Unvented roof assemblies and roof deck vapor permeance
As a long time visitor and few years member, I have tried to absorb the lessons, opinions, science and conjecture presented here. However, I now find myself confronted with conflicting theory regarding unvented roof profiles.
For my own project, (high altitude, dry, -10 to 90 F) I have regarded the “top” layer of sheathing in an unvented assembly to be essentially in dead space – not exchanging air or moisture – the underside in contact or glued to impermeable insulation, the top side sealed off from all contact with the locally required ice and water shield over the entire roof.
Two current discussions present thinking I am now confused by.
and
I have already chimed in on the latter, unfortunately before reading the breathable underlayment one. Unfortunately, builders with years of experience seem to be in conflict in both discussions. The thoughts and theories I have derived from this site’s expert sources are perhaps incomplete at best.
So, how do I interpret the GBA details of unvented, stacked polyiso sandwiches with mixed ice/water shield and ordinary underlayment with no “venting” under both metal and asphalt roofs?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Roger,
For some types of roof assemblies, the vapor permeance of the roofing underlayment matters. For others, it doesn't.
Some types of roofing -- for example, cedar shingles, slate, and concrete tiles -- are vapor-permeable. If you choose to install vapor-permeable roofing, it usually makes sense to choose vapor-permeable roofing underlayment (for example, asphalt felt, which is a "smart" vapor retarder), so that the roof sheathing can dry to the exterior.
In other cases, the roofing itself isn't vapor-permeable, but outward drying is still possible. The example I'm thinking of is through-fastened metal roofing. This type of roofing has ribs that allow eave-to-ridge air flow, and this air flow helps damp roof sheathing to dry out. That's why vapor-permeable roofing underlayment makes sense for this type of roofing, just as it does for cedar shingles or concrete tiles.
A third category of roof assembly is one with a ventilation channel immediately above the roof sheathing or immediately below the roof sheathing. In this type of roof assembly, vapor-permeable roofing underlayment is more desireable than vapor-impermeable roofing underlayment, because the underalyment doesn't interfere with the redistribution of moisture and drying.
It's worth pointing out that most manufacturers of synthetic roofing underlayment (vapor-impermeable underlayment) forbid the installation of their products on unvented roofing assemblies.
A final category of roof assembly is an unvented roof assembly with no possibility of outward drying -- for example, and EPDM membrane roof that is unvented. In this type of roofing assembly, the vapor-permeance of the roofing underlayment is irrelevant.
Is any outward drying possible through asphalt shingles? This question has been debated for some time. The short answer is "possibly, but not much."
Roger,
The roof underlayment and drying concerns can drive one loony. The jury is still out on what will happen when you take a vapor impermeable membrane (Grace peel & stick) and place it on top of the roof sheathing and then have a vapor impermeable material (8"+ of closed cell) beneath the roof sheathing. Like you mentioned, it creates an unventilated assembly and the top layer of sheathing is basically "dead space" since it is not exchanging air or moisture.
In theory, as long as that sheathing remains dry and doesn't get vapor diffusion through it (which it shouldn't since the above and below layers have basically a 0 perm rating), that sheathing should last the lifetime of the building.
Here is my roof stackup:
8" Polyurethane SIP (closed cell) with OSB creating the sandwich
Seams taped and sealed with a flexible gasket
Peel & Stick roof membrane with a standing seam metal roof
The above in theory is vapor impermeable to the exterior with only a very slight vapor permeability to the interior. Will the assembly be dry and intact 50 years from now? Nobody really knows and I will be dead and gone by then so the answer will be found out by the other homeowners :)
Peter: At 8" most closed cell foam is still only a class-II vapor retarder, and not a true vapor barrier (but close, at about 0.15 perms.) If the roof deck is dry when the foam is applied it'll stay dry, and will have negligible seasonal moisture cycling. But if the roof deck got rained on prior to applying the membrane and wasn't given ample time to dry before installing the foam it's a problem- it can't dry very fast either.
OTOH at 8" thickness the foam itself is about as structural as half inch OSB (!).
Most of the time assemblies with as much as 8" of closed cell polyurethane is something of a crime against the planet, since it's nearly universally blown with HFC245fa, at a global warming potential ~1400x CO2, which makes me not such a fan of polyurethane SIPs, despite the fairly good thermal performance. Water blown 2lb foam may only run R5/inch, but it's long term environmental impact is a tiny fraction of typical goods. While low GWP blowing agents for closed cell polyurethane exist (HFO1234yf at about 4x CO2 GWP, under a couple of different trade names, eg: Honeywell Solstice, or DuPont FEA-1100), SFAIK none of the major spray foam vendors or SIP builders are using it.