Field of the Invention

The presented invention relates to fiber-optic sensing systems for the in situ and in operando monitoring of battery cells. More precisely, the invention relates to 24 hours, simultaneous, in situ, in service and in rest monitoring, transmitting and controlling of key physical and chemical parameters of battery cells by one or more fiber-optic sensors embedded into the cell.

Background of the Invention

“Green” energy is widely discussed as a solution to counteract the global warming caused by the emission of carbon dioxide (C0 ₂ ) primary from fossil-fuel combustion. Driven by growing concern about climate changes most industrial countries agreed to reduce the emission of C0 ₂ and to keep the global warming below 2°C. According to the strategic vision the electrification is set to be one of the main technological pathways to reach carbon neutrality. Electricity and transportation are two main sources of C0 ₂ being based on the combustion of fossil fuel for power generation. The most effective way to reduce the C0 ₂ emission is to reduce the fossil fuel consumption. Different “low- carbon” strategies comprising a so called “energy mix” have been adopted to meet the goal. Among them renewable energy sources, associated with energy storage systems (ESSs) and electric vehicles (EVs) draw a big attention of decision-making people. “Green” transportation is already present on the roads worldwide in the form of electric vehicles (EVs), which do not emit C0 ₂ and other pollutants such as nitrogen oxides (NOx), non-methane hydrocarbons (NMHC) and particulate matter (PM). Electric vehicles provide quiet and smooth operation and consequently create less noise and vibration.

Electric batteries and among them lithium-ion batteries (LIBs) are recognized as the key technological enablers to drive both aforementioned fields acting towards a decarbonized society. Lithium-ion batteries are currently the dominant power sources and electrochemical energy storage devices. However, there is a number of shortcomings of current Li-ion battery cells that relate to the safety issue and require urgent improvements. They are rooted in a rather complex chemistry running into the battery cell as well as in different abusive conditions. If a Li-ion battery cell is subjected to operate outside of its design window, it may fail through a rapid self-heating or thermal runaway, which can lead to the catastrophic scenario associated with venting, flames, flying parts, and explosion. This situation becomes even worse, because a damaged single battery cell can trigger other cells inside the battery pack to a runaway incident causing a strong health hazard. To prevent these hazardous situations the current Li-ion battery cells are equipped with a Battery Management System (BMS), which is permanently monitoring and controlling the voltage, current and temperature of the battery cells. In order to minimize the hazardous effects of the already initiated thermal runaway into the cell, some other safety devices are integrated, e.g. the Positive Temperature Coefficient (PTC) and safety vent, which can switch-off the cell if temperature rises over 130°C and inner cell pressure being larger than 10 bar. Despite of the above protect devices problems still exists with the safety of battery cells. There are many dramatic situations reported about such events including brands such as Tesla S and Boing 787 “Dream liner”, wherein the latter had to cancel its mission due to the LIB failure. Such incidents are a direct consequence of the thermal runaway problem and may happen without an accident or crashing. It was also reported that electric vehicles staying at rest, e.g. in the garage, can simply catch fire causing a dramatic fire situation in a residential building. This is due to inefficient batter management systems, which mostly monitor the battery performances at the battery pack level outside of the cell only, instead of in situ into each individual cell.

Therefore, the instant invention aims at developing of a more effective sensing system capable of permanently monitoring each single battery cell in every occasion.

Detailed Description of the Invention

To solve this object, the invention provides a battery block comprising battery cells and a fiber-optic sensing systems for the in situ monitoring of said battery cells, comprising a plurality of fiber-optic sensors that are embedded into the battery cells and an optoelectronic and signal processing unit, to which the fiber-optic sensors are connected for obtaining and evaluating sensor data, wherein the fiber-optic sensors comprise at least one fiber-optic sensor for measuring a chemical composition and a concentration of an electrolyte material of the battery cell by means of low-coherence interferometry, at least one sensor for measuring an internal pressure of the battery cell, at least one sensor for measuring a strain of the battery cell by means of low-coherence interferometry, and at least one sensor for measuring a temperature of the battery cell.

The combination of a plurality of different fiber-optic sensors allows for a complete monitoring of battery cells, so that deviation from normal operating conditions can be detected quickly and precisely. The fiber-optic sensors are chosen to measure the chemical composition and the concentration of an electrolyte material of the battery cell, the internal pressure of the battery cell, the strain of the battery cell, and the temperature of the battery cell. The battery block and the fiber-optic sensing system may be combined with a batter management system (BMS) to form a close-loop control system, as shown in Fig. 1 . The closed loop configuration of Fig. 1 comprises three main blocks schematically presented in the figure.

A battery block, such as those typically used in an electric- or hybrid electric vehicles (EV & HEV), comprises a plurality of individual Li-ion cells, the number depending on their capacity, size and arrangement. Usually, the cells are cylindrical, e.g. 18650 (18mm x 65mm size), hard-prismatic or soft-prismatic, so called pouch cells, which are packed in a separate block.

Figure 2 shows a schematic representation of a battery block in combination with a multiparameter fiber-optic sensing system for in situ monitoring, transmitting and controlling of physical and chemical parameters of the Li-ion battery cells. The fiber-optic sensing system shown in Fig. 2 comprises a battery block 201 , which is composed of several individual cells mutually connected in parallel and/or serial connection depending on the required total power of the whole battery package and on the type and capacity of the individual cells. Multiparameter fiber-optic sensors 202 are embedded into each cell 201a so as to ensure complete hermeticity of the cell housing, regardless of which housing type is applied. The fiber-optic sensors 202 are connected to a common fiber-optic cable 205 that provides a data communication in both directions between the optoelectronic and signal processing unit 204 and the sensors 202 via a multi-fiber push on connector 207 such as a MPO type connector. The optoelectronic and signal processing unit 204 is responsible for sensor signal multiplexing and conditioning, for signal processing and for data communication with the intelligent battery management systems (BMS) 203 associated to each individual cell. The data communication between the optoelectronic and signal processing unit 204 and the sensors 202 on the one hand and the data communication between the optoelectronic and signal processing unit 204 and the battery management systems 203 on the other hand can be done via one and the same fiber optic cable 205, or by means of separate cables, so that a second cable 206 would be required, as depicted in Figure 2. In another embodiment the intelligent BMS may be included as a common multichannel unit into the optoelectronic and signal processing unit 204 that communicates with each individual cell by the same (above mentioned) means. Some other data communication means may be used as well, including standard wiring or wireless data transmission. The fiber-optic cable interconnection between the BMS 203 and the optoelectronic and signal processing unit 204 may be performed by the same MPO connector 207.

