Tropical Hardwood Decking vs. Thermally Modified Wood in Freeze Thaw Climates: Which Actually Performs Better

Freeze Thaw Damages a Deck Through Water Sitting in Checks, Cut Ends and Fastener Holes

Wood does not spall the way concrete does, so freeze thaw failure on a deck is almost always liquid water expanding inside checks, end grain and fastener holes that were opened by ordinary moisture cycling first. Get that sequence straight and the whole species argument changes shape.

Concrete fails in freeze thaw for a clear reason. It is a saturated porous solid, water fills the capillary network, and freezing water expands about nine percent with nowhere to go. Wood is different. Wood is a cellular material with air space in the lumens and a cell wall that holds water chemically bound to the cellulose. Bound water does not freeze at 32 degrees Fahrenheit in any meaningful way. Free water in the lumens does, and in a deck board at 12 to 16 percent moisture content there is essentially no free water present. The board is below fiber saturation. There is nothing to freeze.

So where does the damage come from? Openings. A surface check, an unsealed cut end, a countersunk screw hole, a butt joint against a fascia. Those hold liquid water after rain or snowmelt, and that water freezes and pries. Each cycle widens the opening a fraction. Next cycle takes more water. The failure ratchets. A deck in Minneapolis or Denver or Buffalo can run through dozens of freeze thaw crossings in a single season, and the openings that started as hairlines at year two are visible splits at year eight.

The second mechanism is dimensional. Winter in a northern climate is a dry season indoors and a wet season outdoors. Boards swell with snowmelt, shrink hard in a February cold snap at 20 percent relative humidity, swell again in the March thaw. The USDA Forest Products Laboratory works through the sorption and shrinkage relationships in the Wood Handbook chapter on moisture relations and physical properties. Every one of those swings loads the fastener connection and the board face. The species that survives is the species that either moves less or resists the tearing better.

Thermal Modification Wins on Movement by Removing the Wood's Ability to Hold Water

Heating wood to roughly 190 to 215 degrees Celsius in a low oxygen kiln destroys the hemicellulose, which is the polymer that does most of the water binding, and the result is a board with equilibrium moisture content cut roughly in half and dimensional movement reduced by half or better. That is a real, measurable, laboratory-verified change, and in freeze thaw country it matters.

Hemicellulose is the sponge. Strip it out and the cell wall simply has fewer sites for water molecules to attach to. A thermally modified board that would have settled at 12 percent moisture content in a humid August now settles closer to 5 or 6. Tangential shrinkage that ran near 8 percent green to oven dry commonly drops by roughly half. The board stops chasing the weather.

Fewer moisture swings means fewer opened checks, and fewer opened checks means fewer places for ice to get a purchase. The freeze thaw benefit is indirect but it is genuine. Thermally modified material also gains decay resistance, since the degraded hemicellulose is no longer available as fungal food and the modified lignin structure resists enzymatic attack. Thermory and Abodo Vulcan both produce material that tests into the upper EN 350 durability classes, and unlike solid unmodified wood, those modified products carry manufacturer warranties.

Accoya belongs in this conversation with an asterisk. It is not thermally modified. It is acetylated, meaning the hydroxyl groups on the cell wall are chemically converted with acetic anhydride rather than cooked off. Same outcome on water uptake, different route, and Accoya keeps its strength where thermally modified wood gives some up. The FPL covers both chemistries in its Wood Handbook chapter on specialty treatments.

Density Buys Toughness, and Toughness Is What Survives a Snow Shovel in March

Ipe at a Janka rating around 3,680 lbf and a density near 1,050 kg/m3 arrives with EN 350 Class 1 durability already built in and gives up nothing in strength to get it. No kiln touched its mechanical properties. That matters more in a cold climate than the brochures admit.

A northern deck takes abuse that a Carolina deck never sees. Steel shovel blades. Ice chippers. Frozen patio furniture dragged across the boards in October. Snow load compressing debris into the surface for weeks. Grit tracked in with boots and ground underfoot. Every one of those is a surface hardness problem, and surface hardness is exactly where dense tropical hardwood is unbeatable and thermally modified softwood is weakest.

Thermally modified pine sits somewhere near 600 to 900 lbf depending on the source species and the treatment intensity. Thermally modified ash does better, roughly 1,000 to 1,300 depending on the same variables, which is below the unmodified white ash figure of about 1,320. Ipe is roughly three times harder than modified ash and four to six times harder than modified pine. Push a shovel across thermally modified pine decking in February and you will find the mark. Push it across Ipe and the shovel loses.

Density also means the boards resist water entry in the first place. Vapor moves through Ipe at roughly 1,050 kg/m3 or Massaranduba at 1,150 the way traffic moves through downtown. Slowly. That cuts both directions, and it is the reason dense tropicals need real care during drying and end sealing, but on an installed deck it means snowmelt sits on the surface rather than soaking to depth.

The Two Materials Fail Differently, and the Table Shows Why

No single number decides this specification, so read the comparison across all six columns at once rather than picking the winner in any one of them.

Species figures cross-check against the Wood Database entry for Ipe and its Cumaru and Garapa pages. Modified product values shift with treatment intensity and source species, so pull the current technical data sheet from the manufacturer for any load-bearing decision.

