MDF Formaldehyde: The 0.11 ppm Stamp Tests a Panel at Rest, Not Your Router

There is a stamp on the edge of the MDF I buy, and for a long time I read it the way you read a UL mark on a power tool: somebody tested this, it passed, move on. Then I looked at what the number behind that stamp actually describes. Out here the air is wet most of the year, and a good share of my shop hours go to turning sheet goods into parts on a router. The gap between what gets tested and what I breathe turned out to be the whole story.

The Four Numbers, and the Box They Came Out Of

Under 40 CFR 770.10, as published by the Cornell Legal Information Institute, TSCA Title VI sets four limits: 0.05 ppm for hardwood plywood, 0.09 ppm for particleboard, 0.11 ppm for medium-density fiberboard, and 0.13 ppm for thin MDF. The method written into the rule is ASTM E1333-14, the compliance date was June 1, 2018, and per the California Air Resources Board these limits are identical to CARB's.

Now the box. Capital Testing, an EPA-recognized third-party certifier that runs the test, describes ASTM E1333 as an 834-cubic-foot chamber held at 25 ± 1 °C and 50 ± 4% relative humidity, with 0.50 ± 0.05 air changes per hour. MDF goes in at 66.7 square feet of exposed surface, after seven days of conditioning, and sits 16 to 20 hours before air is drawn off and analyzed.

Read that as a woodworker. Half an air change per hour, room temperature, half humidity, a panel lying still. Nothing in the method involves a blade, an abrasive, heat, or a person. So 0.11 ppm is a property of the material at rest — a steady-state room concentration, not an exposure. The 2010 USDA Forest Products Laboratory paper by Frihart and colleagues puts the same objection to the method, asking whether those conditions "represent all the exposure that interior composite wood products will experience in service."

The trap is the unit. The number on the stamp and the number that matters to your lungs are both ppm formaldehyde in air, which is exactly why people collapse them. They describe two different situations.

What the Stamp Promises, and What It Never Claimed

Per 40 CFR 770.45, a compliant panel label carries four things: producer's name, lot number, the EPA Third-Party Certifier number, and a statement that the product is TSCA Title VI certified. A finished good carries the fabricator's name, a month/year date, and a compliance statement.

Three things that label does not say:

• Finished goods are never tested. EPA's FAQ for regulated stakeholders is blunt: "Finished goods do NOT require formaldehyde emissions testing." The fabricator sources certified panels and labels the result.
• EPA certifies nothing. Certification is issued by third-party certifiers, who test quarterly to ASTM E1333-14 or D6007-14 and inspect producer facilities.
• The rule is silent on machining. It governs panels as manufactured and sold. There is no regulated number for what comes off a router bit.

Per EPA, hardboard, structural plywood, OSB, prefabricated I-joists, and finger-jointed lumber sit outside TSCA Title VI entirely — the pegboard and underlayment on the same rack are not covered, and this is a different mark from the veneer grade stamp besides. EPA also lets NAF and ULEF panels qualify for reduced testing or limited exemption from third-party oversight after three or six months, so a legitimately labeled panel may not be under routine quarterly testing.

Of three dealer listings I checked on August 19, 2026, only Plywood Company of Fort Worth printed "TSCA Title VI Compliant" on the product page. At the counter you are trusting the supply chain, not reading it.

The Manufacturer Draws the Line at the Blade

The strongest support for all this comes from a panel maker, not a critic. ARAUCO North America's safety data sheet for MDF and particleboard, revised September 6, 2022, states that its hazard information "applies only when the product is altered downstream by cutting, sawing, sanding, heating or other means and significant dust or fume is generated. In its shipped and finished form, this product is not considered hazardous."

So one panel carries two GHS classifications. As delivered: not classified. Once machined: eye irritant, respiratory and skin sensitizer, lung damage on repeated exposure, combustible dust, and Carc. 1A, may cause cancer by inhalation. Signal word Danger.

ARAUCO lists the composition as 76–93% wood dust, up to 18% melamine-urea-formaldehyde polymer, up to 15% urea-formaldehyde polymer, and under 0.01% free formaldehyde. The formaldehyde is bound in cured resin, and cured resin is what your cutter atomizes.

What Anyone Has Actually Measured

Less than you would expect. The one real machining study I can point to is Chung and colleagues in The Annals of Occupational Hygiene, 2000, testing four MDF types against pine and oak. They found free formaldehyde in air "less than 0.17 mg m⁻³" while machining the higher-emitting board — roughly under 0.14 ppm converted at 25 °C, my arithmetic and not theirs. Sanding MDF produced substantially more dust than sanding solid wood, sawing showed no significant difference, and the dust itself was "comparable" to hardwood and softwood dust.