The intelligent BMS 203 assures a full supervision and control of each individual cell by unifying the online processed data obtained by the fiber-optic sensors and the voltage, current and outlet temperature data as commonly measured by current BMS. The intelligent BMS provides maximum performance and safety operation of each cell by mitigating possible thermal runaway happening during the fast charging/discharging, over-charging/discharging, or due to extreme weather conditions. In the BMS 203 the signals obtained from the sensors (202) will be analyzed and compared with reference values. Depending on the ratio between measured values and reference values, an action will be executed. If the ratio (in some cases) is out of the tolerated range, a feedback signal will be sent towards the compromised cells. If some of them have a severe drawback, they will be immediately excluded from the battery block in order to avoid possible thermal runaway yet in the early stage. If the failure is not so dramatic, these cells will be charged/discharged under an adjusted voltage/current regime.

Additionally, the intelligent BMS 203 may comprise a logging block for recording the signal history of each individual cell, which assists in the analyzing procedure. Analyzed data together with voltage/current data may be used for estimating a state of charge (SoC), state of energy (SoE), state of power (SoP), state of health (SoH) and for predicting a state of safety (SoS) and an end of life (EoL) of the cell and will ensure increased cell quality, reliability and longevity (QRL). The BMS 203 may also be equipped with visual and sound alert in order to warn a user, such as a car-driver, on time. The BMS 203 may have a wireless data transmission unit for the external communication with portable devices (e.g. mobile phone, notebook, PC, etc.).

The fiber-optic sensing system may comprise different fiber-optic sensors, which are multiplexed for parallel and simultaneous measurement of one or more physical and chemical parameters of Li-ion cells. Basically, each single sensor may consist of an optoelectronic unit and a fiber-optic link terminated with a sensing element. The fiber-optic sensors used in the invention may be designed as extrinsic and/or intrinsic sensors. In case of extrinsic sensors, a miniature sensing head may be made of a transducing element responsible for converting an optical characteristic, e.g. intensity, optical spectrum, luminescence, etc., being a function of the measured parameter, into an electrical value. In case of intrinsic sensors, the optical fiber per se converts an optical characteristic, e.g. phase angle change, spectral shift, etc., being a function of the measured parameter, into an electrical value. In both cases, the sensing transducers or optical fibers are embedded and assembled into the cell housing, including electrodes, separator and electrolyte, thereby providing absolute tightness of the cell. In some cases, optical fibers may be adhesively bonded onto the external walls of the chamber as well.

The fiber-optic sensing system may use different interrogation techniques, depending on the sensing principle. In one embodiment, when using light intensity modulated sensors, a voltage photodiode signal produced by the back reflection of a light beam modulated by a transducer may be measured. In another embodiment, an optical spectral analyzer (OSA), i.e. spectrometer may be used for the measurement of a change of spectral shape of an optical radiation after a transducer. In case of interference-based fiber-optic sensors, a change of the phase angle that is induced by the influence of the physical or chemical parameter to be measured in an optical fiber may be measured. In case of a shift of light spectrum after Fiber-Bragg-Gratings (FBGs) imprinted into the sensing fiber, a spectrometer may be used.

The fiber-optic sensing system may comprise an optoelectronic block, including light sources (lasers and laser diodes, e.g. Vertical Cavity Surface Emitting Laser (VCSEL), super-luminescent diodes (SLD), light emitting diodes (LED) or different lamps), photodiodes (PD) accompanied with appropriate transimpedance amplifiers (TIA) and a digital signal processor (DSP) for on-line signal processing. The raw data of the fiber-optic sensors may be acquired by various photodetectors, e.g. photodiodes, 1D or 2D arrays of complementary metal-oxide-semiconductors (CMOS), charge- coupled devices (CCD), or optical spectral analyzers, e.g. a miniature spectrometer. Signal processing may be performed by execution of one or more different algorithms embedded into the appropriate digital signal processor (DSP), fast enough to provide real-time calculation of all signals. The calculated output signals may be analog- or digital signals, which correspond to physical or chemical parameters of battery cell. A correlation factor between the measured optical signals and the respective physical and chemical parameter may be determined by a calibration procedure.

The fiber-optic sensors may be made of single-, or multi-mode optical fibers, 125pm in glass cladding diameter and 150pm overall plastic coating diameter. The coating material may include poly(methyl methacrylate) (PMMA), polyether ketone (PEK), polyether ether ketone (PEEK), polyvinylidene fluoride or polyvinylidene difluoride (PVDF), polyimide (PI), etc. The main features of the plastic coatings are to provide a good flexibility, mechanical, chemical and thermal stability that spans over -50 to about 300°C corresponding to PI features, and to increase overall robustness of the thin fiber-optic sensing structure. This is advantageous when considering the corrosive environment within the cell and the high temperature (over 300°C) prevailing in case of a thermal runaway. Some other fiber-optic types are possible as well, including 50 and 80 pm glass cladding diameters that contributes to overall miniaturization of optical fibers and their better incorporation into the confined room of the battery cell. In addition, there is a great number of plastic optical fibers (POF), photonic band gap (PBG) fibers, or polarization maintaining fibers, which may be used for sensor applications. In addition, planar optical waveguides (POW) can be used as well that allow to be printed onto the monitored cell structure, e.g. electrode, separator or inner and outer housing walls. Optical fibers are made of dielectric materials (primarily glass and plastic) and are therefore intrinsically explosion proof and safe to be used within the battery cells, thereby demonstrating a high robustness against short circuit, electromagnetic interference (EMI) and radio frequency (RF) interference and high voltage (HV).

Overall sensing configuration Figure 3 schematically represents an overall sensing configuration, which basically comprises an optoelectronic and signal processing unit 204 and a fiber-optic sensing system embedded into the battery cell 201a. For the sake of clarity, only one battery cell is shown, although the optoelectronic and signal processing unit is able to provide conditioning of an entire battery block or even all cells in an entire battery package of an EV or ESS, thanks to multiplexing of fiber-optic sensors, including multiplexing of light sources and photodetectors. Two multiparameter fiber-optic sensors 309 and 310 are embedded into the battery cell. The first sensor 309 functions as a fiber-optic chemical composition and concentration sensor (FOCC) 311 and a fiber-optic pressure sensor (FOPS) 312. The FOCC sensor 311 may additionally serve as a turbidity and gas bubble sensor, as explained later in more detail. The second sensor 310 functions as a fiber-optic strain sensor (FOSS) 313 and a fiber-optic temperature sensor (FOTS) 314. According to the invention, the fiber-optic chemical composition and concentration sensor (FOCC) 311 and the fiber-optic strain sensor (FOSS) 313 are operating on basis of low-coherence interferometry. The fiber-optic pressure sensor (FOPS) 312 is preferably operating on basis of Fabry-Perot (FP) interferometry. The fiber-optic temperature sensor (FOTS) 314 is preferably operating according to the Fiber-Bragg gratings (FBG) technology.

