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
Musings of an Energy Nerd

Sandwiching Roof Sheathing Between Two Impermeable Layers

Should we be scared of a “foam sandwich” with sheathing in the middle?

If you have installed closed-cell spray foam on the underside of your roof sheathing and vapor-impermeable roofing or Grace Ice & Water Shield on the exterior, then your roof sheathing can't dry inward, and it can't dry outward. Is that a problem?

Many builders have heard the phrase, “Walls should be able to dry in at least one direction.” The rule usually makes sense. If you want to install rigid foam on the exterior side of a wall — foam that will prevent outward drying — it’s usually a good idea if the wall can dry to the interior.

What about insulated roof assemblies? Should the same rule apply?

Ways that roof sheathing can dry

If a building has an insulated roof assembly — either a cathedral ceiling or an insulated low-slope roof — there are several ways that the roof sheathing may be able to dry.

1. In a conventional vented roof assembly — one with a ventilation channel between the underside of the roof sheathing and the top surface of the insulation — the roof sheathing can dry by giving up its moisture to ventilation air that enters at the soffit vents and leaves at the ridge vent.

2. In an unvented roof assembly with rigid foam above the roof sheathing, the roof sheathing can dry toward the interior by diffusion, as long as the insulation between the rafters is vapor-permeable, and as long as the ceiling materials are also vapor-permeable. Fiberglass batts, mineral wool batts, and cellulose are all vapor-permeable. Gypsum drywall is also vapor-permeable. This approach works well, as long as the builder remembers not to install any interior polyethylene. (Of course, in the illustration above, the second layer of roof sheathing — the outermost layer — can’t dry inward. But that’s a discussion for another article.)

[Image credit: ATAS International]3. Ordinarily, an unvented roof assembly with closed-cell spray foam installed on the underside of the roof sheathing (for example, a flash-and-batt assembly) does not allow the roof sheathing to dry inward — at least, not much. However, this type of roof assembly may be…

GBA Prime

This article is only available to GBA Prime Members

Sign up for a free trial and get instant access to this article as well as GBA’s complete library of premium articles and construction details.

Start Free Trial


  1. Expert Member
    ARMANDO COBO | | #1

    AAARRRGGGHHH!!! Frustrating...nice job on mix signals. So we've been telling people for years NOT to trap Plwd/OSB between two layers of impermeable or low-permeable materials, but holy and behold, now you, Joe and others say is OK? and later...
    1. "If you’ve got a choice, it’s probably a good idea to design your roof so that it can dry freely in at least one direction."
    2. "Eventually, of course, the roof will leak and get wet; that happens to all roofs. When the roof leaks, it’s time for a roof repair or new roofing."
    3. So if "We have lots of historic experience with commercial roofs that can’t dry out in either direction." we should be OK to follow same with homes?
    4. What about code R806.5.5? Does it need to be changed to reflect that you can't, BUT...
    5. Is anyone accounting for realities on jobsites vs. control lab experiments?

    1. GBA Editor
      Martin Holladay | | #3

      Like any other specification for a building, this takes judgment. Should you buy a siding product made out of sawdust and glue, pressed together to look like clapboard? I dunno. It's up to you.

      Should you sandwich roof sheathing between two impermeable layers? It's up to you. I think there are better ways to design an unvented roof -- for example, by using rigid foam above the roof sheathing, instead of closed-cell spray foam on the underside of the roof sheathing. I'm not a fan of spray foam. But if you love spray foam, I think it's possible to do it safely.

      Q. "So if 'We have lots of historic experience with commercial roofs that can’t dry out in either direction,' we should be OK to follow same with homes?"

      A. Yes.

      Q. "What about code R806.5.5?"

      A. What provisions of the code are you concerned about?

      Q. "Is anyone accounting for realities on jobsites vs. controlled lab experiments?"

      A. Can you be more specific? What are you worried about? Workers make mistakes every day of the week, so it's always best to specify simple assemblies that are easy to inspect. That doesn't mean that workers won't still make mistakes.

    2. bwjames | | #14

      Armando, I agree with your sentiment here. This is akin to somebody saying "this is how we've always done it" as a valid excuse for poor practice.

