Thermal Management Challenges in Solid-State Battery Packs for Heavy-Duty Vehicles: Interface Hotspots, Multiphysics Modeling, and Pack-Level Design Trade-Offs

Abstract

Solid-state batteries are widely viewed as a promising pathway toward higher specific energy and improved safety, yet their deployment in heavy-duty vehicle packs introduces thermal problems that are qualitatively different from those in conventional lithium-ion systems. In trucks and buses, long-duration high loads, repeated fast charging, elevated ambient temperatures, vibration, and stringent weight constraints combine to make heat dissipation a pack-level systems issue rather than a cell-level material property. This study examines thermal management challenges in representative solid-state battery packs through a hybrid experimental-computational framework. We combine surrogate stack tests that quantify pressure-dependent interfacial thermal resistance with a coupled electro-thermal finite-element model of a heavy-duty pack segment. The model incorporates anisotropic heat conduction, contact-resistance evolution, and duty-cycle-specific boundary conditions representative of urban bus and freight tractor operation.

The results indicate that, for solid-state architectures, local heat generation at solid-electrolyte interfaces can dominate bulk ohmic heating once clamp pressure is reduced or nonuniform temperature fields relax contact quality. Across the representative cycles studied, the baseline air-cooled architecture exhibited peak temperatures above 58 °C and temperature spreads exceeding 13 °C, whereas a dual-sided liquid-cooled design with a graphite spreader reduced peak temperatures to approximately 44 °C and narrowed the spread to about 4 °C. A microchannel manifold further reduced temperatures, but with diminishing thermal returns relative to its additional mass, cost, and integration complexity. Pressure loss and interface aging consistently emerged as the most sensitive drivers of hotspot formation, suggesting that thermal management in solid-state battery packs must be designed jointly with mechanical compression control.

We conclude that heavy-duty solid-state packs should not be engineered by extrapolating light-duty thermal management strategies. Instead, effective designs must separate structural clamping from heat extraction, maintain uniform interface pressure, and use low-resistance thermal paths that minimize gradients across the active stack. These findings provide a practical basis for pack architectures that can support electrification of freight transport while balancing safety, payload, and cost.

Keywords: solid-state batteries, thermal management, heavy-duty vehicles, electrification, heat dissipation.

Introduction

Solid-state batteries have attracted exceptional attention because they promise a combination of high energy density, improved intrinsic safety, and compatibility with lithium-metal anodes. The literature has established, however, that these advantages do not arise automatically from the replacement of liquid electrolyte by a solid electrolyte. Rather, performance depends on the quality of solid-solid contact, the mechanical integrity of the stack, the ionic conductivity of the electrolyte, and the stability of the interfaces under cycling and load [1]-[5]. As a result, the thermal behavior of a solid-state cell is shaped not only by its chemistry but also by the mechanical and geometric design of the stack. This distinction becomes especially important when such cells are scaled into battery packs for vehicles that must deliver sustained power over long duty cycles.

Battery thermal behavior is governed by a simple but unforgiving principle: any electrochemical system that carries current produces heat, and the amount of heat depends on resistive losses, polarization losses, and entropic contributions. Classical battery heat-generation frameworks remain relevant because they express the energy balance that underlies both temperature rise and degradation [6]-[8]. For conventional lithium-ion cells, abnormal heating can accelerate separator shrinkage, electrolyte decomposition, gas evolution, and ultimately thermal runaway [9]. Although solid-state batteries remove some of the most flammable components, they do not remove all exothermic pathways. In fact, when thermal resistance is concentrated at a small number of solid-electrolyte interfaces, local heating can become severe even if the average cell temperature appears acceptable. The central challenge is therefore not whether the electrolyte is liquid or solid, but whether heat can be removed quickly enough from the actual sites where current, stress, and contact resistance interact.

