Alumina foam-based electrolyte for thermal batteries: design and performance evaluation Authors: Tae-Young Ahn, Hye-Ryeon Yu, Sang-Hyeon Ha, Jae-In Lee, Hyun-Ki Yoon, Yusong Choi Published: 2026 ABSTRACT Thermal batteries are predominantly used in defense applications owing to their exceptional shelf life and high-power discharge capabilities. However, the electrolytes conventionally used in these batteries are typically fabricated via cold pressing, which imposes significant manufacturability limitations, particularly when aiming for thin and wide formats, and they require labor-intensive processing. Although numerous approaches have been explored to address these constraints, commercially viable alternatives have yet to be realized. In this paper, we present a strategy in which porous alumina foam, infiltrated with molten salt, functions as a high-performance electrolyte suitable for thermal battery systems. We characterized the chemical compositions and microstructures of the ceramic foam and thermal battery unit cells using scanning electron microscopy combined with energy dispersive spectroscopy. Through a systematic evaluation, we identified the pore structure, separator thickness, and porosity of the foam as the key parameters governing the discharge behavior. Our approach presents a paradigm shift in electrolyte fabrication in that it offers a scalable and structurally robust alternative to conventional pellet-pressing techniques.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-56494-6. Introduction Thermal batteries are classified as reserve-type power sources that remain inactive under ambient conditions and become operational only after the rapid internal temperature increase, initiated by the ignition of a pyrotechnic heat source, melts the solid-state electrolyte1,2. In these systems, inorganic molten salts serve as the electrolyte; this class of materials is currently regarded as a highly promising platform for next-generation energy storage due to its exceptional ionic conductivity and thermal stability at elevated temperatures. As shown in Fig. 1, the activation of a thermal battery involves the rapid transfer of thermal energy from the heat source to the cell components. This heat melts the salt electrolyte into a molten state, enabling ion conduction between the Li-Si anode and the FeS2 cathode. Typically, these batteries are designed to operate in a temperature window of 400–550 °C, ensuring stable electrochemical performance during high-power discharge. Fig. 1Schematic illustration of a typical thermal battery unit cell and its operating principle. Schematic illustration of a typical thermal battery unit cell and its operating principle. Owing to their intrinsic reserve-type characteristics and rapid activation capability, thermal batteries have been primarily developed as high-power energy sources for military applications since World War II1. These batteries are uniquely suited for defense systems, such as guided missiles and torpedoes, because they offer an exceptionally long shelf life that often exceeding 10 years without maintenance and provide high power density instantly upon activation. Furthermore, their ability to operate reliably under extreme mechanical vibrations and wide temperature ranges makes them indispensable for mission-critical military hardware1,2. As the technological requirements of defense applications have become increasingly demanding, research efforts have focused on improving the performance of thermal batteries, with significant emphasis placed on developing efficient anode and cathode materials3–6. However, compared with electrode materials, few attempts have been made to improve the performance of electrolytes for thermal batteries, despite the vast potential for these molten salt systems to be expanded into various high-temperature energy storage applications beyond their traditional military use7. The electrolytes in commercial thermal batteries are conventionally fabricated via cold pressing using molten salts with high ionic conductivity2. Leakage at the elevated operating temperatures of thermal batteries (typically 400–550 °C) is mitigated by incorporating ceramic binders such as magnesium oxide (MgO) and aluminum oxide (Al2O3) to stabilize the electrolyte structure2. The surface modification of these ceramic binders has been the primary research focus aimed at suppressing molten salt leakage during the high-temperature operation of thermal batteries8. In several studies, the use of ceramic felts as alternative electrolytes for thermal batteries has been explored; however, these approaches have not yet led to commercial adoption9. According to Guidotti and Reinhardt10, these flexible fibrous architectures often suffer from severe warpage and cracking due to high stresses generated during electrolyte solidification, compromising their structural integrity. Sandia National Laboratories has conducted extensive research to overcome the limitations of cold-pressed electrolytes used in thermal batteries. These