Tracking Li-Ion Batteries Using Fiber Optic Sensors

Abstract

Batteries are being seen as a key technology for battling CO2 emissions from the transport, power, and industry sectors. However, to reach the sustainability goals, they must exhibit ultrahigh performance beyond their capabilities today. So, it is becoming crucial to develop advanced diagnostic/prognostic tools injected into the battery that could nonintrusively track in time and space its physical and chemical parameters, for ensuring a greater lifetime and therefore lower its CO2 footprint. In this context, a smart battery sensing system with high performance and easy implementation is critically needed for the vital importance of safety and reliability in all batteries. Parameters like temperature (heat flow), strain, pressure, electrochemical events from electrode lithiation to gassing production, refractive index, and SoX battery indicators are of high importance to monitor. Recently, optical fiber sensors (OFS) have shown to be a feasible, accurate, and useful tool to perform this sensing, due to their intrinsic advantages and capabilities (lower invasiveness, multipoint and multiparameter detection, capability of multiplexing being embedded in harsh environments, and fast response). This chapter presents and discusses the studies published regarding the different types of OFS, which were developed to track several critical key parameters in Li-ion batteries, since the first study was reported in 2013.

1. Introduction

According to recent COP21, COP25, COP26 Conferences, and EU2030 targets, there is a need for significant reductions in CO₂ and greenhouse gas emissions in a short span period, targeting the reduction of climate warming in 1.5–2.0°C up to 2030 [1]. With the worldwide acceptance of electric vehicles together with the new era of connected objects, ensuring battery reliability, lifetime, and sustainability is becoming a necessity [2]. In this way, batteries are currently seen as important technological enablers to drive the transition toward a decarbonized society. They have recently achieved considerable improvements in terms of technical performance and economic affordability [3]. However, for a successful mass introduction of electrified mobility, renewable and clean energy systems with market competitive performances, fast charging capability, and substantial improvements in battery technologies (autonomy and safety) are required [4, 5].

Currently, to guarantee safe operation, a battery management system (BMS) only measures externally accessible parameters such as voltage, current, and temperature. The scarcity of information regarding the interior of the cell currently hinders the improvement of the accuracy and predicting capabilities of current BMS algorithms and models, while equally limiting attempts to refine the battery thermal design due to the absence of heat-transfer information. This has led to increasing interest in spatiotemporal imaging of the thermal flows within a cell using temperature sensors [6, 7, 8, 9, 10, 11, 12]. Typically, they are used in electronic sensing devices, such as thermocouples (TCs) [13, 14], thermistors [15], IR thermography [16], and resistance temperature detector (RTDs) [17]. However, in addition to short resolution and accuracy, huge measurement setup, or higher volume/size preventing them from being inserted in a cell, they cannot be appropriate to be embedded in batteries due to their electrochemical harsh environment.

Furthermore, batteries are breathing objects that expand and contract upon cycling, with volume changes that can reach up to 10%. These changes, together with the electrode volume expansion associated with the solid electrolyte interface (SEI) growth, lead to important mechanical stress inside the battery materials (like cracks) that are harmful to their performances. Methods, to sense intercalation strain and pressure, are equally critical to control the SEI dynamics affecting their states of charge (SoC) and health (SoH). The methods already used are not acceptable: strain-gauges fall short of providing spatial information and cannot also be embedded to internally sense battery cells [8, 18].

Alternative solutions, due to their full advantages, such as greater precision, multiplexing, immunity to electromagnetic interference, chemical inertness, small size/low invasiveness, and a possibility to be tailored regarding their dimensions and sensitivities, are sensors based on optical fiber technology [2, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59]. Since the first study developed by Pinto et al. in 2013 [19], OFS starts to be integrated into lithium-ion batteries (LIBs) to monitor critical key parameters, such as temperature and/or thermal gradients, strain, gases, pressure, electrochemical events (chemical changes and lithiation), refractive index, and the states of charge, discharge, health, power, energy, and safety (SoX) battery indicators (Figure 1).

Figure 1.

Critical key parameters identified to be tracked in LIBs.

Performing a systematic review, we used two databases to retrieve scientific publications: Web of Science (www.webofknowledge.com, accessed on 16 November 2021) and Scopus (www.scopus.com, accessed on 16 November 2021). A comprehensive search on the use of OFS to monitor LIBs was performed based on a query by topic (title, abstract, and keywords) of the terms: ((optical AND fiber AND sensors*) AND (lithium AND batteries*)); spanning over the years 2013 to November 2021. The search query resulted in a total of 60 papers that were subsequently reviewed by the authors, of which 40 were considered eligible for the present work.

Figure 2A summarizes the number of studies published by year, since the first paper in 2013, regarding the use of OFSs to track LIBs parameters. From a critical analysis, an increase of publications can be observed from the beginning, however, with a lower number in 2020, probably due to the pandemic world situation. In Figure 2B, it is also presented an illustration of the critical parameters tracked in the LIBs. Temperature and strain were the parameters more studied followed by the correlation of the optical fiber signals with the electrochemical events and SoX battery indicators. The tracking of gasing, refractive index, and pressure variations are very recent topics of sensing inside the LIBs. However, due to the difficulty and complexity of sensing, the integration of the OFS inside the battery cells being necessary, they were not yet so explored. In this way, this chapter provides a complete overview of all studies published from 2013 to the present on the use of OFS to track critical key parameters in LIBs. Section 2 describes the theoretical approaches of the OFS used (fiber Bragg grating, interferometric, and evanescent wave sensors) to monitor the critical parameters. In section 3, all critical parameters (temperature, strain, SoX battery indicators, and electrochemical events) tracked so far using fiber optic sensing technology are presented and fully described.

Figure 2.

(A) Statistical summary of the number of papers by year, published since the first study, regarding the using of OFS in LIBs. (B) Percentual distribution of the critical parameters tracked by the OFS in the LIBs.

2. Optical fiber sensors: theoretical approaches

Manfred Börner, a German physicist, developed, in 1965, the first fiber optic patent related to a working fiber-optic data transmission system [60, 61]. Years later, in 1978, the concept of wavelength division multiplexing, where several optic signal chargers are multiplexed into a single optical fiber through different wavelengths was firstly published [62]. Since then, the optical fiber community has expanded and the use of optical fibers as sensing elements attracted a lot of attention. Figure 3 summarizes the different types of fiber optic sensors developed in the last years [63]. Regarding the monitorization of physical and chemical parameters in LIBs, just some of them were already tested. The fiber Bragg grating sensors (FBG) and tilted FBG sensors (TFBGs) were used to track temperature, strain, refractive index, and SoX, inside and outside of batteries, correlating these signals with electrochemical events during their operation. From the interferometric sensors, Fabry-Perot interferometric (FPI) and Mach-Zehnder interferometric (MZI) sensors were tested to monitor and decouple temperature, strain, and SoC signals. Evanescent wave sensors based on surface plasmon resonance and evanescent field fluorescence were also already used to monitor temperature shifts and SoC values in batteries. OFS based on Rayleigh scattering distributed sensing was also already used. However, due to their instrumental complexity, elevated interrogation costs, and low experimental use relative to the other methods, this type of sensor will not be approached in this chapter.

