Failure Mechanism and Residual Stress Analysis of Crystal Materials for the Thermal Battery

Abstract

This paper investigates the thermal battery as a research topic. We conducted an in-depth analysis of various thermal battery aspects, such as the cathode material CoS₂ and electrolyte material morphology, crystal type, and interface state changes before and after service. The aim was to explore the core reaction and main failure mechanisms of the thermal battery. Prior to the reaction, the thermal battery cathode and electrolyte material consisted of pure-phase CoS₂ and a composition of MgO-LiF/LiBr/LiCl. After service, the cathode and electrolyte of the single thermal battery exhibited significant morphological alterations caused by the presence of a molten state. The cathode transformed from CoS₂ to Co₃S₄ and Co₉S₈ together with the presence of a marginal quantity of Co monomers visible throughout the discharge process, which was confirmed by means of XRD and XPS analyses. After the reaction, the electrolyte material was primarily made up of LiF, LiBr, and LiCl while the crystal components remained largely unaltered, albeit with apparent morphological variations. As was deduced from the thermodynamic analysis, the cathode material’s decomposition temperature stood at 655 °C, exceeding the working temperature of the thermal battery (500 °C) by a considerable margin, which is indicative of outstanding thermal durability within the thermal battery’s operational temperature range. Furthermore, the discharge reaction of the positive electrode was incomplete, resulting in reduced CoS₂ residue in the thermal battery monomer after service. The reaction yielded a combination of Co₃S₄, Co₉S₈, and small amounts of Co monomers, indicating possible inconsistencies in the phase composition of the pole piece during the reaction process. In this study, we examine the distribution of residual stress in the thermal battery under various operating conditions. The simulation results indicate that exposure to a 70 °C environment for 2 h causes the maximum residual stress of the battery, which had an initial temperature of 25 °C, to reach 0.26 GPa. The thermal battery subjected to an initial temperature of 25 °C exhibited a maximum residual stress of 0.42 GPa subsequent to a 2-hour exposure to a temperature of −50 °C.

1. Introduction

Thermal batteries are backup batteries that use molten salt and are activated by a heating agent. They are commonly used in defense applications such as guidance systems, radar, and electronic equipment, including missiles and torpedoes [1,2]. The electrolyte of a thermal battery at room temperature is a solid, non-conducting inorganic halide salt, typically LiCl-KCl eutectic. However, upon melting, battery impedance decreases sharply, allowing rapid discharge of the battery through an external circuit [3]. Thermal batteries can be activated either electronically or mechanically. After activation, the reaction of the heated agent in the battery generates heat that melts the electrolyte, enabling the battery to enter the operational state. A thermal battery pack comprises a pressurized powder assembly with a positive electrode, a diaphragm, a negative electrode, and a heating agent [4,5]. This is a brief description of the research progress and demand for positive electrode, negative electrode, and electrolyte materials. A wide variety of materials have been applied to the positive electrodes of thermal batteries. These materials include CaCrO₄, K₂Cr₂O₇, K₂CrO₄, PbCrO₄, V₂O₅, WO₃, CuS, FeS₂, and CoS₂ et al. [6,7]. Some of the applicable criteria for cathode materials include high monomer voltage when paired with a suitable negative electrode, good compatibility with molten halogenated salts, and thermal decomposition temperatures near 600 °C. As military equipment continues to upgrade and modernize, the performance requirements for thermal batteries are increasing. The development objectives for thermal batteries are primarily focused on achieving high power and an extended lifespan. The two critical factors that dictate battery power and longevity are the battery’s discharge voltage and polarization magnitude.

Researchers worldwide have extensively studied improving cathode performance [8,9,10], primarily for the mentioned needs. Among the researched cathode materials for thermal batteries, some notable ones include FeS₂ (pyrite), CoS₂, NiCl₂, NiS₂, and vanadium oxides. FeS₂ and CoS₂, the most frequently used materials in recent years, are frequently paired with lithium negative electrodes. FeS₂ is thermally stabilized at 540 °C while CoS₂ is stabilized at 640 °C. FeS₂ is commonly utilized in the preparation of thermal batteries due to its ample natural resources, its ability to be used in electrode preparation even when unpurified pyrite is utilized, its excellent compatibility with molten electrolytes, and its stable discharge performance. However, the thermal instability of FeS₂ at conditions exceeding 500 °C limits its use in thermal batteries [11].

Therefore, given their use in high-temperature environments, it is essential to investigate potential cathode substitutes. Technical abbreviations will be defined upon their initial use. Among all options, CoS₂ displays the most promise as a replacement for pyrite due to its excellent thermal stability, electrical conductivity, and discharge performance. CoS₂’s discharge voltage exhibits a relatively smooth curve with a resistivity of about 0.002 Ω/cm while FeS₂ has a resistivity as high as 17.7 Ω/cm. The low resistivity of CoS₂ cathode materials is beneficial in reducing ohmic polarization during battery discharge, ensuring a smoother voltage over extended discharge periods. As such, the electrochemical properties of CoS₂ have led to its widespread use in developing new, high-specificity engineered thermal batteries [12].