conventional fabrication methods involve labor-intensive processing, limited scalability for large-area shaping, and restricted control over geometric uniformity11. Moreover, reduced electrolyte thickness can lead to structural deformation when the eutectic salt melts, thereby increasing the risk of leakage under high-shock conditions. Furthermore, enhancing the thermal stability of advanced energy storage systems has emerged as a critical research focus to prevent catastrophic failures under extreme conditions. Recently published comprehensive reviews have demonstrated that incorporating thermally resilient architectures, including innovative ceramic composite separators and inorganic solid-state frameworks, significantly mitigates thermal runaway and improves overall battery safety across various technologies12,13. Building upon these material advancements, alternative fabrication strategies like Pickering emulsion templating, magnesium oxide-based ceramic felts, and conventional ceramic foam manufacturing methods have been explored to overcome the limitations of cold-pressed electrolytes. These approaches aim to enhance the structural integrity and mechanical durability under extreme operating conditions11. Ceramic foams exhibit high mechanical strength, with further improvements achievable by optimizing the design of the porosity, pore structure, and fabrication parameters14. Materials such as Al2O3 and MgO, already used as binders for molten salts in thermal batteries, show excellent compatibility, reinforcing the potential of ceramic foams as next-generation electrolyte frameworks. In this regard, Mondy et al.11 reported initial feasibility studies of thermal batteries incorporating ceramic foams, yet detailed investigations into the dynamic electrochemical behavior across various anode systems remained limited. Recently, Meng et al.15 reported the significance of architectural design in establishing stable electrolyte frameworks to prevent performance degradation. Furthermore, Fan et al.16 demonstrated that morphology oriented regulation of the electrolyte system can facilitate more efficient ionic conduction. While the aforementioned 3D interconnected architectures such as ceramic felts effectively immobilize molten salts9, they inherently rely on flexible frameworks. Such flexible structures can be susceptible to dimensional deformation including warping during the subsequent solidification of the infiltrated salt. Against this background, we developed an electrolyte configuration, based on alumina ceramic foam, for thermal batteries. Building upon these conceptual advancements, our adoption of a rigid, pre-structured alumina foam fundamentally resolves this thermomechanical vulnerability, providing a dimensionally stable, scaffold-stabilized architecture. Using the developed electrolyte, we systematically investigated the effects of the foam thickness, porosity, and pore architecture on the electrochemical behavior of these batteries. Methods Alumina foam electrolyte fabrication process for thermal batteries Alumina was selected as the ceramic material for use in the thermal battery separator. In this study, commercially available porous alumina foams (Seongji Material Tech, Seoul, Korea) were used to ensure high material consistency and reproducibility. The porosity of the foams was as specified by the manufacturer, determined by the ratio of the bulk density to the theoretical density of fully dense alumina (3.9 g/cm3). For further details on the density-based determination and the corresponding values for each sample, please refer to Supplementary S1. (Determination of Porosity). In addition, a reduced thickness of 0.35 mm, the use of which is typically restricted in pressing-type separators owing to its mechanical fragility and the risk of short-circuiting, was evaluated to determine its impact on the performance17. Post-fabrication procedures were conducted in a moisture-controlled dry room. We infiltrated the pores of the alumina foam separator with a ternary LiF–LiCl–LiBr molten salt, an electrolyte known for its relatively high ionic conductivity among those used in thermal batteries2,18. A ternary mixture of LiF (9.6 mass%), LiCl (22 mass%), and LiBr (68.4 mass%) was prepared according to the eutectic composition2. The molten salt was highly sensitive to moisture and was thus melted in a furnace located inside a dry room (dew point below − 40 °C, see Supplementary Fig. S1). The melting and infiltration processes were conducted under a dry air atmosphere to prevent salt degradation. The eutectic composition used herein has a melting point of approximately 440 °C2, and the furnace was preheated at a constant rate of 10 °C/min to 550 °C, approximately 100 °C above the melting point, to ensure complete melting. The crucible containing the salt mixture was placed in the furnace and maintained at this temperature for a dwell time of 30 min. After confirming complete transition to the liquid state (see Supplementary Fig. S1), the alumina foam