A Cell-to-Cell Thermal Barrier Validation Plan

1. Define the claim at the correct test level

Write one bounded claim before selecting a construction. For the specified initiating-cell condition, the target might be no adjacent-cell thermal runaway during the observation window. If propagation cannot be excluded, define a minimum delay and protected-side temperature limit. State whether the evidence covers a material comparison, installed cell stack, module or larger ESS safety case; these are not interchangeable claims.

UL Solutions separates UL 9540A evaluation into cell, module and larger system levels. Cell-level work addresses thermal-runaway characteristics and released gases; module-level work examines propagation tendency, heat and gas release, and ignition or deflagration potential. This hierarchy also helps bound internal engineering tests. A converted barrier is an input to module behavior, not a stand-alone UL 9540A-certified product.¹

• Claim boundary: material, converted part, assembled stack, module, unit or site
• Outcome: no adjacent-cell runaway, minimum propagation delay, or a defined thermal-response limit
• Observation window and post-event inspection criteria

2. Characterize the initiating cell and exposure envelope

Freeze the conditions that control event severity: cell format, chemistry, capacity, conditioning, state of charge, orientation and trigger location. Record how the trigger differs from a credible field failure. Map cell-can contact, radiant heating, conduction through tabs or structure, hot-particle impact and vented-gas paths. A vent jet that bypasses the barrier edge makes a favorable through-thickness coupon result unrepresentative of the installed failure path.

Sandia's multi-cell pouch study shows why propagation must be tied to stack energy and thermal architecture. The researchers examined reduced state of charge and metal interlayers, related propagation limits to stored energy density, and used delays between cell events to study heat transfer. Do not transfer those results directly to another chemistry or geometry; test the actual state of charge, inter-cell contact and heat-capacity conditions.³

• Cell identity, lot, capacity, chemistry, conditioning and state of charge
• Trigger method, location, input energy and trigger-cell orientation
• Nominal and tolerance-limit gap, applied compression and restraint stiffness
• Vent outlet, edge bypasses, busbar paths and cooling plate contacts

3. Convert the hazard map into barrier requirements

Treat the barrier as a set of functions, not one thermal-conductivity number. It may need to reduce transient heat flow, remain positioned as cells expand, resist hot-gas erosion, preserve electrical isolation after deformation and avoid blocking a vent. Thickness, density, facings, adhesive, edge geometry and compression can alter these functions. Rank each against the exposure and distinguish safety-critical functions from packaging needs.

The safety case also needs controls beyond the inter-cell part. Sandia's grid-scale hazard analysis identified unsafe control actions and loss scenarios, then converted them into design objectives intended to reduce thermal-event likelihood or severity. Apply that logic locally: for each heat, gas, electrical and mechanical path, name the preventive control, mitigation, required evidence and consequence if the control is absent.²

• Through-face heating | screen transient thermal response | verify protected-cell response in the installed stack
• Edge or seam bypass | inspect coverage and assembly tolerance | instrument the exposed edge during the module event
• Vent-jet impingement | evaluate erosion, displacement and gas routing | confirm the representative vent path
• Compression or cell growth | measure thickness and force behavior after conditioning | verify that the barrier remains located

4. Use three evidence gates before release

Gate one is material and converted-part screening. Compare exact stack-ups at specified thickness, compression, temperature and exposure duration. Measure responses tied to the failure map, such as transient cold-side temperature, dimensional change, edge integrity and post-exposure electrical condition. Use the results to eliminate weak constructions and understand sensitivities, not as proof of module non-propagation.

Gate two is installed-stack verification. Assemble representative cells or thermal surrogates with production-intent gap, restraint, barrier outline, facings and adhesive orientation. Apply the expected pressure and a justified tolerance corner; instrument faces, edges and structural bypasses. Find integration failures such as curling, incomplete coverage, adhesive degradation, local crushing or heat flow around the barrier before consuming propagation-test hardware.

Gate three is a representative module propagation test using the defined cell and trigger. Record cell voltages, temperatures, event timing, venting, relevant pressure and post-test damage. UL Solutions identifies propagation tendency, heat and gas release, and ignition or deflagration potential as module-level UL 9540A concerns. Agree the formal plan and edition with the responsible test body; development tests do not replace required system evidence.¹

• Gate 1 output: ranked constructions and documented material-level limitations
• Gate 2 output: verified assembly window and identified bypass paths
• Gate 3 output: pass or fail against the predefined module outcome, with event chronology

5. Build a test matrix that exposes sensitivities

Use a controlled baseline and change one primary factor at a time during screening. Cover construction, installed-thickness limits, compression, cell gap, vent orientation and relevant conditioning. Include an unprotected or current-production reference when it can be tested safely. Set the repeat strategy from expected cell variability; one favorable event is not a robust design margin.

Separate acceptance limits from diagnostics. Acceptance might concern adjacent-cell runaway, delay or a protected-surface limit. Diagnostics include time from initiating-cell vent to neighbor response, face and edge temperatures, barrier displacement, gas bypass and restraint deformation. Sandia used gap-crossing and cell-crossing time to describe propagation in its pouch-cell stacks. Similar definitions can aid comparison if set before testing without transferring Sandia's numeric results.³

• Factor matrix: construction, thickness, compression, gap, cell condition, orientation and conditioning
• Acceptance channels: neighboring-cell state, delay and protected-side thermal limits
• Diagnostic channels: face and edge temperature, voltage, event timing, pressure and visible vent path

6. Freeze the validated construction and change controls

The released drawing should identify material grade or controlled specification, thickness, outline, critical edge coverage, cutouts, facing and adhesive orientation, liner or pull-tab needs, assembly direction and inspection points. Link specimens to material lot, conversion route, drawing revision, cell lot and assembly record. Photograph the as-built stack so failures can be separated into material, conversion, fitment and system causes.

Define changes requiring review or revalidation: cell chemistry or supplier, state-of-charge window, gap, restraint, vent direction, barrier thickness, facing, adhesive, cutout or adjacent structure. Ask whether the change alters stored energy, heat capacity, contact resistance, gas routing or the safety-control structure. Sandia's propagation research and hazard analysis support treating those interactions as configuration-dependent.^{2, 3}

• Release only the tested or technically justified construction window
• Keep sample genealogy and test configuration with the validation report
• Route cell, barrier, stack-up and vent-path changes through documented impact review
• Repeat the appropriate evidence gate when a change can affect the failure path