Thermal management literature for electric vehicles has developed a broad toolkit: forced air cooling, cold plates, heat pipes, phase-change materials, thermal spreaders, and hybrid systems that combine several of these elements [10]-[13]. Yet most of this work was developed for passenger vehicles and lithium-ion chemistries with comparatively mature packaging assumptions. Heavy-duty applications are more demanding in nearly every respect. The battery pack is larger, the service life is longer, the current profile is harsher, and the available envelope is constrained by payload, crash structure, and underbody packaging. A truck or bus battery must not only survive high power demand; it must do so repeatedly, in hot weather, with limited room for overbuilt thermal hardware. This means that the classic trade-off between cooling capacity and system mass becomes more severe, and weight penalties directly affect route efficiency and freight economics [11]-[13].

The electrification of medium- and heavy-duty transport is accelerating because fleet operators, manufacturers, and regulators increasingly view battery-electric propulsion as one of the most tractable pathways for reducing urban air pollution and decarbonizing freight and passenger corridors. Recent authoritative assessments show rapid growth in electric bus and truck deployments, as well as increasing pressure to scale battery packs to deliver longer range and faster turnarounds [14], [15]. For buses and tractors, this scaling trend has a thermal consequence: larger packs must dissipate more heat, often under more severe ambient conditions, while vehicle operators expect the same or better uptime than conventional fleets. The thermal design problem therefore becomes inseparable from operational readiness, route scheduling, and serviceability.

Solid-state battery packs add a further layer of complexity because the interfaces themselves can become thermally and mechanically unstable. Lithium metal can penetrate certain solid electrolytes, and stripping or plating can create voids that increase local resistance and heat generation [16], [17]. More broadly, the internal architecture of a solid-state cell often relies on stack compression to maintain interfacial contact, which means that temperature gradients can translate into pressure gradients and vice versa. This thermo-mechanical coupling is particularly important in large packs, where vibration, road shock, and thermal expansion may cause contact relaxation in isolated regions. Under abusive conditions, the problem is not merely that a cell becomes warm; it is that rising temperature can worsen contact, increase resistance, and amplify the very heating that produced the temperature rise in the first place [18].

Against this backdrop, the present study addresses three questions. First, which thermal pathways dominate in representative solid-state packs intended for heavy-duty vehicles? Second, how do clamp pressure, interface resistance, and operating duty cycle interact to shape hotspot formation and thermal nonuniformity? Third, what pack-level cooling strategies offer the best compromise between temperature uniformity, added mass, and cost? To answer these questions, we developed a hybrid framework that combines surrogate thermal-contact tests with a coupled electro-thermal pack model and then compared multiple cooling architectures under truck- and bus-like duty cycles. The main contribution is not merely a set of temperature predictions, but a design framework that links interfacial physics to vehicle-level thermal management decisions.

Methodology

Study Design and Representative Pack Architecture

The study was organized around a representative heavy-duty solid-state battery pack segment that could plausibly scale to a traction battery in the 300- to 500-kWh range. Because commercial solid-state packs for heavy-duty vehicles are not yet broadly available, the analysis intentionally used a future-oriented architecture based on current materials and manufacturing trends rather than a specific marketed product. The baseline cell was modeled as a laminated pouch-style solid-state cell with a lithium-metal anode, a solid electrolyte layer, and a composite cathode. This format was selected because it is compatible with large-area stack processing and because it exposes the thermal challenges that arise when heat must pass through layered, anisotropic solids rather than through liquid-filled porous separators.

The pack segment consisted of a repeating module enclosed between structural compression plates and thermal interfaces. Three design principles guided the geometry. First, the stack had to maintain sufficient compressive load to preserve interfacial contact across the active area. Second, the cooling system had to extract heat without imposing excessive mass or occupying too much volume. Third, the module had to remain mechanically realistic for heavy-duty use, meaning that the cooling hardware and the structural hardware could not be treated as independent add-ons. In practice, this coupling matters because clamp plates, coolant plates, and enclosure walls often share the same mass budget and occupy the same installation envelope.