Figure 3.

Different types of OFS used to track critical parameters in LIBs.

2.1 Fiber Bragg grating sensors

The first FBG, fabricated using a visible laser propagating along with the fiber core, was proposed by Ken Hill in 1978 [64]. OFS based on FBGs has been widely applied in the measurement of physical, chemical, biomedical, and electrical parameters, especially for structural health monitoring in civil infrastructures, aerospace, energy, and healthy areas [65].

Classically, an FBG sensor consists of a small segment of a single-mode optical fiber (with a length of a few millimeters) with a photoinduced periodically modulated index of refraction in the fiber core. The FBG resonant wavelength is related to the effective refractive index of the core mode (n_eff) and the grating period (Λ). When the grating is illuminated with a broadband optical source, the reflected power spectrum presents a peak (with a full width at half maximum of a few nanometers), which is produced by the interference of light with the planes of the grating and can be described through Eq. (1) [66].

$${\lambda }_B=2n_{\mathit{eff}}\mathrm{\Lambda }$$E1

where λ_B is the so-called Bragg wavelength. When the fiber is exposed to external variations of a given measurand (such as strain, temperature, stress, or pressure, among others), both n_eff and Λ can be changed, causing an alteration in the Bragg wavelength, as shown in Figure 4 [66]. In addition to the common advantages of fiber sensors, this wavelength interrogation method offers robustness to noise and power oscillations and also enables wavelength division-multiplexing, by recording numerous FBGs with diverse grating periods in the same optical fiber (see Figure 5). This permits the monitorization of different spots in one structure/surface with only one sensor line, decreasing in this way, the total interrogation costs. The FBG sensitivity toward a given parameter is obtained simply by subjecting the sensor to pre-determined and controlled variations of such parameters and measuring the Bragg wavelength for each step.

Figure 4.

Scheme and operation mechanism of an FBG sensor to external strain and temperature perturbations.

Figure 5.

A) FBGs network inscribed in the same optical fiber. B) Optical spectrum of a network of 10 FBGs inscribed in the same fiber, where different wavelength peaks can be observed.

In the case of a linear response, the sensitivity (k) is given by the slope of the linear fit obtained from the experimental data. The effects of temperature are accounted for in the Bragg wavelength shift by differentiating Eq. (2),

$$\mathrm{\Delta }\lambda =2{\lambda }_B\left(\frac{1}{n_{\mathit{eff}}}\frac{\partial n_{\mathit{eff}}}{\partial T}+\frac{1}{\mathrm{\Lambda }}\frac{\mathrm{\partial \Lambda }}{\partial T}\right)\mathrm{\Delta }T={\lambda }_B\left(\alpha +\xi \right)\mathrm{\Delta }T=k_T\mathrm{\Delta }T,$$E2

where α and ξ are the thermal expansion and thermo-optic coefficient of the optical fiber material, respectively. On the other hand, if the fiber is subjected to strain variations, its response can be determined by differentiating Eq. (3),

$$\mathrm{\Delta }\lambda ={\lambda }_B\left(\frac{1}{n_{\mathit{eff}}}\frac{\partial n_{\mathit{eff}}}{\partial \epsilon }+\frac{1}{\mathrm{\Lambda }}\frac{\mathrm{\partial \Lambda }}{\partial \epsilon }\right)\mathrm{\Delta }\epsilon ={\lambda }_B\left(1-p_e\right)\mathrm{\Delta }\epsilon =k_{\epsilon }\mathrm{\Delta }\epsilon ,$$E3

where p_e is the photoelastic constant of the fiber (∼0.22) and Δε is the applied strain. The strain variations can be determined using the equation Δε = ΔL/L where ΔL is the length variation and L is the fiber length over which strain is applied. On a single measurement of the Bragg wavelength shift, it is not possible to decouple the effect of variations in strain and temperature (for example). Normally, a reference is used for temperature measurement, by using another fiber strain-free or other FBGs and sensing heads that have different strain and temperature sensitivities. Different strategies are being used in the literature, such as FBGs recorded in different fiber thicknesses, FBGs recorded in special microstructured fibers, and FBGs cascaded with other optical fiber sensing configurations (FPI, MZI) [66, 67, 68]. The discrimination of both variables is performed, through the matrixial method by using all the sensitivities of both sensors to each variable. In this way, a sensitivity matrix (4) for simultaneous measurement of strain and temperature can be derived as:

$$\left[\begin{array}{c}∆\epsilon \\ ∆T\end{array}\right]=\frac{1}{D}\left[\begin{array}{cc}-k_{\mathit{FBG}{1}_T}& k_{\mathit{FBG}{2}_T}\\ k_{\mathit{FBG}{1}_{\epsilon }}& -k_{\mathit{FBG}{2}_{\epsilon }}\end{array}\right]\left[\begin{array}{c}\mathrm{\Delta }{\lambda }_{\mathit{FBG}1}\\ \mathrm{\Delta }{\lambda }_{\mathit{FBG}2}\end{array}\right],$$E4

where D = K_FBG1ε x K_FBG2T – K_FBG1T x K_FBG2ε is the determinant of the coefficient matrix, which must be nonzero for possible simultaneous measurement.

The Bragg gratings can be inscribed in an optical fiber core through side exposure; two main types of techniques can be implemented: interferometric and non-interferometric techniques. In the noninterferometric technique, the phase mask method is one of the most commonly used (see Figure 6A). Generally, it is associated with longer laser pulses (near the nanoseconds) in the ultraviolet (UV) region. The phase mask consists of a diffraction grating shaped by small depressions in a silica substrate, separated by a predefined period (phase mask pitch, Λ_PM), which will define the modulation pattern linked to the resonant Bragg wavelength of the fabricated FBG (see Figure 6B).

Figure 6.

A) Schematic representation of the phase mask inscription method using a pulsed laser. +1 and −1 indicate the laser beam diffraction orders used to inscribe the Bragg grating in the optical fiber core. B) Typical phase masks used on FBG sensors recording (from Ibsen®).

