows slowly, which shows that the reversible pressure change of the battery remains almost constant from cycle to cycle. The internal pressure of the battery can be divided into reversible pressure due to the “breathing effect” of the battery, and irreversible pressure, caused by the accumulation of trace gases produced by the battery’s side reactions. As the number of battery cycles increases, the battery ages and its capacity decreases. Therefore, tracking and monitoring the gas accumulation inside the battery allows real-time observation of the battery’s health status and cycling performance.

We propose an MEMS fiber-optic F-P sensor based on Au-Au thermo-compression bonding to effectively achieve in-situ state monitoring of LIBs under actual operating conditions. After performance testing, the sensor demonstrates a pressure accuracy better than 8.6×10^-4f_FS over a wide temperature range of -40 to 60 ℃, a cross-sensitivity as low as 0.091 kPa/℃, and a consistency as high as 99.145%. It survives the battery environment, which is subject to strong redox reactions, and measures the internal pressure of LIBs in real-time and accurately over long cycling periods. A stable and reproducible correlation between the internal pressure of the cells and the electrochemical signals is observed over 40 charge/discharge cycles of LIBs. By extracting the pressure value at the end of each cycle and establishing a baseline for the pressure cycle, it is found that the internal pressure of the battery can be divided into reversible changes due to the battery charging/discharging “breathing effect” and irreversible changes due to trace gas production from side reactions. As the number of battery cycles increases, the battery capacity gradually decreases, and the internal pressure baseline increases. This MEMS fiber-optic sensor provides an effective tool for in-situ monitoring of batteries, offering valuable insights into the internal electrochemical reactions in LIBs and helping improve battery performance and safety.