GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Audio Play Icon Headphones Icon Plus Icon Minus Icon Check Icon Print Icon Picture icon Single Arrow Icon Double Arrow Icon Hamburger Icon TV Icon Close Icon Sorted Hamburger/Search Icon
Q&A Spotlight

Does This Roof Need a Vapor Retarder?

The open-cell spray foam in the rafter bays provides an air barrier but not a vapor retarder

How will this roof perform? The roof of this Boston, Massachusetts, area home is insulated with open-cell polyurethane foam, prompting the homeowner to wonder whether an interior vapor retarder would be a good idea.
Image Credit: David Amenhauser

David Amenhauser is buying a home near Boston, Massachusetts, that’s apparently still under construction, but far enough along to have the roof framed and insulated.

“When I decided to purchase the home, it was already framed and the roof deck/rafter bays were sprayed with open-cell spray foam,” Amenhauser writes in a post at Q&A post at GreenBuildingAdvisor. “I do not know what was used for the roof underlayment, but I know there are asphalt shingles. The rafter bays are made out of 2x12s and are completely filled with foam.”

Because the roof is insulated with open-cell foam, which is vapor-permeable, Amenhauser had been planning to ask the builder to prime the ceiling drywall with a vapor-retarding primer/sealer. Then, he came across an article by building scientist Joseph Lstiburek entitled “Cool Hand Luke Meets Attics.”

Now, Amenhauser isn’t so sure he needs a vapor retarder after all. The article implies that may not be necessary as long as the attic has supply and return air ducts. Amenhauser, however, is still concerned about the area behind the kneewalls where the HVAC equipment and ducts will be housed.

“So my question is,” he writes, “what do the experts here recommend to make sure I don’t have moisture problems in the attic?”

For starters, there’s not enough insulation up there

The photo of the attic that Amenhauser posted along with his question suggests to Dana Dorsett that the roof is under-insulated and probably doesn’t meet code requirements for this Climate Zone 5 location.

“The pictures don’t look anything like a complete fill with foam,” Dorsett writes. “It looks (being generous) closer to an ~8-10-inch average depth which, by the way, does not meet code minimum in Massachusetts, with some spots as thin as 6 inches.”

Even if the rafter bays were completely filled with foam, a total of 11 1/4 inches for a sawn 2×12, the insulation would fall short of the code-required R-49. And, Dorsett adds, when the thermal bridging of the rafters is factored in, the situation looks worse.

“To meet code on a U-factor basis it has to measure out with a whole-assembly average of U-0.026 or less, which is essentially uniform R-38 without the thermal bridging of the rafters,” he adds. “At the 8 to 10-inch average depth, you’d be lucky if it performed at R-38 average even at center-cavity, let alone after calculating the thermal bridging.”

Without an additional layer of rigid foam, at between R-20 and R-24, Dorsett says, there is a risk of moisture accumulation during winter.

Dorsett’s suggested fix is to bring the insulation contractor back to completely fill the rafter bays, then add a layer of MemBrain, a CertainTeed product, as an air barrier, and install 1 1/2 inch of unfaced expanded polystyrene insulation (EPS) to the underside of the rafters. That could be followed by a gypsum drywall layer, bringing the assembly to a center-cavity R-value of about R-45.

There’s another problem here, too

In addition to too little insulation, adds GBA senior editor Martin Holladay, the kneewalls appear to have been framed and the HVAC equipment installed before the drywall and required vapor retarder went in.

“This violates the installation requirements of most brands of open-cell spray foam, including Icynene,” Holladay writes. “Installing the required drywall (or, better yet, rigid foam) now will be awkward, and may require the kneewalls and ductwork to be temporarily removed and later reinstalled.”

Is the insulation contractor stretching the truth?

Amenhuaser has had a chat with the insulation contractor, who tells him that because the town has adopted something called the “Stretch Code,” the attic R-value only has to be R-38.

“This doesn’t quite seem right to me,” he adds. “The foam they used is R-3.9 [per inch]. He claimed that the foam is 10 inches or more in each bay and therefore the [required] R-value was met. He also indicated that gypsum with wallboard would cover the entire roof deck even behind the partition walls and that the gpysum would be primed and painted with a vapor retarding sealer/paint.”

Holladay quoted from the official explanation of the Stretch Code: “The stretch code appendix offers a streamlined and cost-effective route to achieving approximately 20 percent better energy efficiency in new residential and commercial buildings than is required by the base energy code. This is largely achieved by moving to a performance-based code, where developers are required to design buildings so as to reduce energy use by a given percentage below base code, rather than being required to install specific efficiency measures.”

