Underlayment for an Unvented Roof
My questions are related to part of the reply Mr. Holladay made to the “Proper Underlayment for Unvented Asphalt Roof” question posted on 11-14-14.
The response noted the manufacturer’s requirement to use a vapor permeable barrier on sheathing for unvented roofs – in that case with open cell foam under the sheathing. I thought a vapor permeable barrier could put the sheathing at risk for moisture adsorption during cold weather from either side. Even if asphalt shingles are indeed a low permeance barrier, will the potential vapor driven moisture passed by either the underlayment or the open cell foam build up in the sheathing and be trapped far too long to ever escape during the seasonal drying period? Shouldn’t both sides be impermeable to prevent this?
I am very concerned given my current roof plans in zone 6, dry alpine, 7500HDD.
The unvented roof profile planned is truss or frame, sheathing, cover with impermeable underlayment, then 3″ Polyiso, 6″ EPS nailbase, HT impermeable water shield, possible furring sleepers, metal roofing. The overall design intent (like REMOTE) is for the entire structure to be secured inside a WRB with out-sulation providing 2/3 of the total wall/roof values. Unfortunately, by code in my building location, I am required to use the equivalent of high temp ice and water shield from eave to ridge under metal (for insurance in fire area) roofing.
I have considered Grace Triflex and equivalents for the primary underlayment to save costs, but the lack of a seal at penetrations by the numerous long screws for the foam build up, leaves me queasy. Many of the discussions about vented and unvented roofs with outsulation frequently do not even note the presence of a WRB under the insulation sheets. This would seem to invite moisture transport from inside the house to the inevitable voids in the sheathing and foam stack up. Very few, if any, of the many offered diagrams for unvented roofs, here and elsewhere, make specific statements of WRB vapor permeance.
I have searched the extensive range of related prior discussions again and I must say I am now re-confused about the permeable vs. impermeable question. The REMOTE manual itself mentions use of bituthene sheet, poly sheet and Tyvek with generally okay results on walls, but I am unaware of double WRB layers occurring on roofs beyond the inherent properties of the foam stacks themselves.
Are all the superinsulated roofs going to get soggy? Will the foam get soggy? Is air likely to drive outward from living space into the top side foam?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Roger,
Using a vapor-impermeable roofing underlayment in your proposed roof assembly is contrary to the manufacturer's installation instructions, but is unlikely to cause any problems. Your roof assembly is not designed to dry to the exterior, so the vapor permeance of the roofing underlayment is irrelevant.
I think that the likelihood that your local code inspector will realize that you are violating the manufacturer's installation instructions is quite low. This is a low-risk crime, my friend. I think you can sleep well.
Why do manufacturers of vapor-impermeable roofing underlayment forbid the use of their products on unvented roof assemblies? Well, some unvented roof assemblies have damp-sheathing problems. Some types of roofing (for example, cedar shingles, concrete tiles, or slate) allow upward drying; if a roof assembly has one of these types of roofing, a vapor-permeable roofing underlayment makes sense. That's reason enough for the lawyers employed by the roofing underlayment manufacturers to be nervous.
Mr. Holladay,
Thank you for the assurances, but my inspector is both very sharp and very supportive of my attempt to exceed code. The primary concern I have is the longevity of the nailbase OSB if covered with the required High Temp water shield and then bound on the interior side by another impermeable layer. Six inches of EPS is very low on the perm ratings, but the gaps between sheets on the two layers will be very "vapor open". Weather conditions and schedules for the foam require we put on the primary roof sheathing, then a WRB and then build up the foam layers when it arrives and seal that with the HT water shield. Should I be concerned about choosing the primary WRB as either a vapor open or vapor closed type? Should I be concerned about the holes that will occur in a non-sealing underlayment when fastening the foam/nailbase down?
To guide my thinking, I have tried to extend the reasoning of Mr. Lstiburek's comments on golf shoes and vapor barriers under slabs. The holes though great in number represent a tiny percentage of area, so don't sweat it. Which would probably be true when speaking of impermeable underlayments like Grace Triplex. I am less certain about vapor permeable underlayments like Deck Armor. Would vapor diffusion through the primary sheathing and the Deck Armor succeed in migrating a significant amount of moisture into the foam layers?