All sensors are integrated into and arranged in the battery cell, which may have a prismatic form, either in hard- or soft-package, e.g. in the pouch cell, although some other cell forms are also possible, e.g. cylindrical or coin cells.

In the description of Figure 3, the following abbreviations will be used:

BMS: Battery Management System DRV: driver

DSP: digital signal processing TIA: transimpedance amplifier

PD: photodiode

SLD: super-luminescent diode MPO:multifiber connector WDM: wavelength demultiplexer FOTS: fiber-optic temperature sensor FOSS: fiber-optic strain sensor FOPS: fiber-optic pressure sensor FOC: fiber-optic coupler

FOCC: fiber-optic chemical composition/concentration.

The multiparameter fiber-optic sensors 309 and 310 are built around the 3x3 fiber-optic couplers (3x3 FOC) 315 and 316 acting (in some cases) as a modified Michelsons low-coherence interferometer. Such a configuration allows a passive stabilization of raw interference signals by keeping them in so called quasi-quadrature that in turn overcomes a well-known problem of the interference signal fading. On the other hand, the low-coherence interferometry allows absolute measurement of targeted parameter due to a non-periodic Gaussian-shape of the interference pattern. In this way, the second (also well-known) problem of sensor initiation, which is typical for a common high-coherence interferometry, was overcome.

The fiber-optic sensors may be made of chemically and electrochemically resistant single-mode optical fibers made of pure silica glass core, having a cladding of 10pm and a diameter of 125pm. Some other core/cladding diameters may be used depending on the propagating light wavelength. The fibers may be coated e.g. by a polyimide or Teflon layer of about 150pm in diameter enabling survival over a temperature range from -50°C up to 300°C in a harsh chemical environment such as in the battery cell electrolyte. These fibers are designed for use at 850, 1310 and 1550nm or some other light wavelengths. In addition, these single-mode fibers may have FBGs (Fiber-Bragg gratings) 314 written into the structure (core) so that they can be used for distributed temperature measurement as well. The overall diameter of one multiparameter sensor composed of three sensing arms may be as small as a few hundred micrometers.

The entry point of the fiber-optic sensors into the battery cell may be manufactured by sealing the input optical fibers or their stainless-steel housing (a metal tube having an outer diameter of 3mm, or smaller, and a length of 3cm, or shorter, and acting as a feedthrough of the optical fibers) 317 and 318 of the 3x3 couplers 315 and 316 between the pouch cell foils or by welding to the hard metal housing of prismatic cells. Hence, the hermeticity of the cell is assured. On the other hand, such a fiber-optic feedthrough guarantees very high mechanical stability and overall robustness of the sensing structure. Fiber-optic sensors are multiplexed and may be connected to the interrogation optoelectronic and signal processing unit 204 via the MPO multifiber connector 307. The MPO connector 207 is a unique interconnecting device, because it can mate multiple fibers (2-72 pieces) within a single connector housing, in this way significantly reducing overall space and weight of the sensing system. The connector’s design utilizes guide pins and precision holes for ferrule-to-ferrule alignment.

The multiparameter fiber-optic sensing system comprises two low-coherence light sources, e.g. pigtailed super-luminescent diodes (SLD) 319 and 320, emitting light at two different wavelengths, i.e. a first wavelength of e.g. 1310 nm and a second wavelength of e.g. 1550 nm. These diodes may be driven by low-noise and thermally stabilized drivers (DRV) 321 and 322 controlled by a microprocessor (DSP) 323. In this way the long-term reliable and constant optical parameters of the emitted radiation, such as output power and spectrum, are assured. The pigtailed SLD may be connected to the input arms of a 2x2 fiber-optic coupler (2x2 FOC) 324. The output arms of the coupler are connected to the middle output arms of the two 3x3 couplers (3x3 FOC) 315 and 316 via a connector, such as a multifiber MPO connector 207. The output arms of the 3x3 couplers 315 and 316 are terminated with the fiber-optic sensors 309 and 310. The back reflected signals from the sensors 309 and 310 are fed into the optoelectronic and signal processing unit 204 by the four remaining input arms, which are connected to the input arms of four (1x2) wavelength demultiplexers (WDM) 325, 326, 327 and 328 capable of splitting light having the first wavelength, such as 1310 nm, from radiation having the second wavelength, such as 1550 nm. Hence, eight optical signals are simultaneously generated, among them four from low-coherence interferometry (LCI) sensors 311 and 313, one from the FBG temperature sensor 314 and one from the Fabry- Perot (FP) pressure sensor 312. Therefore, there are two couples of LCI signals at a first wavelength of e.g. 131 Onm, which are simultaneously captured by pigtailed (or some other packing) InGaAs photodiodes PD2 329 and PD3 330 (for FOSS) and PD4 331 and PD5 332 (for FOCC) associated with appropriate transimpedance amplifiers TIA2 333, TIA3 334, TIA4 335 and TIA5 336. One input arm of the WDM in the FOSS line 310 is connected to a miniature spectrometer 337 for acquiring the FBG signals at the second wavelength, such as 1550 nm, corresponding to the temperature measurement. Similarly, the remaining input arm in the FOCC line 309 is connected to the photodiode PD1 338 accompanied with the transimpedance amplifier TIA1 339 for capturing the FP pressure signal at the second wavelength, such as 1550 nm. Two remining input arms 340 and 341 of WDMs 326 and 327 at the second wavelength, such as 1550 nm, may be inactive or may be used as redundant channels.

The raw LCI photodiode signals coming from TIA2 333, TIA3 334, TIA4 335 and TIA5 336 are further introduced into the DSP 323, where on-line signal processing is carried out. Simultaneously, the FP pressure signals from TIA1 339 and FBG temperature signals from Spectrometer 337 are also introduced into the DSP. The processed output signals are delivered by a fiber-optic link to the intelligent BMS. A self-testing procedure of sensors operation may be performed by temporary sending a pilot signal controlled by the BMS from optoelectronic and signal processing unit 204 toward multiparameter sensors 309 and 310.