Before the 1980s, most thermal battery designs utilized calcium metal anodes, typically by attaching calcium foils to metal current collectors, such as iron, stainless steel, or nickel. Another commonly used anode material is magnesium metal, which can be made into a bimetallic anode and also pressed or spot-welded onto the substrate. The development of new technology has gradually eliminated this complex process from the market. In the late 1970s, lithium became increasingly prevalent as an anode material for thermal batteries. Two primary forms of lithium anode are utilized: lithium alloy and pure lithium metal. The most commonly used lithium alloy anodes are those comprised of lithium aluminum containing 20 wt% lithium and lithium silicon containing 44 wt% lithium. These powder-based LiAl and Li (Si) alloys are cold-pressed and molded into anode sheets of 0.75 mm to 2.0 mm in thickness. One side of the lithium alloy negative electrode in batteries comes into contact with an iron, stainless steel, or nickel collector [13,14,15].

During thermal battery operation, the lithium alloy negative electrode takes the shape of a solid negative electrode [16]. Therefore, the temperature of the negative electrode must be below its melting point or the temperature at which it is only partially melted. For instance, a lithium-silicon alloy consisting of 44% lithium by weight partly liquefies at a temperature of 709 °C whereas α-LiAl and β-LiAl partially liquefy at 600 °C. If the aforementioned melting points are surpassed, the melted negative electrode will directly touch the positive electrode material, resulting in a monomer short-circuit. The operating temperature of a thermal battery typically exceeds the melting point of lithium metal, which is 180.5 °C. To avert short-circuiting caused by flowing molten lithium metal, a high specific surface area metal powder or metal sponge is essential to adsorb the lithium when utilizing pure lithium metal as the negative electrode [17]. Mainly, the surface tension of the bonding material adsorbs the molten lithium metal. Li-Fe negative electrodes were developed for thermal battery preparation. In the Li-Fe negative electrode, 15–30 wt% Li is primarily adsorbed through the surface tension of Fe powder. These anodes are referred to as “LAN” anodes. Li-Al alloys and Li-Si alloys have since been utilized as negative electrodes in batteries, exhibiting exceptional performance [18,19].

Thermal battery designers have utilized electrolytes made from the eutectic salt LiCl/KCl (wLiCl: wKCl = 45:55; melting point of 352 °C) composed of lithium chloride and potassium chloride [20]. Lithium-containing halide salts are commonly chosen constituent salts due to their high ionic conductivity and compatibility with anode and cathode materials. Compared to numerous oxygen-containing salts with low melting points, the blend of halogenated salts demonstrates less susceptibility to decomposition at elevated temperatures resulting in gas release and other secondary reactions. Recently, researchers have been incorporating bromide into electrolytes to obtain low melting point electrolytes. This extends the operating time of thermal batteries and reduces the internal resistance, thus improving the current loading capacity [21,22].

Bromide-containing electrolytes include LiF/KBr/LiBr (melting point of 320 °C), LiCl/KBr/LiBr (melting point of 321 °C), and the all-lithium electrolyte LiF/LiCl/LiBr (melting point of 430 °C) [23]. Electrolytes containing a mixture of cations, such as Li⁺ and K⁺ instead of all Li⁺, have a tendency to produce a concentration gradient of Li⁺ during the discharge process. Thermal batteries can bias eutectic salt composition due to the Li⁺ concentration gradient at high current output, causing early electrolyte eutectic salt solidification. At the operating temperature of a thermal battery, the viscosity of the molten salt electrolyte is very low. Adding a binder is necessary to prevent the molten salt electrolyte from flowing. Early on, kaolin and gasified silica clays were utilized in the Ca/CaCrO₄ and LiAl/FeS₂ systems. All of these silicate substances cause a reaction to Li (Si) alloys and lithium metal anodes. MgO, which has a high specific surface area and is relatively inert to active metal anodes, is currently the chosen molten salt electrolyte binder for the majority of systems [24].

Therefore, it is highly significant for optimizing the development and processes of thermal batteries to conduct comprehensive analyses of morphology, composition, crystalline structure, interface, and other material aspects of positive and negative electrodes and electrolytes, both before and after use. These analyses lend themselves to exploring the central reaction mechanism of thermal batteries.

Li and colleagues found that the electric capacity of the thermal battery decreased with an increase in the number of accelerated cycles. The thermal images presented indicate that the discharge process of the thermal battery after the equivalent accelerated storage test resulted in a higher surface temperature than the initial discharge surface temperature. The particle filtering algorithm can accurately model the degradation of electric capacity, producing predictions for remaining battery capacity that deviate by no more than 10% from test results [25].

Xing and colleagues examined the impact of high-temperature accelerated storage on thermal battery performance, showing an acceleration in reactions between the lithium-alloy anode and water that resulted in a shortened operating time. Moreover, storage at those temperatures increased the oxidation of activated iron powder, resulting in a longer activation time for the batteries [26].