He suggests that Amenhuaser ask the contractor whether he has balanced the skimpy attic insulation with other features in the house that would make up the difference.

But he adds this: “If the open-cell spray foam is upgraded to R-38, and if the contractor is really able to install drywall on the underside of the rafters — a feat that will be difficult behind the kneewalls, where the ducts are — and if vapor-retarder paint is applied to the drywall, you won’t have any moisture problems.”

The drywall must be completely airtight

Dorsett doesn’t agree. “For a vapor retarder to protect the roof deck it has to be completely airtight,” Dorsett says. “That’s damned hard to achieve with an airtight gypsum approach given where you’re starting, and if the rafter bays are not completely filled you have large convection channels to amplify any air leakage.

“Air-transported moisture is a huge risk here,” he adds, “even if you installed a broad sheet membrane vapor retarder like Intello or MemBrain. “

Holladay thinks the amount of open-cell foam in the roof will be an adequate air barrier even if the drywall is not airtight.

But Dorsett says he would be concerned with any gaps between the back of the ceiling drywall and the foam insulation. A gap of 2-3 inches is a “convection channel bigger than is required for a vented roof,” he says, but in this case it would be on the wrong side of the insulation.

“Dana is wrong for several reasons,” Holladay replies. He makes these points:

  • The drywall layer does not have to be airtight in order for it to be an effective vapor retarder. A vapor retarder addresses vapor diffusion, not air leakage. The installed open-cell foam is a very good air barrier.
  • Convection currents don’t matter. Air in the attic is conditioned, separated from the outside by an air and thermal barrier. “Sure, you’ll get heat transfer through the roof assembly — and that heat transfer rate is a function of the insulation’s R-value,” he says. “No air leakage, though, so convection doesn’t matter.” With drywall and vapor retarder paint applied, no much air is moving behind the drywall because there aren’t any leaks to the exterior.

Our expert’s opinion

Who’s right? Here’s how GBA technical director Peter Yost sees it:

I see two main issues with this project: First, the need for a vapor retarder. “Need” is a funny term; does the attic “functionally” require a vapor retarder, or does it “need” one to meet local code or the local building inspector? Just how much vapor, moving by diffusion, gets into the open-cell spray foam past the painted gypsum wallboard? (Let’s say that this gypsum wallboard layer ends up being a Class III vapor retarder once you include regular primer and two coats of acrylic paint.)

The amount of vapor that gets past the gypsum wallboard (GWB) depends in large part on the interior relative humidity in the conditioned attic space during the winter. My bet is this open-cell spray foam, covered by the painted GWB, will never see a moisture problem due to the lack of a Class II dedicated vapor retarder (that is, vapor-retarder paint or MemBrain).

And so what happens if a vapor-retarder paint is used as a Class II vapor retarder; any problems with that? Well, with nearly vapor-impermeable asphalt roofing shingles on the topside of this assembly, should the assembly get wet, it will need to dry mainly to the interior. Is it better to more effectively promote drying to the interior or more effectively restrict the movement of wintertime moisture into the spray foam? Since I have never seen significant wetting and moisture problems in this type of assembly without that Class II retarder, I would go with the greater drying potential.

[Editor’s note: Joe Lstiburek and Armin Rudd have measured the moisture content of roof sheathing behind open-cell spray foam installed without an interior vapor retarder in Climate Zone 5 and found that in some cases the moisture content is high enough to be worrisome. GBA readers should make their own judgments on the need for vapor retarder paint in this location. Joe Lstiburek advises builders to include vapor-retarder paint on gypsum wallboard under roof assemblies insulated with open-cell spray foam in Climates Zones 5, 6, 7, and 8. When Peter Yost was asked a follow-up question about his recommendations, he wrote, “I’m sticking to my guns after giving this more thought and doing a bunch of WUFI runs. I know that WUFI is only part of understanding the situation and I am also not that sure that if the three of us were discussing this particular assembly we would disagree, but to me, while the interior coatings do affect how much wintertime moisture gets into the assembly and at the roof sheathing, even with no coatings and oriented north, the assembly dries out each year — and by the way, whether the roof is oriented north or south — maybe no surprise — makes a huge difference. And with no drying potential to the exterior, should this assembly get wet from a leak, you will need every ounce of that interior drying potential to pull the plywood sheathing moisture content down to acceptable levels, and even then, that 10 to 11 inches deep of open cell spray foam really slows it down.”]

And incidentally, citing Joe Lstiburek’s “Cool Hand Luke Meets Attics” for Amenhauser’s project is not a real fit. This article is about attics that are insulated at the roof line but not conditioned or occupied space. Amenhauser’s attic is living space so it will be fully conditioned (except for the HVAC closet) and finished with gypsum wallboard on the interior. And even in the HVAC closet, which is unlikely to be supplied and returned, this space is very unlikely to see the accumulation of more buoyant moisture-laden air.