The following description relates to the design and assembling of multiparameter fiber-optic sensors into the battery cell, including a fiber-optic sensor for measuring a chemical composition and a concentration of an electrolyte material of the battery cell by means of low-coherence interferometry, a sensor for measuring an internal pressure of the battery cell, a sensor for measuring a strain of the battery cell by means of low-coherence interferometry, and a sensor for measuring a temperature of the battery cell. A high sensitivity, accuracy and resolution and a wide dynamic range of the measured parameters is assured by combining three mature sensing technologies: low-coherence interferometry (LCI), optical spectrometry and fiber Bragg grating (FBG). Fiber-optic sensor for measuring the chemical composition and concentration by LCI

In the multiparameter sensor 309 represented in Figure 4 the LCI in combination with Micro-Electro- Mechanical-System (MEMS) technology acts as a multipurpose technique dedicated to the determination of the chemical composition and concentration of the liquid electrolyte by measuring the index of refraction (n) based on a high precision measurement of the optical path difference (OPD) in a MEMS interferometer of the modified Michelson type. It comprises a 3x3 fiber-optic coupler 315 made of a single mode optical fiber 10pm in core diameter and 125pm in glass cladding diameter. The central input arm 401 of the 3x3 coupler 315 is connected to one output arm of the 2x2 fiber-optic coupler 324 (the second output arm is left free for the sake of clear presentation) that is further connected to the pigtailed low-coherence light sources, i.e. super-luminescent diodes (SLD) 319 and 320 emitting light with a first and second wavelength of e.g. 131 Onm and 1550nm, respectively.

The chemical composition and concentration sensor 309 operates at the first wavelength, such as 131 Onm, only. The SLD 319 light radiation travels via the 2x2 coupler 324 and the central input arm 401 of the 3x3 coupler 315 and the single-mode fiber (SMF) towards the output arms 311a, 311b and 403 of the 3x3 coupler 315 at the right side. The output arm 311 a is a sensing arm that is set into the MEMS open structure 404 (see inset above in Figure 4), while the output arm 311 b is a reference arm. The third (middle) output arm 403 is connected to the MEMS Fabry-Perot pressure sensor 312 (see middle inset in Figure 4).

The MEMS open structure 404 is designed to have an optical cavity 405 acting as a microcuvette filled by a liquid electrolyte 407, as shown in more detail in Figure 5. In Fig. 5a is shown a fiber-optic sensor for the chemical composition/concentration measurement, and Fig. 5b shows a fiber-optic sensor for the turbidity measurement and the gas-bubbles detection. The MEMS structure is made by chemical etching of Si wafer 404a to have a “V” grove 404b for accepting the sensing fiber 311a that is “locked” in the “V” groove by adhesive bonding 404c or by local fusion splicing the glass plate and fiber cladding. The whole structure is anodically bonded to the glass wafer 404d. The sensing fiber 311 a and the reference fiber 311 b are set in proximity occupying relatively small room in the cell chamber 201a. Therefore, the light beam 406 traveling through the core of the sensing fiber comes out and propagates though the liquid phase of the electrolyte 407, reflects at the cavity wall 408 and travels back again through the electrolyte entering the optical fiber core. At the same time, approximately equal intensity of the light beam 409 travels via reference arm 311b reflects at the metalized fiber tip 410 and propagates back toward the 3x3 coupler 315. Here, the sensing and reference optical beams interfere generating two low-coherence interference signals captured by photodiodes PD4331 and PD5 332. Based on these two signals, the OPD between the sensing and reference arms, and subsequently the index of refraction of the electrolyte can be calculated by Equation (1). The OPD parameter is in direct correlation with a minute change of index of refraction n of the electrolyte given by:

OPD = An L (1) where An is an effective change of the index of refraction of the electrolyte and L is the physical length of the measured liquid.

Fiber-optic sensor for the chemical composition, concentration and humidity measurement by means of optical spectroscopy

Figure 6 represents another embodiment in a form of a single fiber-optic sensor for measuring the chemical composition and concentration of a liquid phase electrolyte. It can be very easily integrated into the basic configuration depicted in Figure 3 and Figure 4 instead of the MEMS based FOCC sensor. Here, the sensing fiber is terminated with a miniature spherical (or other form, e.g. cylindrical, GRIN, etc.) collimator (see Figure 6a) that collimates a broadband radiation supplied via the 1x2 fiber-optic coupler by some lamp, infrared light emitting diode (IRLED) or SLD emitting within or over UV, VIS and NIR, e.g. in the range from about 190 to 2500nm. The collimated light beam propagates through the liquid phase electrolyte, which occupies the empty room made in ferrule and returns back after mirror-reflection into the collimator and further via the same fiber-optic link and 1x2 fiber-optic coupler toward the spectrometer. The optical path length is intentionally extended in order to increase the overall sensitivity of the sensor. Here, the optical path length may be twice the presented gap because light travels in both directions.

The principle of operation is based on the monitoring of one or more absorption peaks in the radiation range of the applied source, which are characteristic for the targeted chemical components of the liquid electrolyte. This sensor allows the early detection of those chemical substances that could develop due to side reactions in the case of a thermal runaway, or due to humidity penetration into the cell due to a compromised hermeticity of the battery cell. In the latter case, HF molecules will occur. Figure of merit of concentration of targeted chemical component into the cell is measurement of magnitude of its absorbance vs. time. Inset in Figure 6b schematically presents a transmission spectrum with characteristic absorption peaks.

In some other embodiment the absorbance may be measured with a modified sensing configuration which utilized one or more photodiodes (or some other photodetectors) at the receiving side instead of a spectrometer. The specificity and sensitivity of such a sensor is ensured by involving interference filters in front of the active area of the photodetector which serve to extract a narrow region around characteristic absorption peaks of targeted chemical components. In case of simultaneous monitoring of more than one chemical component it is necessary to use an (1 x n) fiber optic coupler in the receiving side, where n corresponds to the number of components and determines the number of used photodetectors. Regardless of the number of components it is sufficient to measure the intensity of the corresponding photodetector signal to be able to determine the concentration of the targeted component.