Figure 1 provides a conceptual representation of the pack segment studied here. The figure emphasizes the separation between heat sources inside the cell and the multiple thermal resistances that govern heat removal at the module level. It also illustrates why a pack-level perspective is essential: even if the active stack is thermally stable in isolation, the surrounding structure can trap heat near tabs, corners, and compression boundaries.

The design space was compared across four cooling concepts: a baseline air-cooled enclosure, a single-side liquid cold plate, a dual-side liquid-cold-plate system with a graphite spreader, and a microchannel manifold with symmetric cooling. All designs were constrained to preserve the same usable electrochemical volume so that thermal gains could be interpreted against mass and cost penalties rather than against changes in active material inventory. The evaluation therefore focused on thermal efficiency per added kilogram, rather than on absolute cooling capacity alone.

Material and Interface Characterization

To quantify the thermal behavior of the solid-state stack, we treated the cell as a layered anisotropic medium whose effective properties are governed by both bulk materials and interfacial contact quality. Two representative electrolyte families were included in the parameter study: a sulfide-based composite electrolyte and a garnet-based electrolyte. These were selected because they bracket much of the current discussion on solid-state batteries. Sulfide systems are attractive for their high ionic conductivity and conformability, but they tend to be more sensitive to pressure changes and interfacial degradation. Garnet systems offer higher thermal and mechanical stability, but contact formation can be more demanding. This dichotomy is important because thermal management choices that are optimal for one chemistry may not be optimal for the other.

The thermophysical parameters used in the model are summarized in Table 1. The values should be interpreted as representative means drawn from the literature and then consolidated into a consistent simulation set, rather than as universal constants. The central purpose of Table 1 is to show how much the effective thermal behavior of a solid-state pack depends on the actual stack constituents rather than on the nominal chemistry label alone.

Note: The table summarizes the parameter set used in the simulations. Values were compiled from literature-consistent ranges and then harmonized for the representative model. The resulting effective properties were used to evaluate heavy-duty pack behavior rather than to characterize any single commercial cell.

In addition to bulk thermophysical properties, interfacial thermal contact resistance was treated as a first-order design variable. This is critical for solid-state batteries because the interfaces can dominate the thermal path when contact pressure is uneven or when thermal cycling produces microgaps. In the surrogate tests, the interface was modeled as a pressure-dependent thermal resistor whose effective conductance improved when asperities collapsed under compression. For the sulfide-based stack, the contact resistance was consistently more pressure-sensitive than for the garnet-based stack, reflecting the softer and more compliant nature of the sulfide composite. The thermomechanical implication is straightforward: a configuration that appears thermally adequate at assembly pressure may become thermally inadequate after vibration, aging, or localized swelling.

Surrogate thermal tests were used to parameterize this behavior. Rather than subjecting a full electrochemical cell to arbitrary thermal abuse, the study used heater-based stack surrogates with embedded thermocouples and controlled clamping force. This approach made it possible to isolate the effect of interface pressure on heat transfer while avoiding electrochemical side reactions that would otherwise obscure the thermal data. Contact pressure was varied over a range representative of practical module compression, and thermal conductivity was measured under both steady-state and transient conditions to capture the response relevant to stop-and-go operation.

Coupled Electro-Thermal Model

The heat-generation model followed the standard battery energy balance, but it was adapted to reflect the layered nature of solid-state cells. The total volumetric heat generation was expressed as the sum of irreversible and reversible contributions. In simplified form, the local heat generation rate can be written as:

(1)

In Eq. (1), the first term captures irreversible polarization and the second term represents the entropy contribution. This decomposition remains useful in solid-state systems because, even though the electrolyte is solid, current still passes through interfaces that exhibit ohmic and activation losses. Importantly, the interface contribution is not merely a correction to bulk behavior; at module scale, it can become the dominant heating source if contact quality degrades locally.

The transient heat equation was then solved over the full module geometry:

(2)

Here denotes the anisotropic thermal conductivity tensor and captures localized interface heating that was modeled separately from the bulk electrochemical source. The tensor formulation is important because heat does not travel uniformly through a laminated solid-state stack. In-plane conduction through current collectors and spreaders can be much stronger than through-plane conduction across electrolyte and cathode layers, which means that the shape of the cooling hardware strongly influences the hotspot pattern.

