What’s Actually in a Cabinet Box, and Which Part Can Off-Gas

CARB Phase 2 is the floor everyone is standing on

The rule people mean when they say “CARB2” started in California, as the Air Resources Board’s airborne toxic control measure for composite wood. It became continental when the United States wrote the same limits into federal law as Title VI of the Toxic Substances Control Act, and Canada followed with its own composite-wood formaldehyde regulation built to the same numbers. The practical effect: a panel legally sold here already meets it.

So “CARB Phase 2 compliant” on a spec sheet is a statement that the supplier is obeying the law. That is not nothing — non-compliant panel does still reach this market, which is the reason the rule carries an enforcement programme at all — but it is a baseline claim, not a distinction. Anyone selling it to you as an upgrade is selling you the speed limit.

What the standard actually does is cap how much formaldehyde a bare panel may release in a controlled chamber, by panel category:

Read those as chamber figures, not room figures. The test runs on the bare panel as it leaves the mill, at a fixed loading of board per cubic foot of air, at a set temperature and humidity. Your kitchen is a different room with different air, different weather and a very different amount of board in it. The glossary has the short definitions if the vocabulary is new.

And here is the gap that matters most, because nobody puts it on a label: the standard governs a rate per panel, and what you breathe is a dose per room. A bathroom vanity and a 20′ galley kitchen built from identical certified board are not the same exposure. Twenty-six boxes put roughly twice as much board into the house as thirteen. No certificate can help with that, because the quantity is your decision, not the mill’s.

What they search at midnight “do kitchen cabinets off gas” — usually typed the evening the boxes are unwrapped in the living room, with a baby asleep upstairs and a smell in the house that nobody can name.

Finally, the list of things Title VI is silent on, which is longer than the list of things it covers. It says nothing about strength, screw holding or what happens after a leak. It measures one chemical from one class of product, so every other adhesive in the cabinet — the hot melt behind the edge band, the construction adhesive on site, the silicone around the sink — is outside it, as is every other volatile compound in the room. It says nothing about the finish on the door. And it says nothing about how much of the stuff you buy.

One is a claim about the resin, the other about the testing

Above the legal floor sit two terms of art, and they are not two grades of the same thing. They answer different questions and they appear on a mill certificate in different places.

“Non-toxic” is not a specification. It appears on no certificate, has no test behind it and no threshold attached to it. Neither does “green,” “eco,” “natural” or “chemical-free.” The words that mean something are dull: the panel category, the certifier, the certificate number, the exemption status, the ppm figure and the date.

One more caution, because NAF gets oversold in the other direction too. A no-added-formaldehyde resin is an air-quality decision and nothing else. It says nothing about whether the board holds a hinge screw, or how it behaves when a dishwasher hose weeps for a week. Some early soy-bonded panels were noticeably less happy in the wet. Judge the board on both counts, and the construction argument lives on the materials page, not here.

A cabinet is not one material, and it is not one certificate

A cabinet is an assembly of boards from several different mills — which is the part the advocacy pages and the trade explainers both step over. The ¾″ carcass board, the thin back panel, the door core and the drawer bottom are four separate products, frequently bought from four separate suppliers, each with its own certificate — and not all of them in the same panel category, which means not all of them held to the same ceiling.

So “our cabinets are CARB2” is a sentence about a company, not a document about a panel. The question that gets you somewhere is which board, from which mill, for which part. A supplier who can answer part by part has the paperwork in a drawer. A supplier who answers with one sentence is repeating what their own panel salesman told them.

Here is how a frameless box breaks down. The construction figures are the ones our materials page publishes; the emission column is the point of this page.

Two things fall straight out of that table. The first is that a large fraction of a modern kitchen has no emission story at all: the steel drawer sides, the stone, the hinges, the runners, the handles, the PVC band, an aluminium kick. The parts capable of releasing anything are the board parts, and there are fewer of them than the phrase “non-toxic cabinets” implies.