Read the pattern rather than the rows. Modified wood is stable and soft. Tropical hardwood is tough and moves. Freeze thaw punishes both traits, and which one punishes your project depends on details that have nothing to do with the wood.

Brittleness Is the Hidden Cost of Thermal Modification, and Cold Makes It Worse

The same hemicellulose degradation that stops thermally modified wood from absorbing water also strips 10 to 30 percent of its bending strength and a much larger share of its impact toughness. A stable board that snaps is not a solution.

Thermally modified wood behaves the way overcooked wood behaves. It is stiffer per unit of deflection and it fails suddenly instead of bending first. Modulus of rupture drops. Work to maximum load, the property that describes how much energy a board absorbs before it breaks, drops much further. The Wood Handbook chapter on mechanical properties of wood lays out how these properties relate, and thermal modification pushes every one of them in the same unhelpful direction.

Where does that bite on a real deck? Fasteners. Drive a screw too close to an end without a pilot hole in thermally modified pine and it splits. Cold makes the material more brittle still. A November install crew working thermally modified decking at 25 degrees Fahrenheit will split more board ends than the same crew working in June, and every split end is a water entry point that ice will exploit for the life of the deck.

Impact matters too. Dropped tools, a slipped shovel, a heavy planter set down hard on a frozen board. Dense tropical hardwood absorbs those events. Thermally modified softwood registers them.

"People call us asking which one is better in cold weather and they want a one word answer. It doesn't work that way. If you are putting down a residential deck up in Vermont with 14 foot runs and you care about movement more than anything, thermally modified is a legitimate answer and it comes with a warranty behind it. But when somebody is specifying a municipal boardwalk in Chicago that gets plowed and salted and walked on by ten thousand people a month, I am not selling them a modified softwood. I am selling them Ipe, because in eight years the modified boards will be chewed up on the surface and the Ipe will look like somebody just oiled it. The question is never which material is better. It is what is going to happen to this deck in February."

Norm Moton, Director of Sales, J. Gibson McIlvain

Length Settles More Freeze Thaw Specs Than Any Property on the Data Sheet

Thermally modified and acetylated products top out around 16 feet, and most arrive on metric lengths that land a few inches under, since the modification vessels, kiln or reactor, physically cannot hold longer stock. This is not a supply problem that patience fixes. It is the size of the vessel.

Every butt joint is a freeze thaw liability. Two cut ends meeting over a joist, each one pulling water in far faster than face grain does, sitting in a joint that stays damp for days after the surface has dried. That is the exact geometry that ice pries apart. A 24 foot run in 16 foot material means one joint. In 20 foot Ipe it means none.

Long boardwalks, long dock runs, long commercial decks with continuous sight lines all push toward tropical hardwood for this reason alone, before durability class enters the discussion. J. Gibson McIlvain maintains a vast Ipe inventory across all standard and custom dimensions, including the long lengths that make continuous runs possible, and ships that material nationwide to jobs from New England through the upper Midwest and out to California. Full inventory and dimension detail lives on the Ipe decking page.

Modified material has its own dimensional advantage worth naming. It comes pregrooved and profiled from the factory with consistent tolerances, which speeds a hidden fastener install and reduces the amount of field cutting. Fewer field cuts is fewer end grain exposures. That partially offsets the length penalty on shorter decks.

Snow Removal and Deicing Chemistry Punish the Softer Material Every Winter

Nothing on a deck ages a surface faster than a steel shovel and a bag of ice melt, and the harder the board the less it cares. Maintenance reality decides more freeze thaw outcomes than material science does.

Chloride deicers are hygroscopic. They pull water and hold it against the wood, which keeps the surface wetter for longer and extends the window during which a freeze event finds liquid water sitting in a check. They also accelerate corrosion on anything metallic in the assembly. Calcium chloride and magnesium chloride are both aggressive toward fasteners and hardware. Sodium chloride is less so but still bad.

Sand is the better answer on a wood deck. It does nothing chemically, it provides real traction on ice, and it sweeps off in the spring. Where a deicer is unavoidable, calcium magnesium acetate is the least destructive of the common options.

Snow removal technique matters. Plastic shovel, not steel. Work with the board direction, never across it, so the blade rides the grain instead of catching the raised early wood. Leave a thin layer rather than scraping to bare wood. On thermally modified softwood this is not advice. It is the difference between a deck that looks good at year five and one that looks worn out.

Installation Detailing Carries More Weight Than the Species Choice

A correctly detailed thermally modified deck outlives a badly detailed Ipe deck, and the reverse is equally true, since freeze thaw attacks the assembly and not the material. There is no silver bullet species. There is only a deck that drains and dries or one that does not.

Start below grade. Footings go below the local frost line, no exceptions, and the depth commonly ranges from about 12 inches in the mid-Atlantic to 60 inches or more in northern Minnesota. Frost heave will lift a shallow footing, rack the frame, and open every joint in the deck above it, and no species survives that. Prescriptive deck framing guidance from the American Wood Council codes and standards library covers the framing side, and local amendments adopted through the International Code Council govern the actual frost depth number.