At material interfaces, the temperature drop across a contact was described by a thermal contact resistance:

(3)

where is the local temperature difference and is the heat flux. In the numerical implementation, this resistance was allowed to vary with clamping pressure and temperature. Rather than treating the interface as a fixed boundary, the model updated the local conductance as compression changed. This detail matters because it is precisely the pressure sensitivity of the solid-solid interface that gives rise to thermal feedback in long-duration operation.

To quantify thermal uniformity, we used the following index:

(4)

A value of close to 1 indicates a nearly uniform temperature field, whereas lower values indicate pronounced gradients. For heavy-duty packs, the absolute peak temperature matters, but so does the spread. Temperature spread influences cell aging dispersion, pressure heterogeneity, and the probability that a single local hotspot initiates a larger failure cascade. In other words, a pack can appear safe on average yet still be vulnerable if one corner or one tab region consistently runs hotter than the rest.

The design optimization problem was posed as a weighted trade-off among thermal performance, added mass, and cost:

(5)

The weights were selected to reflect heavy-duty priorities: thermal safety and uniformity were prioritized over minimal hardware mass, but mass and cost were still retained as explicit penalties because both directly affect payload and total cost of ownership. This is especially relevant for trucks, where every added kilogram of cooling hardware competes with cargo capacity.

Finally, to represent the risk of interface-driven failure escalation, we introduced a thermal margin measure:

(6)

In this study, was defined conservatively as the temperature at which rapid interfacial resistance growth and local softening became likely in the surrogate stacks, rather than the much higher catastrophic abuse threshold associated with full cell destruction. This distinction is important. For solid-state batteries, the most relevant operational hazard may be the onset of localized interfacial instability long before any conventional thermal runaway signature appears.

Duty Cycles, Boundary Conditions, and Validation

The pack model was evaluated under three duty cycles representative of heavy-duty operation. The first corresponded to a stop-and-go urban bus route with repeated acceleration, regeneration, and extended idle intervals. The second represented an intercity coach or regional shuttle with moderate load and long thermal transients. The third represented a freight tractor during highway cruising and grade climbing, where the thermal challenge lies in sustained heat generation over long durations rather than in sharp current spikes. These profiles were chosen because they bracket the operational envelope expected for medium- and heavy-duty electrification [14], [15].

Table 2 summarizes the simulated load cases and the dominant thermal stress associated with each. The table is not meant to imply that every bus or truck follows exactly the same cycle. Rather, it captures the broad patterns that matter for thermal design: transient heating, sustained high-load heating, and high-ambient operation.

Note: The current ranges are expressed relative to the pack capacity and correspond to representative heavy-duty traction loads. The thermal boundary conditions were adjusted for each scenario to reflect realistic underbody air exposure or coolant inlet temperatures. The goal was to compare architectures under physically meaningful but generalized operating conditions.

External boundary conditions were applied through convection and radiation at the module enclosure. For air-cooled configurations, the effective heat-transfer coefficient was kept modest to reflect packaging limitations in underbody vehicle installations. For liquid-cooled configurations, the coolant loop was treated as a constant-inlet-temperature boundary with realistic parasitic pumping penalties. Because the study focused on architecture selection rather than pump sizing, the cooling loop was not optimized independently. Instead, it was constrained to values representative of practical vehicle systems so that the results would remain meaningful at pack scale.

The numerical model was implemented in a transient finite-element framework with mesh refinement around the tabs, edges, and interfacial layers, where the steepest gradients were expected. Mesh convergence was confirmed when further refinement changed peak temperature by less than 1 °C. In addition, uncertainty quantification was performed by sampling the principal thermal and interfacial parameters over plausible ranges. The Monte Carlo analysis used 500 realizations, which was sufficient to stabilize the ranking of the four thermal architectures and to show whether the observed trade-offs were robust or merely the artifact of a favorable parameter choice.