      Martin: When Joseph Lstiburek says "Commercial roofs", exactly what does that mean? I don't believe comparing a built up membrane roof to something like an asphalt shingle roof is fair or meaningful. I don't believe comparing a steel structured commercial building with a stick frame house is meaningful either.

      The long term performance to the former does not guarantee performance of the latter.

      1. GBA Editor
        Martin Holladay | | #16

        Q. "When Joseph Lstiburek says 'commercial roofs,' exactly what does that mean?"

        A. In this context, I assume he means something like "low-slope roof assemblies on large commercial buildings like retail establishments, warehouses, and factories."

  2. fourforhome | | #2

    What about a vented layer on top of the foamed-below sheathing?

    1. GBA Editor
      Martin Holladay | | #4

      Read my article again. I discussed that approach in the paragraph that begins with the number "3" in the section under the heading, "Ways that roof sheathing can dry."

      I wrote, "This type of roof assembly may be able to dry outward if one of two approaches is used: ... (b) above-sheathing ventilation channels are provided (either by installing 2x4s, 16 inches on center, perpendicular to the ridge, along with a second layer of sheathing, or by installing metal roofing on 1×4 or 2×4 purlins that are parallel to the ridge). If you choose approach (b), remember that the rate of outward drying will be limited by the permeance of the roof sheathing and the permeance of the roofing underlayment. Board sheathing is more vapor-permeable than OSB sheathing."

      I also discussed this approach in my 2011 article, "How to Build an Insulated Cathedral Ceiling." In that article, I wrote:

      "If you prefer, you can locate your ventilation channels on top of the roof sheathing rather than under the roof sheathing. If you decide to do this, make sure that any roofing underlayment that you install above the roof sheathing is vapor-permeable — for example, #15 asphalt felt, VaproShield SlopeShield, or Solitex Mento — and that your local building department accepts this approach to roof venting. If you install ventilation channels on top of a vapor-impermeable synthetic roofing underlayment, the flowing air won’t be able to help dry out the roof sheathing.

      "If you plan to install ventilation channels above your roof sheathing, it’s best to choose a roof sheathing that is vapor-permeable (for example, fiberboard). If you use plywood or OSB, there’s a small chance that the sheathing can still accumulate worrisome amounts of moisture over the winter; this is especially true for north-facing roofs.

      "You can create 1 1/2-inch-high ventilation channels above the roof sheathing with 2x4s installed on the flat, with the 2x4s located above the rafters, 16 inches or 24 inches on center. Although this approach is less fussy than installing vent baffles underneath the sheathing, it usually costs more, because most types of roofing require a second layer of plywood or OSB on top of the vent channels.

      "In some cases, these ventilation channels are installed above a layer or two of rigid foam. It’s also possible to purchase nailbase (a type of SIP with OSB on one side instead of two) that includes integrated ventilation channels between the OSB and the rigid foam; one brand of these panels is Cool-Vent from Hunter Panels.

      "If you are choosing to build a vented roof assembly, don’t forget to include soffit vents and ridge vents."

  3. Expert Member
    ARMANDO COBO | | #5

    Obviously all depends on installation quality with any product, but since we know that water leaks on the roof will happen eventually, why give options for builders to apply a roof system that sooner or later the owners are going to have to change after the warranty expire... costing them tens of thousands of dollars.
    Somebody is going to take today's blog and say, "hey, if we screwed up in the design, its ok, we'll fixed with a terrible suggestion".
    I going to start working with a new builder to design their new homes, but he took me to visit a jobsite designed by someone else, were the impermeable WRB and tile roof has been installed, and they were going to install 5.5" R50 ocSF under the roof decking. IMO, his options at this point are to install 11" R38 ocSF or 2" R13 ccSF + 7.5" R25 ocSF under the roof decking, he knows that both have risks. What a trap... it seems to follow today's blog.

    1. GBA Editor
      Martin Holladay | | #6

      If you believe strongly that roof sheathing should be able to dry in one direction, that's fine. This article lists several ways to do that -- and I'm sure that you are familiar with all of these methods. Nothing in my article prevents you from designing roofs that allow sheathing to dry in one direction.

      Whether you like it or not, lots of builders install closed-cell spray foam on the underside of the roof sheathing. While it's true that all roofing eventually fails, requiring a re-roofing job, I haven't seen any evidence that the presence of closed-cell spray foam changes the average cost of a re-roofing job. Roof leaks and the eventual need for re-roofing are facts of life, for any building.

