Mineral Wool vs Sheep Wool Insulation

Last updated: August 2026

Both materials are lower-risk than fiberglass or cellulose for indoor air quality. That is not in dispute, and anyone telling you otherwise is selling something. The question is narrower: when you are building new, and you control the assembly, which one do you specify?

Most of what is published on this comparison is written by one of the two manufacturers. Sheep wool comparisons are frequently authored by sheep wool companies. Mineral wool technical data comes from mineral wool companies. Both are useful. Neither is neutral. This page states its criterion up front, applies the same test to both materials, names what we do not know, and labels every vendor-authored source.

The test: terminal, specifiable, non-substrate

The Good Air Standard evaluates insulation against three questions. They are worth stating plainly, because most insulation marketing answers a different question — how low-emitting is this material on the day it is installed — when the question that matters is how the material behaves for the next fifty years inside a wall you will not open again.

1. Is it terminal?

Does the material reach a defined end state where emissions go to zero and stay there? A material that off-gasses hard for three weeks and then stops is preferable to a material that emits very little but never stops, because the first one can be tested, verified, and closed out. You can demonstrate that a terminal material is finished. You cannot demonstrate that an ongoing one ever will be.

2. Is it specifiable?

Is the composition disclosed, bounded, and consistent batch to batch? This matters more than it sounds. A standard is a promise that the same specification produces the same result every time. Materials with natural variation — different flocks, different scouring chemistry, different residual lanolin, seasonal variation in fiber — cannot deliver that promise, no matter how benign any individual batch is. Unrepeatable is a defect in a certification context even when it is not a health defect.

3. Is it a substrate?

Under any moisture condition the building might realistically see, can this material become food for something? This is the question that cannot be engineered away after the wall is closed. Every other property of an insulation can be compensated for elsewhere in the assembly. This one cannot.

Why this page names brands

Elsewhere on this site, the Good Air Standard evaluates materials by category rather than brand, because products change faster than standards can track. See Safer Alternatives for the category-level view.

This page is a deliberate exception, for a reason that is consistent with the rule rather than in violation of it: within the mineral wool category, the binder chemistry differs by product line. One ROCKWOOL product has a phenol-formaldehyde binder; another has a starch-based one. Answering "is mineral wool lower-risk" at the category level would be answering a question the reader did not ask. When the variable that matters lives below the category level, the honest answer has to go there too.

The specific products named here were accurate at the last update date. Verify current composition against the manufacturer's published data before you specify. The reasoning is durable; the SKUs are not.

Side by side

Is mineral wool actually inert?

Closer than most people assume, but you have to specify the small fraction that is not rock.

The fiber itself is 96 to 99 percent of the mass. It is inorganic, it does not off-gas, it does not degrade, and it does not feed anything. ROCKWOOL's published product data for Comfortbatt shows it is non-combustible under ASTM E136, with a flame spread index of 0 and a smoke developed index of 0 under ASTM E84, and that it passes ASTM C1338 for fungi resistance. That last one is worth noting specifically: it is a standardized fungal resistance test, not a marketing claim.

The remaining one to four percent is where product lines differ, and it contains three things.

The binder

Published ingredient data for Comfortbatt, Safe'n'Sound, and Comfortboard lists under 3% phenol-formaldehyde, under 1% starch, and under 0.2% mineral oil. ROCKWOOL's own current Safe Use Instruction Sheet confirms that composition varies by product family — some families list a phenol-formaldehyde binder, others list hydrolysed starch syrups instead.

The critical point about the binder is that it is thermoset. Once cured, it is a cross-linked solid, not a reservoir. Residual free formaldehyde off-gasses quickly and largely completely once the packaging is opened. It finishes. Off-gas the material in a ventilated space before installation and the emission is over. That is the terminal property the first criterion is asking about.

The mineral oil

This is the ingredient almost nobody discusses, and it is the one to be honest about. Mineral oil is used as a de-dusting and water-repellent agent at under one percent. It is a light hydrocarbon. It is the faint smell on a freshly opened batt, and it is the slowest of the three additives to leave. For most people it is negligible. For chemically sensitive occupants it is real, and a minority will react to it. Anyone selling you mineral wool as odorless has not opened a package.

