IXPE vs EVA Underlayment for SPC Flooring: What the IIC Number Actually Measures

Ask ten flooring suppliers whether IXPE or EVA is the better underlayment for SPC flooring and you will get ten versions of the same answer: IXPE, because it scores higher. The comparison tables behind that answer are full of two-letter ratings — IIC 55, IIC 66, STC 71 — presented as though they were properties of the foam, like density or thickness.

They are not. An IIC rating describes a floor-ceiling assembly: a structural slab, whatever sits on it, and the room below. A millimetre of foam does not have one. What an underlayment contributes is an improvement measured against a specified reference floor, and that improvement changes when the floor underneath changes. Once you see the number that way, most of the published IXPE-versus-EVA tables stop being comparable — and a specification decision that looked simple turns into a document request.

Written for flooring distributors, contractors and project buyers specifying rigid-core SPC flooring for multi-storey residential, hospitality and commercial projects.

Key takeaways

• An underlayment has no IIC rating. It has a measured improvement — ΔIIC (ASTM E2179, classified under ASTM E989) or ΔL_w (ISO 10140-3, rated under ISO 717-2) — against a stated reference floor.
• The same pad produces different improvement figures on different floors. ASTM E2179 states that a difference spectrum measured on concrete must not be used to estimate performance on lightweight joist floors.
• ASTM IIC and ISO L_n,w are separate systems that must not be compared directly; the shortcut IIC ≈ 110 − L_n,w holds only under a condition stated in the ASTM documents themselves.
• Laboratory results are measured with flanking transmission controlled. Field results may be lower, or may differ in other ways, because a real building has flanking paths, boundary conditions and workmanship the laboratory does not reproduce.
• IXPE and EVA differ most in how the foam is made and how it recovers under sustained load — not in some universal sound score.
• EVA is a defensible specification in the right project. Where it stops being defensible is decided by the project’s documents — a required improvement figure, a heating declaration, an emissions certificate — not by the material name.
• Never add a second loose layer over a plank that already has a factory-attached pad unless the manufacturer confirms it.

2. How to choose, in one table

The material question is the second question in every row below. The first is always: what does the project require, and can the supplier document it?

Project situation	What usually decides it	Document to obtain before specifying
Multi-storey residential with a code impact-sound requirement	The documented improvement figure for the project’s floor type — not the foam name	ΔIIC (ASTM E2179 / E989) or ΔLw (ISO 10140-3 / 717-2) with the reference floor stated
Hotel or serviced apartment	Recovery under furniture and trolley loads across the service life, plus the acoustic figure	Pad density and thickness, and the plank’s own load and locking-system data
Commercial or retail with heavy point loads	Long-term thickness retention, not the headline rating	Compression and load data for the specific pad, under stated test conditions
Over underfloor heating	Total thermal resistance of the build-up and the manufacturer’s heating declaration	Written heating compatibility, total thermal resistance, maximum permitted surface temperature
Project with an emissions scheme	A certificate for the finished product as supplied	FloorScore, GREENGUARD Gold, AgBB or whichever scheme the specification names
Ground level with no acoustic requirement	Budget, and whether a resilient layer is needed at all	Confirmation of what the plank already carries

1. What the two names actually tell you

The two names do not describe the same kind of thing, and that is where most published comparisons go wrong.

IXPE — irradiation cross-linked polyethylene. The name describes a process: polyethylene is extruded into sheet, cross-linked by electron-beam irradiation, and then expanded into a closed-cell foam. Cross-linking ties the polymer chains into a network, so a cell wall that is squeezed can spring back instead of staying collapsed. Done under good process control, the result is a fine and uniform cell structure.

EVA — ethylene-vinyl acetate copolymer. The name describes a polymer, not one manufacturing route. EVA foam is produced with different formulations, densities and cross-linking processes — cross-linked EVA foam is widely manufactured, and so is uncrosslinked material. Nothing about the performance of a particular pad follows from the letters “EVA” on its own.

That is why the usual shortcut — IXPE is cross-linked, EVA is not, therefore IXPE recovers better, resists heat better and smells less — should not be written into a specification. Cross-linking is a real variable, but so are resin grade, density, foaming control and the finished plank construction. Two EVA pads from different production lines can differ from each other more than either differs from an IXPE pad.