was immersed in the molten salt for infiltration. Following this, the crucible was removed from the furnace and the infiltrated foam was allowed to cool naturally to room temperature under the dry room atmosphere. To ensure the reliability and reproducibility of the reported results, this electrolyte fabrication process was validated through independent infiltration trials(at least duplicate or triplicate, n ≥ 2 or 3). In addition, performance tests were conducted on alumina foam with a closed-cell structure (which exhibits higher mechanical strength compared with open-cell types19 with a thickness of 0.35 mm and porosity of 63% (cell diameter: 56 mm; Finetech, DaeJeon, Korea) using the same procedures as described above. Unit cell discharge Li–Si alloy and FeS2 were used as the negative and positive electrodes, respectively1. The unit cell was assembled by sequentially stacking a Cu plate top lead, a stainless steel (SUS) current collector, a Li-Si anode, the fabricated alumina foam electrolyte, an FeS2 cathode, another SUS current collector, and a Cu plate bottom lead. The actual photograph of the assembled single-cell stack and the overall fabrication procedure are provided in Supplementary Fig. S2 and Fig. 2, respectively. The single cells were discharged at 500 °C while applying a pulse current profile of 0.1 A cm−2 for 8 s, 0 A for 1 s, and 0.2 A cm−2 for 1 s (0.1 Aavg cm−2) under a DC electric load (SL-600, Unicorn) at a force of 100 kgf (approximately 1.39 MPa) to maintain strict interfacial stability and accurately evaluate the intrinsic performance of the proposed alumina foam electrolyte system. This stack pressure, which aligns with the typical assembly range of 1.03–1.72 MPa, ensures structural stability and minimizes interfacial contact resistance20. Additionally, the pulse discharge profile was utilized to calculate total polarization and evaluate response characteristics under high-power conditions characteristic of tactical mission profiles. For the 56-mm closed-cell sample, a compressive load of 250 kgf (~ 1.0 MPa) was applied to ensure a comparable stack pressure, and a slightly modified pulse profile (0.1, 0.2, and 0 A/cm2 for 6 s, 1 s, and 0 A for 1 s) was used. However, the identical peak current density (0.2 A/cm2) and pulse duration (1 s) ensure a reliable comparative analysis of the dynamic resistance. Before the single-cell discharge was performed, the fabricated single cell was rested for 1 min to allow the electrolyte to melt at 500 °C, and the discharge test was terminated when the voltage dropped to 0 V. The overall procedure for the alumina foam–electrolyte fabrication and the final unit cell assembly is schematically illustrated in Fig. 2. Fig. 2Schematic block diagram illustrating the sequential process flow from the eutectic salt melting and molten salt infiltration into the alumina foam to the final single-cell assembly and discharge testing setup. Schematic block diagram illustrating the sequential process flow from the eutectic salt melting and molten salt infiltration into the alumina foam to the final single-cell assembly and discharge testing setup. SEM analysis The chemical compositions and microstructures of the thermal battery unit cells and ceramic foams were characterized using SEM (FEI-QUANTA650) equipped with EDS, specifically examining the surfaces before and after infiltration along with the post-infiltration cross-sections, to confirm the uniform impregnation of the molten salt into the pores. Results and discussion Figure 3a and b show SEM images of the alumina foams with porosities of 55% and 63%, respectively. Both foam types exhibit open-cell architectures with a uniform three-dimensional network of interconnected pores. Supplementary Fig. S3 shows an image of the ceramic separator after impregnation with the LiF–LiCl–LiBr eutectic salt. Alumina exhibits excellent wettability with molten salts, as demonstrated in previous studies2,5. Consequently, the infiltration of the molten LiF–LiCl–LiBr salt into the porous structure of the alumina foam is clearly observed. Post-infiltration SEM analysis, to examine whether the molten salt had fully filled the porous regions of the alumina (see Supplementary Fig. S4), reveals that, after infiltration, the surfaces of the alumina foams with porosities of 55% and 63% are substantially altered compared with the pre-infiltration state shown in Fig. 3. The pore regions that are clearly visible before infiltration are no longer observed, suggesting that they are completely filled with molten salt. Fig. 3SEM images of alumina foams with different porosities: (a) 55% and (b) 63%. SEM images of alumina foams with different porosities: (a) 55% and (b) 63%. Cross-sectional SEM was used to confirm that not only the surface but also the internal regions were effectively filled with molten salt through the impregnation process. Figure 4 presents a cross-sectional SEM image and the corresponding EDS elemental mapping results of alumina foam with 55% porosity after impregnation with