Thought I might try to clarify a few things on the post I wrote.
For those who think all of this is just "spielerei":
Knowing why something works can help us to optimize it.
What's called heat-treating when lumber is concerned, and when bows are concerned is obviously different: the former affects the entire piece of wood or lumber, and typically takes more than two hours at high temperatures (note that I was talking about 200-260 degrees Celcius, not Fahrenheit), whereas the latter just affects the belly and takes maybe fifteen minutes and with a more superficial effect. So the structural degradation when heat-treating bows is probably limited. We also pay attention not to heat the back, as that would affect its tension properties when combustion kicks in. (it's english heat-treatment: toasting on one side only).
But essentially, for the surface of the belly wood, the chemical transformation is very similar: it's (the start of) a combustion process caused by high temperatures. This was already discussed once before on this forum:
http://www.primitivearcher.com/smf/index.php/topic,35327.msg465115.htmlThe longer you expose the wood, the more it is combusted and the lighter in weight and darker in color it becomes. Logically, it also loses its strength, because you are burning wood in a very controlled way. This is not the case for kiln-drying, or steaming or boiling or ammonia-bending: molecular structure remains largely the same as there is no combustion of wood molecules. (note that in kiln-drying modest modifications probably take place, also affecting hygroscopicity).
The very first molecules to start combusting (at the lowest temperatures), are the hemicelluloses, of which the hydroxyl groups (which determine the affinity of these molecules with water) are split off first. “In thermal wood modification … the reduction of accessible OH-groups (hydroxyl) leads to a limited interaction with water compared to untreated wood”
http://www.thermotreatedwood.com/researches/3-5-1.pdfWhen the belly wood starts to turn brown and emits a toasted whiff, that's when we typically move on to the next section. This actually means we’re already breaking down the hemicellulose more than by just removing the hydroxyl groups. The smell is caused by the pyrolysis (combustion) of sugar units split off the hemicellulose molecule by heat, a process known as caramelization, also causing the brown color. If we don’t stop, carbonification sets in, or charring.
On the belly, this start of the combustion process reduces hygroscopicity (affinity for water), thereby reducing the equilibrium moisture content of the belly wood, but not the back wood, as we carefully avoid heating the back. The following two figures should clarify what that does to bow properties:
the first graph shows how tension parallel to grain (A ) changes with moisture content, peaking at 12%. The compression parallel to the grain (C) is highest when the wood is driest. But below 5% MC, any bow is dangerously close to breaking in tension (with spectacular bow explosions).
The second graph shows the equilibrium MC of Scots pine with and without heat treatment (or chemical treatment that also removes hydroxyl groups from hemicellulose in cell walls). Untreated wood has a much higher MC than heated wood at the same relative humidity. (figure from
http://www.thermotreatedwood.com/researches/3-5-1.pdf)
Back to the first graph: a heat treated bow in typical humidity conditions has the compression properties of the dot in red on line C, and the tension properties of the green dot on line A: Heat treating allows the bow back to have a MC of 12%, and the belly a MC of 5%.
“The main advantages of heat-treated wood are a reduction in hygroscopicity and an improved dimensional stability, generally believed to be due to the degradation of hygroscopic hemicelluloses”(
http://www.metla.fi/dissertationes/df134.pdf)
Marc St-Louis first observed that the belly shrinks during heat treatment. Any wood losing moisture shrinks (see for example
http://www.wolfgangbrinck.com/boats/woodlore/shrinkagewp.jpg), but heat-treated wood regains less moisture than it lost (because it has lost its appetite for moisture during the process), so it experiences a net shrinkage. It seems like the density has increased. In some way, that’s right: the wood has shrunk, so for the same relative humidity the volume is smaller and there’s still (nearly) the same amount of dry matter of wood. But the shrinkage, so I think, could maybe be ascribed entirely to a loss of moisture.