Mount Sinai's MDF safety document for carpenters, hosted by eLCOSH, reports 0.01–0.14 ppm for sawing and sanding MDF in a ventilated dust chamber, and 0.19–0.78 ppm for sanding particleboard. Two cautions: no primary studies are cited for either range, and the higher figure is particleboard, though plenty of woodworking coverage repeats it as MDF.

For an upper bound, Thetkathuek and colleagues in Advances in Preventive Medicine sampled 439 workers at an MDF furniture factory in eastern Thailand in 2015 and found personal formaldehyde exposure averaging 2.62 ppm, ranging to 20.85, with dust averaging 7.67 mg/m³. That is a production plant with poor controls, not a garage, and the mean sits more than three times over OSHA's limit. The ceiling is set by ventilation, not by the chamber number.

Ding and colleagues in the Journal of Wood Science (2020) sampled a production sanding line and found 79.6% of particles under 40 μm and roughly a third under 10 μm, in tension with Chung's "comparable" finding. I won't resolve that toward the scarier reading; both agree sanding makes more dust than solid wood does.

The honest gap: I could not find one published measurement for a person machining MDF in a small shop, and no operation-specific data for routing or CNC work at all. Every figure above is a chamber, a factory, or an unsourced summary. Nobody has published your shop.

The Limits That Apply to Your Breathing Zone

OSHA's formaldehyde standard, 29 CFR 1910.1048, sets a PEL of 0.75 ppm as an 8-hour TWA, a 2 ppm 15-minute STEL, and a 0.5 ppm action level that triggers exposure monitoring and medical surveillance. NIOSH's REL is 0.016 ppm TWA with a 0.1 ppm ceiling, roughly 47 times below OSHA's number.

Wood dust is where most coverage goes wrong. Per OSHA's woodworking eTool there is no wood-dust-specific PEL; wood dust falls under Particulates Not Otherwise Regulated at 15 mg/m³ total and 5 mg/m³ respirable, while NIOSH's REL is 1 mg/m³. You will see it written that OSHA sets 1 mg/m³ for hardwood — Dust Safety Science says exactly that — but 1 mg/m³ was an ACGIH value, and OSHA's 1989 wood dust PELs were vacated by the courts in 1992, which is why the permissive figure is the one that applies. The correction cuts against the alarm, and it is still the reason to size collection around NIOSH's number.

Both agents are IARC Group 1. Formaldehyde was reclassified in Volume 100F (2012), with sufficient evidence for nasopharyngeal cancer and leukaemia; the 1995 Group 2A rating is superseded and still widely quoted. Wood dust is Group 1 for nasal, paranasal sinus, and nasopharyngeal cancer.

Humidity Is the Variable a Wet Shop Can't Design Around

The Forest Products Laboratory work is the second argument against reading the stamp as a guarantee. Frihart and colleagues tested two hardwood plywood panels, both independently verified CARB Phase 2 compliant under 0.05 ppm, differing only in adhesive — one ULEF urea-formaldehyde, one soy-based NAF. Their figures are micrograms per gram from a modified bottle method, so they are not ppm and do not belong beside the chamber standard. The ratio is the point: at 35 °C and 100% RH over four days the ULEF panel emitted 178.1 µg/g against 5.3 for the NAF, more than 33 times as much. They are clear 100% RH is an accelerated test, not a room. The mechanism is hydrolysis — free water drives urea-formaldehyde back toward free formaldehyde.

Two panels, identical on paper, nothing alike off-spec. Their table of U.S. city humidity lists Seattle at 83% morning RH, and I take that personally.

Designing With Intent: Buy the Panel, Control the Cut

The certification is worth having, and I would rather buy under it than not. It prices a panel at rest. Everything after the forklift is an engineering problem you own.

Capture at the point of generation, because Mount Sinai is explicit that "filters designed to trap wood dust will not necessarily capture gases, like formaldehyde." A shop air cleaner recirculating through a pleated filter will not necessarily capture the gas fraction. Prefer exhaust that vents outside for sustained MDF work, and never blow down a bench or your clothes with compressed air. For respirator selection, follow the manufacturer's guidance rather than forum consensus.

The FPL data is also the strongest independent case I know for choosing NAF over ULEF in a damp shop — not because the label differs, but because the two behave nothing alike once conditions leave the chamber. Columbia Forest Products says its PureBond panels use soy, PVA, and pMDI adhesives and typically test below reliable detection; that is a company claim rather than an independent finding, but the chemistry is the part that matters.

Ask your supplier for the TPC number on the next MDF order. If nobody in the chain can produce it, you have learned something useful about that particular rack — and either way, the cut is still yours to ventilate.