the field · effusivity

What a seat costs you

A steel bench and a pine bench in the same room are the same temperature. They do not feel the same, and the difference is not in your head — it is thermal effusivity, b = √(kρc), which sets how fast a surface pulls heat out of skin. The interface settles at the effusivity-weighted mean of the two bodies. Steel and pine land about 9 °C apart at the skin.

conditions

The palette

Four axes, and they disagree. Feel is effusivity. Chemistry is what it off-gasses. Shedding is whether it becomes microplastic as it wears. Vapour is whether moisture can leave. Foam used to hold the first column outright; kapok now sits level with it, within the spread of published figures for a fibre that varies with how it is packed. So the case is not that plants feel warmer — it is that one of them matches the best synthetic there is and carries none of the other three costs.

sort
materialrolebat skinchemistrysheddingvapour
kapok fibre
silk-cotton tree floss — the lightest loft here, and the lightest thing here by a clear margin
loft3432.6°cleannonehygroscopic
PU foamexcluded
best effusivity on the board; friable with age, sheds continuously
avoid3532.6°petroleum + additivessheds continuouslysealed
flax / linen batt
lightest loft layer here
loft4732.5°cleannonehygroscopic
cellulose fill
recycled; borate treated for fire and pests
loft5732.4°cleannonehygroscopic
hemp fibre batt
the workhorse loft layer
loft5732.4°cleannonehygroscopic
coconut coir
springy — but check the latex: natural fine, SBR is plastic
only with: natural Hevea latex, not SBR — SBR is styrene-butadiene, petroleum
loft7532.2°binder-dependentcheck binderbreathes
straw
structural at building scale
town8832.1°cleannonebreathes
mycelium composite
grown to shape on agricultural waste, compostable, level with cork on feel — the one material here that is farmed into its final form rather than cut to it
only with: a lignocellulosic substrate — straw, hemp hurd, sawdust; some growers bind with synthetic scrim
surface10232.0°binder-dependentnonebreathes
cork
closed-cell — keeps its effusivity wet, where soil fails
surface10731.9°cleannonesemi-closed
hempcrete
walls, not seats — hemp shiv, but the binder is lime, which is mineral
only with: a plant binder if one exists at this scale; lime is mineral, and honestly it usually is lime
town21731.0°binder-dependentnonebreathes
hemp fibreboard
binder decides it; MDI binder is polyurethane
only with: lignin or soy binder — MDI binder is polyurethane, and dairy binders are not vegan
panel28830.5°binder-dependentcheck bindersemi-closed
pine (soft)
frames
frame31030.3°cleannonebreathes
bamboo culm
whole or split cane — no adhesive anywhere in it, so nothing to qualify. The conditional flag was never about bamboo, it was about the glue in the laminate
frame43629.5°cleannonebreathes
oak / ash
arms and legs — thermally mediocre, fine for structure
frame43629.5°cleannonebreathes
bamboo laminate
strips pressed into board — the standard binder is urea-formaldehyde, and that is the only thing wrong with it
only with: bamboo lignin, or soy/starch — bamboo can bind itself; the standard is urea-formaldehyde
panel49329.2°binder-dependentcheck bindersemi-closed
dry soil
same effusivity as a plastic chair, and never stops drawing
avoid60028.7°cleannonebreathes
polypropyleneexcluded
the stacking chair; abrades
avoid61328.6°petroleumabradessealed
concrete
allowed by the rule and condemned by the numbers — an infinite sink that never saturates. Try it in the calculator and read the verdict
surface168325.4°cleannonesemi-closed
wet soil
worse than the plastic chair; water fills the pores
avoid212124.7°cleannonesemi-closed
steel
allowed by the rule and the worst thing here to touch — burns in summer, freezes in winter. Fine as structure with a thermal break; never as a seat face
frame1353921.1°cleannonesealed

Warm is a failure too

The first version of this page optimised for losing as little heat as possible, which makes zero loss look ideal. It is not. A body at rest produces roughly 100 W and has to put it somewhere. A seat that fills up and then accepts nothing traps that heat against you — and if it is a sealed cell structure, the moisture cannot leave either. That is the whole memory-foam complaint, and an effusivity-only model rates foam best.

So the target is a window rather than a minimum: land somewhere around 29–32 °C at the skin, and keep a vapour path open. This is where the plant fibres earn their place — flax, hemp and kapok are hygroscopic, taking on water vapour without feeling wet and giving it back. Cork is closed-cell, which is why it holds effusivity in rain, and it is also why cork wants to be a thin wear layer over a breathing loft rather than the bulk of a seat.