The second is where the board is concentrated. It is not the doors, which people fixate on because doors are what they look at. By surface area it is the carcass — the sides, shelves and bottoms, repeated across every box on the wall — and that board is the one most likely to be quoted as an afterthought.

If you are still deciding between ready-to-assemble, assembled stock, semi-custom and full custom, the cabinet types comparison sets those four beside each other, and framed beside frameless. The part-by-part question applies to every one of them — only the parts list changes.

The faces are sealed. The edges are the whole story.

What a panel releases into a room is roughly proportional to how much uncoated board surface the room can actually reach, which reframes everything above. Board buried under a barrier is board that mostly cannot talk to your air.

A melamine face is that barrier. It is a resin-impregnated paper pressed onto the board under heat and pressure until it cures — not a coat of paint, not a film sprayed on afterwards, but a thermoset skin fused to the panel. A panel faced on both sides and banded on the edges that show has very little raw surface left anywhere. That is why a well-made melamine box is quieter than its panel category suggests, and it is the reason the interior of a good cabinet looks finished rather than furry.

So what is left is the edges, and there are exactly two kinds.

The first kind is the edge the factory left bare on purpose — typically a back edge or a top edge that will never be seen. That is normal and mostly harmless: it is a narrow strip, it faces a wall, and it is inside a sealed assembly.

The second kind is the one you make. Every site cut is a fresh raw edge, and site cuts cluster in exactly the wrong places. Scribing a filler to a wall that is not plumb. Trimming an end panel. Cutting the back for the waste and the supply. Notching a drawer bottom around a trap. Cutting the toe kick to length. Drilling through a gable for an outlet cable. Count those up and most of them are in or around the sink base, which is the warmest, dampest cabinet in the kitchen.

The fix takes five minutes per cut and no special material: close the edge. An offcut of iron-on banding is ideal, because it is the same material the factory used and it seals against water as well as air. Failing that, two coats of a water-based primer or sealer on the cut edge does most of the job. You were going to want that anyway — a raw particleboard edge under a sink is the classic swelling failure, which is why our materials page rates furniture-grade particleboard as fine if sealed rather than fine full stop.

One last observation, and it explains the most common real-world complaint. A closed cabinet is a small volume with a great deal of board per cubic foot of air inside it — a far higher loading than the room. Shut the doors for a week and whatever is coming off the interior accumulates in there. Then you open the door and get a concentrated lungful. “The kitchen is fine but the inside of the cupboard smells” is not a mystery; it is arithmetic, and it is why the airing protocol further down is mostly about doors and drawers rather than windows.

Plywood is not automatically the low-emission choice

Search for low-emission cabinets and within two clicks somebody will tell you to buy plywood boxes. It is the most repeated advice in this subject and it is half true for a reason that is not the reason people give.

The half that is true: hardwood plywood is held to a tighter ceiling than the other two categories — 0.05 ppm against 0.09 for particleboard and 0.11 for MDF. That is a real difference in the paperwork. But it is a difference in the ceiling, not in your panel, and the two are easy to conflate. A particleboard panel bonded with pMDI sits far below the plywood ceiling. A urea-formaldehyde hardwood plywood can sit right up against its own. Ceilings tell you the worst legal case, never the actual case.

The half that is misleading: the ceilings differ because of how much resin is in the product, not because of any virtue in the species. Plywood is veneers with glue in a handful of lines between them. Particleboard and MDF have resin distributed through the entire mass of the board. More resin per pound, more to release, looser ceiling. That is mass, not morality.

What actually decides emissions is the resin system. Ask for that, and the substrate argument mostly dissolves:

That third row deserves a note for anyone building boxes rather than buying them. Sheathing-grade plywood and OSB are bonded with phenolic resin and are excluded from the composite-wood rule, which surprises people who assume “not covered” means “worse.” It is the opposite: they are outside the rule because phenolic glue is a different and far more stable animal. They are also rough, heavy and void-prone, so that is an argument about emissions only — not a recommendation to build a kitchen out of sheathing.

