Designing carbon wetting layers with inorganic salts as additives for enhanced adhesion and molten sodium wettability in electrochemical application

1 Introduction

Grid energy storage is a crucial component of modern energy infrastructure, enabling large-scale storage within the electrical grid to maintain a stable and reliable balance between electricity demand and supply (Dunn et al., 2011; Yang et al., 2011). On a daily basis, it stores excess electricity generated during periods of low demand and releases it during high demand, helping to offset electricity prices. From a resiliency perspective, as electricity demand continues to grow and grid disruptions from extreme events become more frequent, grid storage plays a vital role in enhancing reliability and resilience by providing backup power for critical infrastructure (Gür, 2018). Among the various battery technologies developed for grid energy storage applications, molten sodium batteries, such as sodium-metal-halide (Na-MH) batteries, are considered one of the most promising due to several advantages. These include the use of abundant and low-cost sodium as the charge carrier, high energy density (∼100 Wh/kg at the module level), long cycle life (>20 years), superior safety, and environmental independence (no need for HVAC systems), etc. (vanZyl, 1996; Dustmann, 1998; Li et al., 2016; Sudworth, 1994). In general, Na-MH batteries (Ahn et al., 2019; Heinz et al., 2020; Gross et al., 2021; Gao et al., 2022; Lan et al., 2025; Ao et al., 2024; Li et al., 2015; Zhan et al., 2021; Zhan et al., 2020; Hueso et al., 2017), including Na-NiCl₂ (ZEBRA), are inherently resistant to thermal runaway, a major safety concern in many high-energy battery systems, such as lithium-ion batteries (LIBs) (Larsson and Mellander, 2014). Unlike typical LIBs and sodium ion batteries (SIBs), which use flammable organic electrolytes, Na-MH systems employ non-flammable ceramic and molten inorganic salt electrolytes, greatly reducing fire risk. Furthermore, their internal chemistry is more stable under abuse conditions such as overcharging or short-circuit (Dustmann, 1998; Benato et al., 2016; Sudworth, 2001). This intrinsic thermal stability makes Na-MH batteries particularly attractive for stationary energy storage where safety and reliability become critical. To expand the market volume of Na-MH batteries in grid storage applications, it is necessary to reduce the cell/module manufacturing cost, which is often linked to the high operating temperature—260 °C–350 °C for conventional Na-S and ZEBRA batteries. It should be noted that in these batteries, solid-state electrolytes, such as β″-alumina solid-state electrolyte (BASE), are utilized to separate the molten sodium anode from the cathode materials (Lu et al., 2010; Lu et al., 2018). Therefore, an important component, known as the wetting layer, is required to facilitate efficient contact between the molten sodium and the BASE surface. Without it, the battery may experience inferior performances from high interfacial resistance between molten sodium and BASE, as well as underutilized cathode materials.

Unfortunately, the intrinsic wetting of molten sodium (Na) on bare BASE is quite poor, which is attributed to the high surface tension of molten sodium and the high surface energy of BASE (Chang et al., 2018; Li et al., 2022). Moreover, the wettability of sodium deteriorates further as the operating temperature decreases (Reed et al., 2013). The application of metallic coatings such as Sn (Reed et al., 2013; Gross et al., 2020), Pb (Chang et al., 2018; L et al., 2021), Bi (Jin et al., 2019a), and Ni (Hu et al., 2013), on the BASE surface has been shown to significantly improve Na wettability. For instance, micron size spherical Pb particles formed through the decomposition of lead acetate trihydrate on BASE, facilitate excellent Na wettability via “sunny-side up” wetting model at temperature as low as 120 °C, effectively reducing interfacial overpotentials (Chang et al., 2018). Also, molten Na alloy anodes, such as Na-K (Ding et al., 2020), Na-Rb (Lu et al., 2014), and Na-Cs (Lu et al., 2014) have been used for improving Na wettability on the BASE at lower temperature. This improvement is achieved by increasing the work of adhesion (Na alloys), as described by the Young-Dupré relation for the wetting model on smooth surfaces (Adam, 1957; Schrader, 1995).