Joists at 16 inches on center for dense tropical decking, tighter at 12 inches for diagonal layouts or where a 5/4 board carries real load. Remember that a 5/4x6 board covers about 2.1 linear feet per square foot when you take off quantities. Thermally modified softwood decking wants 16 inches on center as well, and the manufacturer's span table governs. Do not extrapolate a span from a tropical hardwood table to a modified board. The strength properties are not the same.

Fasteners. 316 stainless for any coastal work and for any deck that will see chloride deicers, which in a freeze thaw climate is most of them. 305 stainless is the inland minimum. Nothing plated survives salt and moisture cycling. Predrill and countersink dense tropical hardwood without exception, and predrill thermally modified material near board ends where brittleness concentrates.

Profile drives the fastener choice, not the other way around. Shiplap and square edge boards take visible fasteners through the face. Only tongue and groove works with a hidden fastener system riding in the groove. Pregrooved decking accepts a clip. On any grooved or T&G material the grooves face DOWN so water drains out of the profile rather than collecting in it, which in freeze thaw country is the difference between a channel and an ice lens.

Material sits on joists or furring strips with air moving underneath. Both faces have to dry at the same rate or the board cups and the cup holds water. Ventilate the skirt. Keep 18 inches of clear space to grade where the design allows. The rain control principles in the Building Science Corporation digest on rain control in buildings apply to horizontal assemblies with more force than most builders assume, since a deck is a roof that people walk on.

Seal every field cut the day it is made. Wax emulsion end sealer, two coats on butt joints. This single step does more for freeze thaw longevity than any species upgrade available.

Deliver at the Right Moisture Content or Winter Will Find the Difference

Tropical decking installs at 12 to 16 percent moisture content, thermally modified material arrives near 4 to 6 percent and stays there, and material delivered outside its window checks in the first cold snap no matter how carefully it gets finished afterward.

The 12 to 16 percent range brackets exterior equilibrium moisture content across most of the United States over a full year. A board delivered at 20 percent into a Chicago November is going to release four points of moisture in a month of dry cold, and it will do it faster at the surface than at the core. That gradient opens checks. Those checks fill with snowmelt. The snowmelt freezes. The ratchet starts.

Use a pinless meter with a species correction setting. Dense tropicals throw uncorrected pin readings by several points, which is the whole margin. Check ten boards per unit at midspan, not at the ends.

Pigmented penetrating oil is the correct finish on tropical hardwood, applied within days of installation while the surface still accepts it. Film forming finishes fail on oily dense species, and once a film cracks it traps water behind it and drives damage from underneath. Thermally modified wood takes coatings more readily and holds them longer, since the substrate is more stable, which is one more legitimate point in its column.

How J. Gibson McIlvain Would Specify This

Sort the question by traffic, span and maintenance access before you sort it by material. Four scenarios cover most of what comes across the desk.

Commercial or municipal, plowed, salted, high traffic, long runs. Ipe. It is not close. The combination of Class 1 durability, a Janka rating near 3,680 lbf, ASTM E84 Class A fire performance verified under the ASTM test standard at a flame spread index of 25 or less, and availability past 20 feet answers every constraint at once. Joists at 16 inches on center, 316 stainless fasteners, predrilled and countersunk, all cut ends sealed.

Residential deck, moderate traffic, runs under 16 feet, an owner who cares about movement and wants a warranty behind the product. Thermory or Abodo Vulcan is a real answer here and J. Gibson McIlvain carries both. Specify to the manufacturer's span table, use the factory profile and clip system, keep grooves facing down, and set expectations honestly on surface hardness.

Rooftop or weight-constrained assembly in a freeze thaw climate. Thermally modified material at roughly 450 to 650 kg/m3 against Ipe at 1,050 is a structural argument that wins on its own terms. Confirm the dead load allowance with the engineer of record before the species conversation goes any further.

Long runs where cost drives the decision more than stability does. Cumaru or Massaranduba deserve a look, since both cost less than Ipe, with the caveat that both move more than Ipe and demand tighter fastening discipline. Jatoba brings an ASTM E84 Class A rating for projects where fire performance drives the spec. Garapa and Red Balau serve lighter-duty residential work where the surface will not see plow blades. Clear vertical grain cedar remains a legitimate choice for a low traffic deck that will never meet a steel shovel, and select tight knot cedar answers a different aesthetic entirely.

Sourcing runs through the same checks either way. FSC chain of custody documentation is available for certified material, and the standards live at FSC International. When using a CITES-listed species, documentation is required and legal-harvest verification has to follow the material through the chain, with current listings published by the CITES Secretariat. Specifiers building the broader assembly case will find the technical library at WoodWorks useful.

J. Gibson McIlvain mills and stocks tropical hardwood decking and carries Thermory, Abodo Vulcan and Accoya, shipping nationwide to commercial and contractor-volume projects. Bring the drawings, the frost depth, the span and the maintenance plan to the conversation and the material answer falls out of it. Call 800-638-9100 for a quote or to talk through a freeze thaw detail with someone who has specified it before. More on the company and its product range at mcilvain.com.