The slag

Pre-consumer recycled slag from steel and copper production typically makes up 16 to 40 percent of the product. Steel slag can contain trace amounts of metals including titanium, aluminum, calcium, and chromium. These are bound in a glassy mineral matrix and are not in a bioavailable form in an installed assembly, but the material is a recycled industrial byproduct and it is more honest to say so than to describe stone wool as simply melted rock.

Is there a formaldehyde-free mineral wool, and can you actually buy it?

Yes to the first. Qualified yes to the second, and the qualification matters.

ROCKWOOL introduced AFB evo with a formaldehyde-free binder in July 2017. It is UL validated as formaldehyde free and holds GREENGUARD Gold certification. The constraint is that AFB evo is a light-density acoustic and fire batt intended for interior partitions — it is not the exterior wall thermal product, and it is harder to source than the standard batts.

Owens Corning's Thermafiber FF line uses a bio-based formaldehyde-free binder, is UL Environment validated as formaldehyde free, holds GREENGUARD Gold, and is listed in the Declare database as Living Building Challenge compliant. It typically has to be special-ordered.

Knauf's ECOSE binder is bio-based and carries Eurofins Indoor Air Comfort Gold and Declare Red List Free status.

So the practical tiering is:

Tier 1 — no organic content at all

Cellular glass board, perlite loose fill, or unbonded mineral wool loose fill. No binder, nothing to off-gas, nothing to test. Limited application range and significant cost. This is the tier for the most severely sensitive occupants and for spaces where nothing else will do.

Tier 2 — formaldehyde-free binder

ROCKWOOL AFB evo, Owens Corning Thermafiber FF, Knauf ECOSE. Accept sourcing friction and lead time. Appropriate when the occupant is known to be formaldehyde-reactive or when the specification has to survive third-party scrutiny.

Tier 3 — standard mineral wool, off-gassed before install

ROCKWOOL Comfortbatt and Comfortboard, GREENGUARD Gold certified. Off-gas in a ventilated space, then install. This is appropriate for the large majority of projects, including most chemically sensitive occupants, and it is the default under the Good Air Standard.

The tier that is right for a given project is a function of the occupant, not the budget. Tier 3 is not a compromise. It is the correct answer most of the time.

Does sheep wool insulation really clean the air?

The chemistry is real. The application is where it falls apart.

Wool is keratin, a protein built from amino acids. Several of those amino acids — lysine and arginine in particular, along with the amide groups of glutamine and asparagine — react with formaldehyde to form a methylene bridge cross-link. That bond is covalent and irreversible, which means the captured formaldehyde does not come back out. This part of the wool industry's claim is sound and is supported by peer-reviewed work.

There are two problems with translating that into a reason to insulate a house with wool.

Problem one: the wall is supposed to be sealed

If your assembly has a continuous air barrier — and it must, if you are building correctly — then the insulation is not in the indoor air path. Air that never reaches the wool cannot be cleaned by the wool.

Read plainly, wool's air-purification benefit is an argument for a leaky wall. The leakier the assembly, the more indoor air passes through the insulation, and the more air cleaning the wool can theoretically do. In a properly air-sealed house, the benefit trends toward zero. You design the feature out by building right.

This is the single most important point on this page and it applies regardless of what you conclude about the chemistry. The diagram below shows the position that matters: insulation sits inboard of the air barrier, on the cavity side, not in the room's airstream.

Problem two: the evidence is about carpet

The widely cited figure is that wool may purify indoor air for up to 30 years. That number traces to New Zealand research on wool carpets and is promoted by the International Wool Textile Organisation, a trade body.

Carpet sits in the room. It is in the airstream, walked on, in direct and continuous contact with the air people breathe. Insulation sits behind an air barrier inside a sealed cavity. These are not the same situation, and moving a performance figure from one to the other is not supported.

Does wool stop absorbing once it fills up, like activated carbon?