Depending on the incident angle of the laser beam on the phase mask surface, different diffraction orders will be predominantly transmitted: the pairs +1/0 or + 1/−1. To attain different wavelength peaks, phase masks with different Λ_PM can be used. Typically, when using a UV laser, a better inscription efficiency is expected for doped or hydrogenated optical fibers [67].

2.2 Tilted FBG sensors

Compared to the normal FBG sensors, TFBG sensors have a special configuration, which leads to enhanced sensitivity to the surrounding refractive index (SRI). Thus, this type of sensor has been employed in many parameters sensing, such as temperature, liquid level, RI, and relative humidity, in biochemical research. TFBGs are short-period gratings in which the modulation of the RI is purposely tilted concerning the longitudinal axis of the fiber, to improve the light coupling between the forward-propagating core mode and the backward-propagating cladding modes (see Figure 7) [68].

Figure 7.

Schematic diagram of a TFBG where Λ_g is the grating period and $\mathrm{\theta }$ is the tilt angle (adapted from [68]).

The wavelength of the coupled i-th cladding mode λ_cl(i) can be expressed as (Eq. 5):

$${\lambda }_{\mathit{cla}\left(i\right)}=\left(n_{\mathit{eff}}^{\mathit{\text{core}}}+n_{\mathit{eff}\left(i\right)}^{\mathit{cla}}\right)\mathrm{\Lambda }=\frac{\left(n_{\mathit{eff}}^{\mathit{\text{core}}}+n_{\mathit{eff}\left(i\right)}^{\mathit{cla}}\right){\mathrm{\Lambda }}_g}{\mathit{cos}\mathrm{\theta }}$$E5

where $n_{\mathit{eff}}^{\mathit{\text{core}}}$and $n_{\mathit{eff}\left(i\right)}^{\mathit{cla}}$are the effective RIs of the fiber core and i-th cladding mode, respectively. Λ and Λ_g are the grating periods along with the fiber longitudinal axis and perpendicular to the grating plane, respectively. The excited cladding modes are limited in the fiber cladding by total internal reflection on the cladding-surrounding medium interface. Each of the cladding modes propagates with a corresponding effective RI value. When the RI value of the surrounding medium spreads the one of a specific cladding mode, the cladding mode will be coupled out of the fiber cladding, resulting in a variation in the grating transmission spectrum. Therefore, the shifts of the SRI can be quantitatively tracked by detecting the variations in the grating transmission spectrum of the TFBG [69]. These TFBGs can also be fabricated in line with other normal FBGs to simultaneously decouple different parameters, such as RI and temperature, as they have different sensitivities.

2.3 Interferometric sensors

Since the first study, published in 1897 by Charles Fabry and Alfred Perot, about the FPI principle [70], the OFS based on this methodology was used in numerous applications, such as biological, chemical, and various physical parameters, including temperature, strain, pressure, and RI [63, 71]. Literature shows that these sensors are used also like candidates to improve the discrimination of strain and temperature in batteries [44]. An FPI sensor is performed by considering two parallel reflecting surfaces divided by a certain physical length of the cavity (L). FPI sensors can be classified as extrinsic or intrinsic, as can be seen in Figure 8a and b, respectively. The intrinsic FPI sensor has reflecting components inside the fiber itself [70]. In the extrinsic FPI, the air cavity is designed by an auxiliary structure. Due to the optical phase difference between two reflected signals, the reflection spectrum of an FPI can be defined as the wavelength-dependent intensity modulation of the incident signal spectrum. The phase difference of the FPI (δ_FP) can be given as (Eq. 6):

Figure 8.

a) Extrinsic FP sensor performed by forming an external air cavity, and b) intrinsic FP sensor formed by two reflecting components, R1 and R2 (adapted from [70]).

$${\delta }_{\mathit{FP}}=\frac{4\mathit{\pi nL}}{\lambda }$$E6

where n is the RI of the cavity material, and λ is the wavelength of the output signal. Consequently, an external perturbation to the FPI sensor (such as strain, temperature, or IR), will promote a length variation of the FPI cavity, resulting in wavelength changes. By tracking the wavelength shift of the spectrum, and after an experimental pre-calibration to each specific parameter, to determine their sensitivities, a linear conversion of the data signals of the measured parameter values can be performed by analyzing the spectrum produced.

Another type of interferometer is the MZI sensor. They are usually applied for sensing parameters such as temperature, strain, curvature, and RI, among others [71], due to their advantages of high RI sensitivity and flexible designs, as shown in Figure 9.

Figure 9.

Different types of MZIs configurations; using: (A) a pair of LPGs, (B) core misalliance, (C) air-hole collapsing of PCF, (D) MMF section, (E) small SMF core, and (F) tapering fiber regions (adapted form [71]).

An MZI is designed due to the formation of an optical path difference between the fundamental core mode and the higher-order cladding modes in optical fiber. Subsequently, in the interference spectrum, dips or peaks can appear [72]. These peaks or dips values are used as tracking signals because they change with external perturbations (such as temperature, strain, pressure, and RI). For simplicity and spectral data analysis, only the core mode (I₁) and one cladding mode (I₂) are considered. The transmitted interference signal, I, can be expressed as (Eq. 7) [73]:

$$I=I_1+I_2+2\sqrt{I_1{I}_2}cos\left(\varphi \right)$$E7

where $\varphi =\frac{2\pi \left(n_{\mathit{eff}}^{\mathit{\text{core}}}-n_{\mathit{eff}}^{\mathit{cla}}\right)L}{\lambda },$is the phase difference, being $n_{\mathit{eff}}^{\mathit{\text{core}}}$ and $n_{\mathit{eff}}^{\mathit{cla}}$ the effective RIs of the fiber core and cladding mode, respectively. The λ is the input optical wavelength in vacuum, L is the interferometric MZI length, and ϕ = 0 is the initial interference phase. When I₁ = I₂, the fringe visibility reaches its maximum value. From Eq. (7), when $\varphi =\frac{2\pi \left(n_{\mathit{eff}}^{\mathit{\text{core}}}-n_{\mathit{eff}}^{\mathit{cla}}\right)L}{{\lambda }_m}=\left(2m+1\right)\pi$, the output intensity dips will appear, where m is an integer. Specifically, the phase difference between two adjacent minimum intensity dips is $\frac{2\pi \mathrm{\Delta }{n}_{\mathit{eff}}L}{{\lambda }_m+1}-\frac{2\pi \mathrm{\Delta }{n}_{\mathit{eff}}L}{{\lambda }_m}=2\pi$. Therefore, the difference between two adjacent interference wavelengths, as well known as the free spectral range (FSR) can be calculated as $\mathit{FSR}={\lambda }_m-{\lambda }_{m+1}=\frac{{\lambda }_m{\lambda }_{m+1}}{\mathrm{\Delta }{n}_{\mathit{eff}}L}$, and the theoretical cavity length is $L=\frac{{\lambda }_m{\lambda }_{m+1}}{\mathrm{\Delta }{n}_{\mathit{eff}}\left({\lambda }_m-{\lambda }_{m+1}\right)}$. Using these formulas, the theoretical values of MZI length can be compared with experimental results to reduce errors, and also, it can be seen that when the L increases or the Δn_eff increases, the FSR decreases. Note that when Δn_eff changes, it indicates that RI of the external environment changes, promoted by the external parameters, while RI of the optical fiber core is constant. When the external parameters, such as temperature or RI, around the MZI is different, which will lead to the changes of Δn_eff, the wavelength of interference dip will also shift. So, the surrounding environment can be analyzed through spectra after a pre-calibration process to each external parameter to which the optical sensor will be submitted.