Two more things to watch when reading quotations. IXPE is sometimes used loosely to cover chemically cross-linked polyethylene as well as irradiation-cross-linked material, and the two are not the same process. And irradiation cross-linking says nothing by itself about how the foam was expanded: IXPE is commonly produced with a chemical blowing agent that decomposes at temperature to release gas. Ask which route was used, and ask for the pad’s measured properties, rather than inferring them from the name.

3. The IIC number on a data sheet belongs to an assembly, not a pad

This is the point that separates a usable specification from a plausible-looking one.

What IIC actually is

Impact Insulation Class is a single-number classification defined in ASTM E989, derived from a laboratory measurement made to ASTM E492. In the test, a standard tapping machine — five hammers, each dropped from a fixed height at a fixed rate — strikes the floor of a test assembly while microphones measure the sound pressure level in the room below across sixteen one-third-octave bands from 100 Hz to 3,150 Hz.

Those sixteen values are compared with a standard contour. The contour is shifted until the sum of unfavourable deviations reaches as much as 32 dB without any single band deviating by more than 8 dB. The IIC value is the contour value at 500 Hz in its final position. Higher means better.

Every element of that description belongs to an assembly: a floor, a structure, and a receiving room. Nothing in it describes a sheet of foam.

So what does an underlayment have?

An improvement figure — and there are two of them, depending on which system your project is specified under.

ΔIIC is the ASTM quantity. ASTM E2179 is the test method: it covers the laboratory measurement of the effectiveness of floor coverings in reducing impact noise from a standard tapping machine through concrete floors. ASTM E989 then classifies the result, and it is explicit about naming — the rating is called impact insulation class (IIC) when it comes from Test Method E492, change in impact insulation class (ΔIIC) when it comes from Test Method E2179, and AIIC, ISR or NISR when it comes from the field method E1007. The name tells you which measurement was made, which is exactly why quoting “IIC” for a sheet of foam is a category error.

ΔL_w is the international counterpart. ISO 717-2:2020 defines single-number quantities for the impact sound reduction of floor coverings and floating floors, calculated from measurements made to ISO 10140-3. Both systems measure against a defined reference floor; ISO 717-2 devotes a clause to specifying that reference floor, which is a heavy concrete slab in the standard’s procedure for bare heavy floors.

One further limitation belongs in the specification. ASTM E989 states that its rating strictly applies only to excitation by the standard tapping machine, and that it does not deal with sound below 100 Hz — the range that lightweight joist floors typically generate when they are walked on. A ΔIIC measured on concrete is therefore not a description of how the same pad will behave under footfall on a timber floor.

The practical consequence is narrow, and worth stating precisely:

That is all the figure says. It is a difference measured between two conditions of the same assembly, and it should not be lifted off one floor construction and added to another in order to predict a project result. ASTM E2179 makes the same point in its own significance-and-use section: the same floor covering in combination with different basic floor assemblies will not always give the same IIC ratings. The standard goes further, warning that a difference spectrum measured by this method shall not be used to estimate impact sound levels for floors comprising only one or two lightweight layers such as oriented strand board or plywood, because such estimates would be very inaccurate.

That is why two suppliers can quote honest, correctly measured improvement figures for similar pads and still arrive at different system ratings. It is also why a ΔIIC measured on a heavy concrete slab is not a promise about a joist floor in a timber-frame building.

It also explains something you will notice if you read widely: published supplier comparison tables place the same generic material — EVA — anywhere from about IIC 55 to the low 70s. A spread of more than fifteen points for one material is not a measurement disagreement. It is what happens when numbers that belong to different assemblies are printed side by side as though they belonged to the foam.

4. Four acoustic numbers that get mixed up

Sound behaves in several different ways in a building, and each has its own metric. Mixing them up is the most common error in flooring specifications.