None of which means plywood is a bad buy. It holds a screw better in a cut edge, it is lighter for its stiffness, and it recovers better from a soaking — all good reasons, and box material heads the list of six things that separate a good cabinet from a bad one. Buy it for those. Do not buy it believing the substrate name settled the emissions question, because the name is the one thing that does not.

Week one is solvent. It is almost never formaldehyde.

The smell that fills a house in the days after a kitchen goes in is, overwhelmingly, solvent flashing off wet materials — and formaldehyde from board is neither wet nor fast, so it does not behave that way at all.

Run through what was actually applied that week: silicone around the sink and along the backsplash; construction adhesive under a countertop; paint on the walls and ceiling; a new floor with its underlay and its adhesive; caulk at the toe kick; and a living room full of shrink wrap and corrugated cardboard, which soak up solvent and re-release it for days. That is five or six wet products and a lot of packaging, and every one of them is louder in week one than the board is.

It also matters how the doors were made. A factory-pressed laminate — melamine or PET bonded under heat — is cured before it leaves the plant. There is no wet coating curing in your house. On a kitchen built that way, the first-week smell is somebody else’s trade: the caulk, the adhesive, the paint, the floor. A sprayed-lacquer or site-painted kitchen is a genuinely different proposition, and there the doors really can be the source for a few weeks.

All of it smells different, one source from the next, and the difference is learnable:

• Sweet, heady, paint-shop or nail-polish Solvent from a coating or an adhesive. You can walk around and point at the object it is coming from. Strongest for a day or two, largely finished in one to two weeks, longer inside a closed cabinet or where the film is thick.
• Sharp vinegar Acetoxy-cure silicone, which releases acetic acid as it sets. Loud for a few hours, gone in a day or two. Harmless to diagnose and impossible to mistake once you know it.
• Warm, papery, faintly plastic Packaging and PVC film. It leaves when the packaging leaves, which is a good reason to get it out of the house the day the boxes are unwrapped.
• Sharp, sour, faintly pickled — and a sting The formaldehyde signature. It tends to register as irritation at the eyes and the back of the nose before you consciously identify a smell. It is worse in a closed room, worse on a hot humid day, and worst in the first second after you open a cabinet that has been shut.

Two caveats on your own nose, because this is where people mislead themselves in both directions. Odour thresholds vary enormously between individuals, and irritation and odour do not track each other neatly — so smelling nothing is not proof of anything, and smelling something is not an identification. A nose tells you to investigate. Only a measurement tells you a number.

The curves are different shapes too, and that is what governs how long to be patient:

• Solvent falls off a cliff — most of it is gone in days, essentially all of it inside a few weeks. A smell still going at month three is very unlikely to be solvent, unless something thick-filmed is curing under a floor or behind a panel.
• Formaldehyde decays slowly with a long tail — the steep part is the first weeks, but the decay modelling the regulators publish for pressed-wood products puts the half-life in years rather than months, and it never reaches zero because wood emits a trace on its own.
• Heat and humidity drive the rate hard — a hot, humid, closed-up house in July is the worst case; the same kitchen at a normal winter temperature is measurably quieter.
• Ventilation changes the air, not the board — open the windows and the concentration drops fast; close them and it climbs back, a little lower each time. “We aired it out and the smell returned” is the normal shape of this, not a failure.
• Panel age helps you before you do anything — board that was pressed three months before it reached you has already run down the steep part of its own curve in somebody’s warehouse.

The airing protocol, and why the order matters

Given the curves above, airing a new kitchen is worth doing properly — and getting the order right is worth more than keeping it up for longer. Six steps, in this sequence.

