A lithium-ion cell is not a material; it is a laminate — dozens of repeats of metal foil, porous electrode and separator, wound or stacked — and its thermal behavior belongs to the architecture more than to any ingredient. Along the layers, metallic current collectors give heat a highway; across the stack, every porous, electrolyte-soaked, imperfectly-contacted interface adds resistance in series — and the two directions commonly differ by one to two orders of magnitude — direct measurements on cylindrical cells and full-cell reviews put the contrast squarely in that range12. That anisotropy is not a curiosity: it decides which cooling surfaces actually cool, how steep internal gradients grow at fast charge, and how quickly a local fault’s heat reaches its neighbors — a front whose passive diffusion timescale is set by diffusivity while the event itself is also driven by reaction kinetics and gas transport3. This article maps the property set that battery thermal design runs on — k in both directions, α, and cp — how each is honestly measured on real cells and components, and where the numbers move with state of charge, temperature and age.
| Measured directly | Effective cell- or component-level response (TPS on a face, flash on extracted stacks) |
|---|---|
| Derived | Effective k∥, k⊥, α and cp for the architecture as tested |
| Required inputs | Stack geometry, density, calorimetric cp, wetting and state of charge |
| Direction | In-plane and through-plane reported separately |
| Main corrections | Contact to the cell face, layer contact resistance, state and temperature |
| Reported uncertainty | Design-, state- and age-dependent — measure the working design |
- 1Why a cell is a thermal laminate
- 2The four numbers and what each one governs
- 3Interactive: the runaway front is a diffusivity race
- 4How battery thermal properties are actually measured
- 5Numbers that move: state, temperature, age
- 6Frequently asked questions
- 7Keep exploring the knowledge hub
- 8References
Why a cell is a thermal laminate
Unroll a cylindrical cell or open a pouch and the thermal problem announces itself: alternating micrometer-scale layers of aluminum and copper foil, particulate electrodes bound in porous scaffolds, polymer separators wet with electrolyte. In-plane, the metal foils dominate — continuous high-conductivity sheets that turn each layer into a lateral heat highway. Through-plane, heat must cross every porous electrode and every soaked separator in series, each contributing its low intrinsic conductivity plus contact resistance at its interfaces — the same series arithmetic our interface article formalizes, repeated dozens of times per millimeter45. The measured consequence across cell formats: in-plane conductivities of order tens of W m⁻¹K⁻¹ against through-plane values near or below one — large-format pouch measurements report roughly 26.6 in-plane against 0.52 through-plane on one commercial LFP cell — a built-in anisotropy that every pack-level thermal model inherits12.
The four numbers and what each one governs
k∥ governs where cooling works. Tab and edge cooling exploit the in-plane highway; large-face cooling of a wound cell fights the through-plane bottleneck — the geometry of thermal management is chosen by this one ratio6. k⊥ governs internal gradients. Fast charge deposits heat throughout the jelly roll; the through-stack resistance sets how steep the core-to-skin gradient grows, and with it the aging spread between inner and outer layers. cp is the buffer. Volumetric heat capacity converts watts of loss into degrees per second of drift; it also links the other numbers through k = αρcp, the triangle every derived datasheet value silently uses7. α sets the passive clock. When a cell faults, the question of neighbors is a transient one, and thermal diffusivity sets the timescale on which released heat is redistributed by conduction — the k-versus-α distinction our foundation comparison makes precise. The actual propagation front is not diffusivity alone: it also depends on exothermic reaction kinetics, gas and ejecta transport, interfaces and pack architecture, which is why runaway models couple all of them rather than solving a heat equation38.
Interactive: the runaway front is a diffusivity race
The simulator stages the safety question as the race it physically is: a local heat pulse ignites at one end of a cell strip, and the warm front advances as √(αt) toward a neighbor at the far end. Drag the effective through-stack diffusivity — one of the knobs cell architecture, interlayer contact and barrier materials turn — and watch arrival times stretch quadratically; the conductivity-flavored slider deepens how much heat floods in without changing who arrives when. Treat this as a diffusion-only baseline: it deliberately omits reaction heat, venting and phase change, so it shows the conduction clock rather than a runaway prediction.
The model is the same normalized one-dimensional diffusion solution used throughout this hub — a teaching shape for the √(αt) law and the k/α role split; real runaway adds exothermic reaction heat, venting and ejecta transport, and pack geometry — engineering models couple those to the diffusive backbone rather than scaling it3.