Fiber-optic sensor for the detection and counting the gas bubbles into the electrolyte of the battery cell

Figure 5b shows a gas bubbles sensor that is de facto a sensing part of the low-coherence interferometer set into the MEMS structure 404 (Fig. 5a). It allows an early detection of the gas bubbles 411 based on a sudden and rapid change of the OPD due to a dramatic decrease of index of refraction, such as a decrease of about 28%. For example, the index of refraction of 1 M solution of a LiPF ₆ electrolyte of a typical Li-ion cell in ethylene carbonate (EC) and diethyl carbonate (DEC) organic solvents is equal to 1 ,4010. A simple calculation shows that the expected change of OPD should be of about 40pm for an optical cavity length (L) of 100pm. This corresponds to a phase shift of about 400 rad that is much larger than the sensitivity of the applied LCI technique, which is of about ±70 prad/VHz. In addition, any bubble occurrence in front of the optical beam 406 of the sensing arm 311 a of the interferometer will lead at first to an abrupt drop of the intensity of the photodiode signal due to reflection at the spherical interface and the subsequent sudden increase of the signal due to the focusing effect. Such a signal allows not only the detection, but also the counting of the gas-bubbles. This intensity variation of one photodiode signal has no big influence on the resultant interference signal that makes the index of refraction and chemical composition measurement possible and accurate.

Figure 7 presents another embodiment of a fiber-optic sensor for the detection and counting of gas bubbles in the electrolyte of the battery cell. It is depicted as single sensor that can be easily integrated into the basic sensing configuration depicted in Figure 3 and Figure 4 instead of the MEMS sensor. The overall construction is rather simple, and comprises an optoelectronic unit containing any type of light source, e.g. LED, SLD or laser diode LD and receiving photodetector, such as Si or InGaAs photodiode depending on the wavelength of the used light source. Light radiation travels via the 1x2 directional fiber-optic coupler and the single- or multi-mode optical fiber toward the fiber tip of the sensing fiber serving as a very simple sensing head. The sensing fiber is immersed into the electrolyte.

The principle of operation is based on a change of the light intensity back reflected from the sensing fiber tip, as mentioned above. This change is caused by an alternation of the Fresnel reflection at the interfaces of glass-liquid in case A and glass-gas in case B (see inset Figure 7a). The reflection coefficient in case B is of about 4% and decreases at about 0,1%, if the index of refraction of the electrolyte is about 1 ,4. Such a large change of index of refraction will be manifested as a sudden increase of voltage photodiode signal when a bubble comes in contact with the tip of the fiber. It is schematically demonstrated in diagram in Figure 7b.

The width (x) of the pulse corresponds with retention time of one single bubble before it leaves the sensing fiber tip. The width (y) shows the time duration between the two events of the gas bubble detection. Finally, the number of pulses illustrates how many gas bubbles are detected. This will be used for counting the number of bubbles and after calibration for calculation of delivered gas over the electrolyte. Here presented rectangular pulses are just for the sake of illustration and may not correspond with real pulses.

The kinetics of gas generation in the electrolyte may be used for following up the kinetics of the electrochemical reaction in the cell and to predict state-of-charge (SOC) and state-of-health (SOH) of the cell.

Fiber-optic sensor for turbidity measurement of the electrolyte in the battery cell

In addition, the LCI sensing setup, depicted in Figure 5b, simultaneously may serve as a turbidity sensor capable of measuring the back reflected signal into the sensing arm. The first traces of an anode or cathode decomposition associated with the generation of particles 412, presented as black points in Figure 5b, will cause a turbidity increase in the battery cell and subsequently in the liquid electrolyte in front of the sensing fiber of the MEMS sensor. The information about the turbidity of the liquid electrolyte will depend on the reflection and scattering of light hitting the flowing particles in the sample solution that eventually will induce a variation and a drop of the photodiode signal. A correlation between the photodiode signals and different turbidity levels of the electrolyte may be determined by previously calibrating the sensor before its embedding into the battery cell. The calibration turbidity samples may be prepared by mixing known concentrations of e.g. graphite particles in the liquid electrolyte.

Figure 8 schematically shows another embodiment of a possible sensing configuration for performing a fiber-optic turbidity measurement of an electrolyte. It is configured as a single sensor that can be easily adapted into the basic sensing configuration presented in Figure 3 and Figure 4 instead of the MEMS sensor. Basically, this is an intensity-based fiber-optic sensor, which is composed of an optoelectronic unit and a sensing head made of a miniature ferrule terminated with a side-looking window. Such an open structure allows the electrolyte to get into the room in front of the sensing fiber (see inset in Figure 8). The optoelectronic unit consists of a broadband light source, e.g. IRED emitting at 850nm, 131 Onm or 1550nm connected to the input arm of a 2x2 multimode fiber-optic directional coupler. Light travels via the fiber-optic coupler and the sensing fiber toward the sensing head. The head is immersed into the liquid electrolyte. The principle of operation of this sensor is a transmission measurement of liquid phase in the gap between the sensing fiber tip and the mirror (see inset in Figure 8). An increase of turbidity of the electrolyte shall induce the light transmission decrease through the electrolyte, and further a decrease of the light intensity captured by the photodiode PD1 at the receiving side. In order to avoid any parasitic effect, caused by a possible intensity variation of the light source, the PD1 signal shall be divided to a PD2 signal. Therefore, the turbidity level of the electrolyte may be defined as a function of the ratio of PD1 to PD2 voltage signals.

Fiber-optic Fabry-Perot pressure sensor

The same sensing configuration 309, which is presented in Figure 4, may be used for the simultaneous pressure measurement into the cell by a fiber-optic Fabry-Perot (FP) interferometer realized by MEMS technology. This is a robust pressure sensor capable of measuring internal pressure in the cell with high sensitivity and in a very short response time of about 1 ms. This is of paramount importance since the peak pressure caused by a charged 18650 battery can reach 1 .080 x10 ⁷ Pa (108.0 bar) with a pressure increasing rate of about 1.036 x10 ⁹ Pa min ¹ (1.036 x10 ⁴ bar min ¹ ), demonstrating a very fast pressure increases that likely leads to the battery rupture.