For the air-cleaning function, yes — and the carbon analogy is exactly right. For moisture, no. These are two different mechanisms and conflating them is the source of most of the confusion.

Moisture buffering does not deplete

That is physical sorption. Wool takes up water vapor and releases it, and it cycles indefinitely. It does not wear out. The carbon analogy does not apply.

Formaldehyde binding does deplete

That is chemisorption. It consumes a fixed number of reactive amino acid sites by forming permanent covalent bonds. When the sites are used, they are used. There is no regeneration. This is how activated carbon behaves, and the instinct that it must saturate is correct.

Here is the part worth knowing about the "unlimited" framing: the peer-reviewed literature states that no reports have identified a saturation point for sheep wool products. Read carefully, that says the saturation point is unmeasured. It does not say it is unlimited. Marketing routinely converts the first into the second. It is a finite number of chemical bonds, and a finite number that nobody has counted is still finite.

There is also a partially reversible fraction. By the sheep wool industry's own account, up to about 30 percent of what wool captures is held by physisorption rather than chemical bonding, and that portion can be re-emitted given sufficient change in relative humidity, temperature, or concentration gradient. So even on the favorable reading, roughly a third of the capture is not permanent.

Can sheep wool insulation grow mold?

Wool is a protein. At elevated moisture content, in a dark enclosed cavity, it can support fungal growth. Mineral wool does not support mold growth, because there is nothing in it to metabolize — which is what the ASTM C1338 fungi resistance result means.

This is the point that vendor comparisons of these two materials tend to omit, and on a site about indoor air quality it is the one that matters most. For the construction-phase side of this problem, see Mold During Construction.

Note the uncomfortable symmetry: wool's forgiveness and wool's risk are the same property. The material that absorbs 33 percent of its weight in moisture without losing R-value is, at that moisture content, a protein at elevated water activity inside a sealed cavity. You cannot have the buffering without the substrate. They are the same physical fact described two ways.

Wool insulation is normally treated with borate to resist pests and add fire resistance, which may also reduce fungal growth. That treatment is protective and it is real. It is also a chemical additive, which sits awkwardly alongside marketing that positions wool as the additive-free option — and its presence and concentration are not consistently disclosed across products. See the open questions section below.

If the house is built right, do you need moisture-buffering insulation?

Largely no. And specifying insulation for its ability to absorb moisture is designing around a defect rather than eliminating the defect.

The building science order of operations is bulk water first, then air, then vapor, then thermal. Rain screen, flashing, continuous air barrier, correct vapor control layer for the climate zone, enough exterior insulation to keep the sheathing above dew point, and balanced ventilation with dehumidification. Get those right and the cavity does not see problem moisture. Insulation hygric buffering is a fourth-order effect. If you are relying on it, something upstream has already failed.

There is a second reason it does not deliver what people think, and it is the same reason as the air-cleaning argument: buffering behind an air barrier does not change the humidity you actually breathe. Indoor relative humidity is governed by ventilation rate, dehumidification, and interior finishes — drywall, plaster, exposed wood, textiles. Insulation in a sealed cavity is on the wrong side of the air barrier to buffer room air.

The counterargument, stated fairly

Buffering is insurance against the failure you did not design for. The supply line that lets go in year twelve. The window flashing detail that ages out. Construction moisture that got closed in. Wool gives the assembly forgiveness for the mistake nobody planned. That is a real argument and it deserves a real answer.

The answer

If the cavity does get wet, you want a material that drains, dries, and does not feed anything — not one that holds a third of its weight in water and is made of protein. The forgiveness you are buying comes bundled with the substrate you do not want.

And if the year-twelve leak is the concern, the correct responses are leak detection, an assembly that can dry in at least one direction, and access for inspection. Not an organic sponge. Design for the failure directly instead of hoping the insulation absorbs it.

Which is better for MCS and chemical sensitivity?

Mineral wool, off-gassed thoroughly before installation, in a well-sealed assembly. And the reason is not that wool is dangerous. It is that mineral wool is predictable over decades and wool is not.