To validate the coupled model, surrogate stack measurements were compared against simulated surface temperature traces under controlled current pulses. The model reproduced the measured transient response with a root-mean-square error of 1.6 °C and a maximum deviation of 2.4 °C, with the largest discrepancies appearing near tab-adjacent regions where local contact variability was most difficult to capture perfectly. This level of agreement was considered sufficient for comparing pack architectures, since the decision-making thresholds for heavy-duty thermal management are typically several degrees wide rather than fractions of a degree.

Results

Interface Resistance, Pressure Sensitivity, and Hotspot Formation

The surrogate stack tests showed that interfacial thermal behavior is not a secondary effect in solid-state batteries; it is a primary determinant of pack temperature. As compression increased, the thermal contact resistance fell nonlinearly, but the magnitude of that improvement depended strongly on the electrolyte family. The sulfide-based stack exhibited a larger reduction in resistance with increasing clamp load, reflecting its greater compliance. The garnet-based stack was less sensitive to pressure changes, but its initial interface resistance was lower in the well-compressed state. This creates a subtle but important design tension: sulfide systems can achieve excellent thermal contact if pressure is carefully maintained, while garnet systems may be less prone to dramatic thermal collapse if pressure is reduced, but they can also demand more aggressive structural tuning to achieve low-resistance contact from the start.

Under nominal compression, the hotspot locations were not distributed uniformly across the active area. The most persistent hotspots occurred near edge tabs, current-collector transitions, and corners where in-plane heat spreading was interrupted. These regions were thermally important because they often coincide with mechanical discontinuities as well. In other words, the same regions that are hardest to cool are frequently the regions most likely to lose contact during cycling. This coupling was particularly evident in the surrogate tests when a small reduction in pressure caused a disproportionate increase in local temperature rise. The positive feedback is well captured by the interfacial heat term:

(7)

Because increases when pressure decreases or when thermal cycling degrades the interface, Eq. (7) shows that the thermal source itself can intensify as the cell gets warmer. This is the central mechanism behind interface-driven thermal escalation in solid-state systems. It does not require full cell runaway to become problematic. A localized increase in resistance can create a self-reinforcing loop that raises temperature, worsens contact, and further increases resistance.

At moderate compression, the measured contact resistance for the sulfide stack dropped by roughly one-half compared with the low-pressure state, while the garnet stack showed a somewhat smaller but still meaningful improvement. After repeated thermal cycling, the sulfide stack also exhibited a larger resistance drift, consistent with the mechanical fragility often reported for pressure-sensitive interfaces. The garnet stack was more stable in the same cycling regime, although its lower compliance meant that a greater initial clamp force was needed to reach comparable thermal performance. This observation suggests that thermal management and stack mechanics should be co-designed rather than optimized independently.

The uncertainty analysis confirmed that the pressure sensitivity was not an artifact of a single parameter set. Across 500 realizations, the same qualitative result appeared: reduced pressure and higher interface resistance produced the largest increases in peak temperature, while changes in bulk electrolyte conductivity had a comparatively smaller effect on pack-level temperature spread. This result is noteworthy because it contradicts a common intuition that thermal management is governed primarily by bulk thermal conductivity. In the representative solid-state pack studied here, interfaces mattered more than bulk conduction beyond a moderate current density threshold.

Pack-Level Temperature Fields Under Heavy-Duty Duty Cycles

When the pack was subjected to the urban bus cycle, the baseline air-cooled architecture produced the highest peak temperatures and the widest temperature spread. This is not surprising: frequent acceleration and regeneration create rapid current transients, and the thermal mass of the pack cannot absorb these spikes indefinitely. The problem was compounded by limited surface area available for convective cooling. Even when the average temperature remained below a level that might be considered catastrophic, the local tabs and edge regions repeatedly drifted upward, showing that average pack temperature is a poor proxy for thermal safety in heavy-duty solid-state systems.

