Industries / Food, farming and forestry / Forestry and logging / Timber and sawmills / Lumber treating
Modified wood
Changing the wood's own cell walls, with heat or with chemicals that bond to or harden inside them, so that it takes up less water and decay fungi cannot use it, without adding a poison.
5
Milestones achieved, from 1946 to 2024.
1
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1946
Acetyl groups stop swelling
Acetylation reacts wood with acetic anhydride, a chemical related to the acetic acid in vinegar, so that acetyl groups take the place of the water-attracting hydroxyl groups in the cell wall. By Roger Rowell's account, Fuchs in Germany first acetylated wood in 1928, to study lignin; at FPL, H. Tarkow showed in 1945 that acetylated balsa resisted decay, and in 1946 Tarkow, Alfred J. Stamm and E. C. O. Erickson first described using it to keep wood from swelling in water. Attempts to sell acetylated wood failed in the United States (Koppers, 1961), Russia (1977) and Japan (Daiken, 1984) because of high production costs, according to George Mantanis's 2017 review; Rowell writes that Daiken began making acetylated flooring in the late 1980s and was still making it.
LessonThe chemistry was proven decades before anyone could make the wood cheaply enough to sell.
Credit
H. Tarkow, Alfred J. Stamm and E. C. O. Erickson, Forest Products Laboratory
1993
Heat in place of chemicals
Thermal modification heats sawn wood in a kiln, to as much as 212 °C for the more durable grade, with no chemicals, which changes its chemistry so that it takes up less water and resists decay better. By the International ThermoWood Association's (ITWA) account, the first commercial plant, in Germany in the early 1980s, never reached industrial scale; the breakthrough came in Finland in 1993, when VTT, the Technical Research Centre of Finland, developed an industrial-scale process with wood companies, patented in 1994. ITWA, founded in 2000, counts ThermoWood production rising from 18,799 m³ in 2001 to 312,267 m³ in 2025. Its handbook says the wood is not sold strength-graded, must not be used for load-bearing structures, and is not recommended in contact with soil or water.
LessonHeat buys durability and stability at some cost in strength, which keeps the wood to cladding, decking and joinery.
Credit
VTT and Finnish wood companies; the International ThermoWood Association
2009
Furfuryl alcohol at full scale
Furfurylation soaks wood in furfuryl alcohol, a liquid made from farm wastes such as sugar cane and corn cobs, and heats it so that the alcohol turns into a polymer inside the wood. Research began with Goldstein and Stamm from 1959; Marc Schneider in Canada made a better product in 1995 with new catalysts, and M. Westin's group at SP in Sweden developed stable treating solutions, according to Mantanis's 2017 review, which reports that furfurylation raises wood to durability class 1, the highest. Kebony, formed in Norway in 1997 when it bought the patent, opened a pilot plant in 2003 and its first full-scale plant in 2009, moved production to a new factory in Belgium from 2018, and had centralized all of it there by 2025, by the company's own account.
Credit
Marc Schneider; M. Westin and SP Sweden; Kebony AS (formerly Wood Polymer Technologies)
2011
BASF drops its resin process
DMDHEU, a chemical first used in the textile industry, can be pressure-impregnated into wood and cured at about 100 to 120 °C in humid conditions, so that it hardens inside the cell walls, based on early work by Holger Militz in 1993, by the account of a 2019 European review from COST Action FP1407. BASF ran a pilot plant for the process under the name Belmadur. On 28 April 2011 the trade publication EUWID reported that BASF was dismantling the pilot plant and ending its involvement in the technology, with production already stopped.
LessonWorking at pilot scale is not the same as finding a market; the 2019 review put Europe's output from all chemical methods other than acetylation and furfurylation at about 35,000 m³ a year.
Credit
Holger Militz (University of Göttingen); BASF
2024
Acetylated wood made in Tennessee
Accsys Technologies, which makes acetylated radiata pine under the name Accoya at Arnhem in the Netherlands, said on 4 September 2024 that Accoya USA, its joint venture with Eastman Chemical Company in Kingsport, Tennessee, was in commercial production with an initial capacity of 43,000 m³ a year. Eastman had launched its own acetylated southern pine, Perennial Wood, from a Kingsport pilot plant in 2012 and stopped in 2014, Rowell writes. On 19 September 2024 Accsys discontinued a plant at Hull, England, meant to acetylate wood chips for its Tricoya panels, taking a charge of about €20 million. Accsys reported record sales of 77,237 m³ in the year to March 2026, its own count including the joint venture. Mantanis's review credits acetylated wood with the top durability class and cites an 11-year test in which highly acetylated pine resisted marine borers better than pine treated with chromated copper arsenate (CCA).