This is worth stating carefully, because the usual framing of this question is wrong. People ask which material has the lower emission on installation day. That is the least important question. A sensitized occupant is going to live with this wall for thirty years and will never open it. What matters is whether the material's behavior is knowable for that entire period.

Mineral wool's emissions have an end date. You can air the material out, test it, install it, and know the chemistry is finished. It will be the same material in year thirty as in year one, because there is no mechanism by which stone changes. And because it is manufactured to a tight process specification, the batch you buy next year behaves like the batch you bought this year.

Wool offers none of those guarantees. It is a reactive protein rather than an inert solid, it varies by flock and processing, its treatment additives are inconsistently disclosed across products, and at elevated moisture it can become a substrate. Every one of those is a source of variance in a material you will never inspect again.

For a sensitized occupant, the reliability is the health argument. A material whose future you can predict is lower-risk than a material with a marginally better starting number and an open-ended future.

There is a narrower point that also holds: a minority of people react to residual phenol-formaldehyde or to the de-dusting mineral oil in stone wool. That is real, and individual variation in chemical sensitivity is not predictable from a specification sheet. Where it occurs, move to Tier 2 formaldehyde-free lines, then to Tier 1 binder-free products such as cellular glass or perlite, before considering wool.

Whatever the material, the occupant should test a sample before the whole house is committed. No specification sheet substitutes for that. The category-level substitutions are listed on Safer Alternatives, and the materials this standard rules out are on Materials to Avoid.

When sheep wool is the right choice

Two situations. They are narrower than the marketing suggests, and stating them honestly is the point of this section.

Historic and solid-wall retrofit

A pre-1900 timber frame or solid masonry wall where you cannot install an air barrier or a vapor control layer without causing harm, and where the assembly has to stay vapor-open and dry in both directions. Building conservation practice genuinely favors hygroscopic natural insulation in these walls, because impermeable and non-buffering materials have a documented failure history in them. This is wool's strongest legitimate case and it is the one nobody advertises, because the market for it is small.

Documented reactivity to mineral wool components

Where an occupant has tested and reacted to residual formaldehyde or de-dusting oil, and Tier 2 and Tier 1 mineral options are unavailable or also fail testing. Wool is a genuine option here, not a consolation prize — provided the substrate trade-off is accepted knowingly and the assembly is detailed for it.

Two things wool is credited with that do not survive scrutiny

Installer exposure. Wool installs with bare hands and no mask, and mineral wool fiber is a mechanical irritant requiring gloves, sleeves, and eye protection. This is a real difference, but it trades one week of installation discomfort against decades of assembly performance, and it is addressed by PPE. One clarification is owed on the health framing: modern stone wool fibers are biosoluble, and in 2002 the International Agency for Research on Cancer moved rock, slag, and glass insulation wools from Group 2B to Group 3, not classifiable as to carcinogenicity in humans. The installation irritation is real. The cancer implication that vendor comparisons sometimes gesture at is not supported.

Acoustics. Wool absorbs mid-to-high frequencies well and mineral wool's density favors low frequency. The difference is audible in a music room. It is not a health argument and should not be weighed as one.

What about van and RV builds?

This case is worth its own answer, because sheep wool is the near-consensus recommendation in van build content and we think that consensus is wrong.

The argument for wool in a van is moisture. Condensation inside a van is not a risk, it is a certainty — occupants breathing and cooking in a very small volume, against a single-skin metal wall with almost no thermal break. Wool absorbs that moisture and releases it later, so the reasoning goes, and you can skip the vapor barrier.

Here is the problem. In a van, the condensation forms on the inside face of the bare metal skin. A material that absorbs that water and holds it in fiber contact with unpainted steel is not solving a rust problem. It is a mechanism for one. Independent comparative testing by van builders has found exactly this tension: wool showed less visible standing condensation than competing materials, precisely because the water had gone into the wool. Whether that counts as moisture management or as insulation holding water against your van body is the entire question, and the answer is not obvious in wool's favor.

