Compression Pad Selection

Pad suppliers often frame foam selection as a material choice: EVA or silicone, small pores or large, V-0 or not. That gets the order the wrong way round. The pad’s job is to hold the cell inside a pressure window from build to end of life, and the material follows from that window. This post is a companion to our Compression Pads page and builds directly on the worked numbers in LFP Cell Clamping Pressure.

Hand-drawn sketch of a cell stack between two end plates held by a tie rod. At beginning of life the foam pads between the cells are lightly compressed; at end of life the same stack length holds cells that have grown and bulge on both faces, pinching the pads between them hardest at mid-face and raising the end-plate force.
Same stack length, more cell: as cells grow and bulge over life, the pads take up the difference and the end-plate force climbs. Schematic, not to scale.

Start with the pressure window, not the material

A pad sees two kinds of movement: the small reversible breathing of each charge and discharge, and the slow irreversible growth over life, which can reach 10 to 20% of cell thickness (see Cell Expansion and Cell Mechanics). Pressure has to stay above a lower bound so the layers stay in contact, and below an upper bound set by the cell maker and by the end plates.

The two ends of that window fail differently. Cannarella and Arnold found that a small stack pressure improved capacity retention over unconstrained pouch cells, while higher pressures shortened cycle life [1]. Wang et al. went further and held pressure constant with pneumatic bellows: too little pressure drove cathode cracking, too much drove lithium plating [2]. We cover that paper in What’s the right stack pressure?, including why its 12.5 bar optimum should not be lifted straight into a specification.

So the first deliverable is not a foam name. It is four numbers: beginning-of-life (BOL) pressure, end-of-life (EOL) pressure limit, the irreversible growth per cell, and the maximum sustained temperature the pad will sit at.

The three properties that actually decide it

Compression force deflection (CFD). This is the stress-strain curve of the foam. What we want is a long, flat plateau, so the force changes little while the cell grows. Rogers, for example, describes its EVExtend grades as having a flat CFD curve intended to meet both BOL and EOL limits [3]. Two practical points: read the curve at your preload strain, not at the headline 25% value, and note the test rate, since datasheets quote a slow quasi-static rate rather than anything resembling cell breathing.

Compression set. This is the share of the deflection a foam does not recover after being held compressed and hot. A pad that takes a large set stops pushing, so BOL pressure quietly decays. The test condition matters as much as the number: a figure at 23°C tells us little about a pack that soaks at 50 to 60°C. Battery-grade PU datasheets typically quote set at 70°C, with Rogers specifying a 5% maximum on selected EVExtend 43 grades [3] and Saint-Gobain quoting under 10% at 70°C and 70% compression for its PF27 series [4].

Stress relaxation. This is the force lost over time at a fixed strain. It is the property that most directly describes what happens in a bolted or banded module, because the stack is held at constant length, not constant load. It is also the property least often published. If a supplier cannot give a relaxation curve at temperature, that is worth asking for before anything else.

A worked example: sizing a pad for the EVE LF50F

Using the cell from our LFP clamping post: the EVE LF50F face is 148.3 x 127 mm (0.01883 m²) and the specified starting force is about 1.96 kN, which gives roughly 104 kPa, or about 1 bar.

Step 1, pick the grade by preload. The 192 kg/m³ EVExtend 43 grade has a typical CFD of 61.1 kPa at 25% deflection [5]. That is well short of 104 kPa, so to reach the cell’s preload it would have to be pushed far up its curve, which uses up the travel we need for swelling. A denser grade is the better starting point: the higher density EVExtend 43 range quotes 48 to 186 kPa at 25% across 224 to 288 kg/m³ [6]. Our reading is that a grade near the middle of that range would put 104 kPa somewhere around 25% strain, but this needs checking against the full curve on the datasheet.