the molten salt. The SEM image in Fig. 4a shows a densely impregnated morphology in which the porous framework is no longer visible. The homogeneous distribution of Cl, a principal component of the molten salt, across the entire cross section provides direct evidence for the effective infiltration of the molten salt into the internal pore structure. Supplementary Fig. S5 presents a magnified cross-sectional region of the alumina foam with 55% porosity after impregnation with the molten salt, offering a more detailed view of the resulting internal morphology. The O and Cl elemental maps shown in Fig. S5 reveal reduced Cl intensity in regions where O is present, indicating that the molten salt has entered and filled the internal pores. Fig. 4Cross-sectional SEM and corresponding elemental distributions of alumina foam with 55% porosity after molten salt impregnation: (a) SEM micrograph; (b-d) Al, Cl, and O maps. Cross-sectional SEM and corresponding elemental distributions of alumina foam with 55% porosity after molten salt impregnation: (a) SEM micrograph; (b-d) Al, Cl, and O maps. Figure 5 presents the results of the same set of analyses conducted on the alumina foam with 63% porosity. Cross-sectional observations confirm that, similar to the sample with 55% porosity, the molten salt has effectively impregnated the internal pores. Supplementary Fig. S6 shows that, despite the structural variations associated with the higher porosity, the molten salt is uniformly distributed within the internal pores, demonstrating impregnation behavior equivalent to that of the alumina foam with 55% porosity.Fig. 5Cross-sectional SEM and corresponding elemental distributions of alumina foam with 63% porosity after molten salt impregnation: (a) SEM micrograph; (b-d) Al, Cl, and O maps. Cross-sectional SEM and corresponding elemental distributions of alumina foam with 63% porosity after molten salt impregnation: (a) SEM micrograph; (b-d) Al, Cl, and O maps. To quantitatively verify the reproducibility of this infiltration process, multiple independent trials were conducted, and the salt loading variability is summarized in Table 1. As shown in the table, the absolute infiltrated mass remains highly consistent across all evaluated specimens. Although the ultra thin 0.35 mm samples naturally exhibit a higher standard deviation when converted to relative weight%, the highly reproducible absolute mass loading, paired with the complete internal filling verified by SEM, confirms the physical reliability of the system. Consequently, this established physical consistency provides a robust foundation for the subsequent electrochemical evaluations, ensuring they precisely reflect the intrinsic capabilities of the proposed alumina foam electrolyte system. Table 1Reproducibility of molten salt infiltration into alumina foamsOpen cell porosity (%)Thickness (mm)Weight before infiltration (g)Weight after infiltration (g)Molten salt loading (wt.%)552.002.779 ± 0.0124.048 ± 0.03131.347 ± 0.540550.350.442 ± 0.0040.892 ± 0.09850.090 ± 5.253632.002.01 ± 0.0034.06 ± 0.0450.513 ± 0.467630.350.287 ± 0.0030.768 ± 0.04862.602 ± 2.740 Reproducibility of molten salt infiltration into alumina foams In the case of pressed pellets, which are fabricated via cold pressing at room temperature, the eutectic salt remains in its discrete particulate form during compaction. In contrast, the molten salt that has impregnated the ceramic foam solidifies within the interconnected pore network during cooling to form a continuous rigid phase that reinforces the structure and contributes to improved mechanical strength14. Supplementary Fig. S2 shows an image of the single cell used for testing in this study. The single cell, comprising a Li−Si anode, the alumina foam electrolyte, and an FeS2 cathode, has a diameter of 30 mm. Figure 6 presents the results of a discharge test of unit cells using the alumina foams with porosities of 55% and 63%. The phase transitions in Li–Si alloys at discharge typically involve a series of reactions that include the formation of the following intermediate phases:1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathrm{Li}}_{{{\mathrm{22}}}} {\mathrm{Si}}_{{\mathrm{5}}} \to {\mathrm{Li}}_{{{\mathrm{13}}}} {\mathrm{Si}}_{{\mathrm{4}}} \to {\mathrm{Li}}_{{\mathrm{7}}} {\mathrm{Si}}_{{\mathrm{3}}} \to {\mathrm{Li}}_{{{\mathrm{12}}}} {\mathrm{Si}}_{{\mathrm{7}}} \to {\mathrm{Si}}$$\end{document} Note that the extreme moisture sensitivity of Li22Si5 presents significant handling challenges even under dry-room conditions (relative humidity below 3%)21,22. Consequently, this study employed Li13Si4 as the initial material for anode fabrication. Owing to the smaller quantity of anode material compared with the cathode material, the observed plateaus correspond to the sequence of phase transformations of Li13Si4 → Li7Si3 → Li12Si7 → Si. Fig. 6Discharge curves of single cells tested at an average current of 0.1 Aavg/cm2 using the alumina foam electrolyte with different porosities and thicknesses: (a) 55% porosity, 2.0 mm thickness; (b) 63% porosity, 2.0 mm thickness; (c) 55% porosity, reduced thickness of 0.35 mm; and (d) 63% porosity, reduced thickness of 0.35 mm. Inset in each panel: magnified view of voltage drop under pulse-discharge conditions. Discharge curves of single cells tested at an average current of 0.1 Aavg/cm2 using the alumina foam electrolyte with different porosities and thicknesses: (a) 55% porosity, 2.0 mm thickness; (b) 63% porosity, 2.0 mm thickness; (c) 55% porosity, reduced thickness of 0.35 mm; and (d) 63% porosity, reduced thickness of 0.35 mm. Inset in each panel: magnified view of voltage drop under pulse-discharge conditions. Since thermal batteries are primarily used as military power sources, reliability is of vital importance. Accordingly, when utilizing the Li–Si anodes as described, thermal battery systems are typically designed conservatively. In practice, the operational life of the battery is strictly limited to the initial voltage plateau(Li13Si4 → Li7Si3) to ensure the highest stability and prevent potential system failure caused by subsequent voltage drops. The characteristics observed in this region play a decisive role in determining the overall performance of the thermal battery. Figure 6a shows that the sample with 55% porosity and thickness of 2.0 mm exhibits a significant voltage drop of approximately 0.18 V at a current density of 0.2 A/cm2. The dynamic resistance of the battery was measured by applying a pulse current, and its value was obtained using Eq. (2)22:2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${R}_{t}= \frac{{V}_{OC}-{V}_{CC}}{I}$$\end{document}where Rt is the dynamic resistance (ohm), and Voc, Vcc, and I are the open-circuit voltage (V), closed-circuit voltage (V), and discharge current (I), respectively. According to this equation, the observed voltage drop suggests that the internal resistance of the cell increases proportionally. By contrast, Fig. 6b illustrates that the 63% porosity sample with the same thickness of 2.0 mm exhibits a clearly reduced voltage drop of approximately 0.09 V under the same current density, suggesting lower internal resistance. The experimental results demonstrate that the dynamic resistance can be evaluated based on the voltage drop during the initial high-current pulse. Alumina foam maintains its structural integrity during discharge owing to the inherent rigidity of ceramic foams. This structural advantage contributes to reducing the dynamic resistance and improving the performance of the thermal battery by shortening the ion transport distance, as reported in previous studies23–25. The single-cell performance of the thermal batteries was evaluated using alumina foams with a reduced thickness of 0.35 mm at porosities of 55% and 63%. As shown in Fig. 6c, the voltage drop observed during the first high-current pulse is approximately 0.08 V, indicating that the performance is comparable to that of the 63% porosity sample with 2.0 mm thickness. However, it clearly demonstrates a significant improvement in the dynamic resistance compared with the original 55% porosity sample with 2.0 mm thickness. As shown in Fig. 6d, the voltage drop during the first high-current pulse is approximately 0.04 V, which, based on our analysis, confirms a substantial reduction in dynamic resistance. For a direct comparison, this performance was evaluated against literature data for conventional pellet-pressed separators. Under an identical pulse current density (0.2 A/cm2), the evaluated 0.35-mm-thick open-cell foam exhibits a voltage drop of ~ 0.04 V. This value is highly competitive with the voltage drop of 0.06–0.07 V typically observed in conventional pellet-type configurations11. This electrochemical efficiency can be attributed to the interconnected pore networks in the foam structures, which facilitate efficient lithium-ion transport26. By providing continuous and unobstructed pathways, this structural feature effectively minimizes ion-transport resistance and significantly reduces the internal resistance of the cell. The low dynamic resistance observed experimentally is consistent with these structural characteristics. Due to the stringent voltage regulation required for these applications, the analysis of dynamic resistance was focused on the first discharge plateau (Li13Si4 → Li7Si3)21. Fig. 7Dynamic resistance of alumina foam electrolytes over time during discharge. Comparison of dynamic resistance profiles for open-cell alumina foam electrolytes with porosities of 55% and 63% and thicknesses of 2.0 and 0.35 mm. Dynamic resistance of alumina foam electrolytes over time during discharge. Comparison of dynamic resistance profiles for open-cell alumina foam electrolytes with porosities of 55% and 63% and thicknesses of 2.0 and 0.35 mm. As presented in Fig. 7, among the controlled conditions, the lowest dynamic resistance during this plateau is evident for the open-cell structure with a thickness of 0.35 mm and a porosity of 63%. This excellent characteristic is primarily