Quantity	What it describes	Direction	Where it is defined
NRC (Noise Reduction Coefficient)	How much sound a surface absorbs rather than reflects back into the same room	Higher absorbs more	ASTM C423 (North America); αw under ISO 11654 (international)
STC (Sound Transmission Class)	Airborne sound — speech, television — blocked by a partition	Higher blocks more	ASTM E90 (test) and E413 (classification); Rw under ISO 717-1
IIC (Impact Insulation Class)	Impact sound — footsteps, dropped objects — transmitted through a floor-ceiling assembly	Higher blocks more	ASTM E492 (test) and E989 (classification)
ΔIIC (change in Impact Insulation Class)	The improvement a floor covering makes to a defined reference concrete floor — the figure a pad can actually be given	Higher improves more	ASTM E2179 (test) and E989 (classification)
Ln,w (weighted normalised impact sound pressure level)	The international counterpart to IIC for impact sound	Lower is better	ISO 10140-3 (test) and ISO 717-2 (rating)

Two things follow. First, an acoustic wall panel’s NRC tells you nothing about a floor’s impact performance, and a floor’s IIC tells you nothing about how much sound a wall surface absorbs — they are different physical questions. Second, IIC and STC are also separate: a floor assembly can satisfy an airborne requirement while failing the impact requirement, because the vibration path is structural rather than airborne.

For an SPC flooring project in a multi-storey building, the impact figure is usually the one that governs occupant satisfaction — footfall is the noise residents actually complain about — but the airborne figure may also appear in the same code clause. Check both against the current local requirement rather than assuming one covers the other.

5. ASTM and ISO: two systems that do not convert

North America rates impact sound with IIC. Most of the rest of the world uses L_n,w. They look interchangeable because they are computed from the same kind of measurement, and they are not.

The relationship is often quoted as:

That expression comes with a condition attached, and the condition is stated inside the ASTM documents themselves rather than in any supplier’s brochure. ASTM E2179 records that it closely follows the procedure described in ISO 10140-1 (Annex H), except that the single-number rating used is the IIC of Classification E989 and that the description of the standard concrete floor also differs. It then adds, in a note: because Classification E989 requires that no deviation above the reference contour may exceed 8 dB, there is no simple relationship between ISO 10140-1 (Annex H) test ratings and those generated by this method.

So the arithmetic shortcut holds while that 8 dB requirement is not invoked, and diverges when it is. The two scales also run in opposite directions — a better floor has a higher IIC but a lower L_n,w — which is where a lot of copied comparison tables quietly go wrong.

What this means for a buyer: name the metric your project is specified under, then ask for the measurement that produces it.

• For IIC — an ASTM E492 measurement, classified under ASTM E989.
• For ΔIIC — an ASTM E2179 measurement, classified under ASTM E989.
• For L_n,w or ΔL_w — an ISO 10140-3 measurement, rated under ISO 717-2.
• For field verification — ASTM E1007 (AIIC, ISR or NISR) or ISO 16283-2.

Do not accept a figure converted from the other system, and do not convert one yourself to check a code threshold.

6. Laboratory IIC and field IIC are not the same number

Laboratory measurements are made in facilities designed to suppress flanking transmission — sound travelling around the test specimen through walls, structure and services. ISO 10140-3:2021 states this in its own scope: measurements are performed in laboratory facilities in which sound transmission via flanking paths is suppressed, and the results are not applicable directly to the field situation without accounting for other factors such as flanking transmission, boundary conditions and loss factor. That is precisely why separate field methods exist — ISO 16283-2 internationally, ASTM E1007 in North America.

Field results may therefore be lower than laboratory results for the same assembly, or they may differ in other ways, because a real building introduces flanking paths, boundary conditions and workmanship that the laboratory does not reproduce. How large that difference is depends on the building, so it has to be measured on site rather than assumed from a rule of thumb. This is why separate field methods exist — ISO 16283-2 internationally, and ASTM E1007 in North America, whose ratings are named AIIC, ISR and NISR precisely so that they are not confused with laboratory IIC. A specification built on laboratory figures alone can pass on paper and still generate complaints.

Two site details do more damage than most material choices:

• Bridged perimeter gaps. A floating floor relies on being free to deflect. If the expansion gap is filled with rigid sealant, bridged by grout, or pinched by tightly fitted skirting, the assembly mechanically couples to the wall and the resilient layer can no longer work as intended.
• Long unrestrained runs. Large rooms need transition profiles so the floor can move in manageable sections. Where the product instructions specify one and it is omitted, both dimensional stability and acoustic behaviour suffer.

Neither is a reason to change the pad. Both are reasons the specified pad never got to perform.