Zhong and his colleagues developed an accelerated life testing methodology for thermal batteries, utilizing the temperature’s impact on reaction rates. They fitted the experimental data using mathematical analysis software, establishing a fourth-order regression equation. This equation was subsequently used to design the thermal batteries, thereby facilitating the control and prediction of their storage life. The study discovered that the thermal battery’s operational lifespan decreases to a certain threshold with the increase in storage time. However, it then recovers to a certain extent, which deviates from the anticipated linear or asymptotic decline [27].

Ye and colleagues addressed the challenge of accurately measuring temperature increases on thermal battery structural surfaces during operation by developing a digital thermal battery and component model utilizing the COMSOL 6.0 software. They then performed a triple-coupled simulation of temperature changes on the structural surface based on the physical fields of radiation–convection–conductivity. Finally, an epoxy glass cloth plate made of 3240 material was utilized as a mask for constructing the thermal battery structure for measuring the real temperature. This was then compared to the simulation results [28].

Addressing the issue of low efficiency and accuracy in manually detecting defects in thermal battery assembly, Zhou and colleagues propose a method for detecting thermal battery assembly defects through image recognition. First, the battery and template images undergo enhanced grayscale normalization. The single battery stack region is accurately extracted through the template library and adaptive template matching methods. Secondly, this study examines the three common defects that arise during the assembly process of thermal batteries: inversion, omission, and incorrect ordering. Structural features of single cells are extracted for self-comparison and combined with the battery defect library matching method to classify the results. As a result, X-ray imaging can be used for defect detection in thermal batteries. Test results indicate a high level of accuracy and robustness using this method, with a 96% detection rate [29].

This paper investigates the thermal battery pack and conducts research on its mechanisms. Using materials science techniques, such as SEM, XRD, EDX, and XPS, we conducted a thorough analysis of the positive and negative electrodes, electrolyte, and key components of the faulty thermal battery, as well as the key interfaces to explore the main failure mechanism and core reaction mechanism of the battery. Additionally, we used finite element modeling to analyze the residual stress distribution of the thermal battery under different operating conditions.

2. Experimental Methodologies

2.1. Equipment for the Experiment

The experiment was carried out using the following devices: X-ray Diffraction Analysis (XRD), Scanning Electron Microscopy (SEM), Energy Dispersive X-ray Spectroscopy (EDX), X-ray Photoelectron Spectroscopy (XPS), and Thermogravimetric Analysis (TGA).

X-ray Diffraction Analysis (XRD) is a non-destructive method that characterizes the crystal structure and chemical composition of substances. It allows for the quantitative and qualitative analysis of materials. The instrument utilized for this experiment was a Bruker D8a X-ray diffractometer. The machine was manufactured by the Bruker Company and operated under the following working conditions. Cu-Kα radiation was produced by a source with a wavelength (λ) of 0.15406 nm whilst employing a working tube current of 40 mA and a tube voltage of 40 kV, with the continuous scanning mode at 4°/min. Data were gathered between 10° and 80°. The physical phases were identified and macrostructural information was analyzed using MDI Jade 5.0 software.

Scanning Electron Microscopy (SEM) is a technique used to observe the surface morphology of a sample. A high-energy electron beam is directed toward the sample’s surface, which produces physical signals that are detected and imaged on a fluorescent screen. Additionally, the distribution and content of various elements within the sample are analyzed using X-ray Energy Dispersive Spectroscopy (EDX). This paper utilizes a German ZEISS Merlin microscope at operating conditions of 5 kV and an objective distance of 6 mm.

Energy Dispersive X-ray Spectroscopy (EDX) analyzes the composition and element distribution in a micro-region of a sample using X-rays generated by an electron beam. It provides insight into the sample’s surface composition and other attributes. The model Apollo from EDAX was used in this thesis.

X-ray Photoelectron Spectroscopy (XPS) can analyze the chemical elements on a material’s surface qualitatively and semi-quantitatively. The instrument used in this paper is the Axis Ultra DLD spectrometer, which indirectly analyzes the electrochemical reaction process by measuring changes in the valence state of the chemical elements on the material’s surface in various states.

Thermogravimetric Analysis (TGA) involves heating a sample under temperature control to produce a curve of mass changes with temperature. This method is used to analyze thermal stability, decomposition temperature, decomposition products, and the composition of substance components. The SDT Q600 TG-DTA instrument with a temperature range from room temperature to 800 °C and a rate of 10 °C/min was used in this experiment.

2.2. Characterization of Cathode and Electrolyte Materials for the Thermal Battery

2.2.1. Characterization and Analysis of the Cathode Materials

In this study, comprehensive material characterization of the thermal battery cathode and electrolyte material pre-reaction was conducted using SEM, XRD, EDX, and other analytical tools to explore the physical morphology, composition, and crystalline type of materials [12,30]. This information was gathered prior to the reaction to obtain an accurate baseline for future analysis.