How battery thermal properties are actually measured
Multiscale by necessity, and reviewed as such in the battery-thermal literature9. Component level first, because cells are laminates of measurable parts: electrode coatings and separators meet the film toolbox — through-plane by flash on stacked discs or by steady sandwich methods, in-plane by suspended-strip electrothermal transients, with wet-versus-dry state changing answers enough that the state belongs on every report1011. Cell level next: the transient plane source pressed against a cell face reads effective through-plane properties with minimal preparation12; flash serves extracted layer stacks as discs10; cp comes from calorimetry, whose product closes the k = αρcp triangle for every derived value7. Direction discipline is the whole game: a through-plane TPS number quoted for an in-plane cooling calculation is the classic battery-datasheet failure, and the four-field reporting habit — direction, state, temperature, method — is the antidote this hub’s datasheet guide generalizes13.
Numbers that move: state, temperature, age
Battery thermal properties are operating points, not constants. Electrolyte state: wetting fills pores with a conducting liquid; dry component data understate the assembled stack, and the difference is measurable, not cosmetic4. Temperature: polymer separators, electrolyte and interfaces all drift across the −20 to +60 °C service window, so single-point room-temperature values under-specify a pack’s winters and summers6. State of charge and age: lithiation changes electrode lattices; cycling changes porosity, contact and gas content — effective properties evolve over life, which is why serious programs measure aged cells rather than extrapolating fresh ones613. The reporting consequence is the sentence this hub keeps writing: value, direction, state, temperature, method — and for safety-relevant numbers, a measured α rather than a triangle-derived one, because the runaway clock deserves its own experiment.
Frequently asked questions
What is a typical cell’s thermal conductivity?
Reported values cluster around tens of W m⁻¹K⁻¹ in-plane and well under one through-plane — one large-format LFP pouch cell measured 26.6 and 0.52 respectively — but they are format-, chemistry-, state- and temperature-dependent, which is why full-cell reviews stress method selection alongside the numbers12.
Which direction matters for my cooling concept?
Follow the heat path: tab/edge cooling rides k∥; face cooling fights k⊥ plus interface resistances to the cold plate. The ratio between them is the first number a pack architect should demand6.
Does diffusivity govern runaway propagation?
It governs the passive part: conduction redistributes released heat on a √(αt) timescale. Propagation as a whole is a coupled problem — reaction kinetics, gas and ejecta transport, interfaces and module geometry all participate — so α is a necessary input, not the whole answer3.
Can I measure a cell without disassembly?
Effective through-plane properties, yes — contact transients on the cell face are standard practice; direction-resolved and component-resolved data require opening the architecture, done safely at component or dummy-cell level1210.
How do thermal barriers work?
Through several mechanisms at once: lowering effective conductivity and diffusivity along the path, adding heat capacity, absorbing energy through endothermic decomposition or phase change, raising interface resistance, and managing vent gases. The simulator shows only the first of these — barrier selection is evaluated against the coupled event, not a diffusion curve3.
Keep Exploring the ACS Thermal Metrology Knowledge Hub
This article is one chapter of the ACS thermal metrology knowledge hub. To keep going:
- Thermal conductivity & diffusivity testing: the pillar guide — methods, samples and a buyer’s framework in one place.
- Thin-film thermal conductivity: in-plane vs cross-plane — why one number is never enough.
- Interfacial thermal resistance — the invisible wall inside every stack.
- Thermal conductivity vs thermal diffusivity — why the two rankings disagree and which methods time which.
- Transient plane source (Hot Disk TPS) explained — the spiral sensor for bulks, powders and pastes.
- Laser flash analysis (LFA) explained — the rear-face transient workhorse, and where it fails.
- ACS thermal testing services — the team that turns these distinctions into data on your specimen.
References
This article discusses battery thermal properties for educational purposes. Quoted magnitudes are representative literature ranges for orientation; actual values are architecture-, state-, temperature- and age-dependent and require measurement on the specific design. The runaway-front simulator is a normalized diffusion-only teaching model that omits reaction heat, venting and phase change; it is not a safety prediction and must not be used to assess propagation risk. Consult cell datasheets and safety documentation, and contact our thermal testing team for measured properties.