2.4 Evanescent wave sensors

Other types of OFS, which are being used to track specific parameters in Li-ion batteries, were the evanescent wave sensors. This type of sensor is created on the interaction of the evanescent field in the cladding with the fiber surroundings, resulting in fluctuations of the transmitted spectrum. It follows that they hold the capability of translating a discrepancy of the target analyte into optical signals so that they are widely applied to chemical and biosensing [74]. As shown in Figure 10, the evanescent field E_ew(d) decays exponentially as (Eq. 8):

Figure 10.

Design of a fiber evanescent wave spectroscopy sensor with the standing wave pattern and exponentially decaying evanescent wave.

$$E_{\mathit{ew}}\left(d\right)=E_{0}exp\left(\frac{-d}{d_p}\right)$$E8

where E₀ is the magnitude of the field at the fiber core-cladding interface, d is the distance from the core-cladding interface, and d_p is the distance where the evanescent field decreases to E₀/e and is described as the penetration depth which is given by (Eq. 9):

$$d_p=\frac{\lambda }{2\pi \sqrt{n_{\mathit{\text{core}}}^2{\mathit{sin}}^2\theta -n_{\mathit{cla}}^2}}$$E9

where λ is the wavelength of the incident light, θ is the angle of incidence at the fiber core-cladding interface, and n_core and n_cla are the RIs of the fiber core and cladding, respectively.

This optical fiber methodology of sensing can also be modified by depositing specific film materials (metal-dielectrics) in the fiber cladding surface and interacting between them. In this way, the surface plasmon resonance (SPR) technique can be used. The SPR is a collective oscillation of free electrons excited by light at the metal-dielectric interface. The electromagnetic field decays exponentially into both metal and dielectric, the propagation constant of SPR can be given as (Eq. 10):

$$k_{\mathit{sp}}=\frac{\omega }{c}\sqrt{\frac{{\epsilon }_m{\epsilon }_d}{{\epsilon }_m+{\epsilon }_d}}$$E10

where ω is the angular frequency of the incident light, c is the speed of light in space, and ε_m and ε_d are the dielectric constants of the metal and dielectric, respectively. The propagation constant of the evanescent wave parallel to the planar metal film surface can be expressed as $k_{\mathit{ew}}=\frac{\omega }{c}\sqrt{{\epsilon }_{\mathit{\text{fiber}}}}\mathit{\text{sinθ}}$, where ε_fiber is the dielectric constant of the fiber. The SPR occurs when both propagation constants are equal, it exhibits high sensitivity to even slight oscillations in the dielectric constant of the dielectric material. Therefore, SPR-based sensors can successfully track diverse variables due to the location of the resonance shifts with the varying the RI of the nearby dielectric.

3. Critical parameters tracked in LIBs by OFS

In the almost last 10 years, OFS was introduced and started to be used as a useful and precise tool to monitor critical key parameters inside and outside the LIB. Significant advantages instead of other sensing technologies can be reported comparatively with the electronic technology, in the monitorization of temperature (TCs, RTDs,) and strain (strain-gauges), regarding the low intrusiveness (fibers thickness of ∼125 microns), higher capacity of multipoint and multiparameter discrimination, and essentially the capability of to be embedded in their harsh electrochemical environment, by tracking in loco and real-time specific parameters with good accuracy and reliability, and without damage the batteries performances. Until now, it was almost impossible to access the internal behavior of these batteries in operation, to know how they behave in terms of physical and chemical performance.

This optical fiber technology presented above has been successfully integrated into battery sensing, allowing their smart sensing of LIB safety aspects, such as temperature and/or thermal gradients, strain, RI, and pressure variations, internal gassing evolution, electrochemical events (chemical reactions, and SEI composition), and their correlation with SoX battery indicators. This section divides and reports all studies presented in the literature since the first paper reported by Yang et al. in 2013, regarding the use of OFS to monitor all these parameters in LIB.

3.1 Temperature tracking

Of all safety problems in LIBs, thermal runaway is a vital issue, which is reproduced by the fast increase of temperature. This rise produces heat energy at a rate faster than heat can be dissipated followed by a failure of the LIB internal separator components, resulting in local short circuits and critical situations, to their explosion [75]. Moreover, accumulated heat in the batteries takes worries of performance drops and safety risks. Temperature can affect the LIBs lifetime and energy, and therefore, it should be within an ideal range of temperature, to ensure better performance and long life, both for use and storage [17]. The ability to quantify and evaluate the mechanism of thermal runaway generated during the electrochemical processes which happen will create beneficial information regarding their behavior, as well as an active tool to promote their safety [76, 77, 78].

Typically, in the real context of the LIB, this parameter is monitored through external electronic devices, such as TCs and RTDs, by detecting just single points on their surface. Optical fiber sensing technology was used as an alternative method to realize multipoint external and internal temperature measurements on LIBs, during their operation, also performing in different types of LIBs, thermal gradients characterizations, and evolutions. Of the different types of OFS, the FBGs were the mainly used due to their inherent advantage of multipoint monitoring and fast response time.

In 2013, Yang et al. [19], integrated by the first time, FBGs in a coin LIB to measure real-time temperature changes during the battery’s operation under normal and abnormal conditions. The FBG sensors exhibited good thermal responses to dynamic loading when compared with the TCs. Novais et al. [26], in 2016, presented the integration of four FBGs in lithium-ion cells for in situ and in-operando temperature monitoring during galvanostatic cycling at C-rates ranging from 1 to 8°C (Figure 11A). In the internal FBG locations, the fiber was covered by an exterior silica tube to free the sensors of any external stress promoted by the surrounding materials and, in this way, only detect temperature variations.