In addition to the aforementioned processes, various carbon-based materials are of increasing interest due to their potential for low cost and scalability in interfacial engineering, aimed at addressing the poor Na wettability at low temperatures (Weller et al., 2024; Hu et al., 2012; Kim et al., 2015). Jin et al. introduced sparked reduced graphene oxide (rGO), which enabled complete Na wetting via capillary effects in nanogaps, significantly improving performance with high Coulombic efficiency at 175 °C, facilitating low temperature operation (Jin et al., 2019b). Hu et al. demonstrated a low-cost, soft template method using glucose and PMMA to create porous carbon films on BASE, enhancing Na wettability and reduced interfacial polarization at 350 °C, although relatively high processing temperatures were needed (Hu et al., 2012). Lai et al. developed a 3D cross-linked carbon fiber network with Sn particles, improving both wettability and guiding uniform Na deposition. This structure achieved high cycling stability and reduced interface resistance to 6.6 Ω cm² at room temperature (Lai et al., 2022). Landmann et al. demonstrated stable Na plating/stripping at >1000 mA/cm² and >10 Ah/cm² using carbon-coated BASE at 250 °C (Landmann et al., 2020). Li et al. proposed a hybrid interface nanoparticle-decorated porous carbon layer containing PbO_x, achieving stable cycling over 6000 cycles with minimal (∼3wt%) Pb loading, and suggested Sn as a safer alternative (Li et al., 2022). More recently, Weller et al. extended the nanoparticle-decorated porous carbon layer to a metal-free intrinsic sodiophilic carbon black wetting layer, which provides non-toxic, and inexpensive while still have good Na wettability and low area specific resistance at the low temperature (∼110 °C) (Weller et al., 2024). It is important to note that the method for preparing carbon layers by Weller et al. involves the use of organic surfactants and binders to achieve a well dispersed carbon suspension or ink in an aqueous solution. During heat treatment, these surfactants and binders are typically burned off, leaving meso- and macroporous structures in the carbon layers. These porous structures have been shown to be critical for enhancing Na wettability at lower temperatures (Weller et al., 2024). While burning off the surfactants and binders is necessary, it inevitably weakens the adhesion between the carbon layers and the BASE surface, primarily due to the reduction of van der Waals forces caused by the loss of organic-based binders. Therefore, it is highly desirable to develop new types of binders that can maintain strong mechanical adhesion on the BASE surface without compromising Na wettability, even after the high-temperature heat treatment that burns off the organic surfactants and binders. To address the challenges associated with organic binders, a series of inorganic materials, such as conventional salts, have been selected as a secondary binder in addition to the organic binder for testing. The selection of inorganic salts was based on several criteria: (1) The salts should contain sodium as the cation, potentially enabling some level of sodium ion conduction while preventing undesirable ion exchange with Na ions in the BASE. (2) Salts in a more simple or conventional chemical formula are preferred to minimize cost increases. (3) The salts should exhibit relatively good solubility in water to use aqueous based carbon inks for spray coating.

In this study, various conventional sodium inorganic salts—such as NaCl, Na₂CO₃, NaNO₃, and Na₆(PO₃)₆ (sodium hexametaphosphate, SHMP)—were investigated as additives in carbon inks applied to different substrates, including molybdenum (Mo) and BASE, to examine their effects on molten sodium wetting/spreading and on the performance of molten sodium symmetric cells. The results indicate that the No-salt, SHMP/carbon, and NaCl/carbon layers all exhibit good molten sodium wettability, characterized by instantaneous Na wetting. When comparing Na spreading speeds, the SHMP/carbon layer shows a 14% faster Na spreading rate than the No-salt layer, whereas the NaCl/carbon layer exhibits a 13% slower spreading rate than the No-salt layer. The other salts, NaNO₃ and Na₂CO₃, either underwent side reactions with the carbon layer during high-temperature treatment or disrupted the porous structure of the carbon layer due to preferential salt precipitation on the surface, thereby impeding molten sodium wettability. Notably, the SHMP/carbon layers also demonstrated better mechanical adhesion compared to the No-salt layers, which can be attributed to the formation of fewer cracks and voids that are otherwise commonly observed in the No-salt carbon coatings. The molten sodium symmetric cell tests for the SHMP-containing cells are consistent with the Na wetting/spreading results, showing lower overall resistance (both ohmic and charge-transfer components) compared to the No-salt cells.