The second point is stronger. A van cavity is permanently inaccessible. You would have to demolish cabinetry to inspect it. Inaccessibility is an argument for a material that does not become a substrate, not for one that forgives moisture. The logic that favors mineral wool in a house applies more forcefully in a van, not less.

The third point is the one van build guides rarely raise, and it may be the most consequential. A vehicle subjects its walls to constant vibration for the life of the build. Insulation in a vertical cavity under sustained vibration can settle — and settlement in an enclosed wall is not a gradual loss of comfort, it is the creation of a void at the top of the cavity. That void is precisely where warm moist interior air collects and meets cold metal. So settlement does not merely cost R-value. It manufactures the condensation site that the entire insulation strategy was meant to prevent, in a location you cannot see and cannot reach.

Mineral wool is better positioned here. Comfortbatt is a semi-rigid batt with a density above 2 lb/ft³ that friction-fits and holds its position, and rigid board such as Comfortboard removes the question altogether. Sheep wool batts are lower density and rely on needle-punch cohesion or a binder fiber to hold form.

Be aware also that a large share of van insulation content is affiliate or sponsored, including claims that mineral wool traps condensation and breeds mold — which is incorrect on both counts, since stone wool is vapor-open and contains nothing that fungi can metabolize.

Our position: for a van you are building from scratch, you control the envelope. Detail it correctly, ventilate it properly, and use a rigid or semi-rigid material that will not settle and does not feed anything.

Does insulation settle or slump over time?

This question matters well beyond vehicles, and it is a fair test of the "terminal and specifiable" standard applied to physical form rather than chemistry.

An insulation batt has one job it must keep doing for decades: stay where you put it, in full contact with the cavity, with no gaps at the top or the edges. A material that slumps even slightly creates a void, and a void in an enclosed assembly is a cold spot, a condensation site, and a thermal bypass at once. Because the cavity is sealed, the defect is invisible. Nobody finds it until something else fails and the wall gets opened.

Mineral wool is engineered against this. Comfortbatt is semi-rigid at a density above 2 lb/ft³ and is designed to friction-fit — it is held by its own stiffness against the framing rather than by staples, adhesive, or a facing. Rigid mineral wool board is dimensionally stable by definition. Neither depends on a binder remaining intact to keep its shape.

Sheep wool batts are lower in density and hold their form through needle-punch entanglement or a small fraction of binder fiber. That is adequate in a stable wall. It is a more open question in an assembly subject to sustained vibration, repeated thermal cycling, or moisture cycling that swells and relaxes the fiber.

What we do not know

Stating this openly rather than papering over it.

Wool's saturation point is unmeasured, not unlimited

No published study has established the formaldehyde saturation capacity of sheep wool insulation in a building assembly. The chemistry indicates it is finite. Nobody has published the number. Any claim of indefinite air cleaning is an extrapolation.

Sheep wool product composition is inconsistently disclosed

Public sources conflict on whether specific sheep wool insulation products contain a polyester bicomponent binder, and on whether and how much borate treatment is applied. Some manufacturer materials describe a product as 100 percent wool bonded by needle punch with no chemicals; retailer listings for the same brand describe roughly 90 percent wool with a recycled polyester binder; independent trade sources describe borate treatment. These claims cannot all be simultaneously true of the same SKU.

We have contacted the manufacturer for clarification and will publish the reply here verbatim. Until then, treat composition claims for any sheep wool product as requiring direct verification with the manufacturer for the specific SKU and production run you are buying.

Long-term settlement has not been comparatively tested

We are not aware of published long-term settlement or compaction testing comparing sheep wool and mineral wool batts in vertical cavities, in buildings or in vehicles. The position taken above follows from density and material structure rather than from measured results, and we have labeled it as such on the page. If comparative testing exists or is published, we will update this section.

Installed cost varies too widely to state a single figure

Sheep wool carries a meaningful premium over mineral wool, but the multiple depends heavily on region, product form, and freight. Get local quotes rather than trusting a published ratio.