facilitated by the exceptionally efficient ion transport pathways provided by its highly interconnected pore structure. Fig. 8(a) Photograph of the unit cell after discharge and (b) cross-sectional SEM image of the alumina electrolyte after discharge, showing the microstructure. (a) Photograph of the unit cell after discharge and (b) cross-sectional SEM image of the alumina electrolyte after discharge, showing the microstructure. One of the most critical factors associated with the electrolytes of thermal batteries is the prevention of molten salt leakage. When using pressed pellets, non-uniformity at the edges and boundaries can lead to risks such as leakage and internal short-circuiting11. In addition, studies have reported that, during thermal battery operation under high-shock conditions, the deformation of the electrolyte can lead to the slippage of the entire battery stack11. Therefore, a structure that does not undergo deformation even after the molten salt is liquefied would be expected to enhance the durability under high-shock conditions. Because the alumina foam effectively serves as a separator, it offers reliable discharge performance under the high-shock conditions typical of thermal battery applications. Furthermore, as this approach avoids the limitations of pressing, it is a promising approach to significantly improve the production process of thermal battery electrolytes. A post-discharge analysis was conducted to verify whether molten salt leakage occurred during discharge. Figure 8a presents an image of the assembled single cell immediately after discharge. The absence of molten salt around the cell periphery confirms that no leakage has occurred during the single-cell discharge test. This may be attributed to the superior wettability between the alumina foam and the molten salt. Figure 8b presents an SEM cross-sectional analysis performed after discharge to examine whether the alumina foam electrolyte maintains its original structure. The porous alumina foam maintains its original shape, with a measured thickness of approximately 350 μm, indicating no structural deformation during discharge. Notably, this structural integrity was preserved even under a continuous compressive load of 100 kgf, (approximately 1.39 MPa), which was applied to simulate the high-stress environment of an actual battery stack. This mechanical robustness ensured a stable separator layer, effectively preventing internal short circuits and active material leakage. High porosity enhances ionic conduction but necessitates trade-offs with mechanical strength and electrolyte retention. In defense applications involving extreme shocks, excessive porosity affords thinner alumina ligaments, compromising compressive strength and the capillary forces required to immobilize molten salts. We found that 63% porosity provides sufficient resilience to withstand a stack pressure of 1.39 MPa without leakage. While this confirms stability under static loads, future research will address dynamic high-G environments to ensure operational reliability in mission-critical systems. Furthermore, as reported by Guidotti et al.1, minimizing internal resistance is critical to avoiding premature voltage cutoff. Accordingly, the low dynamic resistance observed in Fig. 7 indicates high effective ionic conductivity, allowing for enhanced capacity utilization. Guidotti et al.2 reported that Al2O3 is a chemically inert material conventionally used as a binder in thermal batteries. Therefore, its application as a pre-structured foam represents a significant advancement in ensuring both chemical and mechanical reliability under extreme operating conditions. The effect of the pore morphology and structure on the discharge performance was verified by testing ceramic foam with a closed-cell configuration, a porosity of 63%, and an identical thickness of 0.35 mm. Closed-cell structures generally exhibit higher mechanical strength compared with their open-cell counterparts, which makes them suitable for use in mechanically demanding environments19. Fig. 9(a) SEM image of the alumina foam with a closed-pore structure and 63% porosity. (b) Discharge profile recorded at an average current density of 0.1 A/cm2. The electrochemical performance is significantly inferior compared with the open-pore structure. (c) Evolution of dynamic resistance during discharge, illustrating a continuous increase in internal resistance that correlates with limited ion transport within the isolated pore framework. (a) SEM image of the alumina foam with a closed-pore structure and 63% porosity. (b) Discharge profile recorded at an average current density of 0.1 A/cm2. The electrochemical performance is significantly inferior compared with the open-pore structure. (c) Evolution of dynamic resistance during discharge, illustrating a continuous increase in internal resistance that correlates with limited ion transport within the isolated pore framework. Figure 9a presents an SEM image of alumina foam fabricated with a closed-cell