2 Experimental section

2.1 Chemicals

The following chemicals were used in this work as received. Carbon black (Vulcan XC-72), Tergitol 15-S-9 (Sigma-Aldrich), hydroxypropyl methylcellulose (HPMC, Thermo Scientific), de-ionized water (MilliQ NanoPure system >18MΩ·cm), ethanol (Koptec 200 proof anhydrous), sodium hexa meta phosphate (Na₆(PO₃)₆ or SHMP, Matheson Coleman & Bell, SX583), sodium chloride (NaCl, Sigma-Aldrich), sodium nitrate (NaNO₃, Sigma-Aldrich), and Sodium carbonate (Na₂CO₃, Thermo Scientific))

2.2 Preparation of carbon-based wetting layers with salt additive

Carbon-based inks with different salt additives were prepared from a suspension ink containing carbon black (0.33g), Tergitol 15-S-9 (0.1g), hydroxypropyl methylcellulose (0.05g, HPMC), de-ionized water (18 mL) mixed with ethanol (2 mL) (Weller et al., 2024), and various sodium salts (0.11g), including SHMP, NaCl, NaNO₃, and Na₂CO₃. The mixture was homogenized by gently stirring with a PTFE stir bar overnight. The prepared carbon ink was deposited onto target substrates (glass, BASE, or Mo foil) using a shadow mask and spray-coating (Iwata nozzle, ∼25 psi, ∼10 cm distance) on a hot plate at 90 °C. After drying under ambient conditions, the films underwent heat treatment in a box furnace (KSL-1100X, MTI Corporation) at 350 °C in air. The heating protocol included a ramp to 120 °C at 3 °C min^-1 with a 1 h dwell, followed by heating to 350 °C at 3 °C min^-1 and holding for 5 h. Cooling was carried out at 2 °C min^-1–75 °C, after which the samples were immediately transferred to an Ar-filled glove box (MBraun LabMaster).

2.3 β″- alumina composite preparation

Sodium β″-alumina solid electrolyte disks (BASE, 70 vol% β”-Al₂O₃, 30 vol% YSZ) were obtained using the vapor-phase conversion method described in our previous study (Jung et al., 2018). The typical size of BASE disk is 27 mm in diameter with 500 mm thickness.

2.4 Peeling test

Peeling tests were performed on carbon wetting layers (circular area ∼3 cm²) containing different salts, prepared on glass substrates using the method described above (spraying coat/heat treatment). A Scotch tape was applied to each layer under a ∼350 g load and then gently peeled off to evaluate the residual carbon ink on the substrate and the amount detached onto the tape.

2.5 Sodium wetting test

Using the same process, rectangular carbon-based wetting layers (3 cm × 0.4 cm) with different salt additives were prepared on glass slides with shadow masks for sodium spreading speed tests. The samples were placed on a hotplate (Chemglass Life Sciences) inside an Ar-filled glove box, and molten sodium was applied to one end of the wetting layers using a preheated glass pipette to measure the spreading of the Na droplet along the rectangular shaped wetting layers.

2.6 Scanning Electron Microscopy (SEM)

Carbon wetting samples with different salt additives were prepared on molybdenum (Mo) substrates for SEM measurement. Mo was chosen to provide electrical conductivity while minimizing potential reactions between the carbon layers and the metal substrate during high-temperature heat treatment. Plasma focused-ion beam (PFIB)-SEM (Thermo Scientific Helios Hydar UX) was conducted using a Xe-ion beam on a JEOL IT-200 SEM with energy dispersive X-ray spectroscopy (EDS) detector.

2.7 X-ray Photoelectron Spectroscopy (XPS)

XPS measurements were performed using a Nexsa ThermoFisher Scienetific spectrometer, using focused Al Kα monochromatic x-ray source (1486.6 eV) operated at 72 W and a high-resolution spherical mirror analyzer using 50 eV pass energy. The data acquisition was carried out using a 300 μm diameter x-ray beam and emitted photoelectrons were collected at the analyzer entrance slit normal to the sample surface. The chamber pressure was maintained at ∼5 × 10⁻⁹ Torr during the measurements. All the XPS peaks were referenced using C 1s binding energy of 284.8 eV.