Step 2, check travel. At 10% growth the cell thickens by 2.67 mm (10% of 26.7 mm). If one pad sits between each pair of cells, each pad must absorb roughly that much. The minimum pad thickness is:

tmin=Δεmax−εpret_{min} = \frac{\Delta}{\varepsilon_{max} – \varepsilon_{pre}}

where Δ is the irreversible growth to be absorbed per pad (mm), εpre is the strain at preload and εmax is the highest strain before the foam stiffens sharply. Taking εpre = 0.25 and εmax = 0.70 (our assumption, chosen because the PU datasheets run their compression set tests at 70% [4]) gives tmin = 2.67 / 0.45 ≈ 5.9 mm. At 20% growth it doubles to about 11.9 mm.

That is the useful result. EVExtend 43 is sold from 1 to 3 mm [3], so a single 3 mm pad per cell has only about 1.35 mm of useful travel by this estimate, half of what the 10% case needs. The options are a thicker or stacked pad (costing stack length and energy density), accepting that pressure will climb into the foam’s stiff region late in life (so the end plates must be sized for it), or a different stack strategy altogether. Nigel’s comment on the LFP post quotes EOL module forces around 50 kN, which shows where that last option ends up.

Prismatic cans bulge mostly at the centre of the large face while the edges stay close to nominal (Zlatko Dilberović makes this point in the LFP Clamping Pressure post comments). That means the 10 to 20% figure overstates the travel needed at the pad edges and understates local strain at the centre.

PU, silicone and EVA compared

The supplier post that prompted this piece compared EVA and silicone. Microcellular polyurethane belongs in that list too, since it is the material most compression pad ranges are built on: both Rogers and Saint-Gobain offer PU and silicone families specifically for cell stacks [3][7].

PropertyMicrocellular PUSilicone foamEVA foam
Example gradesRogers PORON EVExtend; Saint-Gobain Norseal PFRogers BISCO; Saint-Gobain Norseal F-seriesNot found (see text)
Compression set (supplier test)≤5% (EVExtend 43, selected grades) [3]; <10% (PF27) [4]<5% (Norseal F-20) [7]Not found
Rated temperature90°C constant, 121°C intermittent (EVExtend) [3]; 70°C (PF27) [4]-55 to 200°C (BISCO cellular) [8]Not found
Low temperatureEmbrittlement -20°C (EVExtend) [3]Flex pass at -55°C [8]Not found
Typical UL94HBF on cell-pad grades [3][7]V-0 [7]Grade dependent
Watch-outNarrower temperature range; check grade, not familyVolatile siloxanes near contactsCreep and set at temperature
Values are supplier datasheet figures for named grades, not family-wide properties. Test conditions differ between suppliers, so compare like with like.

PU is the default for a reason: it is the material the dedicated battery-pad grades are built on, with compression set quoted at 70°C and flat CFD curves. The catch is the temperature ceiling, which varies a lot by grade: one PU series is rated to 70°C constant use [4], another to 90°C [3].

Silicone holds its properties over a much wider temperature range and reaches V-0 more easily. The penalty that rarely appears in pad comparisons is outgassing. Low molecular weight siloxanes can migrate into relays and switches, decompose under arcing to silica and cause contact failure; Omron advises against silicone coatings even on the outside of a relay [9]. In a module that shares an enclosure with contactors or low-current signal relays, ask for the supplier’s volatile siloxane content.

EVA is the cheap option, but I could not find a battery-grade EVA datasheet that quotes compression set or stress relaxation at 70°C. The nearest evidence is a footwear-foam patent that reports better compression set than its EVA comparison foam, with the new foams below 35% at 50°C [10]; that is outside our application and should be read only as a hint that EVA creep at temperature needs testing, not as a figure. Absence of data is not proof of poor performance, but it does mean the burden of proof sits with the supplier.

Over to you: have you seen EVA used as the main compression pad in a production module, rather than as an insulator or spacer? If so, in what kind of application, and how did it hold its preload over life? Let us know in the comments below.