Formaldehyde-free mineral wool availability changes

Which formaldehyde-free lines are stocked in residential thermal applications shifts with manufacturer rollout schedules. Confirm current availability with your distributor rather than assuming the tiering above is still accurate.

The conclusion: what to specify

Stated plainly, because a page that argues for four thousand words and then hedges at the end has wasted the reader's time.

The specification

Cavity insulation: semi-rigid mineral wool batt, friction-fit, density above 2 lb/ft³. Continuous exterior insulation: rigid mineral wool board. Off-gas all material in a ventilated space before installation, and do not skip this step — it is what takes residual emissions from low to finished.

Where an occupant is known to be formaldehyde-reactive, specify a formaldehyde-free binder line. Where an occupant has tested and reacted to mineral wool itself, move to a binder-free product such as cellular glass or perlite before considering wool.

Pair any of these with a continuous air barrier and correct vapor control for the climate zone. The insulation is not the air barrier and was never going to be.

Why this is not close

Three reasons, in order of weight.

It is predictable. Mineral wool's emissions reach a defined end state and stop. Its composition does not drift. The batch you buy next year behaves like the batch you bought this year. You are choosing a material for a cavity you will seal and never reopen, and under that constraint predictability is not a convenience — it is the entire risk argument. A material whose future you can know beats one with a marginally better first day and an open-ended rest of its life.

It does not become a substrate. This is the only property in the comparison that cannot be compensated for elsewhere in the assembly. Every other variable — moisture, vapor, air, thermal bridging — has a detail that addresses it. Organic material in a sealed cavity does not.

It holds its shape. Insulation has to stay where you put it for decades. Voids from settlement are invisible, unreachable, and create condensation sites in exactly the wrong place.

And why wool is still a good material

Because this needs saying, and because a page that could not say it would not deserve to be believed: sheep wool is a well-made, genuinely low-toxicity product, and it is lower-risk than fiberglass and cellulose on nearly every measure that matters to indoor air quality. The people making it are not selling snake oil.

The problem is not the material. It is the fit. Every advantage wool offers is a forgiveness advantage — moisture buffering forgives a vapor control failure, air cleaning forgives an air sealing failure. Forgiveness is genuinely valuable where correctness is unavailable, which is why wool is the right answer in a 1780 stone farmhouse whose wall predates the concept of an air barrier.

But when you are building from scratch, correctness is available. You can simply build it right. And at that point the forgiveness has nothing left to forgive, while the substrate risk remains — permanently, invisibly, and behind a wall nobody is going to open.

That is the whole argument. It is not that wool is unhealthy. It is that a house built properly does not need what wool is selling, and still has to live with what wool brings. For the full specification, decision worksheet, and contractor spec packs, see the $79 Build Package.

Sources and disclosure

Manufacturer-published technical data

• ROCKWOOL Comfortbatt product data and test results (ASTM E136, ASTM E84, ASTM C1338, GREENGUARD Gold certification)
• ROCKWOOL Safe Use Instruction Sheet, version 2.2, issued April 2026 (product family composition by weight)
• ROCKWOOL AFB evo Health Product Declaration
• Owens Corning Thermafiber formaldehyde-free product documentation
• Knauf ECOSE binder certification documentation

Independent and academic

• International Agency for Research on Cancer, Monograph Volume 81 (2002), reclassification of insulation glass wool, rock wool, and slag wool
• Peer-reviewed research on formaldehyde absorption by wool keratin, including chemisorption via amino acid side chains and methylene bridge formation
• Published review of sheep wool thermal insulation and indoor air quality, noting the absence of any established saturation threshold
• Green Building Advisor, independent ingredient analysis of insulation products

Vendor-authored sources — disclosed as such

• Havelock Wool company publications on wool and formaldehyde. Authored by the manufacturer of the product described.
• Comparative article on sheep wool versus mineral wool authored by the founder of a sheep wool insulation company. Useful on the chemistry; the omissions are as informative as the content.
• International Wool Textile Organisation materials. Industry trade body.
• Mineral wool manufacturer technical literature. Manufacturer-published.