2.8 Cell assembly and electrochemical characterization

The assembly process of sodium symmetric cells followed that of sodium metal–halide cells described elsewhere (Weller et al., 2024). First, carbon wetting layers were spray-coated onto both the anode (bottom) and cathode (top) sides of the BASE electrolyte. The BASE disks were sealed with an ethylene propylene diene monomer (EPDM) ring to α-Al₂O₃ fixtures prior to cell assembly. For the cathode side, a steel mesh coated with the same wetting layer composition was inserted as the current collector to promote uniform Na spreading and ensure full active area utilization (3 cm²). On the anode side, a stainless-steel shim with concentric grooves—previously used in planar-type Na battery tests—was employed to contain molten sodium. The cell was then enclosed with stainless-steel endcaps and PVDF O-rings on both sides. Symmetric cell performance, including current staircase and long-term cycling, was tested using a Landt battery tester (CT3004A). Electrochemical impedance spectroscopy (EIS) was performed with a Gamry Reference 600+ integrated into the Landt system, over a frequency range from 1 MHz to 0.1 Hz.

3 Results and discussion

3.1 Peeling and Na wetting tests

We prepared carbon layers with different salt additives on the substrates using the spray coating method, as briefly illustrated in Figures 1a, b) (see the experimental section for details). First, carbon black (Vulcan), surfactant (Tergitol 15-S-9), and binder (hydroxypropyl methylcellulose, HPMC) were mixed thoroughly using a stirring bar on a stirrer at ambient temperature. When employing the spray coating method, the selection of solvents for the spray ink is crucial since it influences the morphology and thickness of the wetting layer. The key consideration for solvent selection in this work was to balance the drying rate—ensuring it is fast enough to prevent forming droplets from re-dissolving underlying layers, but not so rapid that it disrupts uniform film formation. To achieve this balance, we used a solvent mixture of 90% deionized (DI) water and 10% ethanol, and this solvent combination (90% DI water/10% ethanol) seemed to promote the development of a more homogeneous morphology (Supplementary Figure S1) compared to that of 100% DI water-based ink. After the homogeneous carbon inks were made, a spray nozzle was connected to house compressed air flow at 2 atm pressure to apply a uniform coating on the substrates (glass slides for peel/wetting tests and BASEs for Na symmetric cell tests) on a hotplate at 80 °C. We used two different shapes of magnetic shadow masks, circular wetting layers were created for adhesion tests (Figure 1), while rectangular layers were used for wettability tests. The substrates with carbon layers were then transferred to a furnace and heat-treated in air at 350 °C. It is worth noting that heat treatment at higher temperatures (>450 °C) causes significant carbon loss due to burnout.

Adhesion strength of carbon layers with different salts addition SHMP, NaCl, NaNO₃, and Na₂CO₃) after the peel test, which was conducted using Scotch tapes, are shown in Figure 1. The typical weight ratio of salt to carbon powder in the ink is 1:3. The carbon layers with SHMP (Figure 1d) and NaCl (Figure 1e) exhibited significantly less detachment of carbon layers from the glass substrate compared to the pristine carbon layer without salt addition (Figure 1c), which indicates stronger adhesion to the substrate. In contrast, the amount of detached carbon layers with NaNO₃ (Figure 1f) is slightly higher than that from the pristine carbon layer. The carbon layer with Na₂CO₃ behaves quite differently compared with all others. As shown in Figure 1g, it exhibits a much lighter color and significant material loss after heat treatment. It is known that carbon and Na₂CO₃ can react at elevated temperatures to form gaseous products such as, CO or CO₂ (Kim and Lee, 2001), which aligns with the material loss observed in Figure 1g. In the tape peel test shown in Figure 1g, most of the carbon layer with Na₂CO₃ detached from the glass substrate, indicating the weakest adhesion among the samples tested in this work. The relative adhesion strength of the carbon layers with SHMP and NaCl is higher than that of the layers without salt (Figure 1h). In contrast, NaNO₃ and Na₂CO₃ do not improve their adhesion strength.