The crystal structure analysis of the cathode material was conducted, with the X-ray Diffraction Analysis of this material provided in Figure 1. The figure shows that the (200) and (311) crystal planes of CoS₂ can be attributed to the main peaks around 32.32° and 54.98° while the crystal planes of (111), (210), (211), and (220) can be attributed to the main peaks at 27.89°, 36.26°, 39.86°, and 46.36°, respectively. These results suggest that CoS₂ displays favorable octahedral crystallinity (JPSD no. 89-1492). The set of diffraction peaks is well defined and contains almost no extraneous peaks, indicating that the cathode material is primarily composed of pure-phase CoS₂. A minimal amount of LiF-LiCl-LiBr was detected via XRD while a small quantity of Li₂O was also present (JPSD no. 77-2144).

Figure 2 depicts the SEM images of the cathode material. Notably, Figure 2a,b portray CoS₂ in the form of lumps that exhibit a particle distribution ranging from 2 to 20 μm. Upon close examination of the enlarged image, we can observe that the bulk CoS₂ is composed of small spherical particles that cohere and form aggregates. Moreover, the granular form of CoS₂ is clearly visible on the surface, as portrayed in Figure 2c.

Figure 3 displays the X-ray Photoelectron Spectroscopy (XPS) images of the cathode materials along with the comprehensive spectrum of XPS presented in Figure 3a. This figure showcases the Co 2p, F 1s, O 2p, C 1s, S 2s, S 2p, Cl 2p, and Br 3d photoelectron spectra in a clear and concise manner without the use of any convoluted terminology. The spectra for Li 1s, S 2p, and Co 2p were obtained and displayed individually in Figure 3b–d. The cathode materials revealed the presence of LiF, LiCl, and LiBr, which were identified based on characteristic peaks located at 49.6, 54.9, and 56.8 eV, respectively. Furthermore, the spectra also demonstrated the presence of Li-O (53.1 eV) and Li-N (53.4 eV) peaks.

The X-ray Photoelectron Spectrum (XPS) of sulfur 2p presented in Figure 3c exhibits peaks centered at 167, 162.2, and 161 eV, corresponding mainly to the S-O, S-S, and Co-S bond types. Figure 3d displays the XPS of cobalt 2p in CoS₂, which contains two spin bimodal peaks (Co 2p3/2 and Co 2p1/2) and two recombinant satellite peaks. The latter have centers at 778 eV. The peaks at 2 and 793.3 eV indicate the presence of Co in its trivalent state (Co³⁺) whereas the peaks ascending at 780.2 and 795.4 eV indicate the presence of Co in its divalent state (Co²⁺). Additionally, there are broad satellite peaks located at 782.6 and 801.0 eV. The Co³⁺/Co²⁺ ratio obtained by XPS is 126.8%.

For the stability of the battery’s cathode material, Figure 4 displays the thermogravimetric analysis outcomes of cathode material specimens, exemplifying that the cathode material typically decomposes at 655 °C, which is the same as CoS₂’s decomposition nature. Significantly, this cathode material’s decomposition temperature is approximately 100 °C higher than FeS₂’s. The thermal battery for electrolysis typically operates at 500 °C and the decomposition temperature of CoS₂ exceeds this threshold. This indicates exceptional thermal stability of the cathode material. Therefore, the cathode material will not decompose at the working temperature of the thermal battery.

2.2.2. Characterization and Analysis of the Electrolyte Materials

The diffraction peaks on the MgO planes were obtained and the primary peaks were located at 36.86°, 42.82°, 62.20°, 74.51, and 78.44, corresponding to the (111), (200), (220), (311), and (222) crystalline planes, respectively. The X-ray Diffraction Analysis of the electrolyte is displayed in Figure 5 (JPSD no. 75-0447). It is worth noting that the diffraction peaks are highly evident, suggesting that the electrolyte includes MgO as the binder. LiF (JPSD no. 72-1538) and LiBr (JPSD no. 74-1973) were also found in the electrolyte, with the main peaks of LiF being located at 38.77° and 45.07° and the main peaks of LiBr being located at 28.14° and 32.60°.

The diffraction peaks of the electrolyte are clearly visible, with MgO, LiF, LiBr, and LiCl as the primary substances detected. While LiCl’s signal was faint and not observed, the remaining substances registered corresponding peaks. Furthermore, minute amounts of Li₃OBr, Li₂O, Li₂O₂, and Li₃N were produced during the detection and operational processes’ exposure to air and these compounds were also detected.

Figure 6 displays the SEM image of the electrolyte, which exhibits that the electrolyte material contains small flaky particles that are clustered together. Furthermore, a considerable amount of tiny particles are absorbed on the surface of the flaky material. Additionally, Figure 6a,b visually illustrate the above-mentioned facts. Upon closer inspection of the block’s surface, Figure 6c reveals that MgO functions as a binding agent among the electrolytes to join tiny particles, specifically LiF, LiBr, LiCl, etc., and creates bigger agglomerated flakes.