Saturation — the number that decides it

Effusivity governs the first seconds. Over a real sit what matters is whether the thing can fill up. A thin shell reaches your temperature and then draws nothing; ground is an infinite sink and never does. At these settings, over one hour, dry ground pulls 528 kJ/m² and wet ground 1867 kJ/m² — against a stack below that stops at its capacity and then costs nothing.

build a stack

at the skin
31.9 °C
effusivity
107
it can absorb
70 kJ/m²
then stops after
36 min
verdict
in the window

Compare: dry ground 528 kJ/m² and still going, an hour in.

Outdoors: the span is the argument

Run the same formula with the surface as the hot or cold body and it explains why children burn themselves on playground equipment and why nobody sits on a metal bench in February. Push the surface slider above to 65 °C for a bench in full sun, or to −10 °C for a winter street, and watch the table.

surfacein sun, 65°winter, −10°span
cork36° 29° 6°
kapok fibre34° 32° 2°
pine (soft)40° 24° 15°
oak / ash42° 21° 20°
polypropylene44° too hot18° 25°
concrete52° burn8° painful44°
steel62° burn-7° freeze69°

The span column is the whole case for building a town this way. Across a 75 °C swing in the weather, cork moves about 6 °C at the skin and steel moves about 69 °C. A low-effusivity surface is nearly season-proof; a high-effusivity one just relays the weather into whoever touches it. That is one material choice buying you a bench that is usable in August and in February, with no mechanism and no maintenance.

The rule, and where it actually bites

Two things, neither of them about performance: nothing artificial, and nothing animal. Artificial means synthesised from petrochemicals — foam, polypropylene, PVA, MDI, polyester. It does not mean processed. Steel is smelted from ore and concrete is calcined limestone; both are worked hard and both are earth rather than invention.

Which means the rule and the numbers are separate judges, and that is the point. Steel and concrete are allowed here. They are also the two worst things on the board to sit on, and the calculator will say so if you pick them — a concrete bench never saturates and a steel one puts 62 °C on your skin in the sun and −7 °C on it in February. Rejecting them on the physics is a better argument than banning them, because it survives someone who does not share the rule.

The rule took five passes to name — no petroleum, then vegan, then plant based, then mycelium had to be in, then steel and concrete came back. The middle correction is the one worth keeping: plant excluded mycelium on taxonomy while every actual reason for the rule included it. The material was not the exception. The name was wrong. Both halves still do work: an animal material is not artificial, and foam is not animal. Each half catches something the other lets through.

Choosing cork over foam is the easy part. The rule bites in the parts nobody specifies: a chair whose cutting list is cork, hemp and ash is still a plastic chair if it is sewn with polyester, sprung on elastic webbing and sealed under polyurethane. Every line below is a place petroleum re-enters a build that looks entirely natural on paper.

partspecifywhat it would be otherwise
threadlinen, hemp or cottonpolyester thread is the default in every workshop
webbingjute or linen webbing, or wood slatselastic webbing is rubber over polypropylene
finishlinseed or tung oil, carnauba or candelilla waxpolyurethane varnish is plastic, and several traditional finishes are not vegan — check the tin
gluelignin (incl. bamboo), soy, starch, latexPVA and MDI are plastic, and most traditional wood glues are not vegan — read the label
fastenerwood pegs, or steel with a thermal breakno fastener passing through the contact stack
fillhemp, flax, kapok, coir, cellulosenever foam, never polyester wadding, and confirm the fill is plant
coverlinen, hemp, organic cotton canvaspoly-blend upholstery sheds with every sit; confirm the cloth is plant fibre

One thing the rule gave back, unexpectedly. Looking for the best plant loft turned up kapok — the floss from a silk-cotton seed pod — at about 34 against PU foam’s 35. Those are close enough to call level rather than a win, and kapok varies with packing, so the page will not claim more than parity. But the material that was unbeaten on feel is no longer unbeaten, and the thing that caught up is a seed pod.

Where the plant lofts are honestly weaker is resilience rather than feel. Flax, hemp and kapok are hygroscopic and sit low on effusivity, so a seat built on them lands in the window and stays there. What they do less well is recover loft after years of being sat on. The answer is a thicker layer than the numbers alone would suggest, and a cover that opens so the fill can be topped up rather than the whole seat replaced.

Four materials in the palette are conditional rather than admissible — natural things that are normally made with a petroleum binder. Each names the substitution that qualifies it. Coir is the sharpest example: the fibre is a coconut husk, and the latex binding it is usually styrene-butadiene.