  4. Expert Member

    "Builders who end up with roof sheathing that can’t dry inward are usually victims of their own poor planning."

    Like any other part of a building, consciously opting for something risky when designing an assembly just doesn't make sense. Especially with roofs, where the consequences of failure can be so catastrophic.

  5. PAUL KUENN | | #8

    Would you trust any roofer? I don't! In two months I've had to strip shingles and replace OSB on 5 homes. OSB had been placed as solid sheathing on top of 100+ year old boards. A few misplaced nails in vents and flashing is all it took to completely rot the sheathing because all 5 homes had closed cell foam sprayed on the interior. Obviously it never dried at the leaks. Two of the 5 had been done within 5 years, one at 10 years and the worst two were done 15 years ago. When I find my charge cord I can upload some great photos. None of these used the same roofer. Keeps me busy!

    1. GBA Editor
      Martin Holladay | | #9

      We all know that a stupid roofer who puts nails in the wrong place can cause roof leaks. That's not news. The real question is: Do you think that leaks cause more extensive sheathing rot before they are noticed when the roof assembly includes closed-cell spray foam?

      When I was a roofer, we saw lots of leaky roofs, many caused by stupid roofers, that had extensive sheathing rot -- and this was back in the day before spray foam was common. Roofers have been replacing rotten sheathing for hundreds of years.

      1. bwjames | | #15

        I really think this hits on the main concern I have with the idea of sheathing being sandwiched: "Do you think that leaks cause more extensive sheathing rot before they are noticed when the roof assembly includes closed-cell spray foam?"

        In my mind it comes down to the distinction between a catastrophic roof failure vs. a transient roof failure. If water is coming in each and every time it rains, and goes undetected for any length of time, I think the end result is mostly the same regardless.

        If water only trickles in under just the right conditions, whether that be rain fall rate, wind, or some other factor - which anybody who's ever spent time in attics know is a bit of a fact of life with most human built roofs - which assembly will be safer? My bet would be on the assembly that allows free-drying in some direction.

        But that said, is it possible that closed cell foam, while potentially accelerating sheathing rot may help prevent collateral damage as well? It wouldn't really surprise me either way I don't think.

        1. GBA Editor
          Martin Holladay | | #17

          Clearly, there are a lot of factors here. I agree that many owners -- myself included -- feel more comfortable if they can enter their attic and inspect the underside of the roof sheathing. Almost everyone agrees that this type of roof assembly -- a vented unconditioned attic with insulation on the attic floor -- is particularly forgiving of roof leaks.

          I still advise GBA readers to build this kind of roof assembly if possible.

          That said, architects continue to design buildings where the approach that I prefer is impossible. We get questions every week of the year from owners of buildings designed by the type of architect I'm talking about. And GBA needs to advise owners of these buildings on how they can insulate their convoluted, valley-infested, hip-plagued roofs, many of which have roof sections that are either low-sloped or cathedralized.

  6. Mike321 | | #10

    Hi Martin,

    I've read all these roof insulation articles and I'm getting more confused the more I read. In previous articles, you warn not to use open cell foam on the underside of a low slope roof because moisture can reach the sheathing from the inside of the house. Now it seems like using closed cell under the roof is bad when there is TPO on top because there is no path to dry. Can you help me out here?

    Edit: I have a 100 year old brick row house with a lofted front facade. There is no possible way to install all of the rigid foam needed to meet code above the sheathing. It would look crazy on the front of the house.

  7. GBA Editor
    Martin Holladay | | #11

    First of all, can you tell us your name? (I'm Martin.)

    If you are seriously interested in installing a continuous layer of exterior rigid foam above your roof sheathing, but are worried that your fascia will look too thick, see the illustration below to see how thick rigid foam can be disguised. [This illustration came from a Building Science Corporation document.]

    As the article on this page makes clear, you don't have to lose sleep over the use of closed-cell spray foam on the underside of roof sheathing, even if your roof sheathing can't dry to the exterior, as long as the roof sheathing is dry on the day that the spray foam is installed.

    You are correct that closed-cell spray foam on the underside of roof sheathing is safer than open-cell spray foam on the underside of roof sheathing -- because the use of open-cell spray foam is associated with damp roof sheathing.