According to the Figure 4, the FP pressure sensor is connected to the middle output arm 403 of the 3x3 coupler 315. It is designed as a MEMS side-looking Fabry-Perot interferometer, although some other designs are possible as well, including front looking pressure sensor based on the Fizeau interferometry principle. The optical radiation having a first wavelength of e.g. 1550nm emitted out of a pigtailed SLD 320 propagates toward the pressure sensor (see detail in the middle right inset) that may be made by chemical etching of a Si wafer 420 and anodic bonding of such a wafer to the glass plate 421 . The MEMS structure allows the introduction of the sensing fiber tip 403 into the “V” grove 422. The input orifice may then be sealed with an epoxy adhesive 423. Another option to “lock” the sensing fiber into the MEMS structure is fusion splicing of the sensing fiber 403 glass cladding with glass plate 421 in the region of the entrance. The light beam 424 comes out of the fiber core and hits the 45° inclined metalized facet 425 and further propagates through the glass wafer 421 entering the optical cavity 426 and then reflects from the Si membrane 427 and travels back to the sensing fiber 403 and the 3x3 coupler. The pressure modulated optical signal splits at WDMs 325 and 326. The former is active, and it is connected to the photodetector PD1 338 for acquiring a pressure signal. This is an interference signal obtained due to an alternation of the optical cavity 426 depth caused by the Si membrane 427 deflection that is induced by external pressure. Dynamic range of the pressure sensor is determined by the membrane thickness and its overall size. Fiber-optic Fizeau pressure sensor

In another embodiment, a fiber-optic Fizeau pressure sensor, schematically presented as a single (front-looking) sensor in Figure 9, may easily be integrated into the basic structure depicted in Figure 3 and Figure 4 instead of the Fabry-Perot interference sensor. It is composed of a broadband light source, e.g. a super-luminescent diode (SLD) connected to an input arm of a single-mode 1x2 fiberoptic coupler (FOC). Light propagates out of the coupler via an optical fiber toward the sensing head depicted in the inset in Figure 9a. The sensing head is embedded into the cell for the sake of pressure measurement. The main part of the sensing head is an elastic membrane, which displaces to the sensing fiber tip in dependence of the external pressure. Different materials can be used for manufacturing the elastic membrane, e.g. stainless steel, silicon, etc. The sensing membrane may be produced by various microsystem technologies (MST), e.g. the Si-membrane may be produced by MEMS technology and then anodically bonded to the hole terminated fiber-optic tip (see inset in Figure 9a).

There are different sensing techniques for measurement of membrane displacement. For this purpose, the most suitable are fiber-optic interference techniques based on “all-in-fiber” interferometry utilizing 1x2, 2x2 or 3x3 fiber-optic directional couplers. In Figure 9, a low-coherence interferometric technique is presented, which uses a Fizeau interferometer at the receiving side for the position recovering of the elastic membrane with respect to the sensing fiber tip (Figure 9a). The low-coherence interferometric signal, depicted in the diagram in Figure 9b, was obtained by electronically scanning of OPD into the sensing head by glass wedge set in front of the linear CCD array (spectrometer). The back reflected light beam from the sensing head comes out of the second input arm of the 2x1 coupler and illuminates an optical set composed of a cylindrical lens, a glass wedge and a CCD array. The low-coherence interference pattern depicted in Figure 9b moves along the CCD array in dependence of the membrane displacement, i.e. the pressure intensity. The membrane position was determined in respect of the position of the central fringe of the low- coherence interferometric pattern. The diagram in Figure 9c shows the central fringe position of the LCI pattern on the applied pressure.

Structural Flealth Monitoring (SHIM) of battery cells by strain and damage detection

A number of side reactions are associated with the degradation of a Li-ion battery cell especially in the case of the thermal runaway such as electrode particle cracking, solid electrolyte interface (SEI) film formation and thickening, transition metal dissolution, cathode electrolyte interface (CEI) formation, lithium plating and dendrite formation, graphite exfoliation, binder decomposition, structure disordering, corrosion of current collectors, electrolyte decomposition, etc. Most of them are a potential source of acoustic emission (AE) events, which belongs to its own and specific frequency range usually spreading over a few hundred kHz.

On the other hand, in the course of charging and discharging of the Li-ion battery, the graphite anode demonstrates a volume change of about 10% due to lithium ions intercalation and deintercalation. The volume change of e.g. the LiMn ₂ 0 ₄ (LMO) cathode also changes dramatically of about 16%, which could induce the structure damage and eventually loss of the active material. A large material deformation of about 300-400% caused by dramatic expansion due to huge swelling during the lithiation phase occurred in silicon-based anode batteries. These mechanical deformations belong to a quasi-low frequency range of about hundreds of Hz that can be easily discriminated from the above-mentioned AE events. In a first embodiment, the fiber-optic low- coherence interference technique based on a single-mode 3x3 optical coupler is proposed as a powerful technique for simultaneous damage and deflection detection of battery cell structures with high resolution.

The second sensing configuration 310 of the overall configuration (Figure 3), depicted in Figure 10, combines a fiber-optic FBG and LCI for temperature measurement and for Structural Health Monitoring (SHM) of strain and damages of the anode, cathode, separator or housing of the Li-ion cell 201 a by FFT spectral analysis of an interference signal. For the sake of clarity, the common sensing configuration 310 in the following text will be separated in two fiber-optic sensors, SHM by strain measurement and damage detection and FBG temperature sensor.

The SHM sensor, or fiber-optic strain sensor (FOSS) 510 and 511 , is composed of a 3x3 fiber-optic coupler 316 that may preferably be made of single-mode optical fiber 10pm in core diameter and 125pm in glass cladding diameter, identically as described in the previous sensor depicted in Figure 4. The central input arm of the 3x3 coupler is connected to a 2x2 fiber-optic coupler 324 via the optical fiber 402. The input arms of the 2x2 fiber-optic coupler 324 are further connected to the pigtailed low-coherence light sources, e.g. super-luminescent diodes (SLD) 319 and 320 emitting light at a first and a second wavelength, such as at 1310nm and 1550nm, respectively. In this embodiment the SHM may operate at the first wavelength, such as 1310 nm, although some other wavelengths are possible as well. The low-coherence light beam is launched into the 2x2 coupler 324 and further split into three output arms. Two of them, the sensing arm 310a and the reference arm 310b, of about the same length, terminated in the coils 510 and 511 , respectively, act as an intrinsic sensor. Some other sensor shapes are also possible, including straight, spiral, elliptical or meander. The shape may be selected depending on the required sensitivity, geometry and size of the monitored structure. In a real application, the sensing 510 and reference 511 coils may be packed in a thin sandwich structure, depicted in the upper inset in Figure 10. The sensing coil 510 may be adhesively bonded onto the monitored structure, e.g. electrode 520, while the reference coil may be separated from the sensing coil by a thin foam layer 521 , acting as an acoustic isolator.