2.3. Characterization of the Thermal Battery after Service

This chapter conducts a thorough material characterization of thermal battery monoliths after use, utilizing SEM, XRD, and EDX analyses to examine the physical morphology, composition, and crystalline shape of failed post-reaction components [31,32]. The cathode side and electrolyte side were labeled based on EDX composition distribution as specific information on the stacking method of the single battery was not obtained.

2.3.1. Characterization and Analysis of the Cross-Section of the Thermal Battery after Service

The cross-section of the single thermal battery post-service is displayed in Figure 7a, with the cathode, electrolyte, anode, and heating agent layers fused together. Based on the SEM lining depth and pore size, it is distinctly separated into two layers: the upper layer is the electrolyte layer and the lower layer is the positive electrode layer. Notably, the electrolyte layer demonstrates a loose and porous morphology, as depicted in Figure 7b. Compared to the flaky microstructure of the electrolyte prior to the reaction, the shape of the electrolyte significantly altered post-reaction. The flaky pattern disappeared entirely, giving rise to a more porous and loose consistency. This phenomenon can be attributed to the electrolyte remaining in a molten state throughout the reaction and ultimately recondensing from that state at the conclusion of the reaction.

Figure 8 displays the SEM image of the lower cross-section of the thermal battery after use, with Figure 8a indicating the observation position of the SEM selected in the middle. The main component of this position is the CoS₂ cathode material product after undergoing a reaction. Figure 8b illustrates that the post-reaction product manifests itself in the form of a denser “coral” in comparison to the post-reaction electrolyte. Compared to the CoS₂ particles’ adherent aggregation before the reaction, the reacted Co₃S₄ and Co₃S₄ particles are more tightly bonded and dissolve with each other to form a coral-like structure. The surface morphology of the enlarged coral is presented in Figure 8c, displaying a smoother and glossier surface with tiny particles visible.

As illustrated in Figure 9, the cross-section underwent Energy Dispersive X-ray Spectroscopy (EDX) to observe the element distribution in the thermal battery’s cross-section. Figure 9a displays an electron micrograph of the sample, with a distinct difference between the top and bottom sides of the yellow line. Figure 9b,c exhibit the distribution of cobalt and sulfur elements in the sample, signifying that the cathode material CoS₂ is the reaction product in the middle region. B-elements are infrequently dispersed within the battery with overlapping sample stages that cannot be explained. The distribution of Mg and O in samples is depicted in Figure 9d,e, implying that MgO is spread across the battery but is more prevalent in the region below the cathode side. Cl and Br are substantially uniformly dispersed and F is highly concentrated on the electrolyte side of the thermal battery.

Elemental content in the thermal battery post-service is depicted in Figure 10. The thermal battery includes reacted elements such as B, C, F, O, Mg, Br, S, and Cl, among others. The most abundant elements are B, C, and F, making up 28.86 wt%, 30.70 wt%, and 14.66 wt% of the thermal cell, respectively. Technical term abbreviations have been explained upon first use. Additionally, O, Mg, Br, S, Cl, and Co have a significantly lower contents of 5.48 wt%, 2.40 wt%, 2.37 wt%, 2.72 wt%, 1.77 wt%, and 7.96 wt%, respectively.

2.3.2. Characterization and Analysis of the Cathode Side

The X-ray Diffraction Analysis of the positive surface of the thermal battery after maintenance is depicted in Figure 11. The thermal battery composition post-maintenance is quite intricate and initial inferences are derived through the meticulous comparison of peak positions. First, the main reaction products on the surface of the positive electrode in the post-service thermal battery were Co₃S₄ (JPSD no. 74-0138, with the main peak near 33.87° and small diffraction peaks near 24.57° and 25.26°) and Co₈S₉ (JPSD no. 75-2023, with the main peaks at 29.88° and 52.19°). The reaction yielded Co₃S₄ as the dominant product, with only a small quantity of incompletely reacted CoS₂ observed. There is a very small amount of metal Co (JPSD no. 88-2325) and a small amount of CoO (JPSD no. 75-0419) produced during air exposure in the reaction products. Also, the product contained impurity peaks of CoS (JPSD no. 75-0605), S (S6 and S, JPSD no.72-2402, JPSD no. 76-0183), and CoSO₄ (JPSD no. 72-0219).

More distinct diffraction peaks of electrolyte materials were found in the X-ray Diffraction Analysis (XRD). Technical term abbreviations are explained on their first occurrence. The identified materials primarily comprise LiF (JPSD no. 72-1538) and LiBr (JPSD no. 74-1973). LiCl was nearly imperceptible but a small quantity of Li₂O (JPSD no. 77-2144) formed during air exposure was detectable. This issue may arise due to the molten state of the electrolyte during the operation of the thermal cell, leading to partial fusion of the cathode material with the electrolyte. Figure 9 shows a significant amount of elemental distributions containing Mg, Cl, and Br still present on the bottom surface of the cathode material.

During the reaction process of the thermal cell, a large number of intermediate products, such as Co₃S₄ and Co₉S₈, are generated, along with a small amount of metal Co, CoS, S, and CoSO₄.