As mentioned above, the Na wetting test was conducted using rectangular carbon layers (Figure 2a, 0.4 mm wide × 3 cm long) to quantitatively compare Na wettability on various carbon layers with different salt additives. First, a single droplet (∼25 µL) of molten sodium was dropped to one end (left) of the rectangular wetting layer using a glass pipette at 120 °C. Then, Na wettability was assessed based on two key factors: instantaneous wetting area and spreading speed. The instantaneous wetting area, denoted as L1 in Figure 2b, was measured immediately after the Na droplet was placed. The spreading speed was then determined by dividing the spreading length, denoted as L2 in Figure 3b, by the elapsed time until the liquid Na front reaches the other end (right) of the rectangular shape carbon layer. For a more detailed explanation and schematic view of L1 and L2 can be found in Supplementary Figure S2. The reason for adopting this unconventional method is that conventional contact angle measurements could not distinguish the Na wettability of the No-salt and SHMP-treated BASE, as Na wetting occurs almost instantaneously and the contact angle rapidly approaches near 0° in both cases. Therefore, tracking the Na spreading speed over a fixed distance provides a more meaningful and quantitative means to compare Na wettability in this study. In general, the carbon layers containing SHMP as an additive exhibit better Na wettability, as indicated by the larger L1 and spreading speed in Figures 2c,d, compared to No-salt. More accurately, the L2 value of carbon layers containing SHMP is 0.128 mm/s, which is 14% faster than that of No-salt (0.112 mm/s). In comparison, the carbon layers with NaCl shows a molten Na spreading speed (0.097 mm/s), which is 13% slower compared to No-salt, as shown in Figure 2d, despite improved adhesion. The carbon coatings containing NaNO₃ or Na₂CO₃ demonstrate poor Na wettability, showing neither instantaneous wetting nor spreading.

3.2 Top morphology

To understand the possible effects of salt additives to the morphology of carbon wetting layers, which could eventually affect Na wettability, SEM measurements were conducted for the carbon layers with or without salt additives. SEM images shown in Figure 3 depict top view of carbon wetting layer. The high-magnification SEM image of the carbon wetting layer without any salt additives (Figure 3a) reveals a porous network with both meso- and macropores. Such features, as we reported in our previous work (Weller et al., 2024), provide an interconnected structure that facilitates electrolyte infiltration and wetting. In sharp contrast, when NaCl is introduced as an additive (carbon/NaCl, Figure 3b), the morphology changes dramatically. Numerous submicron-sized cubic NaCl particles, often appearing as aggregates, are dispersed throughout the carbon layer, effectively interrupting and blocking the original pore structure. A similar trend is observed when Na₂CO₃ is used as an additive (carbon/Na₂CO₃, Figure 3c). In this case, the particles or aggregates are somewhat larger than those seen in carbon/NaCl, but the overall effect remains comparable—Na₂CO₃ deposits fill or obstruct the pores, thereby modifying the open porous morphology that characterizes the additive-free carbon layer. Finally, the carbon layer with SHMP additive (carbon/SHMP, Figure 3d) shows no evidence of surface passivation or pore blockage, and its overall morphology closely resembles that of the additive-free carbon layer. Photographs of the carbon layers sprayed in circular shapes and low-magnification SEM images of all the aforementioned carbon layers are provided in Supplementary Figure S3 to offer a broader view of their surface structures. Overall, the SEM results indicate that maintaining the meso- and macroporous structure of the carbon layers—regardless of the presence of salt additives—is essential for good Na wettability. Any disruption or surface passivation of this porous network, as seen in carbon layers containing NaCl or Na₂CO₃, can substantially hinder Na spreading/wetting. Considering the combined evidence from adhesion tests, Na wetting experiments, and SEM observations, SHMP stands out as exhibiting uniquely superior performance compared to the other samples tested.

3.3 Na symmetric cell test

Therefore, two types of carbon layers—SHMP and No-salt—were selected to evaluate the impact of Na wetting on cell performance using Na-symmetric cells. The detailed assembly procedure is provided in the Experimental Section. Briefly, molten sodium was dropped onto the surface of BASE at 110 °C, which had been spray-coated with carbon ink either containing SHMP or without any salt, followed by heat treatment. The Na-symmetric cell, with a schematic shown in Figure 4a, was then assembled in a glove box and subsequently placed in a furnace for testing outside of the glove box. Figure 4b shows the voltage response of these 2 cells during a ‘current-staircase’ experiment, in which the applied current density was increased from 1 to 15 mA cm⁻² at 120 °C. Both batteries—with SHMP or No-salt—delivered stable voltage profiles during the current-staircase experiment as well as during long-term testing at 5 mA cm⁻². The main difference, as clearly seen in Figure 4b, is the overpotential: the SHMP cell with BASE treated with carbon/SHMP layers exhibits significantly lower values compared to that marked as No-salt. Figure 4c provides a more quantitative comparison by showing voltage versus current density plots for the 2 cells. The voltage at each current density was calculated as follows:where V _*c* is the midpoint of the voltage profile during the charge process at a given current density, and V _*d* is the corresponding midpoint during the discharge. Using this averaged value obtained from Equation 1 helps eliminate voltage offsets that may arise from various factors, including instrumental artifacts. Both data sets, No-Salt and SHMP, can be accurately fitted using linear regressions, suggesting that no non-linear behavior, such as mass transfer limitations, is present during the current-staircase experiment. The slope values of the linear fits, which represent the overall resistance, are notably different between the two measurements. As shown in Figure 4c, the Na-symmetric cell with SHMP exhibits a resistance of 2.8 Ω, approximately half that of No-salt cell (5.5 Ω).