Second, let’s address the need for an air barrier. I agree with Martin on this one. The open-cell spray foam functions as the air barrier and in order to get significant air leakage in any space between the GWB and spray foam you need a driving force and there isn’t one. And remember, vapor movement by diffusion, a field effect, is a function of the overall percentage of surface coverage as compared to air leakage, a point effect, where lots of air and moisture can be forced through relatively small imperfections.

However, notice in the photo that the attic walls are not insulated. This is curious. I decided to contact Amenhauser about this he said the walls will be insulated with fiberglass batts. He added that the insulation contactor assures him that the drywall contractor will be able to install the GWB behind the HVAC equipment in the framed closet. I am afraid this information means three things:

  • Amenhauser needs to ask the drywall contractor just how he or she intends to install the GWB in the framed HVAC closet, not trust the intuition of the insulation contractor.
  • We don’t know what the wall air barrier will be, but if it is going to be the GWB, then there is no way that getting the GWB installed airtight in the HVAC closet is going to happen without taking out the framing and the ducts, at a minimum.
  • Amenhauser needs to make sure that his builder and trades understand the importance of air barrier continuity; its lack is much more of a moisture concern than the particulars of his vapor retarder. And the construction sequence in the attic space is not exactly inspirational in this regard.


  1. Kohta Ueno | | #1

    Vapor Retarders and Open Cell Foam
    Just to give folks a reference point--a key point of the IRC section on unvented roof is:

    In climate zones 5 through 8, the “air impermeable insulation” must be a vapor retarder (Class II; 1.0 perm dry cup or less), or have a vapor retarder coating or covering in direct contact with the underside of the insulation. For instance, an air impermeable but vapor permeable spray foam (0.5 pounds/cubic foot) would not meet this requirement, unless a vapor retarder coating were applied (R806.4.4).

    IRC FAQ: Conditioned Attics

    That being said: are there probably plenty of roofs in Climate Zone 5 without any vapor control that are operating just fine? Yes, that is very likely true--there are not endemic failures of that assembly. But there are enough risks that this code provision is on the more conservative side.

    For an excellent case study of what can go wrong with open cell foam in Zone 6 (near Toronto), and bypassing of a sheet good vapor barrier (polyethylene), see this presentation by John Straube:

    2015-03-24 LAB Event SPF SprayFoam

    He stated that there were not exceptional humidity loads in that house. However, there was... wait for it... a white roof membrane. This reduced how much the roof was "baked down" during the summer--thus, the OSB sheathing that you can step through.

  2. moopy62 | | #2

    Wish I could find someone to help. We are looking at reshingling a cathedral ceiling but incthe 70’s using only 2x6 rafters. It is poorly sealed on the interior as ceiling is old t&g barneood though i believe palced against 1/4” ply. Not enough money to do this right so the plan is to reinsulate with rockwool during re-shingle. I know sealing air gaps is 1st priority but do I use a vapor retarded before putting incthe rockwool then baffle then new decking. Spent many hours trying to find the answer but my problem seems very specific.

  3. User avater GBA Editor
    Martin Holladay | | #3

    1. No one can help you with specific advice unless you share your geographical location or climate zone. Here is a link to the climate zone map: Climate Zone map

    2. "Not enough money do to this right" is an unpromising way to start. Bad details can cause roof rot, especially in a cold climate. If you need a bank loan, you need to get a bank loan -- that's preferable to rotting your roof.

    3. Here is a link to an article that explains your options: "How to Build an Insulated Cathedral Ceiling."

    1. moopy62 | | #4

      Thanks. Im in zone 4a. I did read the article (and many others) I read enough to know I cannot afford to do it right. I did get several qoutes to sister the rafters to essentially raise the roof and create enough insulation space. $20k- just cannot do that. The cost to just reinsulate from above about $8k.
      Anyway I would have the same question regarding using a smart vapor retarder between the interior ceiling and the rock wool and after using (can foam?, etc) to fill any air gaps

      Suggestions are very welcome

  4. User avater GBA Editor
    Martin Holladay | | #5

    A smart vapor retarder can be installed on the interior if you want, but a smart vapor retarder is only one component of a functioning insulation system. You will also need soffit vents, ventilation baffles that create a continuous air gap between the top of the insulation layer and the roof sheathing (from soffit to ridge), a ridge vent, and an interior air barrier -- just to name some of the necessary components of a safe system.

Log in or create an account to post a comment.


Recent Questions and Replies

  • |
  • |
  • |
  • |