2.3.3. Characterization and Analysis of the Anode Surface

The X-ray Diffraction Analysis of the negative electrode surface of the thermal battery after use is presented in Figure 12. Despite being used, the negative electrode surface of the thermal battery still displays distinct and intense diffraction peaks of LiF, including the main peaks of LiF (JPSD no. 88-2298). The crystal planes of (200) and (220) can be attributed to the peaks located at 44.98° and 65.51°, respectively, as evidenced by their high intensity. Additionally, the negative electrode surface displayed clear detection of MgO (JPSD no. 75-0447) and LiBr (JPSD no. 74-1973), with MgO’s main peaks located at 36.86°, 42.82°, and 62.17° and LiBr’s main peaks located at 28.13° and 77.74°.

Li₂O (JPSD no. 77-2144), Li₂O₂ (JPSD no. 74-0115), and Li₃N (JPSD no. 78-2005) were also found on the negative surface due to air oxidation. The major peaks of Li₂O were at 33.58° and 67.16° while those of Li₂O₂ were at 40.66° and 58.72°. The main peaks of Li₃N were located at 28.26° and 76.69°.

3. Finite Element Methods

3.1. Geometric Modeling of the Thermal Battery

The thermal battery’s geometric schematic is displayed in Figure 13. In terms of creating the ANSYS Parameter Design Language (APDL) file,

Table 1. Designations and reference values for geometric parameters of the single thermal battery.

	Geometric Parameter	Designation	Reference Value (mm)
Positive electrode	Radius	R1	50
Electrolyte		R2
Negative electrode		R3	3
Collector plate-4		R4	50
Iron powder		R5
Positive electrode	Height	H1	1
Electrolyte		H2	1.5
Negative electrode		H3	0.5
Collector plate-4		H4
Iron powder		H5	1

and

Table 2. Designations and reference values for partial geometrical parameters of the thermal battery pack.

	Geometric Parameter	Designation	Reference Value (mm)
Asbestos cushion-1	Radius	R6	50
Upper fixing plate		R8
Zircon powder heater-1		R9	3
Under the side insulation layer		R12	50
Asbestos cushion-1	Height	H6	2
Upper fixing plate		H8	3
Zircon powder heater-1		H9	1.5
Lower fixing plate		H12	2
Upper insulation layer		H18	6
Under the side insulation layer		H19	2
Air layer	Thickness	T16	1.5
Side insulation layer		T17	3
Battery case		T20	1

present the variable names and reference values of each component’s geometrical parameters.

The 3D computer-aided design (CAD) model of the thermal battery is constructed using the reference values of the geometric parameters of each component presented in Table 1 and Table 2. Initially, the 3D CAD model of a single battery is developed and its meridional slice diagram, as shown in Figure 14a, includes a positive electrode layer, an electrolyte layer, a negative electrode layer, and a current collector layer. Based on the 3D CAD model of a single cell, we developed a 3D CAD model of a battery pack comprising N1 single cells and other components (with N1 being set to 17 for this study). Figure 14b illustrates a cross-sectional view of this model, including all the components presented in Figure 13.

The simulation model comprises various material parameters, including the modulus of elasticity, Poisson’s ratio, density, thermal conductivity, specific heat capacity, and coefficient of thermal expansion. The technical documents provide physical parameters of the materials, primarily specific heat capacity, density, and thermal conductivity. For the unspecified material parameters, we obtained most of them through a review of the related literature and information [33,34]. For the few parameters with no available information, we estimated them conservatively. The simulation model can be adjusted according to needs during subsequent iterations and applications. Please refer to

Table 3. Parameters of physical properties of materials for simulation modeling.

Material		Modulus of Elasticity (GPa)	Poisson’s Ratio	Density (g/cm3)	Thermal Conductivity [W/(m·°C)]	Specific Heat Capacity [J/(kg·°C)]	Coefficient of Thermal Expansion (10−6/°C)
Positive electrode		10	0.33	2.76	1.02	740	1
Electrolyte				2.49	1.22	870
Negative electrode				1.00	27.4	3330
Stainless steel (current collector, upper fixing plate, lower fixing plate, fixed connection strip, battery case)		197	0.27	7.86	21.64	480	13
Iron powder		150	0.25	3.75	21.85	710	10
Asbestos cushion		1	0.33	1.00	0.2	1000	1
Zircon powder heater	1	10	0.33	1.07	1	2000	1
	2			1.45
	3
	4			1.60
Air layer		10−9	0.33	1.21	0.0257	1005	10−54
Side insulation layer		10	0.33	0.26	0.03	1100	1
Upper and lower insulation layers				0.86	0.05	1200

for the values of specific physical property parameters.

3.2. Modeling of Finite Elements

When utilizing finite element analysis software, the initial step typically involves characterizing the CAD model of the material component to be simulated, as well as discerning its material properties, imposed loads, constraints, and other relevant data. This collection of information generally entails an exceptionally high degree of accuracy and can prove instrumental in forecasting real-world phenomena. Indeed, the level of accuracy achieved based on any given FEA model is directly linked to the finite element mesh utilized. A finite element mesh partitions a CAD model into small domains called cells. Then, a set of equations is solved on these cells to approximate the desired governing equations through polynomial functions defined on each cell. Refining the mesh results in smaller cells, approaching a solution closer to the true value [35].