7. Where IXPE and EVA genuinely differ

With the rating question cleared up, the material comparison becomes more honest — and less one-sided than the published tables suggest.

Aspect	IXPE	EVA	How much it depends on the specific product
What the name tells you	A manufacturing route: irradiation cross-linking	A polymer: ethylene-vinyl acetate, made by various foaming and cross-linking routes	Low — this part is fixed by definition
Cell formation	Cross-linked, then expanded; cell structure is fine and uniform under good process control	Depends on formulation and process; cell uniformity varies more widely between producers	High — foaming control varies by line
Recovery under sustained load	At comparable density and thickness, a cross-linked network generally resists permanent collapse better	Uncrosslinked grades are more prone to retaining compression under sustained load; cross-linked grades behave differently	High — density and thickness dominate
Behaviour at elevated temperature	Cross-linked structure generally holds shape better as temperature rises	Softening behaviour depends on vinyl acetate content, density and whether the foam is cross-linked	Medium — formulation dependent
Odour and emissions	Determined by the blowing system, formulation and process control, not by the name	Determined by the same variables; a well-made product can be low-odour and a poor one can carry residual odour	High — verify with product-specific emissions testing
Moisture behaviour	Closed cell; generally low water absorption, not zero — cut edges and surface defects matter	Closed cell; generally low water absorption, with more variation in cell uniformity	Medium — ask for the absorption figure for the product
Cost	Higher per square metre	Lower per square metre	Low

The honest summary: IXPE is more often specified on higher-specification SPC products, and there are defensible engineering reasons for that — a cross-linked network recovers better from sustained load, and an irradiated line usually runs under tighter process control. But the material name on its own does not establish acoustic, thermal, emissions or long-term load performance. Those are properties of a specific pad at a specific density and thickness, in a specific plank construction, and they are established by test reports rather than by the label. Use the table to decide what to ask about, and let the project’s own documents give the answer.

8. The compression-set figure does not mean what the data sheet implies

Compression set is frequently quoted as the laboratory proof that IXPE outlasts EVA. It is worth knowing what the standard actually covers.

ISO 1856 specifies methods for determining compression set of flexible cellular polymeric materials: a specimen is compressed to a defined proportion of its thickness, held for a defined time at a defined temperature, released, and the permanent loss of thickness is reported as a percentage. Three methods are given — Method A at 70 °C for 22 hours, Method B at the conditioning temperature for 72 hours, and Method C under conditions agreed between the parties.

Two scope limits matter for a flooring underlayment:

• The current edition states its scope is limited to latex and polyurethane foams. Polyolefin foams — which is what IXPE and EVA are — are outside the stated scope as it is written.
• The method is built around a 25 mm specimen. For thinner material the standard requires stacking pieces, separated by glass slides, to reach a total pre-compression thickness of at least 25 mm. A 1.0 mm or 1.5 mm pad has to be stacked many layers deep to be tested at all.

Neither point makes the test useless. Method C exists precisely so parties can agree conditions, and a compression-set number produced under clearly stated conditions still tells you something about how two samples behave relative to each other under those conditions. But a compression-set percentage quoted without the compression ratio, temperature, duration and recovery time is not a comparable figure, and a percentage measured on a 25-layer stack is not a direct prediction of how a single 1 mm layer will behave under a table leg for a decade.

If long-term thickness retention is genuinely important to the project, ask for the test conditions along with the number, or ask instead for the finished plank’s own performance data — dimensional behaviour and locking-system performance under load are closer to what the floor actually has to survive.

9. When EVA is the right specification

Almost every published comparison recommends IXPE, which says more about the price of IXPE than about EVA. There are projects where EVA is the correct call:

• Short-hold or tenanted assets where the floor will be replaced on a known cycle and long-term resilience is not being paid for.
• Ground-level or single-storey applications with no impact-sound requirement at all — the acoustic argument simply does not arise.
• Dry interior areas with no underfloor heating, no heavy point loads and no emissions programme.
• Budget-led programmes where the choice is between EVA-backed SPC and no acoustic layer at all. In that comparison, EVA wins comfortably.

