Let me see if I can clarify it a bit for you.
Radiant heat transfer involves two surfaces. The 'hotter' one radiants
infrared energy, the 'cooler' one absorbs it. How well a particular surface
emits infrared energy when heated is measured by its emissivity.
So to reduce radiant heat transfer, we can coat the cold surface with
something that reflects infrared energy so it doesn't absorb as much. -OR-,
we could coat the hot surface with something that doesn't radiate/emit
infrared energy as well. Either one will reduce the amount of infrared
energy that gets from the 'hot' surface to the cold surface.
Now, it just so happens, that with very few exceptions, surfaces that are
poor at absorbing infrared are poor at emitting infrared. And surfaces that
are good at emitting infrared are also good at absorbing infrared. Polished
metal and metal foils are very poor at absorbing and emitting infrared. So
radiant barriers have a metalized/foil surface. There emissivity is quite
low (< 0.2, some as low as 0.05). Non metalic materials (wood, plaster,
glass wool, etc...) are good absorbers/emitters (emissivity > 0.8, often
So, in the case of under-floor radiant barriers, if we cover the 'hot'
surface with a material that is a poor absorber of infrared (and hence is a
poor emitter of infrared), we get about the same overall affect as if we had
covered the 'cold' surface with it. We could cover either one and get about
the same affect, at least in the short term.
But once the poor absorber/emitter is covered with dust, the heat can travel
from the foil to the dust by conduction (a very good transfer mechanism).
And household dust has a very high emissivity, so it absorbs/emitts infrared
quite well. So the dust layer completely circumvents any savings of the
radiant barrier. So we *really* want to keep the radiant barrier clean.
And by putting the poor absorber/emitter on the underside, we have it in a
position (facing downward) where dust and dirt are less likely to settle on
I've been think about that as far as my staple up radiant goes. It
looks to me that I want to cover most of the area (perhaps all) between
joists with flashing thickness aluminum to spread the heat out. That's a
lower operating temperature for the working surface. I don't think the
emissivity is as low as foil, but probably around .08. It seems to me
that part of the joist should also be covered in a radiant barrier.
It looks like you can gain a good bit of insulation value just from
having a dead air space with an IR opaque boundary.
I see from my "Passive Solar Energy Book" That a horizontal foil
surface with heatflow down has an R value of 4.55. That would seem to
imply face nailing 1" polyiso foil covered, with the foil facing down
onto the joists. Perhaps R17 total. Does that sound about right?