ANSYS offers two meshing methods: free meshing and mapping meshing. Mesh division for mapping mesh follows strict rules with a demanding process that necessitates targeted human input. When the geometric size or shape changes, there is often a need to readjust the mesh division. In contrast, free mesh division has a stronger adaptive ability and is easier to modify, especially for complex structures. With high automation, mesh division is controlled by software algorithms at specified size levels, reducing the need for specialized division skills. This results in easy quality control of the resulting mesh, even when the geometric size or shape changes. After specifying the desired size, the software algorithm controls the mesh division process, which does not demand advanced division skills and offers easy control over mesh quality. Even when geometric size or shape changes, the mesh division method requires minimal adjustments, allowing for automatic mesh division for the same topological structure. In parametric simulation models, the free meshing method presents more benefits [36].

The ANSYS study employs the intelligent free meshing method. To initiate meshing, the intelligent mesh size level must be specified, an integer between 1 and 10, where mesh size increases from 1 (fine mesh) to 10 (coarse mesh). In this study, it is set to 8 and applied to all parts.

Due to differing material properties among components, each must have its mesh properties specified before generating the finite element mesh. The initial step is to generate the finite element mesh for a single cell, as illustrated in Figure 15. Due to the small height of the positive and negative electrode sheets, electrolyte sheet, and current collector, it is necessary to regulate the mesh size to prevent it from having a large aspect ratio. Therefore, the initial mesh size must be specified carefully. This study sets the initial grid size to R1/10, with R1 representing the positive electrode sheet’s radius. The same setting is used throughout the entire process of finite element mesh generation.

The finite element mesh of all N1 single cells is generated sequentially from the bottom to the top through a cyclic method based on a single cell. Then, we generate finite element meshes for the iron powder heating sheet, asbestos pad-2, zirconium powder heating sheet-3, lower fixing plate, zirconium powder heating shee-4, asbestos pad-1, zirconium powder heating sheet-2, upper fixing plate, zirconium powder heating sheet-1, and fixing connecting strips, successively, to complete the finite element mesh construction of the basic skeleton illustrated in Figure 16a.

Based on the finite element mesh of the foundational skeleton, the other components, including the upper insulation layer, asbestos mat-3, lower insulation layer, air layer, side insulation layer, and battery shell, are incrementally generated to produce the outer envelope finite element mesh of the thermal battery pack. Refer to Figure 16b for details.

4. Results and Discussion

4.1. Reaction Mechanisms of the Thermal Battery

In this study, Figure 1, Figure 11, and Figure 13 depict the X-ray Diffraction Analysis conducted on the cathode surface of the thermal battery before and after it was in use. The composition of the cathode material underwent significant changes during the discharge process, transitioning from a nearly pure phase of CoS₂ to composite products, such as Co₃S₄, Co₉S₈, and Co, along with impurities like CoS, S, and CoSO₄.

The equation for discharging a thermal battery is as follows:

$$\mathrm{C}\mathrm{o}{\mathrm{S}}_2+\frac{4}{3}{\mathrm{e}}^{-}\to \frac{1}{3}{\mathrm{C}\mathrm{o}}_3{\mathrm{S}}_4+\frac{2}{3}{\mathrm{S}}^{-2}$$

(1)

$${\mathrm{C}\mathrm{o}}_3{\mathrm{S}}_4+\frac{8}{3}{\mathrm{e}}^{-}\to \frac{1}{3}{\mathrm{C}\mathrm{o}}_9{\mathrm{S}}_8+\frac{4}{3}{\mathrm{S}}^{-2}$$

(2)

$${\mathrm{C}\mathrm{o}}_9{\mathrm{S}}_8+16{\mathrm{e}}^{-}\to 9{\mathrm{C}\mathrm{o}}^0+8{\mathrm{S}}^{-2}$$

(3)

During the thermal battery reaction process, CoS₂ experiences two discharge platforms. The first platform produces the Co₃S₄ phase, which has a higher resistivity than CoS₂, resulting in a significant drop in voltage on the discharge curve. Soon after, the second discharge platform generates both Co₉S₈ and the Co singlet phase. Technical abbreviations will be explained as they are introduced.

The changes of elemental valence states in X-ray Photoelectron Spectroscopy indicate the simultaneous evolution: the Co³⁺ peak weakens before and after service, the low-valence Co²⁺ peak strengthens, and the XPS peaks of Co monomers emerge. Additionally, there is a significant shift in the main composition of S 2p from sulfur elements in the initial state of CoS₂ and the Co-S bond reforms into S²⁻. The components of LiF-LiCl-LiBr and MgO crystals in the electrolyte remain largely unaffected; although, significant morphological transformations occur upon melting.