• Unwrap outside, and get the packaging out the same dayShrink wrap, foam and corrugated board absorb solvent in the factory and re-release it in your living room. They are also the easiest source to remove completely.Do it whenThe delivery arrives — not after the install, when there is a mountain of it in the hall.
• Open every door and every drawer, and leave them openThis is the single highest-value action and the one everybody skips. The air inside a closed cabinet does not ventilate itself, and the cabinet interior has far more board per cubic foot than the room does. An open window with the cupboards shut is airing the wrong volume.How longContinuously for the first few days, then a few hours a day for the first month.
• Cross-ventilate, on two sides of the houseOne open window is a hole. Two, on opposite faces, is a current. A few hours of genuine cross-flow does more than a day of one cracked sash.TimingIn summer, do it in the cool of the morning — you are trying to move air without cooking the board.
• Keep it at living temperature and normal humidityEmission climbs with both heat and moisture, so a house held at a sensible winter temperature with relative humidity in the usual 30–50% band is the simplest control you have. A humidifier running at 60% is working against you — and against the boxes.Do notTry to “bake it out” by heating the house hard. You risk the finishes, and a sealed hot room concentrates exactly what you were trying to remove.
• Vent the other trades separatelyCaulk, adhesive, paint and flooring all want to cure with air moving. If they are curing behind closed doors, you are keeping the loudest sources in the house for no reason.The hoodA range hood that ducts outside genuinely exhausts the room. A recirculating one pushes the same air back into it through a charcoal cartridge far too small to matter — the same saturation problem as the carbon pre-filter on a purifier.
• Wipe the interiors once before food goes inA damp cloth through every box takes out sawdust and cutting residue, which hold and slowly release whatever settled on them. Then store flour, cereal and open packets in sealed containers rather than loose on a shelf — sensible anyway, and it removes the question.AndIf you cut anything on site, seal that edge before it goes into service. That is the one physical change worth making. The finishing chapter runs the same wipe-down as its hand-over step.

Equally useful is the list of things people buy for this that do not work:

• Shelf liner is cosmetic, not a barrier — the faces it covers were already sealed at the factory, and it does nothing for the cut edge you actually made.
• “Formaldehyde sealing” sprays over finished melamine — you are coating a cured laminate that was already the barrier. If you want to seal something, seal the raw cut.
• A HEPA purifier filters particles, not gases — formaldehyde is a gas and passes straight through. Removing gases needs a deep bed of treated carbon media, which is a different and much larger machine; the thin carbon pre-filter on a domestic HEPA unit saturates quickly.
• Houseplants are a lovely idea and a rounding error — the headline experiments ran in sealed chambers. The review that converted all of them to a like-for-like air-cleaning rate found you would need somewhere between one and ninety plants per square foot of floor to match the air the house already exchanges on its own.
• Ozone generators make the problem worse — ozone is itself a respiratory irritant and reacts with household VOCs to produce new ones, including more formaldehyde.

Ventilation is a control, not a cure. It lowers what is in the air right now while the board works down its own curve underneath. Both things are happening; only one of them is under your hand.

What to ask a supplier, and what a straight answer looks like

These are written to be pasted into an email without editing. The value is less in any single answer than in the shape of the reply: a supplier who has the documents answers specifically and quickly, and one who does not answers with adjectives.