    1. Mike321 | | #12

      Ah, I see the username is anonymous. I'm Mike, thank you for your reply.

      Your explanation makes sense and I still plan to use closed cell under dry sheathing. Also, thanks for the explanation on the new fascia. That also makes sense.

      One final question: Does the recommend advice for polyiso thickness above the sheathing apply in the case where there is a TPO-polyiso-sheathing-closed cell foam sandwich?

      I'm in Zone 4A and every roofer I've talked to uses TPO on top of 1" of polyiso regardless of interior insulation options.

      Thanks again!

      1. GBA Editor
        Martin Holladay | | #13

        Q. "Does the recommend advice for polyiso thickness above the sheathing apply in the case where there is a TPO-polyiso-sheathing-closed cell foam sandwich?"

        A. The rules about the minimum thickness (R-value) for polyiso installed above roof sheathing always apply, regardless of the type of roofing that you intend to install. For a full explanation of code requirements for minimum R-values for polyiso installed on the exterior side of roof sheathing -- code requirements that align with the recommendations of building scientists -- see this article: "How to Install Rigid Foam On Top of Roof Sheathing."

        According to almost all building codes, if you install polyiso on the exterior side of the roof sheathing in Zone 4A, the polyiso needs a minimum R-value of R-15 -- equivalent to about 3 inches of polyiso. Installing only 1 inch of polyiso would be a code violation, regardless of what type of roofing (TPO, or something else) is installed.

        The relevant code provision is section R806.4 of the 2009 IRC (or later versions of the code).

        All of that said, if you add closed-cell spray foam on the underside of the roof sheathing in addition to a continuous layer of polyiso on the exterior side of the roof sheathing, then the R-value of the closed-cell spray foam can be added to the R-value of the polyiso to determine compliance with the R-15 minimum requirements.

        So if you have R-6 polyiso above the roof sheathing, and R-9 closed-cell spray foam under the roof sheathing, the combined layers of air-impermeable (foam) insulation would comply with the requirement to install at least R-15 of foam insulation. The remainder of the required insulation (either R-23 or R-34, depending on whether your target for total R-value is R-38 or R-49) would need to be some type of fluffy insulation installed in direct contact with the cured spray foam.

        [Later edit: Just because this type of polyiso / sheathing / spray foam sandwich is possible, doesn't mean it's a good idea. It's always better to install an adequately thick layer of rigid foam on the exterior side of the roof sheathing, rather than a layer of rigid foam that is too thin.]

        1. Mike321 | | #18

          Wow ok thanks again! I want to say I understand. I had to read the code a few times but it does make sense for the most part. I may still be missing the underlying reason that the code is written the way it is despite reading your articles and the IRC.

          I believe that you changed your advice between what you wrote in the above comment and the 2015 Article "How to Install Rigid Foam On Top of Roof Sheathing" where you stated:

          "Option 2: You can install some of the insulation above the roof sheathing, and the rest of the insulation underneath the roof sheathing (and in direct contact with the roof sheathing)."

          "If you follow Option 2, a wide variety of insulation materials can be installed under the roof sheathing. Among the possibilities: fiberglass batts, mineral wool, cellulose, or open-cell spray foam. (Closed-cell spray foam is not recommended for this purpose, since closed-cell spray foam prevents the roof sheathing from drying toward the interior if it ever gets damp."

          As well as:

          "One more caveat: If you are planning to thicken the insulation installed under the sheathing in order to achieve a total R-value that exceeds code-minimum requirements, you’ll need to also thicken your above-sheathing foam layer to keep the ratio of above-sheathing insulation to below-sheathing insulation in the proper proportion."

          In your comment in the May 2019 article you said I can get away with 1" polyiso, sheating, R-9 of closed cell (hitting R15 total) then following up with batting to get to R49.

          I can't use more closed cell instead of batting to get to R-49? (I might not have enough physical room for the batting vs closed cell which is higher R per inch)

          I'm still missing something. Sorry! You're a patient dude it seems!

          1. GBA Editor
            Martin Holladay | | #19

            I stand by the advice I gave in the older article, "How to Install Rigid Foam On Top of Roof Sheathing."