Binders are not interchangeable

“Use a plant binder” is not a specification. Each one bonds through a different mechanism and therefore suits a different substrate: lignin and furfuryl alcohol crosslink into lignocellulose, soy and starch want a porous absorbent face, latex stays elastic and so belongs where the fibre must keep moving. Pairing them wrongly is how a plant-bound panel delaminates — and then the rule gets blamed for the failure.

binderbonds bysuitsnot for
bamboo / wood ligninthermoset crosslink into lignocellulosebamboo laminate, wood fibreboard, particleboardnothing porous-free — it needs a woody face
furfuryl alcoholin-situ polymerisation, from biomass pentosansbamboo, wood — also hardens and stabilises the substrate itselfthin veneers; it darkens and stiffens
soy proteinprotein denaturation and interlockingwood, paper, light boardanything that will stay damp — poor wet strength
starchgelatinised polysaccharide filmpaper, card, light laminatesstructural joints and outdoor use
natural latex (Hevea)stays elastic, grips fibre mechanicallycoir, hemp and flax batts — anything that must keep loftrigid panels; it never fully hardens
myceliumthe organism grows through the substratestraw, hemp hurd, sawdust — it binds by colonising themanything it cannot eat, and anything sterilised late
pine rosin / pitchthermoplastic, remeltsgap-filling, repairs, reversible jointsload-bearing joints and hot climates

A material that binds itself

The conditional flag on bamboo was never about bamboo. A culm is whole cane — split it, weave it, peg it, and there is no adhesive anywhere in the piece. Only the pressed laminate needs a binder, and the standard one is urea-formaldehyde. So the palette carries them separately: the culm is admissible outright, the laminate is conditional on its glue.

And the glue can be bamboo. Lignin is the polymer that stiffens every woody plant, recoverable from the same biomass, and lignin-based adhesives made for bamboo composites reach around 8.3 MPa dry and 5.8 MPa wet— formaldehyde-free, and stronger than most other bio-based adhesives tested against them. Of the lignins compared for panel work, bamboo’s gave the lowest formaldehyde emission of the set.

Which closes a loop the rest of the palette only gestures at. Every other conditional material needs something brought in from elsewhere to qualify — soy, starch, latex from a different plant. Bamboo can be bound with bamboo. The panel and its adhesive are one species.

The ground state chair

Perforated cork, 8–12 mm, over 30 mm hemp batt on an ash or oak frame. Lands near the top of the window at the skin, absorbs its capacity and then stops drawing — and, unlike foam, keeps a vapour path so the heat it stops absorbing can still leave. The hemp is laid thicker than the numbers alone would suggest, to make up in depth what a plant fibre gives away in resilience, and the cover opens so the loft can be refilled rather than replaced.

  • The frame must not bridge. Bolt that pan to steel or give it a continuous path to a cold floor and the frame becomes the infinite sink. Every gram of cork is then wasted — you have rebuilt the ground. Wooden cleats, no fastener passing through the stack.
  • No foam anywhere. Best on the board thermally, worst on the other three. It gets warm precisely because it is good at the first axis and sealed on the fourth, and as it ages it goes friable — the degradation dust is the documented route by which furniture flame retardants reach house dust, and from there skin and lungs.
  • Nothing animal, anywhere in the assembly. This is easy to get wrong at the last step, because several traditional finishes and most workshop wood glues are not vegan and nothing on the tin says so. Finish with linseed or tung oil and a carnauba or candelilla wax; glue with soy, starch, lignin or natural latex.
  • Perforate the wear layer. Cork closed cells are an advantage outdoors and a liability under a sitting body. Perforate it, or keep it to wear points and let the hemp take the field.
  • The plastics hide in the joinery, not the cushion. A cork-and-hemp chair sewn with polyester thread, on elastic webbing, under a polyurethane varnish, is a plastic chair with a cork lid. Linen thread, jute webbing, and a hardwax or plant-oil finish.
  • Outdoors, cork over wood. Closed cells keep their effusivity wet — which is exactly where soil fails, jumping from 600 to 2121 as water fills the pores.

What this is not

Contact thermal comfort is ordinary physics and everything above is calculable from published material properties. The step from this seat is colder to this seat harms you is not taken here, because the evidence does not support it. The one well-established thermal–cancer link runs the other way and is digestive: the IARC classifies drinking beverages above 65 °C as probably carcinogenic for oesophageal cancer — thermal injury from heat, not cold. Chronic cold, where it has been studied directly, suppressed tumour growth in mice by activating brown fat.

Contact of two semi-infinite solids: Carslaw & Jaeger, Conduction of Heat in Solids. Cold and tumour growth: Seki et al., Nature 608 (2022). Hot beverages: IARC Monographs vol. 116 (2016). Material figures are typical published values; mycelium and hempcrete vary widely with process, so trust the ordering rather than the third digit.