UL94 is not a propagation rating

UL94 is a small-flame ignition test on a material sample. It says nothing about what a pad does when the neighbouring cell goes into thermal runaway. Recent lab work is a useful reminder: in a three-cell module test, a conventional 3 mm silicone foam between cells delayed propagation but did not stop it, while a ceramifying silicone composite with glass fabric reinforcement did [11].

In practice this is why suppliers sell separate thermal runaway protection (TRP) pads that bond a compressible layer to mica or an encapsulated aerogel, for example Saint-Gobain’s PU-and-mica and aerogel products alongside its cushion foams [12]. If propagation is a requirement, it should be designed and tested as its own function. Choosing the foam with the better flame rating does not cover it.

What about pore size?

Pore size is a means, not a specification. Density and cell structure set the CFD curve and compression set, and closed cells keep water uptake low. What we should specify is the curve, the set and relaxation at temperature, the thickness tolerance and the dielectric strength. If a supplier meets those, the pore structure has done its job.

A selection checklist

  1. Define BOL pressure, EOL pressure limit, growth per cell and maximum sustained temperature.
  2. Pick candidate grades whose CFD curve passes through the BOL pressure at low strain (around 20 to 30%).
  3. Size thickness from the travel equation above, and check where the EOL strain lands on the curve.
  4. Ask for compression set and stress relaxation at your temperature, not 23°C.
  5. Check thickness tolerance in the stack-up: ±10% on a 3 mm pad is ±0.3 mm per cell, which adds up over a long stack.
  6. Check dielectric strength, outgassing (silicone) and adhesive or liner compatibility.
  7. Treat propagation as a separate requirement with its own barrier and test.
  8. Validate on cells in a fixture of representative stiffness. As Nick Milanovic notes on the LFP post, a cell maker’s ageing data may come from a single end-plate stiffness.

The best overall option for most pouch and prismatic stacks comes out as a battery-grade microcellular PU, sized from the pressure window, with silicone reserved for packs that genuinely run hot or cold, and a separate barrier layer where propagation matters. The pad that looks cheapest per square metre is rarely the cheapest once lost preload and extra end-plate mass are counted.

References

  1. Cannarella, J. & Arnold, C. B. Stress evolution and capacity fade in constrained lithium-ion pouch cells. J. Power Sources 245, 745-751 (2014).
  2. Wang, H. et al. The interplay between stack pressure, mechanical expansion and degradation pathways in lithium-ion batteries. Nat. Energy (2026).
  3. Rogers Corporation, PORON EVExtend product page (vendor data).
  4. Saint-Gobain Tape Solutions, Norseal PF27 Series technical data sheet (vendor data).
  5. Rogers Corporation, PORON EVExtend 43 data sheet, via Marian Inc. (vendor data).
  6. Rogers Corporation, PORON EVExtend higher density 4701-43 data sheet, via JBC (vendor data).
  7. Saint-Gobain Tape Solutions, Tape Solutions for EV Battery Pack Protection brochure (vendor data).
  8. Rogers Corporation, BISCO silicones product page (vendor data).
  9. Omron, Relay FAQ: siloxane and contact failure.
  10. US Patent 9,527,974, Ethylene-based polymer compositions and foams.
  11. Constructing Intrinsically Safe Lithium-Ion Battery Energy Storage via Gradient-Laminated Ceramifiable Silicone Foams, Nano-Micro Letters (2026), read via the authors’ summary.
  12. Saint-Gobain Tape Solutions, EV battery compression pads and TRP pads (vendor data).

See also: Compression Pads, LFP Cell Clamping Pressure, What’s the right stack pressure?, Stack Pressure & Interfacial Resistance for Pack Designers, Cell Electrode Pressure, Cell Expansion, Thermal Runaway

Get the next teardown in your inbox

New teardowns, benchmarks and design data from BatteryDesign.net, free by email. Unsubscribe any time.

Leave a Comment