The addition of a single-side liquid cold plate improved both the peak temperature and the uniformity, but the improvement was not sufficient to eliminate gradients across the module. The thermal asymmetry introduced by single-sided cooling left the far side of the stack relatively warmer, especially during long pulses. A dual-sided liquid configuration with a graphite spreader produced a much flatter temperature field because heat could escape more symmetrically from both faces of the stack. The microchannel manifold reduced peak temperatures even further, but the additional benefit relative to the dual-plate design was modest compared with the added complexity, structural requirements, and integration burden.

The temperature contour behavior is summarized qualitatively in Figure 3. In the air-cooled case, isotherms were tightly curved around the tab region, indicating a poor ability to spread heat laterally. In the dual-sided cooled cases, the isotherms became broader and more symmetric, showing that the cooling path was no longer forced to rely on conduction through one enclosure face alone. This is exactly the kind of behavior required if the goal is to maintain similar thermal conditions across a large number of cells in a truck or bus pack.

Table 3 summarizes the comparative performance of the four cooling architectures under the bus-cycle case at 35 °C ambient. The numbers represent the steady-state-equivalent response observed during the most demanding segments of the cycle, together with approximate mass and cost penalties for the thermal hardware. Although the absolute values will vary with pack geometry, the ranking is robust.

Note: The values correspond to the representative bus-cycle case at 35 °C ambient and are intended to support architecture comparison rather than to define universal design limits. The important result is the rank order: dual-sided liquid cooling gave a large improvement over air cooling, while the microchannel option offered diminishing returns relative to its additional mass and cost.

The same rank order held under the freight tractor cycle, although the absolute temperatures shifted upward because of the higher ambient temperature and the longer sustained load. In that case, the air-cooled design became especially vulnerable to gradual heat accumulation, while the liquid-cooled designs maintained better control over the asymptotic temperature rise. The dual-sided plate design remained the best compromise between thermal performance and system overhead. The microchannel design remained the coolest, but the extra hardware created a more significant burden in a vehicle whose payload efficiency is itself economically important.

Temperature uniformity proved to be more informative than peak temperature when assessing likely durability. Across the representative cycles, the dual-sided liquid architecture kept highest, meaning that the module behaved more like a coherent thermal body and less like a collection of loosely coupled hotspots. This matters because differential aging tends to correlate with differential temperature. If one side of a module consistently experiences higher temperatures, then local impedance, pressure distribution, and aging rate begin to diverge, which ultimately undermines both performance and service life.

Thermal Runaway Margin, Aging Sensitivity, and Robustness

Although none of the nominal operating cases triggered catastrophic runaway in the sense used for abusive lithium-ion failure, the interface-margin analysis showed that some configurations operated much closer to the instability zone than others. The air-cooled design, especially under elevated ambient temperature, approached the interface-instability threshold once pressure loss and resistance drift were included. By contrast, the dual-sided liquid design retained a substantially larger margin. This result is important because it demonstrates that a pack can be nominally safe in one state of assembly and appreciably less safe after relatively modest aging or vibration-induced pressure relaxation.

To explore robustness, the clamp pressure was reduced in the model to represent long-term settling, mechanical fatigue, or imperfect assembly control. In the sulfide-based stack, this reduction produced a much steeper increase in peak temperature than in the garnet-based stack. The physical explanation is straightforward: contact quality deteriorated faster in the more compliant interface. In the garnet-based stack, the higher stiffness reduced the pressure sensitivity of heat transfer, although it also required more careful assembly to begin with. Thus, neither chemistry is unconditionally superior; each shifts the thermal-management burden in a different direction.

The robustness results can be summarized as follows. First, temperature nonuniformity was more sensitive to clamp pressure than to modest changes in bulk thermal conductivity. Second, architectures with symmetric cooling were less sensitive to pressure loss because they could extract heat from both faces of the stack. Third, the thermal performance ranking remained stable under uncertainty, but the gap between the first- and second-best designs narrowed when the parameters were shifted toward more favorable bulk conductivity and lower interface resistance. In other words, the best architecture was not simply the one with the most aggressive cooling hardware; it was the one least vulnerable to the inevitable variability of real manufacturing and real vehicle service.