Based on the previous characterization analysis, it is evident that the discharge reaction was incomplete and resulted in some residual CoS₂. Additionally, reaction products were found in the first, second, and third platforms, indicating a possible inconsistency in the reaction stages within the pole piece during the reaction process.

4.2. Analysis of Residual Stress

The research conducted an analysis of residual stress under two working conditions and presented the respective parameters in

Table 4. Working conditions for the analysis of thermal stress in the thermal battery.

	Working Condition 1	Working Condition 2
Initial temperature (°C)	25	25
Ambient temperature (°C)	70	−50
Duration of the test (h)	2	2
Coefficient of convective heat transfer [W/(m2·°C)]	13.8	13.8

.

To conduct a residual stress analysis, the initial step is to solve for the temperature field under specified initial margin conditions. Following this, structural calculations are executed, utilizing the temperature field results for a specified time period, to obtain the residual stress distribution.

It is essential to solve for the temperature field before proceeding. The solution process implements automatic time-step tracking, which sets the initial sub-steps to one-hundred, the maximum sub-steps to one-thousand, and the minimum sub-steps to two (number of sub-steps = test time/time step).

The results of the temperature field solution for Condition 1 are presented in Figure 17. The simulation results of the temperature field indicate that the thermal cell’s initial temperature of 25 °C increased after being subjected to a 70 °C environment for two hours. The temperature of the entire thermal battery increased and the increase progressed radially from the axis toward the outer surface. The axial temperature experienced an increment of 53.3 °C while the outer surface temperature reached 64.7 °C.

The temperature field solution outcomes for Condition 2 are displayed in Figure 18. The simulation results indicate that the thermal cell’s initial temperature is 25 °C. After subjecting the battery to an environment with a temperature of −50 °C for 2 h, the temperature of the entire battery decreases. The reduction increases radially from the axial center to the outer surface. The axial center’s temperature is approximately −22.1 °C and the outer surface’s temperature is about −41.1 °C.

After analyzing the temperature field, the residual stresses are resolved by transforming the thermal battery into a structural one and applying the Z—direction displacement constraint to its bottom surface. To start the residual stress solution, the nodal temperature from the thermal analysis results is read in and the reference temperature is set to 25 °C.

The nephogram depicting the distribution of residual stress (Von Mises stress) for Condition 1 is exhibited in Figure 19. According to the residual stress simulation results, the maximum residual stress is 0.26 GPa for the thermal battery that was initially at 25 °C after being exposed to an environment with a temperature of 70 °C for 2 h.

Similarly, the distribution of residual stress (Von Mises stress) for Condition 2 is illustrated in Figure 20. The simulation results indicate that the highest residual stress attained is 0.42 GPa in the thermal cell, with an original temperature of 25 °C, after being exposed to a temperature of −50 °C for 2 h.

5. Conclusions

This paper investigates the thermal battery pack and conducts research on its mechanisms. Using materials science techniques, such as SEM, XRD, EDX, and XPS, we conducted a thorough analysis of the positive and negative electrodes, electrolyte, and key components of the faulty thermal battery, as well as the key interfaces to explore the main failure mechanism and core reaction mechanism of the battery. Additionally, we used finite element modeling to analyze the residual stress distribution of the thermal battery under different operating conditions. Our conclusions are as follows:

• The cathode material is made up of pure—phase CoS₂, exhibiting a decomposition temperature of 655 °C, which is higher than that of FeS₂ and similar cathode materials. Moreover, this temperature is far higher than the operational temperature of the thermal battery (500 °C), indicating that the cathode material is highly thermally stable. The electrolyte material comprises mainly MgO, LiF, LiBr, and LiCl. The particles aggregate in a flaky morphology that is dominant. MgO primarily solidifies the molten salt electrolyte under the thermal battery’s working temperature and does not change significantly during the working interval;
• The analysis mechanism indicates two main discharge platforms for CoS₂ thermal battery discharge. The first platform yields the Co₃S₄ phase while the second platform produces phases of Co₉S₈ and Co monomer. The discharge process of this sample transitions from CoS₂ to Co₃S₄ and Co₉S₈, along with a small quantity of Co monomers, as confirmed by XRD and XPS analysis. After the reaction, the electrolyte material consists mainly of LiF, LiBr, and LiCl and the crystal components remain largely unchanged. However, noticeable porous and loose morphologies are present, indicating significant changes in morphology during the melting process;
• The discharge reaction of the positive electrode was incomplete, resulting in reduced CoS₂ residue in the thermal battery monomer after service. The reaction yielded a combination of Co₃S₄, Co₉S₈, and small amounts of Co monomers, indicating possible inconsistencies in the phase composition of the pole piece during the reaction process;
• The simulation results indicate that exposure to a 70 °C environment for 2 h causes the maximum residual stress of the battery, which had an initial temperature of 25 °C, to reach 0.26GPa. The thermal battery subjected to an initial temperature of 25 °C exhibited a maximum residual stress of 0.42 GPa subsequent to a 2—hour exposure to a temperature of −50 °C.