It stops being a specification you can defend with documents in four situations: where an IIC, ΔIIC or L_n,w requirement has to be demonstrated with a report; where underfloor heating is installed and a written compatibility declaration is required; where heavy furniture or equipment will sit on the floor for years and long-term thickness retention matters; and where a low-emission certificate is required by the client or the scheme. In those four the decision moves off the material name and onto the test file — so specify whichever pad the supplier can document, and ask for that file before the pad is chosen.

10. Attached pad, loose-laid pad, and the double-layer mistake

Underlayment reaches the site in one of two forms, and the difference affects cost, programme and risk.

Configuration	How it arrives	What it buys you	What it costs you
Factory-attached pad	Laminated to the back of the plank in production, usually with a hot-melt adhesive	One less site operation, consistent adhesion, no seam taping, no risk of the pad shifting or overlapping	Higher unit price; the pad choice is fixed at the time the plank is ordered
Loose-laid underlayment	Rolled out on the subfloor before the floor is installed	Flexibility to upgrade the acoustic layer without changing the plank, and the option to reduce freight by shipping the two separately	Additional labour; performance depends on the installer; seams must be sealed as specified

The mistake to avoid: doubling up. If a plank already carries a factory-attached pad, adding a loose layer underneath is generally not recommended. Two resilient layers increase the total deflection under load, which can overstress the click joint, produce a soft or springy feel underfoot, and — because it departs from the tested and documented configuration — put the manufacturer’s warranty at risk.

So the first question to ask before specifying any underlayment is not “IXPE or EVA” but “does this plank already have a pad?” If it does, and the assembly does not reach the required rating, the correct move is to change the plank specification or the structure — not to add a second layer.

11. Underfloor heating

Where underfloor heating is installed, the underlayment matters less for sound than for heat and stability.

Any layer between the heat source and the room adds thermal resistance. At the thicknesses used for SPC underlayment the added resistance is small, and what the heating designer actually needs is the total thermal resistance of the whole floor build-up — plank, pad, and any other layer — checked against the system’s limit.

What is less obvious is that the build-up is not a fixed object. A resilient layer that compresses under load becomes thinner — and a thinner layer, all else being equal, has lower thermal resistance, not higher. The real risks on a heated floor are therefore not that “the floor stops heating”. They are: loss of pad thickness under the locking system, dimensional movement of the plank through repeated heating cycles, any change in the bond between pad and plank, and the fact that the total resistance of the build-up as constructed may differ from the figure the heating designer assumed. None of these can be read off a material name, and none should be predicted without product-specific data.

Where heated floors are specified with an IXPE pad, the argument is a durability argument rather than an acoustic one: a cross-linked foam is less likely to soften and lose thickness across repeated heating cycles. That is a reasonable engineering preference, not a test result — so it still has to be confirmed for the product being supplied. Ask the supplier for three things, in writing: an underfloor-heating compatibility statement for the specific construction, the total thermal resistance of the build-up, and the maximum permitted floor surface temperature. Do not rely on a general claim that the product is “suitable for underfloor heating”.

12. What to request from your supplier

Every item below is a document request, not a specification. The point is that a pad decision can be checked before it is purchased.

1. The material and the manufacturing route — irradiation cross-linked, chemically cross-linked, or chemically blown. The trade name alone is not sufficient.
2. Thickness and measured density of the pad as supplied.
3. Whether the pad is factory-attached or loose-laid, and for attached pads, the adhesive method.
4. The commercial parameters that affect the shipment as well as the floor — the thicknesses and densities available, whether the pad can be factory-attached to the plank being purchased, plank thickness and wear layer, locking-system compatibility, packing format and container quantity, and whether printing or colour on the pad can be customised.
5. The improvement figure — ΔIIC or ΔL_w — together with the reference floor it was measured on.
6. The test method, the classification and their editions — ASTM E492 with ASTM E989 for IIC, ASTM E2179 with ASTM E989 for ΔIIC, or ISO 10140-3 with ISO 717-2 for L_n,w and ΔL_w — stated explicitly rather than implied.
7. Laboratory name, report number and test date. An improvement figure with no report behind it is a claim, not a measurement.
8. The emission certificate if the project has one — FloorScore, AgBB, Greenguard Gold or the scheme named in the specification. Our comparison of Greenguard Gold and FloorScore covers what each one tests.
9. An underfloor-heating declaration where heating is present, including total thermal resistance and maximum surface temperature.
10. Confirmation on double layering — whether the plank may be installed over an additional loose layer, in writing.
11. A sample from a production batch, so the pad on the sample matches what will be shipped.

