GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

Exterior Roof Insulation and Interior Vapor Barrier

MALCOLM TAYLOR | Posted in General Questions on

In his blog on cathedral ceilings  Martin notes that assemblies with both exterior foam and permeable cavity insulation should not have an interior vapour-barrier so they can dry to the inside.

Presumably that’s because otherwise the rafter bays are sandwiched between two impermeable layers and have no drying path. But if the exterior foam is thick enough to prevent condensation, do the cavities need a drying path?

So my question is whether it is simply desirable or a necessity that the assembly be vapour-open to the interior? The context is an existing roof where the owner would rather not disturb the drywall or existing roof sheathing, and simply add foam to the top side.

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. Expert Member

    No takers?

  2. iwatson | | #2

    Isn't the idea to assume that water WILL find a way? In other words, at some point water will get into the assembly (leak around a roof penetration during a big storm?) so it needs a path out, even if we're not specifically worried about water from condensation.

  3. maine_tyler | | #3

    I don't know and am not qualified to say, but to spit ball a bit:
    It seems like it comes down partly to risk tolerance? I wonder if the assembly could be made 'safer' by using unfaced EPS (or mineral wool or wood fiber) for the exterior insulation, then adding a vent channel over the insulation. That would provide some outward drying not to mention roof drainage plane redundancy.

    Other than roof leaks introducing moisture (which seems to be a question of whether we should design for that worse-case scenario) the other thing I would worry about is convective currents bringing moisture via air into that 'sandwich' if the interior poly wasn't detailed very tight. The (air) tighter is is, it seems the less likely that would be a problem.

    Is the current roof vented, and would need the vent channel filled?

    1. rlangley | | #4

      Hi Maine-Tyler, and thanks Malcolm for raising this issue on my behalf...I missed your posting hence the slow response.
      The current roof from the inside out....drywall, poly vapor barrier, 2x6" rafter with R12 fiberglass insulation and 1" airgap, 1X10" rough sawn sheathing, felt paper, asphalt shingles. The new roof will be steel.
      In my earlier post a suggestion I received was to work from the top and remove sheathing/air gap/insulation, install R22 batts, (close off soffit and ridge venting) add min 4" polyiso, strapping to produce an air gap, plywood or OSB, roofing.
      Although Malcolm said in his intro that I did not want to remove the sheathing I am in fact prepared to do that. My concern in not doing that would be the air gap between the two insulations...not to mention the lost opportunity of replacing the thin insulation with a batt sized to fill the cavity. I don't want to remove the intricate, messy and disruptive project.
      Based on the suggestion above I asked if I could forego the airgap and plywood/OSB and instead strap the polyiso and install metal roofing (Galvalume) directly to the strapping. I am not clear what the benefit of the plywood/OSB would be in this instance. If the polyiso has the foil facing and is taped it will be water tight effectively negating the need for a roofing underlayment...I think.
      I also asked if I needed to cut out the poly vapor barrier since I would be installing thick enough Polyiso to be non-permeable, above the rafters. Malcolm has brought the second question forward but I received no recommendations in my original post to the first.
      With regard to the poly, I suspect the installation was not super tight. I believe the seams were taped but there are cross braces that would be very difficult to seal. The drywall is taped and painted and could be repainted with a vapor barrier paint if vapor diffusion is the concern.
      I hope this additional information helps clarify my situation. The end result that I am going for is an unvented, well insulated cathedral ceiling. My concern is the risk of rot/mold if this assembly is not designed properly....and the mix of information I have found so far makes me unsure of what the ideal design would be.
      Any help is appreciated.

      1. maine_tyler | | #5

        Do you have the link to that original discussion?

        I don't think there is an easy/obvious right answer for your situation.

        My suggestion of using slightly more permeable exterior insulation was just so that there was some increased drying. Foil faced iso obviously isn't that, but I dont know if it matters or not.

        I'm not a roofer, but if I had a stack up like yours with essentially no drying, I'm not sure I would be comfortable trusting taped iso seams as a roof wrb. I think I'd want a lapped membrane. But that's just a personal conservative mentality not backed up by anything.

        1. rlangley | | #7

          Hi maine_tyler
          Here is the original discussion
          I wonder if there is a benefit in terms of the foil reflecting heat back to the steel roofing and possibly reducing condensation....? probably not and if that is the case then I agree that the underlayment would make sense. It makes me uncomfortable when I can't find definitive answers to a construction issue that I can't be the only person to have run into. Thanks for your input on this. I see that Martin has commented so will read the link now. Cheers

  4. MartinHolladay | | #6

    Here's a link to my article that discussed a slightly different version of your question: "Sandwiching Roof Sheathing Between Two Impermeable Layers."

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |