I just watched this video http://www.youtube.com/watch?v=2-kG5D-GSL0 which shows 4 material samples sat on boiling water jars for 30 mins, and three of them transmitting significant amounts of vapour.

The 4 materials were foamglass, PIR, perlite and rockwool. Only the foamglass didn't transmit significant vapour.

Now this surprised me a little as I understood PUR/PIR to be 'fairly' vapour resistant. Is that in fact wrong? I'd certainly explect it to be _much_ better than rockwool, but maybe the nature of this test is such that you can't tell the difference between 'a bit' and 'lots' of condensation?

Clearly the foil covering is the bit that guarantees vapour-tightness (and keeps in the insulative blowing gas), but for example we recommend filling edges round boards with PUR foam to stop air circulation and vapour getting behind, but that only works if it's reasonably vapour-tight or the foam is covered with plastic/foil/tape.

Perhaps it depends on temp?

Aha. This paper: http://www.nrc-cnrc.gc.ca/obj/irc/doc/pubs/nrcc30890/nrcc30890.pdf (hooray for the intertubes - what did we do without them?) shows that the transmissivity rises rapidly with temp above 21C for both PUR and PIR. It's 1.12 and 1.03 (x10-7 g/m2.s.Pa) respectively, almost doubling to 2.09 and 1.62 at 32C and getting up to 2.66 at 50C (for PIR - PUR not measured). The increase seems to be linear, sugesting it'll be around 6-8 at 100C.

They also found that water absoption in samples stopped after between 40 and 64 days - equilibrium was reached then, although with about 40% water content on the warm side, but not penetrating more than 3cm into the material.

So the 100C test is not really indicative of performance in a real building.

Whilst looking for other relevant data I came across this BRE report on materials properties, which collects data from many labs round the world and synthesises them to give data on thermal conductivty, density, specific heat capacity, vapour resistance, and absorbtivity and emissivity.

It also has some fascinating discussion (if you are a building physics spod), e.g. noting that the vapour transmissivty of plywood becomes non-linear above 60% RH, being 20 times as much at 95% RH as at low RH.
http://www.google.co.uk/url?sa=t&source=web&cd=8&sqi=2&ved=0CEkQFjAH&url=http%3A%2F%2Fwww.esru.strath.ac.uk%2FDocuments%2F89%2Fthermop_rep.pdf&rct=j&q=materials%20vapour%20transmissivity&ei=Ir5ATbXRGIqKhQfIhsXNCA&usg=AFQjCNFxvR-vr0IvX7KLknIWZ3S-fRsSLQ&cad=rja

This report gives vapour resitivty in MNs/gm and I don't know how to convert that to the above x10-7 g/m2.s.Pa. (I guess I could work it out if I sat here long enough, but it's bedtime :-)

Ah yes - sorry about the excessive link. One gets so used to them these days one sometimes forgets to check for needless wrappers.