• “Which standard are the panels certified to, and can you send me the certificate?”The opening question, and the one that sorts the field in a single reply.A straight answerNames the standard — EPA TSCA Title VI, CARB Phase 2 — and attaches or offers the mill’s third-party certificate with a certifier name and number. A dodge: “all our products are eco-friendly and green certified.”
• “Can you break that down by part — carcass board, back panel, door core, drawer bottom — and tell me which mill each comes from?”The question this whole page is built around, because these are four different products.A straight answerFour lines, four products, four certificates — or a frank “the back panel comes from a different supplier, let me get it.” A dodge: one sentence covering “the cabinets.”
• “Is any of it NAF or ULEF, and if so, which parts specifically?”You are testing whether they know the difference, as much as the answer itself.A straight answerNames the part and the resin, or says plainly that none of it is and it meets the standard on conventional resin — which is a perfectly respectable answer. A dodge: “formaldehyde free,” full stop. Nothing made of wood is at zero, so an absolute claim means somebody is selling rather than specifying.
• “Are the panels faced on both sides and banded on all exposed edges — and which edges are left raw?”The exposure question, which nobody volunteers and which decides more than the certificate does.A straight answerConfirms both faces, confirms the banding, and cheerfully names the edges that are bare — usually a shelf’s back edge or the top of a back panel. A dodge: “fully finished throughout.”
• “What is the finish on the doors: a factory-pressed laminate, or a coating applied wet? If wet, what is it and how long does it cure?”This is the question that predicts your first fortnight indoors.A straight answerNames the construction — a pressed melamine or PET face on a board core, or a specific coating with a cure time. A dodge: “premium finish.”

If they are still answering patiently, there is a sixth worth asking: “roughly when were the panels manufactured?” Emission decay starts at the press, not at your delivery date. Board that sat three months in a warehouse arrives with the steep part of its curve already behind it, and board pressed last week does not. Nobody will guarantee a date, but the answer tells you whether you are buying from stock or from a container that has just landed.

On our own paperwork, the position is the one the materials page already publishes and this page will not embellish: the door cores are specified as EPA Title VI compliant, that is the North-American limit for composite wood panels, a panel described as compliant has been certified against it, and it is worth asking any cabinet supplier — us included — to confirm it for the parts you care about.

Where the paperwork stops and your own clinician starts

Everything above is about materials and documents, not medicine. Here is where this page stops.

Formaldehyde is a recognised irritant of the eyes, nose and throat at low concentrations, and the International Agency for Research on Cancer classifies it as a Group 1 human carcinogen. That is a statement about the classification, not about your kitchen. Health Canada publishes residential indoor-air guidance for formaldehyde with separate short-term and long-term values; look those up at the source rather than trusting a figure quoted in any article, this one included, because guidance gets revised and articles do not.

What determines the dose in a real house is a list the spec sheet knows almost nothing about: how much board went in, how large the room is, how often the air changes, how warm and how humid it is kept, how old the panels were on delivery, and how much raw edge was created during the install. The certificate governs one of those. You govern the other five.

If somebody in the household has asthma, a diagnosed chemical sensitivity, a respiratory condition, or is pregnant, immunocompromised or very young, that is a conversation with their own doctor and not a decision to take from a guide page or a supplier’s brochure. We are cabinetmakers. We can tell you what the board is and where the raw edges are; we cannot tell you what any of it means for a particular person.

On testing, two practical notes before you buy anything. A passive badge sampler left out for a stated period and posted to a lab will give you a real number, and it is the only consumer route that does. But an absolute number without its conditions attached — the season, the indoor temperature, the humidity, how long the house was shut — is close to meaningless, because those variables move the result more than most material choices do. A comparative test is far more useful than a single reading: same room, same conditions, before and after a change. And treat direct-reading handheld meters with suspicion — cross-sensitivity is a property of the individual sensor, not of the price. In a 2025 laboratory comparison of two low-cost electrochemical units, one read methanol as more formaldehyde than there was methanol; the other shrugged the alcohols off. On the wrong meter a spray-and-wipe cleaner spikes the display and tells you nothing at all.

• Ask part by part — carcass, back, door core, drawer bottom. Four products, four certificates.
• Ask for the resin, not the substrate — pMDI, MUF, PF or UF tells you more than “plywood” ever will.
• Count the exposed raw edge — the factory sealed the faces. The cuts are yours, and so is closing them.
• Decide how much cabinetry you actually need — quantity is the variable no certificate controls.
• Air it properly for a month — doors and drawers open, cross-ventilated, at normal temperature and humidity.
• Take any specific health question to a clinician — a compliance document is not a medical opinion.