V = V c − V d 2 (1)

Electrochemical impedance spectroscopy (EIS) measurements were performed on two Na symmetric cells, SHMP and No-Salt, across a frequency range of 1 MHz to 0.01 Hz. As shown in Figure 4d, EIS measurements for the SHMP cell were conducted at four temperatures: 115, 120, 125, and 130 °C. Generally, the Nyquist plots (Figure 4d insert) obtained from EIS measurements consist of a high frequency interception, depressed semicircles, and Warburg tail in the low-frequency region. The resulting Nyquist plots were analyzed using an equivalent circuit, with the fitting results presented in Supplementary Table S1. The values of the high-frequency intercept (R0) and the charge transfer resistance (R1) decrease with increasing testing temperature for the SHMP cell. This observation aligns well with the expectation that the bulk resistance, primarily attributed to BASE, and the Na/BASE charge transfer resistance decrease as the temperature rises. By comparing the EIS plot of the No-salt cell with the SHMP cell at the same temperature (e.g., 120 °C, as shown in Figure 4d), it is evident that both R0 (5.16 Ω) and R1 (0.40 Ω) for the No-Salt cell are significantly higher than those for the SHMP cell (R0 = 2.58 Ω, and R1 = 0.2 Ω). Also, these fitting results are consistent with the overall resistance values, 5.5 Ω for No-salt cell and 2.8 Ω for SHMP cell, shown in Figure 4C.

3.4 Cross section and wetting mechanism

Results from the Na wettability test and the Na symmetric cell test reveal an intriguing observation: the overall resistance of the No-salt Na symmetric cell is significantly higher than that of the SHMP-treated cell, despite both surface treatments exhibiting good Na wetting, with SHMP showing only slightly better wettability. To understand the origin of the increased resistance in the No-salt cell compared to the SHMP cell, XPS measurements and morphological investigations using plasma focused-ion beam (PFIB)–SEM were conducted. For these analyses, two types of inks (No-salt and SHMP) were sprayed onto molybdenum (Mo) foil instead of glass slides to ensure better electrical conductivity and to facilitate sample handling for PFIB–SEM. As shown in Figure 5a, the O 1s spectrum from the No-salt sample displays a single O 1s peak assigned to carbonyl (C=O) or (C-O^-) groups near 532.6 eV (P2) (Zhang et al., 2019), which can be attributed to residual functionalities arising from the decomposition of binders and surfactants in the inks. In comparison, the O 1s spectrum of the SHMP sample exhibits distinct peaks originating from hexametaphosphate (molecular structure shown as the insert), which contains two different P-O bonding environments: P-O^-/P=O (P1) and P-O-P bridging (P3) (Tsuchida et al., 2011; Raskar et al., 2008). The higher binding energy peak (P4) is most likely attributed to the Na KLL Auger peak (Shchukarev and Boily, 2008; Zhang et al., 2024). Importantly, the SHMP sample also shows a comparable P2 peak to the No-salt sample, indicating that the chemical environment of the carbon black is not significantly altered by the addition of SHMP as an additive. The C 1s spectra in Figure 5B are also very similar for the SHMP and No-salt samples, further supporting this hypothesis.