Figure 11.

A) Schematic diagram of internal and external FBG sensors positions [26]. B) Schematic of a LiB test setup with the location of TCs and FBGs. Copyrights 2016 and 2017.

The internal sensors registered higher temperature variations at 8°C and in the center of the active area of the one-layer pouch cell. The authors concluded that the low invasiveness and high tolerance to the chemically aggressive environment make them a motivating option for integration into the LIBs. This study also contributes to the detection of a temperature gradient in real-time inside an LIB, thanks to the different locations of the sensors inside the battery. Nascimento et al., one year later [30], attached FBGs, all recorded in the same single fiber and TCs on a commercial LIB surface, to perform a comparative study between their signal responses (Figure 11B). The response rates were 4.88°C/min and 4.10°C/min for the FBG and TC, respectively. The results also demonstrate that the FBGs were able to sense temperature fluctuations with a ∼ 1.2 times higher response rate than the K-type TCs. The rise time obtained for the FBG was 28.2% lower than the TC, making the FBGs a better choice for the real-time temperature tracking on a LIB.

In 2017 and 2019, Nascimento et al. [29, 43] has developed two studies about the thermal distribution on a surface of a prismatic LIB by a network of five FBGs in a single fiber, to assess in real time and operation, the impact of different environmental conditions, temperature, and relative humidity, on batteries performance. These studies provided a real-time thermal mapping to elucidate which areas of the battery needed to be cooled faster when it was exposed to dry, temperate, and cold climates. Faster variations of voltage usually translated in higher temperature variations at the LiB surface, and this effect is evidenced when the LiB operates under abnormal conditions. After a pre-calibration step, the FBGs were calibrated to convert the wavelength shift peak to the correspondent temperature values based on their calculated sensitivity. These temperature values are tracked by following the FBGs peaks in the spectrum response. Complete temperature values of 30.0 ± 0.1°C, 53.0 ± 0.1°C, and 65.0 ± 0.1°C were achieved on the top location (near electrodes) during the higher discharge rate, when exposed to the cold, temperate, and dry climates, respectively. The higher temperature shifts detected by the optical sensors in the temperate and dry environments are related to the superior performance of the LiB in terms of discharge capacity and power capabilities. This study demonstrates also which are the best environmental conditions to run the LiBs, in order to extend their lifetime and safety, and is also helpful for the next generation of batteries, showing which areas require faster cooling to reduce accumulated heat.

Bhagat’s group, in 2018, performed three studies by embedding FBGs, in cylindrical LIBs, to monitor in situ and in-operando temperature variations. The sensors were resistant to the strain imposed during the battery instrumentation procedure and their harsh chemical environment. The results presented a temperature difference between the core and the can temperatures (monitored by K-type TC) of up to 6.0°C during the discharging process, while a temperature difference of 3.0°C was obtained during the charging process. The zones nearer to the anode presented a higher temperature during discharge while the location closer to the cathode performed higher temperature values during charge [38]. The authors demonstrate that FBGs produce reliable core temperature data, while their small mechanical profile allows for a low-impact instrumentation method [41, 42].

Nascimento et al., in 2018, proposed a network of 36 FBGs for real time, in situ, and operando multipoint monitoring of the surface temperature distribution on a pack of three prismatic LIBs, performing a spatial and temporal thermal mapping of all pack interfaces (Figure 12). In total, four optical fibers were used to monitor all locations of the LIB pack. The results show that in general, a thermal gradient is identified from the top to the bottom surface locations. Due to the higher current density of the Li⁺ ions near the positive tab collector, the presence of hot spots between two of the three batteries was identified [36].

Figure 12.

Surface LIB pack thermal mapping performed by a network of 36 FBGs, during discharging at 1.4C [36]. Copyright 2018.

Peng et al., in 2020 [53] and 2021 [55] proposed an OFS to monitor temperature variations in the external LIB electrodes, during cycling tests. The sensor consists of a metal ring and an FBG. The FBGs were gloved on the external electrodes, and PT100 sensors were also attached to the electrodes as a comparison measurement. The FBGs calibration test presents good linearity and high sensitivity. From the results, during all the cycles, the sensors placed on the positive electrodes recorded higher temperature variations instead of those on the negative electrodes. Even this year, Alcock et al. developed an accessible method to attach FBGs on cylindrical LIB surfaces to in situ thermal sensing. This study differs from the others by using a “guide-tube” to decouple the temperature and strain variations on the LIB surface [54]. Recently, Li et al. developed an optical fiber temperature sensor for battery temperature monitoring based on fluorescence intensity ratio technology [58]. In this study, β-NaYF₄:Er³⁺/Yb³⁺@NaYF₄ nanoparticles were used to design the optical sensor in the fiber tip. After the fiber functionalization, this sensor was preliminary calibrated to temperature in function of their fluorescence intensity response to correctly convert their spectral response in temperature values during battery operation. The maximum relative sensitivity obtained by the optical sensor was 1.62% K⁻¹ at 293 K, the temperature detection limit was within ±0.5 K, and high-temperature changes were registered under a higher discharge rate.

3.2 Strain tracking

Along with the cycling processes of LIBs, strain evolution is also an important parameter to be tracked in order to identify possible cracks in their internal materials or the occurrence of some swelling in case of a wrong operation through a gasification production. In this way, OFS has also been recently used to monitor this parameter. From all studies reported so far, the FBGs were the sensors selected to perform this sensing. Li-ion pouch cell configuration is the most used in tests while coin cell configuration is only employed to demonstrate the preliminary tests.

In 2016, Bae et al. [27] developed two approaches to track strain and stress evolution in the graphite anode of a Li-ion pouch cell using FBGs. In one approach, the optical sensor was attached between the graphite anode and the separator, while in the other implanted approach, the sensor was embedded totally within the anode material. Measurements of strain and stress states of the graphite anode were run over cycling tests. Reproducible peak shifting in both attached and embedded FBGs was observed at different states of charge and discharge. Specifically, an embedded sensor that is completely surrounded by graphite particles simultaneously suffers accumulated longitudinal, as well as transverse strains associated with the expansion or contraction of the negative electrode. Additionally, the embedded FBG showed 3× higher sensitivity than the attached FBG sensor at 100% SoC. The process to detect and convert the FBG wavelength peaks to strain measurements is the same as used for temperature monitoring, through the strain sensitivities of each sensor and using free FBGs just to decouple temperature variations.