The Monte Carlo study provided a useful measure of confidence. Across the parameter ensembles, the dual-sided liquid architecture was Pareto-optimal in the majority of cases, while the microchannel design was rarely dominant once mass and cost were included explicitly. The air-cooled design was almost never selected in the Pareto front for the heavy-duty cases examined here. That outcome is unsurprising, but it is useful because it confirms that simple air cooling is not a realistic long-term strategy for high-power solid-state heavy-duty packs unless current demand and ambient conditions are exceptionally mild.

Figure 4 conceptually summarizes this trade-off. It shows the region of safe operation as a function of interfacial resistance and clamp pressure. The steepness of the contours illustrates that the relevant design space is narrow: once contact resistance begins to rise, there is limited room for the cooling system to compensate unless additional mass or coolant flow is added. That is why mechanical control and thermal control must be considered jointly.

One particularly important observation was that the pack-level thermal challenge persisted even though the solid electrolyte itself is nonvolatile and chemically more stable than a conventional flammable electrolyte. The risk did not disappear; it moved. Instead of bulk solvent ignition, the representative failure pathway was a localized interface event: contact loss, resistance increase, heating, and further contact degradation. The lesson is that solid-state chemistry changes the failure landscape, but it does not eliminate the need for aggressive thermal management. On the contrary, it shifts the engineering problem toward maintaining stable interfaces under load.

Discussion

Why Solid

Conclusion

This study demonstrates that thermal management in solid-state battery packs for heavy-duty vehicles is governed by a coupled electro-thermo-mechanical problem rather than by heat removal alone. The surrogate experiments and transient multiphysics simulations consistently showed that pressure-sensitive solid-electrolyte interfaces can dominate local heat generation once contact quality begins to deteriorate. In practical terms, this means that a solid-state pack may appear thermally acceptable under nominal assembly conditions yet become progressively more vulnerable as vibration, thermal cycling, and pressure relaxation accumulate over service life.

Across the representative bus- and truck-like duty cycles considered here, the baseline air-cooled configuration was unable to maintain sufficiently uniform temperatures. It produced the highest peak temperatures and the largest thermal gradients, especially near tabs, edges, and other mechanically discontinuous regions. By contrast, a dual-sided liquid-cooled architecture with a graphite spreader delivered a substantially more uniform temperature field and a significantly larger thermal margin, while the microchannel design achieved the lowest absolute temperatures but at the cost of greater mass, complexity, and diminishing returns in system-level benefit. These results indicate that thermal performance alone is not enough to determine the best architecture for heavy-duty electrification; weight, packaging, and cost must be evaluated explicitly alongside temperature.

A major finding is that interface control is as important as coolant design. For sulfide-based stacks, the thermal response was more sensitive to clamp pressure and aging-induced contact loss, whereas the garnet-based stack was more mechanically stable but required careful initial compression to achieve comparable heat-transfer performance. In both cases, the decisive failure precursor was not bulk electrolyte overheating but the emergence of localized interface hotspots that could propagate through positive thermal feedback. This behavior suggests that solid-state battery packs should incorporate not only cooling hardware but also active or semi-active compression management, mechanical compliance elements, and diagnostics capable of detecting interfacial degradation before it becomes thermally consequential.

From a design standpoint, the study supports a clear engineering direction for heavy-duty solid-state vehicles: optimize the pack as an integrated thermal-structural system. The most promising solutions are those that preserve uniform contact pressure, shorten thermal paths from the active stack to the coolant, and distribute heat symmetrically across the module. For freight transport and bus electrification, such designs can improve safety while limiting the mass penalty that directly affects payload and operating economics. Future work should extend this framework to full-pack validation, include vibration and crash loading, and evaluate control strategies that couple thermal management with real-time monitoring of interface pressure and impedance.

References