The interferometric signals generate because of the superposition of two optical beams propagating through the sensing and reference arm. The sensing arm is exposed to an external stress induced by the structure deformation or degradation, associated with the generation of acoustic emission events as mentioned above (see Figure 10), and will experience a minute change of the index of refraction and of the length of the glass-core of the optical fiber. This is the reason for an alternation of the OPD of the interferometer in Figure 10 in the same way as previously explained by Equation 1 , because the reference coil does not suffer these effects in the same extent as the sensing coil. Therefore, the interference signals are captured by two InGaAs photodiodes, PD2 329 and PD3 330, which are connected to the two other input arms of the 3x3 coupler 316. By signal processing of these raw signals, followed by amplitude and frequency analysis by FFT of the signal, the strain magnitude over the monitored structure and the damage severity can be obtained based on the recorded AE signals.

In another embodiment, a fiber-optic Structural Health Monitoring (SHM) sensor, presented in Figure 11 as a single sensor, can be applied onto the external surface of the battery cell housing. Such a sensor can easily be integrated with the rest of the system demonstrated in Figure 3 and Figure 10. Basically, it operates in the same way as it was described above. The fiber-optic strain and damage sensor (FOSS) 510 and 511 is able to “hear” how the cell “breath” due to swelling as schematically depicted in Figure 11a. The diagram in Figure 11 b (for the sake of illustration) demonstrates the induced strain (me) of a monitoring structure after calculation of OPD generated in the fiber-optic SHM sensor.

Fiber-optic FBG temperature sensor

The fiber-optic temperature sensor, which is also depicted in Figure 10, is composed of a 3x3 fiberoptic coupler 316 and a FBG array 314. The basic principle is a shift of the back reflected light spectrum, i.e. Bragg wavelength l _B , in dependence of the temperature. For this purpose, the sensing configuration is supplied by a broad band light radiation, emitted from a low-coherence light source, e.g. a pigtailed superluminescent diode (SLD) 320 at 1550nm. The light propagation through the structure is performed by connecting the SLD 320 to an input arm of the 2x2 fiber-optic coupler 324 and further via the middle input 402 arm and the output arm 310c of the 3x3 coupler to the FBGs array 314 for multipoint temperature measurement through the liquid electrolyte 407 within the cell 201 a. The reflected light travels backward through the same fiber-optic link, but splits at three input arms of the 3x3 fiber-optic coupler 316. Two lateral arms are connected to the WDMs 327 and 328, which separate the light beam at 1550nm from 131 Onm. The former wavelength is the carrier of the temperature signal. One of the output arms of the WDM 328 is connected to the miniature spectrometer 337 to measure the spectral shift in dependence of the temperature. Another output arm of WDM 327 is inactive.

Such a configuration allows the simultaneous temperature measurement in so many points within the cell as the number of imprinted FBGs arrays has. The temperature influence on the Bragg wavelength shift is about 10 times larger than the strain effect. This technique allows accurate temperature measurement in a wide dynamic range from cryogenic to more than 300°C. Since the FBG optical fiber is immersed into the liquid electrolyte 407 only, without any firm contact with the electrode 520, the separator 521 or the battery cell housing 201a, it is not exposed to stress and there is no induced strain in the FBG. Hence, it is not necessary to apply any compensation technique, because the strain contribution due to stress in the cell structure does not take place at all and possible parasitic artifacts in respect to the temperature measurement are entirely avoided.

Fiber-optic temperature sensor based on spectral transmittivity of a broadband semiconductor

In another embodiment, fiber-optic temperature sensor based on optical transmission of a broadband semiconductor, schematically shown as a single sensor in Figure 12, may easily be integrated in the basic structure depicted in Figure 3 and Figure 10, instead of a FBG sensor. It is made of a single-, or multi-mode optical fiber, typically 125pm of glass cladding diameter and 150pm overall diameter of the PM plastic coating. A rectangular (or circular) sensing dice, smaller than (250x250x100) pm ³ in size, may be manufactured of a broadband semiconductor crystal, e.g. CdTe or GaAs, depicted in the inset in Figure 12a. In one solution the sensing head is adhesively bonded to the fiber-optic tip and encapsulated into the plastic sealant made of, for example, 2K epoxy adhesive, polyimide or Teflon. The rear surface of the rectangular dice may be coated with a metal or dielectric mirror such as Au, Al, Ag or Ti0 ₂ to increase the back reflected signal. The sensor is supplied via a 2x1 fiber-optic coupler by a broadband radiation from a light source, e.g. LED or SLD emitting light having a wavelength of e.g. about 850nm. The light propagates toward the sensing head and reflects back to the spectrometer for analyzing the spectral shape.

The main characteristic of a semiconducting crystal, e.g. CdTe and GaAs, used for the sensing purpose, is the change of energy gap with temperature. This means that light transmission through a semiconductor is wavelength dependent with a sharp rise in photon absorption occurring when photon energy exceeds the band gap energy E _g (T). The transition wavelength above which light transmission substantially increases is given by A _g (T)=hc/E _g (T), where h is Plank ^' s constant and c is the speed of light. The band gap, typically, drops monotonically with the temperature increase. As a result, the transition wavelength shifts to the longer wavelengths as the temperature increases, and for CdTe it is defined by the ratio of d A _g /dT=3,4 x 10 ^~4 pm/K (see Figure 12b). Out of transmission spectrum (Figure 12b), the threshold point that corresponds with the start of the energy gap of CdTe is determined. Diagram 12c shows the threshold point shift vs. temperature. Temperature range is defined by the central wavelength and the spectral width of the light source. However, it can be extended from cryogenic to high temperatures as well, by using more than one light source, the radiations of which are combined by a multiple fiber-optic coupler (see Figure 13). Such a system, when driven with several LEDs of different peak wavelengths, behaves as a temperature sensor sensitive over below -50°C to more than 230°C, or even higher for a specially prepared sensing head without adhesive bonding, e.g. setting the sensing dice into the Si housing manufacturing by MEMS technology.