Before performing PFIB cross-sectional etching, a tungsten (W) cap was deposited onto the carbon layer to prevent the occurrence of waterfall patterns during the PFIB etching process. Secondary ion SEM images for the No-Salt and SHMP are presented in Figures 5c,d, respectively. The corresponding overlaid EDS mappings, displaying the distributions of C, Na, P, O, and Mo, are provided in Figures 5e,f, and individual EDS mappings for each individual element can be found in Supplementary Figure S4. Firstly, both the SEM image (Figure 5a) and the EDS mapping (Figure 5b) for No-Salt show the presence of cracks within the carbon layer and voids at the carbon layer/Mo substrate interface. These imperfections are significantly larger than the porous network of the carbon layer. While these cracks and voids may not necessarily affect Na wettability to a certain extent, they are highly likely to compromise adhesion to the substrate, as evidenced by the peeling test shown in Figure 1c, where the significant amount of carbon was detached by the tape. In contrast to the cracks and voids observed in the No-Salt layers, Figures 5c,d reveal that the SHMP-treated layers exhibit significantly fewer cracks and voids. This improvement can likely be attributed to the microscale aggregates or particles of SHMP, which may act as structural reinforcements or binders within the carbon layer, contributing to its mechanical stability. These aggregates/particles enhance both the internal cohesion of the carbon layer and its adhesion to the substrate, resulting in a more uniform and robust interface. However, it is worth noting that despite the presence of cracks and voids in the No-Salt-treated layers, the majority of the carbon layer retains its porous network. This porous structure is vital for facilitating Na wetting by enabling effective interaction at the material’s surface. Nevertheless, the SHMP-treated layers demonstrate a noticeably better Na wetting performance (Figure 2d), which is likely attributed to their reduced number of defects, such as cracks and voids. The diminished prevalence of these imperfections results in a more favorable surface morphology, enhancing both Na wetting and adhesion properties.

If we extend this observation to the No-Salt Na symmetric cell, it is highly possible that similar types of defects, particularly voids formed at the interface between the carbon layer and the BASE, could adversely affect the transport of Na across the BASE. These voids, depending on their size and distribution, disrupt the continuity of the porous carbon network that is essential for effective Na wetting and transport. Molten Na, which relies on adhesion and wetting properties to permeate through the porous structure, may fail to properly wet these voids due to the absence of carbon material within them. This lack of wetting in void regions could result in localized barriers to Na transport, reducing the overall area available for Na ion transfer and introducing additional resistance during Na symmetric cell testing. As a result, the performance of the No-Salt Na symmetric cell is expected to exhibit a higher overpotential, which is consistent with the results observed in the Na symmetric cell testing (Figures 4b–d). In these tests, the overall resistance of the No-Salt cell was found to be nearly twice that of the SHMP-treated cell, highlighting the significant impact of defects, such as voids, on the efficiency of Na transport and ion transfer. Therefore, SHMP as a binder creates a uniform and interconnected structure of carbon layers that facilitates the seamless interaction of molten Na with BASE, providing better overall performance in experimental applications.

4 Conclusion

To summarize, this study explores the fabrication and optimization of carbon wetting layers sprayed with various salt additives to enhance adhesion and wettability for molten sodium-based applications. Using a spray-coating method with custom carbon ink formulated from Vulcan carbon black, Tergitol 15-S-9 binder, HPMC binder, various sodium inorganic salts, and a 90% deionized water/10% ethanol solvent mixture, carbon coatings were deposited on substrates and examined for adhesion strength, wettability, microstructural properties, and cell performance. Adhesion tests demonstrate superior substrate binding from coatings with SHMP exhibiting the strongest overall performance in terms of adhesion and resistance to mechanical peeling. Wettability experiments revealed that SHMP-enhanced carbon layers significantly improved instantaneous wetting and spreading speed for molten sodium compared to layers with No-salt or alternative additives like NaCl, NaNO₃, and Na₂CO₃. Performance evaluations via Na symmetric cells indicated lower resistance and overpotential for cells treated with SHMP carbon layers compared to No-salt layers. EIS confirms that SHMP layers reduce charge transfer and ohmic resistance, correlating with improved molten sodium wetting and transport. Morphological analysis using plasma-focused ion beam (PFIB)-SEM identifies reduced cracks and voids in SHMP vs. No-salts layers, enhancing adhesion and mechanical stability while mitigating defects that compromise sodium transport efficiency. Collectively, these findings demonstrate the multifaceted benefits of SHMP as a structural binder, offering transformative implications for the design of carbon-based interfaces in molten Na electrochemical systems for lowering its operating temperatures.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fbael.2026.1759353/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Amjad ALI, Nanjing Forestry University, China

Reviewed by: Jamile Mohammadi Moradian, University of Science and Technology of China, China

Cuijuan Zhang, Tianjin University, China