There is a pattern here worth noticing: almost none of these questions is “which foam is better”. They are questions about what was measured, on what, by whom, and under which version of which standard. Those are the questions that survive a value-engineering review.

13. Homax manufacturing note

Frequently asked questions

Sources and technical references

• ASTM E492, Standard Test Method for Laboratory Measurement of Impact Sound Transmission Through Floor-Ceiling Assemblies — the laboratory measurement underlying IIC; supports the tapping-machine procedure and the 100–3,150 Hz one-third-octave range.
• ASTM E989-21, Standard Classification for Determination of Single-Number Metrics for Impact Noise — classifies the measurement and assigns its name. §1.4 states that the single-number rating is called impact insulation class (IIC) when it comes from Test Method E492, change in impact insulation class (ΔIIC) when it comes from Test Method E2179, and AIIC, ISR or NISR when it comes from Test Method E1007. Its contour-fitting procedure supports the 32 dB total and 8 dB single-band deviation limits and the 500 Hz reading point. §4.2 states that the rating strictly applies only to excitation by the standard tapping machine and does not deal with low-frequency sound below 100 Hz.
• ASTM E2179-21, Standard Test Method for Laboratory Measurement of the Effectiveness of Floor Coverings in Reducing Impact Sound Transmission Through Concrete Floors — the test method that produces ΔIIC. §5.1: the same floor covering in combination with different basic floor assemblies will not always give the same IIC ratings. §5.2: the ΔIIC rating is used to compare the effectiveness of different floor coverings on concrete floors. §5.5 (warning): difference spectra measured using this method shall not be used to estimate impact sound pressure levels for floors comprising only one or two lightweight floor layers such as oriented strand board or plywood. §5.6 and Note 2: the method closely follows ISO 10140-1 (Annex H) except that the single-number rating used is the IIC of Classification E989 and the description of the standard concrete floor differs — and because Classification E989 requires that no deviation above the reference contour may exceed 8 dB, there is no simple relationship between ISO 10140-1 (Annex H) test ratings and those generated by this method.
• ASTM E90 and ASTM E413 (STC), and ASTM E1007 (field impact measurement) — support the airborne-versus-impact distinction and the laboratory-versus-field distinction.
• ISO 10140-3:2021, Acoustics — Laboratory measurement of sound insulation of building elements — Part 3: Measurement of impact sound insulation (Edition 2, published April 2021) — the international laboratory counterpart to ASTM E492. Its scope statement that measurements are made with flanking transmission suppressed and are not directly applicable to the field situation supports the laboratory-versus-field section.
• ISO 717-2:2020, Acoustics — Rating of sound insulation in buildings and of building elements — Part 2: Impact sound insulation (Edition 4, published December 2020; confirmed current) — defines L_n,w and, in clause 5, the weighted reduction in impact sound pressure level by floor coverings on bare heavy floors, including the specified reference floor; supports the ΔL_w argument and the ASTM/ISO comparison.
• ISO 16283-2:2020, Acoustics — Field measurement of sound insulation in buildings and of building elements — Part 2: Impact sound insulation (Edition 3, published July 2020) — the international field method referred to above.
• ISO 1856:2018, Flexible cellular polymeric materials — Determination of compression set (Edition 4, published June 2018) — supports the compression-set section. Its scope states that the document applies to latex and polyurethane foams of thickness greater than 2 mm, which is the basis for the caution about applying it to polyolefin underlayment; the specimen requirements and Methods A, B and C are as described there.
• ASTM C423 (NRC) and ISO 11654 (α_w), ISO 717-1 (R_w) — the absorption and airborne ratings in the comparison table. They are separate quantities from impact ratings, are specified against separate documents, and are not further cited here.

The general comparison of IXPE and EVA reflects the two manufacturing routes and is not drawn from a specific test report. Statements about the standards are cited to the documents listed above. Where a claim could not be attributed to a published standard or to a test report, it is stated as a general tendency rather than as a measured value. Homax Decor does not publish acoustic improvement figures for its own products where the figure cannot be attributed to a test report for the specific tested configuration.