LessonBuilding and filling plants, more than the chemistry, has decided which acetylation ventures lasted.
Credit
Accsys Technologies; Eastman Chemical Company
Watch for
Class 1 from heat alone
ABODO, a New Zealand maker of thermally modified radiata pine, says its wood has "Class 1 durability", which its managing director Daniel Gudsell calls "the highest form of durability based on lab testing", in the launch of ABODO Vulcan siding and decking by its US partner Pioneer Millworks. Gudsell adds: "It's critical that you have a field test to support that." The International ThermoWood Association's handbook does not recommend heat-treated wood in contact with soil or water.
Achieved whena laboratory with no tie to the company publishes field test results for the product, in the ground or above it, that place it in durability class 1.
Claimed by
ABODO and Pioneer Millworks, May 2025
Open
01
Field results for modified wood
An open table of every published outdoor test of modified wood (acetylated, furfurylated, thermally modified and resin-modified) with product, chemical loading, test type (stakes in the ground, joints above ground, or panels in the sea), site, years exposed and decay ratings over time, plus a model of years to failure by type, loading and site, set beside the makers' warranties.
The briefCredit · unclaimed02
Checking heat treatment by camera
A phone app that estimates how strongly a board was heat-treated from a photo of a freshly sanded spot next to a printed color reference card, trained on boards treated at known temperatures and checked against laboratory color readings.
The briefCredit · unclaimed
Open 01 · the brief
Field results for modified wood
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An open table of every published outdoor test of modified wood (acetylated, furfurylated, thermally modified and resin-modified) with product, chemical loading, test type (stakes in the ground, joints above ground, or panels in the sea), site, years exposed and decay ratings over time, plus a model of years to failure by type, loading and site, set beside the makers' warranties.
Why it is still open
The results are scattered across papers, a few products at one site at a time, and they do not all point the same way. Acetylated strand board at 20.4% acetyl content had decayed completely after 102 months in the ground in western Greece (Papadopoulos, 2011), while Accsys offers warranties of up to 50 years on Accoya. Brischke and Rapp's model linking decay to wood moisture and temperature, built from 23 European field sites, used untreated pine and Douglas fir. The IRG/WP Durability Database, the International Research Group on Wood Protection's open platform for raw durability results, has sections for thermally and chemically modified wood but holds one thermally modified record and no chemically modified one among its 178 (count of its index, November 2026). Searches of GitHub, Zenodo and Hugging Face found no other open compilation.
Built with what exists
Crossref to find papers; an AI document reader to pull ratings from PDFs, every value checked by hand (the Note "Seven more flagged" covers an AI reviewer of long documents); WebPlotDigitizer for charts; Python with pandas and the lifelines survival-analysis library; Zenodo to host the data.
First step
Put every field result cited in Mantanis's 2017 review and the 2019 COST FP1407 report into a spreadsheet with fixed columns.
Done looks like
At least 500 records of specimen groups from at least 40 sources, covering at least three kinds of modification and ten test sites, openly licensed, with median years to failure in ground contact by type and loading and their uncertainty.
Open 02 · the brief
Checking heat treatment by camera
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A phone app that estimates how strongly a board was heat-treated from a photo of a freshly sanded spot next to a printed color reference card, trained on boards treated at known temperatures and checked against laboratory color readings.
Why it is still open
Brischke and colleagues showed in 2007 that color, measured as CIE Lab* values, tracks heat-treatment intensity for spruce, pine and beech, but they recommended measuring milled wood, because board surfaces vary more, and aimed the method at factory quality control. ThermoWood is sold in two grades, Thermo-S for stability and Thermo-D for durability, made at different temperatures. Searches of GitHub, Zenodo and Hugging Face in November 2026 found no open tool or image set for checking heat-treated wood from photos.
Built with what exists
A phone camera; a printed color reference card; a borrowed benchtop colorimeter for reference readings; offcuts from a thermal-modification plant at several temperatures; Python with OpenCV for color correction and scikit-learn for a first model, or a small open-weight vision model fine-tuned on one graphics card (see the Note "One GPU to fine-tune").
First step
Ask one ThermoWood producer for offcuts of each grade in pine and spruce, and photograph them with a color card under three kinds of light.
Done looks like
On at least 200 boards from at least five producers, none used in training, the app's grade (untreated, Thermo-S or Thermo-D) matches the producer's documented grade at least 90% of the time, with the photos and results published.