In another embodiment, the same sensing configuration, depicted in Figure 12, may be connected to a silicon or InGaAs photodiode in the receiving unit instead of the optical spectral analyzer (spectrometer). In such a case the main output signal is a voltage signal of the receiving photodiode amplified with accompanied transimpedance amplifier (TIA) that directly corresponds with the intensity of light transmitted through the semiconductor dice (e.g. CdTe crystal). Figure 14 illustrates a block diagram of the receiving unit comprising two photodiodes and accompanied TIAs for acquiring of light intensity vs. temperature of CdTe semiconductor. This is an auto- referencing technique, called dichroic ratio, where the transmitted light, returned back into the receiving unit, is split into the short- and long-wavelength components by means of a dichroic mirror. Each light component is directed to a corresponding photodiode, and the obtained photocurrents are used to form a ratio-metric signal. The main benefit of such an assembly is to avoid the influence of parasitic effects, such as intensity variation of light sources, possible power loss along the optical link, etc.

Fiber-optic temperature sensor based on LCI

Figure 15 schematically presents another embodiment of the fiber-optic temperature sensor utilizing the low-coherence interferometry as a basic principle. For the sake of clarity, it is represented as a single sensor, although it may be very easy incorporated into the basic sensing configuration depicted in Figure 4 and Figure 10 instead of the FBG sensor. It consists of a sensing head embedded into the Li-ion cell, a fiber-optic link including a 2x2 coupler and an optoelectronic unit composed of a light source, e.g. SLD, photodiode (PD), a transimpedance amplifier (TIA), a control unit and a scanning mechanism, either mechanical (e.g. PZT, voice coil, DC motor), or electronical (e.g. 1 D or 2D CCD array).

The sensing head may be of different designs; here (in the insets in Figure 15) two types: A (Figure 15a) and B (Figure 15b) are presented. The A-type is a Si dice (although some other optically transmissive material may be used as well) adhesively bonded to the sensing fiber tip. This construction is suitable for manufacturing a fiber-optic temperature sensor based on the thermooptic principle of semiconducting materials. Silicon has a rather large thermo-optic coefficient ( 1/n cfn/cfT=4-10 ⁵ K ¹ ), where n denotes the index of refraction of silicone, at a light wavelength of 1300nm. On the other hand, the thermal expansion coefficient of Si is about 3-10 ⁶ m/mK, so both properties contribute to a large change of optical path difference (Dc) in the interferometer. Additionally, the relatively high thermal conductivity of Si of about 160W/mK yields to the shorter response time of the sensor.

The B-type sensing head consist of ceramic housing (e.g. a small diameter fiber-optic ferrule OD < 1 ,25mm) with an adhesively bonded sensing fiber at one side and an expandable rod arranged on the opposite side, which is adhesively bonded to the housing at the rear side only. The rod may be made of dielectric material of high thermal expansion coefficient, e.g. thermoplastic. It is possible to use some other materials including metal wires (e.g. Al and Cu) as well, since they also have a high thermal expansion coefficient. The electrical isolation is preserved since the wire is into the ceramic ferrule enclosed. The front surface may be polished and metalized (e.g. by Al, Ag, Au or some other thin metal or dielectric film) to increase the intensity of the back-reflection signal.

The fiber-optic link is made of an “all-in-fiber” Michelson interferometer using a 2x2 fiber-optic coupler. All fibers in the link are single mode of 10/125pm (core/cladding diameter) optimized for light wavelengths of 1300nm and 1550nm, although some other fiber types and accordingly light sources and wavelengths (e.g. 850nm) can be used as well. Instead of a 2x2 coupler, a 3x3 coupler may be used for the generation of “quasi-quadrature signals” for the sake of avoiding of possible signal fading, as was pointed out at the beginning of this description.

Referring to Figure 15, a broadband light source, e.g. SLD emitting light having a wavelength of e.g. about 1300nm, is a low-coherence source, which is connected to the one of the input arms of the 2x2 coupler. The second input arm is connected to the photodiode accompanied with a TIA. The output arms of the 2x2 coupler are acting as sensing and reference arms. The sensing arm is terminated with the sensing head, e.g. Si dice (A-type) or thermal expanding rod (B-type), while the reference arm is directed versus a scanning mirror. The scanning mirror recovers the position of the orthogonal (in respect to the light beam) front and rear surfaces of the Si dice (A type), or separation between the sensing fiber tip and the front surface of the rod (B type). Back reflected optical beams from the front and rear surfaces of the Si dice (A type) and back reflected beams from the scanning mirror (B type) make a set of interferometric patterns marked from 1 to 4 (see Figure 15c) that are captured by a photodiode. Similarly, back reflected light beams from fiber tip and front surface of the rod (B type) also make corresponding interference patterns. For illustration purposes, the diagram in Figure 15a shows a raw interferometric signal acquired during temperature measurement of Si dice. Out of this result, by simple signal processing, a calibration diagram of the optical path difference (Dc) vs. temperature may be determined (Figure 15c). Separation between the 2 ^nd and 4 ^th interference patterns corresponds to the thickness change of the Si dice on temperature (A type). Similarly, separation between the 1 ^st (the largest one) and the 2 ^nd interference pattern corresponds to the gap change between the fiber tip and the front surface of the dice, i.e. rod on temperature (B type).

The optical path difference, Dc (A-type), is defined by the contribution of the temperature dependent thermo-optic ( k _n ) and thermal expansion coefficient (kj) of Si by: Dc=h I DT (k _n +ki), where n is the refractive index of Si, / is the thickness of the Si dice and DTίe the temperature difference. The optical path difference, Dc (B-type) is practically defined by the physical separation between the fiber tip and the front surface of the rod only. The Dc in both cases depends on the temperature and is determined as a mutual separation between the two characteristic patterns.

Fiber-optic luminescent temperature sensor

Figure 16 schematically shows an additional embodiment demonstrating a fiber-optic luminescent temperature sensor that may also be easily integrated into the basic sensing configuration depicted in Figure 4 and Figure 10, instead of a FBG sensor.

This is an extrinsic sensor based on the measurement of fluorescence decay time of some thermosensitive phosphorescent materials, called phosphors, versus temperature. The sensing material is adhesively bonded to the tip of the fiber and then encapsulated in some high-temperature plastic, e.g. polyimide or Teflon. The photoactive material is excited by visible light emitted from a LED, or some other light source. The phosphor material (e.g. Y ₂ 0 ₃ :Eu, YAG:Eu, YAG:Dy, YAG:Cr, YP0 ₄ :Eu,Dy, etc.) starts to emit a broadband light in the near infrared region, which comes back toward photodetector. The fluorescence decay time of the phosphor directly depends on temperature and this dependence is firstly determined by sensor calibration.