Mass Timber Self-Extinguishes Only If the Char Stays Glued On

The number everyone quotes

Ask how mass timber survives a fire and you will hear one figure: about 38 millimetres an hour, or one and a half inches, the rate at which a solid timber section chars. Eurocode 5 pins the notional charring rate of softwood at 0.65 millimetres a minute for spruce and pine above roughly 290 kilograms per cubic metre; North American codes round it to one and a half inches an hour. From there the arithmetic is reassuring. After two hours of standard fire a glulam column has lost only about 65 millimetres of its face to char, and the designer simply oversizes the section so that enough unburned wood remains to carry the load. The trick is that the char is not damage so much as armour. As the outer fibres pyrolyse into a porous carbon crust, that crust, around 85 percent carbon and a poor conductor, insulates the wood beneath, so the burn front slows itself down. Engineers even add a seven-millimetre zero-strength layer below the char to account for the heated, weakened wood, and design the rest of the section as if it were cold.

Where the number stops being true

In a real compartment the story is more demanding. Once furniture and contents have burned out, a concrete building simply cools. A timber building will only do the same if its exposed surfaces stop feeding the fire, which is what fire engineers call self-extinction. Full-scale compartment tests have mapped the condition with some precision. The timber surfaces self-extinguish once the incident heat flux at the wall falls below roughly 45 kilowatts per square metre, and compartments reliably burned out and cooled when less than about half of their timber surfaces were left exposed, with a stricter limit, an unprotected ceiling, once more than 38 percent of surfaces showed bare wood. The char layer is doing the work. As it thickens it throttles the heat reaching virgin wood, and measured pyrolysis rates fall to somewhere between 0.4 and 2.3 millimetres a minute, faster on the ceiling than on the walls. Left alone, a well-detailed room of exposed timber can peak and then quietly go out.

The failure is a glue line, not a flame

That self-limiting behaviour rests on a single assumption: that the char stays where it formed. In cross-laminated timber it often does not. CLT is a stack of boards glued in alternating directions, and every glue line is a plane of weakness in fire. Heat-induced delamination happens when a lamination lets go before the char front has even reached its bond line. The still-solid board drops off the panel, and two bad things happen at once. The fallen timber is fresh fuel on the compartment floor, capable of driving a second flashover, a regrowth of the fire, long after the movable contents are spent. And the panel it left behind now presents a clean, uncharred face to the flames, resetting the protective clock. This char fall-off is exactly why some CLT compartments that should have self-extinguished instead reignited and burned on toward structural failure.

Polyurethane versus melamine

What decides whether a bond line holds is chemistry. The two adhesives that dominate CLT and glulam behave very differently when hot. One-component polyurethane, fast, formaldehyde-free and much loved by fabricators, softens and delaminates at elevated temperature; tests show PUR-bonded panels suffering heat-induced delamination and, with it, higher effective charring rates. Melamine-urea-formaldehyde adhesives hold their line far better. MUF-bonded beams char at close to the code notional rate, almost as if the glue were not there, even though their own shear strength collapses to essentially zero by around 280 degrees Celsius. The point is not that MUF is strong when hot. It is that it keeps the char attached long enough for the char to do its insulating job. Second-generation Eurocode 5 fire rules and product approvals are increasingly written around this distinction, demanding adhesives that maintain bond-line integrity in fire rather than merely at room temperature.

Designing for the crust to stay on

So the real fire-safety levers in a tall timber building sit upstream of the charring rate. Four times matter, and none of them is the wood's burn speed: how long the movable fuel takes to burn out, how long until the char falls off, how long any protective encapsulation lasts, and how the compartment sheds heat once the fuel is gone. Encapsulation, a layer or two of fire-rated gypsum board over the timber, buys the single biggest margin, delaying the onset of charring by an hour or more. That is why the tallest code categories, the IBC Type IV-A and IV-B that permit timber towers up to eighteen storeys, require most of the structure to be covered, reserving fully exposed wood for the lower IV-C tier. Where the wood is left proud for its warmth, the specification that matters is invisible: a bond line qualified to hold its char, laminations detailed so that one dropped board cannot cascade into others, and a compartment sized so that even a worst-case fall-off keeps the exposed fraction under the self-extinction threshold.

The least interesting number in the room

The comforting figure of 38 millimetres an hour is true, and it is also the least interesting number in the room. It describes a slab of solid wood burning tidily in a furnace. A building is not a furnace. It is a compartment full of adhesive-bonded panels that will either hold their protective crust and go out on their own, or shed it and burn a second time. The material that makes tall timber possible is not, in the end, the wood. It is the glue that keeps the char where the fire put it.