Peng et al., in 2019, have reported two papers regarding an external and novel strain sensor based on FBGs for LIBs [45, 47]. The structure of the strain sensor consisted of two FBGs, a sensitization structure and a protective cover, which contained two symmetrical lever mechanisms and an installation platform, in which the rotating pairs of levers were replaced by flexure hinges. Enhanced strain sensitivity of 11.55 pm/με was obtained, with good linearity and repeatability. From the cycling tests, the drift in strain is analogous to different C-rate charge-discharge cycles. The strain rises evidently close to the end of discharge with an evolution in the C-rate. However, the proposed sensor cannot be embedded inside the LIBs due to their bulky structure, providing higher invasiveness.

Rente et al., in 2021, reported the tracking of strain shifts, also through FBGs, on a surface of cylindrical LIB, under cycling tests [57]. In this study, a simple machine-learning algorithm based on dynamic time warping (DTW) was used to estimate the SoC of representative LIBs. The FBG data obtained were shown to be reliable and sufficiently reproducible to serve as the input for the DTW algorithm used. The use of a model train has proved to be very effective as a proof-of-concept study for future BMS, especially in electrical vehicles.

3.3 SoX battery indicators tracking

The SoX battery indicators are crucial factors reflecting the state of batteries, in which they are commonly estimated under the assistance of the evanescent wave sensors in LIBs. Additionally, the FBGs are combined with them to improve the sensing performance and used as parameters discrimination. In 2015, an integrated OFS technology for monitoring charge steps in LIB cells was studied by Alemohammad et al. [23]. The sensor consists of an optical fiber encapsulated inside a LIB with direct interaction with the cell electrochemical environment. The sensor operates on the basis of the changes in the optical properties of the LIB cell electrodes, that is, variations in optical absorption and reflection at different charge levels, that will change the spectral response in terms of wavelength and also optical power losses, showing the SoC in battery and providing information about aging and stabilization following charge/discharge cycles.

Ghannoum et al., reported in 2016, a reflectance study of commercial graphite anodes in LIB and the optical fiber evanescent wave spectroscopy of electrochemically lithiated graphite [28]. A substantial rise in the reflectance of the lithiated graphite in the near-IR band (750–900 nm) as a function of SoC and similar SoC tendency in the transmittance when the fiber was embedded in the battery was observed. The same authors, one year later, developed the fabrication and integration of the OFS, by using similar sensing technology, into cylindrical LIBs as well as a Li-ion pouch cell [31]. The sensitivity of the sensor increased along with increasing the contact area of the sensor within the graphite anode and the optical fiber evanescent wave sensor integrated into the graphite anode demonstrated the potential use to track the both SoC and SoH of LIB, by correlating the optical data with the voltage and current signals of the LIBs.

Lao et al., in 2018, designed an innovative method, named TFBG-based SPR sensor, for in situ tracking the SoC of supercapacitors, for the first time [39]. This new configuration is based on a 50-nm-thick gold layer of high surface quality deposited on the TFBG. The FBG recording angle was 18° and an additional gold coating was deposited on the fiber end to achieve a single-ended sensor with interrogation in the reflection scheme (Figure 13A). The proposed plasmonic TFBG sensor was attached to one of the electrodes of the supercapacitor to monitor the electrochemical activity. The charge density and SoC measurement were demonstrated, and the results showed that the spectral response of the SPR mode of the TFBG was directly related to the charge density and the SoC of the supercapacitor. Basically, the wavelength peak of the TFBG and the SPR mode changes with the SoC level at which the supercapacitor is. Then, the variations of the charge density and the SoC during the cycling steps could be tracked by following the shifts of the place and the intensity of the reflection spectrum.

Figure 13.

A) Electrochemical SPR sensing principle and experimental demonstration with a gold-coated TFBG sensor [39]. B) OFS principle. Multicolored light is guided through an optical fiber embedded in the electrode [59]. Copyright 2018, 2021.

Modrzynski et al. [46] presented an SoC measurement technique based on an optical fiber sensing system, in 2019. In this system, two optical fibers were etched to increase the interaction between the light propagation inside the fiber core with the surrounding fiber environment, detecting in this way RI changes in real time. The fibers were integrated into both graphite anode and lithium iron phosphate with the addition of indium tin oxide cathode of a Li-ion pouch cell. The SoC was monitored in real time by simultaneously detecting the light transmission through both fibers. The results showed that the SoC correlated transmission behaved equally for both electrodes. However, diverse relaxation and wavelength-dependent behaviors were identified during the charge and discharge cycling steps. The study proved that the OFS process was able to estimate the SoC independent of the electrical measurement methods.

In 2020, Hedman et al. used an OFS based on evanescent waves for monitoring the charge/discharge cycles of lithium iron phosphate batteries in real time [49]. The sensor is fully embedded within the positive electrode in a customized Swagelok cell in both a reflection- and transmission-based OFS configuration. Both constant current cycling and cyclic voltammetry were employed to associate the optical spectrum response with the cycling processes of LIBs. From the results, the optical signal correlates well with the SoC in the positive electrode in real time, and it is reproducible over various cycles. Furthermore, the optical signal detected does not rely on other usually estimated parameters in SoC estimation, such as current, voltage, and temperature. Rittweger et al., in 2021, present measurement results based on transmitted light intensities through the optical fiber as an indicator for the SoC (Figure 13B) [59]. The work also purposes to present an explanation of how to use the measured transmission intensity to decrease cross effects, such as temperature, pressure, or aging LIBs parameters. For that, a referencing methodology based on transmission intensities from light with different wavelengths is approached. Due to the reduced fiber cladding by a preliminary etching process, the light interacts with electrode material surrounding the fiber. So, transmission losses can be sensed, which depend on the lithium concentration in the electrode. From the results, the calculated transmission ratios are in good agreement with the SoC for various C-rates.

3.4 Electrochemical events tracking

Electrochemical events, such as gassing production, electrode lithiation, and chemical changes of the electrolyte, are fundamental issues that enable the battery manufacturers to identify degradation mechanisms that currently limit the lifetime and capacity of these energy-storage systems.

In 2014, Lochbaum et al. measured the evolution of gaseous CO₂ inside lithium-ion pouch cells during overcharge tests with optical fiber colorimetric sensors (the chemical sensing fiber used comprises a silica core surrounded by a fiber cladding, which is permeable to the chemical to be detected (analyte) and functionalized such that it changes its optical characteristics with analyte concentration) to examine the dynamics of electrolyte decomposition reactions [20]. For the ratiometric read-out principle used, the averaged intensity between 570 nm and 600 nm (CO₂-sensitive band) was normalized by the averaged intensity between 800 nm and 820 nm (CO₂-insensitive band). The results indicate a nonreversible gas evolution inside the LIBs during overcharge, in which the beginning of gas evolution is delayed in time relative to the overcharge condition.