It would seem that cold weather conditions might create a gradient from inside to just under the metal roof that favors moisture migrating to the OSB layer, never to escape. Or it may be that the behavior of moisture by diffusion is limited by the intermediary materials ability to transport moisture.
EPS and Polyiso may not have good ability to move moisture by diffusion, but I am still left with the issue of imperfect edge to edge contact between sheets of foam providing a speedier pathway to the nailbase layer. As I noted before, many of the over-insulated roof diagrams show no WRB at the primary sheathing level.
So do I have a tempest in a teapot, or do I have a slow invidious rot problem?
Roger,
The traditional way to detail a PERSIST roof is to install a vapor-impermeable air barrier on the bottom (interior side) of the assembly. In other words, you first install a peel-and-stick layer like Grace Ice and Water Shield on top of your roof sheathing. That prevents any interior moisture from migrating outward.
The next step is to install your insulation layers -- in your case, 3 inches of polyiso topped with 6-inch nailbase.
On top of that, you would install a layer of roofing underlayment. If I were building the roof assembly, I would choose a vapor-permeable roofing underlayment at this location -- something like asphalt felt -- but I don't really think it matters very much what you install here, as long as the work is performed on a sunny, dry day. Clearly, you don't want to cover the OSB layer of your nailbase with a vapor-impermeable roofing underlayment if you are working in damp weather.
For a wonderful roof assembly that will last for a very, very long time, the next step is to install 2x4s on the flat, either 16 inches o.c. or 24 inches o.c., from soffit to ridge, in order to create ventilation channels. Then you install one more layer of sheathing (plywood or OSB), followed by one more layer of roofing underlayment (either vapor-permeable or vapor-impermeable -- it doesn't matter), followed by your roofing.
That's not a cheap roof, but it will perform very well.
Martin, your sense of wit of late adds a bit of joy. You're moving closer to "the dark side." We should get that solar system up and running for your bro someday. Have high pressure pump will travel.
Aj-bitter nemesis of all that live life risk.... Averse.
Sorry for restarting a seven year old thread.
I'm building a house next spring with vaulted ceilings and unvented roof assembly. I plan for 3" of CC spray foam on the underside of the roof deck with batt filling the rest of the joist cavity. Roof material will be asphalt shingles.
My question is regarding shingle underlayment. Martin stated that you can't use underlayment that isn't vapor permeable, but in the Fine Homebuilding article "A Crash Course in Roof Venting" by Joseph Lstiburek, he states "Asphalt shingles require special attention when installed on unvented roof assemblies in hot-humid, mixed-humid, and marine climates due to inward vapor drive. To keep moisture out of the roof assembly, a roofing underlayment with 1 perm or less (class-II vapor retarder) must be installed under the shingles."
I am in coastal Virginia.
Is the difference here that Martin is talking about a situation with open cell foam and Joseph is talking about a situation with closed cell foam?
Open to recommendations. I planned to use Zip sheathing for the roof deck but it's 12 perm, so if I need less than one perm I will just use regular decking a apply a < 1 perm underlayment.
Thanks in advance.
I'd read the later "Inward drive is not an issue" and consult your local code.
Zip certainly isn't 12 perms - probably around 1 perm.
Thanks for the link, good to see some updated info. The zip panel as a whole is around 1 perm but the protective layer is over 12 perm, and protecting the deck itself is what I'm thinking about. Thanks again.
A small point - Zip doesn't have much R value. Which means that it's about the same temperature as the exterior. So water vapor from the exterior isn't a concern for the Zip itself. The concern is when exterior water vapor makes it all the way to the cool interior side - where it may condense/sorb. In your case, very little will make it to the interior side and what does should pass through into the interior.
I design with SIP roofs which I would prefer to dry to the outside in Oregon and Washington both sides of the mountains and at the coast.
I have been recommending a vapor permeable underlayment for years. With metal and asphalt shingles, it probably doesn't help much. I have not met the client that has an extra $20,000 to $35,000 to build a true cold roof. I think Benjamin Obdyke cedar breather might be a good solution. At $100 per square it is not inexpensive, for the material. The labor would be a lot less than plywood and stickers.