structure. In contrast to the previously described open-cell configuration, a large fraction of the pores are isolated rather than interconnected. The closed-cell alumina foam was immersed in molten salt heated to 550 °C in a dry-room furnace to evaluate its infiltration behavior under the same conditions as the open-cell structure. Although the molten salt also infiltrates the closed-cell structure, the open-cell foam demonstrates comparatively superior infiltration characteristics. Figure 9b shows the discharge result obtained from the single cell with the closed-cell structure, where the voltage drop during the initial high-current pulse was measured to be approximately 1.09 V. Figure 9c illustrates the corresponding time-dependent absolute dynamic resistance profile of the closed-cell structure. Based on Eq. (2), the absolute dynamic resistance during the initial plateau was calculated to be approximately 0.221 Ω. Despite its larger geometric dimension (56 mm diameter), which theoretically lowers absolute resistance, its dynamic resistance substantially exceeded that of the 2.0-mm-thick open-cell sample with 55% porosity, which exhibited the poorest performance among the open-cell structures studied. To accurately evaluate the intrinsic polarization behavior independent of the cell dimensions, dynamic resistance profiles were further normalized to area-specific resistance (ASR) as a function of time, as detailed in Supplementary Fig. S727. The closed-cell configuration consistently exhibits an ASR that is orders of magnitude higher than those of the open-cell counterparts across all thickness and porosity variations throughout the discharge process. This prominent discrepancy confirms that the isolated pores in the closed-cell configuration act as physical bottlenecks that disrupt continuous Li+ ion-transport pathways, leading to progressive resistance growth over time. The fundamental cause of this resistance difference arising from distinct ion transport pathways is visually elucidated in Fig. 10. The open-cell foam provides a continuous three-dimensional network for unobstructed Li+ ion transport, whereas the isolated pores in the closed-cell foam act as physical barriers that sever these pathways. Although closed-cell foams are generally known for their superior structural integrity, the high dynamic resistance originating from the closed-cell configuration limits their ability to satisfy the electrochemical performance requirements of high-power thermal battery applications. Fig. 10Schematic illustration of Li+ ion transport pathways in closed and open cell alumina foam electrolytes. Schematic illustration of Li+ ion transport pathways in closed and open cell alumina foam electrolytes. Based on our experimental results, open-cell alumina foams are identified as more suitable for use as electrolytes in thermal batteries. The findings indicate that the discharge performance is highly dependent on the microstructural characteristics of the foam, particularly the pore structure type and porosity. Precise control over the porosity, microstructural architecture, and thickness enables a shift from conventional pressed-pellet fabrication toward next-generation electrolyte design. Conclusion This paper reported the first successful application of alumina foam as an electrolyte material for thermal batteries, and stable discharge performance was demonstrated under the typical operating conditions. Notably, reducing the thickness of the foam significantly enhanced the electrochemical performance even with identical pore morphology. Pore structure variations also induced substantial changes in the discharge behavior. Only the open-cell structures proved suitable for application as the electrolyte, underscoring the importance of interconnected porosity. These results highlight that electrolytes based on alumina foam with a precisely tuned porosity and thickness offer a viable alternative to conventional pellet-pressed electrolytes, which have poor uniformity and mechanical limitations. In particular, this work presents a paradigm shift in electrolyte fabrication by transitioning from the conventional cold-pressing of powder-based mixtures to a pre-structured, scaffold-stabilized architecture. The findings underscore the pore architecture, porosity, and separator thickness as key parameters governing the electrochemical behavior. However, this study has inherent limitations, because the current evaluation was restricted to the single unit-cell level under static compressive loads. To fully realize the potential of this approach, these structural factors must be systematically optimized and the operational reliability of fully assembled thermal battery stacks under dynamic high-G shock and vibration environments. Furthermore, in-situ high-temperature electrochemical impedance spectroscopy will be necessary to systematically decouple and analyze the intrinsic interfacial kinetics under realistic operating conditions. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1