Ghannoum et al., in 2017, presented the application of an innovative optical fiber-based sensing system for the lithiation of graphite within a lithium-ion pouch cell in real-time using a narrow-band spectrum concentrated around 850 nm [31]. For that, a polymer optical fiber was used and etched for the fiber core to directly interact with the surrounding materials. The main results show that the sensor signal can be correlated with the lithiation of graphite anode over multiple full and partial cycles.

More recently, in 2020, the same authors show an analysis of the interaction between the optical fiber evanescent wave sensor and the graphite particles within a LIB [50]. The proposed sensor was sensitive to lithium concentration at the surface of graphite particles; then, it was able to monitor the capacity fade of LIBs. In the same year, photonic crystal fibers were used by Miele et al. to monitor chemical changes within LIBs under real working conditions [52]. The technique used was based on optofluidic single-ring hollow-core fibers, which uniquely allow light to be guided at the center of a microfluidic channel. The signal analysis was performed by background-free Raman spectroscopy to identify early signs of battery degradation. From the results, the Raman peaks related to ethylene carbonate and the important battery additive vinylene carbonate, offer a direct vision in the formation of the SEI, the main buffer layer that largely forms during its first electrochemical cycle, and whose stability is key to the longevity of the LIBs.

3.5 Simultaneous tracking of temperature, strain, pressure, and RI

The main challenge in tracking critical parameters inside the LIB, such as thermal gradients, strain, pressure, and RI changes, is that due to its electrochemical environment, the LIB presents a very dynamic behavior. The temperature variation influences the thermal expansion of the materials that compose the LIB, promoting strain changes. The electrochemical behavior also promotes internal gassing production, which will affect the pressure variation and RI changes on the electrolyte. LIBs primarily employ liquid electrolytes to ensure rapid ion transport for high performance of the variation in the RI of the electrolytes is related to the variations in the conductive salt concentration. Thus, the RI shifts can be treated as an indicator of the degradation evolutions.

As some of the OFS are sensitive to more than one parameter simultaneously, they suffer from large cross sensitivity, such as strain, RI, and temperature. In this way, solutions to decouple these parameters should be considered by the researchers. As the LIBs are very complex systems with dynamic and diverse physical and electrochemical behaviors, in which many parameters are linked and correlated between them, such as temperature, strain, gas formation, and pressure, several studies were already reported by sensing and decoupling simultaneous parameters in LIB since 2013.

Sommer et al. have reported many studies concerning the use of FBGs to simultaneously decouple strain and temperature variations in LIBs [21, 22, 32, 33]. In 2014, the authors start by externally attaching LIB pouch cell FBGs to monitor additional informative cell parameters (strain and temperature) and using other FBGs as a reference to perform this parameters discrimination, as described by Rao et al. [66]. Two FBGs were employed in the experimental setup, one, bonded at two points to the surface of the pouch cell with epoxy, sensing both strain and temperature variations; while the other one, loosely attached to the cell skin with a heat-conducting paste, only detecting temperature variations. Several charge and discharge cycles were performed to examine the repeatability of the measured signals and compared with conventional strain and temperature sensors to verify the accuracy of these sensors. In 2015 [22], the same authors examined the excess volume change at the end of charge and the volume relaxation in the subsequent rest phase by monitoring the strain variations of externally attached FBGs of a lithium-ion pouch cell. The strain was instigated by the alteration of electrode volume, due to the constant Li⁺ oscillation and intercalation from and to the positive electrode, and thermal expansion/contraction during cycling charge/discharge steps. A strain relaxation was observed at higher SoC levels, especially strain signal relaxed by ∼30% at an SoC level of 100%, and the ratio of Li⁺ in the external electrode region to Li⁺ in the internal electrode region was larger at a higher SoC level. The association between them was also explored at various room temperatures. It concluded that the residual strain increased with decreasing temperature for a certain SoC level, and the alteration between the residual strains was higher for superiors SoC levels.

In 2017, two-part papers about embedded fiber optic sensing for accurate internal monitoring of cell state in advanced BMS by monitoring temperature and strain shifts inside of a pouch cell LIB were developed by Raghavan et al. (part 1) and Ganguli et al. (part 2), belonging to the same research group [32, 33]. Part 1 focuses on the embedding method details and performance of LIBs. The seal integrity, capacity retention, cycle life, compatibility with existing module designs, and mass-volume cost estimates indicate their suitability for electric vehicles and other advanced battery applications. One of the two FBGs was enclosed in a special tubing to make it selectively sensitive to thermal variations alone. The tracked wavelength peak values of the “reference” FBGs in the tubing are subtracted from the total wavelength shift of the adjacent FBG sensor, which is sensitive to strain so that temperature variations are compensated. The second part focuses on the internal strain and temperature signals got under different conditions and their use for high-accuracy cell state estimation algorithms. In particular, the measured strain is used to estimate the battery capacity and predict the capacity up to 10 cycles.

Nascimento et al. have also reported many studies regarding the simultaneous decoupling temperature and strain variations in LIBs through FBGs and interferometric sensors (Figure 14). Different type of LIBs was tested on this discrimination. The prismatic and cylindrical configurations were tested externally and pouch cell configurations were tested both internally and externally [25, 37, 40, 44]. In 2015, FBGs were attached to the surface of a cylindrical LIB to track its thermal and strain fluctuations during charge and different discharge C-rates (Figure 14A). The tests were repeated twice for each discharge C-rate applied (0.25 C and 1.33 C). The FBG1 and FBG2 only measured temperature variations, while FBG3 was fixed to the battery edges and was subjected to strain and temperature variations. Temperature measurements made by the FBG2 sensor were used to compensate for thermal effects on FBG3, allowing in this way to measure the longitudinal strain variation along the battery length [25]. In 2018, a network of FBGs was attached at a prismatic LiB to sense its temperature and bi-directional (x- and y-directions) strain variations during normal charge and two different discharge C-rates (1.32 C and 5.77 C). The discrimination method used by the OFS was also the reference FBG method [66]. Maximum temperature variations were detected close to the positive electrode side, and higher strain values were sensed in the y-direction (Figure 14B). One year later, fiber optic hybrid sensors were embedded in a Li-ion pouch cell to internally monitor and simultaneously discriminate in situ and operando strain and temperature shifts in different locations (Figure 14C). The hybrid sensing network was constituted by FBGs and FP interferometers. Due to the different strain and temperature sensitivities attained by both types of optical sensors, it was possible to decouple the strain and temperature values by using the matrixial method. Galvanostatic cycles by using different C-rates were applied to correlate the temperature and strain signals with electrochemical processes in the LIB.

Figure 14.

Temperature and strain discrimination by OFS in different LIBs configurations. A) Cylindrical B) Prismatic C) Pouch cell [25, 37, 44]. Copyrights 2015, 2018, 2019.

In 2017, Fortier et al. also tracked internal strain and temperature variations in the coin cell configuration [35]. However, how this decouple was performed is not explicit in the manuscript. The batteries were evaluated at a cycling C/20 rate, and the FBGs were placed between electrodes and separator layers, near the electrochemically active area. Results show a stable strain behavior within the cell and a near of 10.0°C difference was registered between the interior of the coin cell and room environment temperature over time during cycling steps.

In 2019, a novel-designed OFS, about self-compensating FBGs, to monitor the separator internal status of a LIB by detecting the RI of the battery electrolyte, was proposed by Nedjalkov et al. [48]. The proposed sensor consisted of two FBGs recorded of the same length but in different fiber layers (one on the core and the other near the surface of the cladding, by using a femtosecond laser system). The cladding, near the FBG region, was also softly etched to increase the sensitivity for RI variations. Between the surface FBG, an additional waveguide positioned at half the distance between the fiber core and cladding surfaces in the radial direction was integrated into the inner cladding at the same axial position. Both the influences of the longitudinal strain and temperature could be compensated with this arrangement, so the remaining variable of the measurement was the influence of the effective RI, which was relative to the reflected Bragg wavelength shift. The proposed FBG configuration was embedded centrally between two separator layers of a 5 Ah lithium-ion pouch cell. The results obtained, show that the optical signal was dominantly influenced by the effective RI of the battery electrolyte.

Huang et al., in 2020, published one study, about operando decoding of chemical and thermal events in commercial LIB, by discriminating and sensing temperature and pressure variations through FBGs and microstructured optical fibers (MOF) [51]. The sensing of different parameters was performed, thanks to the different sensitivities of both optical sensors at each parameter (temperature and pressure), in which the matrixial method was applied. These findings allowed to detect chemical events such as the SEI formation and structural evolution in the LIBs. The authors also demonstrate how multiple sensors are used to determine the heat generated by converting the optical data to heat flux values. In the last year, they also demonstrate the feasibility and diversity of TFBGs to operando access the chemistry and states of electrolytes [2]. They show how a single TFBG can simultaneously sense temperature and RI evolutions inside LIB, which is correlated with the chemical electrolyte behavior (see Figure 15). From the time-resolved RI signals, the feasibility of monitoring electrolyte deteriorations while accessing their turbidity via particulate-induced optical scattering and absorption was studied as well. These unraveled electrolyte characteristics by TFBG help to determine the electrochemical reaction pathways, being strongly correlated with the batteries’ capacity loss.

Figure 15.

Schematics and spectra of the TFBG used by Huang et al. [2]. A) TFBG inserted in the core of a cylindrical LIB. B) A gold layer was deposited on the end of TFBG to obtain the reflection probe. C) The temporal voltage, temperature, and RI of two cells during the first two formation cycles at C/10.

4. Conclusions

This chapter fully describes all main optical fiber sensing techniques used and developed for tracking critical key parameters in LIBs since the first study in 2013. According to the operating principles, FBGs, FPIs, optical fiber evanescent waves, and optical fiber photoluminescent sensors are being used so far. Regarding all the studies selected to perform this overview, the principal parameters presented in the literature were temperature (heat flow), strain, pressure, electrochemical events (such as electrode lithiation and gassing production), RI, and SoX battery indicators (such as SoC, SoD, and SoH). In a general overview, the FBGs, FPIs, and photoluminescent sensors are mostly used to track the physical parameters instead of the evanescent wave sensors are most used to detect the electrochemical events in LIBs due to the necessity of measuring RI values from the surrounding materials that interact with the optical fiber surfaces, in this case.

Between all OFS used in the battery sensing applications and with an easier correlation with BMS, the FBGs coupled with other types of sensors (interferometers and/or evanescent wave sensors), seem to be the most advantageous in the future battery applications, due to their intrinsic characteristics, of the possibility of multipoint and multiparameter monitor, and easy interrogation, operating in a reflection system. Those factors detected have a good alliance with the battery SoX, thus can greatly reflect the battery failure condition. However, the development of sensors for battery tracking is still not consistent with the goal of massive processing, low cost, and daily applications. Some problems, such as excessive data treatments, and the high fragility of some optical fibers (reduced thickness) still exist. Generally, the optimal result of in situ/operando sensing is to real-time monitor and interpret each shift of wavelength into a concrete chemical reaction, so that the precise battery SoX estimation of BMS can be achieved and corresponding actions can be taken place from the external to sustain the continuing operation of batteries.

Comparatively, with other sensing tools or instruments that were also used to monitor critical parameters in LIBs, such as TCs, RTDs, thermography, and, strain gauges, the OFS presents several advantages. They can be embedded in the electrochemical environment of the cells, detecting with elevated accuracy and simultaneously, in multipoint and multiparameter, which are until now, completely unknown, such as the internal pressure and RI variations, which are directly correlated with electrochemical cells events (SEI layer formation).

This advancement in sensing internal and operational batteries using OFS will allow for the improvement of their performance and safety and will help in understanding and improving the lifetime and behavior of the next generation of LIBs to be developed.

Acknowledgments

The authors gratefully acknowledge the European Project “Innovative physical/virtual sensor platform for battery cell” (INSTABAT) (European Union’s Horizon 2020 research and innovation program under grant agreement No 955930), website: https://www.instabat.eu/ . Carlos Marques acknowledges the financial support from FCT through the project DigiAqua (PTDC/EEI-EEE/0415/2021) and CEECIND/00034/2018 (iFISH project). The authors also acknowledge the financial support within the scope of the project i3n, UIDB/50025/2020 & UIDP/50025/2020, financed by national funds through the FCT/MEC.

Conflict of interest

The authors declare no conflict of interest.

Appendices and nomenclature

BMS

Battery management system

DTW

Dynamic time warping

FBG

Fiber Bragg grating

FPI

Fabry-Perot interferometer

FSR

Free spectral range

IR

Infrared

LIB

Lithium-ion battery

OFS

Optical fiber sensors

MOF

Microstructured optical fiber

MZI

Mach-Zehnder interferometer

RI

Refractive index

RTD

Resistance temperature detector

SEI

Solid electrolyte interface

SoC

State of charge

SoD

State of discharge

SoH

State of health

SMR

Single-mode fiber

SPR

Surface plasmon resonance

SRI

Surrounding refractive index

TC

Thermocouple

TFB

Tilted fiber Bragg

G UV

grating Ultraviolet