            To summarize the advice:

            1. If you want to install rigid foam above your roof sheathing, make sure that the rigid foam is thick enough to prevent moisture accumulation in the roof sheathing. (The contractors in your hypothetical example who install only 1 inch of polyiso aren't following my advice.)

            2. Once you have installed an adequately thick layer of exterior rigid foam, the remainder of the needed R-value of your roof assembly should consist of vapor-permeable insulation installed in direct contact with the underside of the roof sheathing. Examples of vapor-permeable insulation include fiberglass, cellulose, mineral wool, or open-cell spray foam.

            In the answer to your specific question about a sandwich of polyiso, roof sheathing, and closed-cell spray foam, I explained whether and how the sandwich would meet code requirements. But I didn't recommend this approach.

  8. Mike321 | | #20

    Mike here again...I'm probably clogging up the board at this point, so feel free to delete if need be.

    I don't understand why it matters if the roof sheathing is below the dew point if it's entirely encapsulated by TPO on the top and closed cell spray foam underneath. The top surface of the sheathing may be below the dew point but it's not in contact with outside air or water vapor. Maybe it sneaks in from where the perimeter edges of the roof somehow?

    The ceiling rafters are buried into double wythe brick on both sides and the roof deck sits on top of them, encapsulated below with closed cell spray foam. The roof deck is on top of the joists and encapsulated above in a 1" polyiso TPO layer. (100 year old brick row house).

    Again, newbie here but reading a lot.

    1. GBA Editor
      Martin Holladay | | #21

      If you are the type of homeowner or builder who has no qualms about building a roof assembly with roof sheathing that can't dry out in either direction, you can go ahead and sandwich your roof sheathing between TPO-covered polyiso and closed-cell spray foam. As I wrote in this article, builders have been doing it for years, and these roofs work.

      That said, what if you are the type of builder who thinks ahead, and comes to GBA for advice before you finalize your details? You might write, "I live in Climate Zone 4, and I plan to install polyiso above my roof sheathing. How thick should my polyiso be?"

      I would advise, "Choose polyiso that is at least 3 inches thick, to keep your roof sheathing above the dew point in winter. Then make sure that whatever insulation you install on the interior side of the roof sheathing is vapor-permeable."

      Evidently, you've already got a roof with polyiso that is too thin. (It's 1 inch thick.) That forces you to supplement the too-thin foam with interior closed-cell foam, and you end up with roof sheathing that can't dry inward. It's not ideal, but you've run out of options -- because your polyiso is too thin.

      1. Mike321 | | #24

        Thanks. I'm in the process of finalizing roof details and will install within a week.

        You wrote: I would advise, "Choose polyiso that is at least 3 inches thick, to keep your roof sheathing above the dew point in winter. Then make sure that whatever insulation you install on the interior side of the roof sheathing is vapor-permeable."

        Except open cell spray foam apparently as the advice on that has changed? So we're down to air AND vapor permeable options (fiberglass, mineral wool, cellulose) on the interior side of the assembly. Your article above says open cell is fine but your comments in this thread say to avoid it.

        1. GBA Editor
          Martin Holladay | | #25

          When open-cell spray foam is installed on the underside of the roof sheathing, you can get problems with damp sheathing. (The sheathing gets cold in the winter, and the open-cell spray foam is vapor-permeable -- so interior moisture can pass through the open-cell foam by diffusion, reaching the cold sheathing.)

          Instead of allowing your sheathing to get cold in the winter, you can install an adequately thick layer of rigid foam on the exterior side of the roof sheathing. This approach keeps the roof sheathing above the dew point for most of the winter. The sheathing is now warm and dry. (In your climate zone, this approach requires at least R-15 of exterior rigid foam.) If the sheathing isn't cold, you won't have moisture accumulation.

          Once you've protected the sheathing by installing an adequately thick layer of continuous rigid foam on the exterior side of the sheathing, you are now free to install vapor-permeable insulation on the interior side of the roof sheathing -- for example, fiberglass, cellulose, mineral wool, or open-cell spray foam -- without worrying about moisture accumulation.

          1. Mike321 | | #30

            I finally get it. I'm kinda slow. Thanks!

  9. yavaid | | #22

    Hi Martin,
    This is Naveen. Have got Prelim-plan and had requested a cathedral ceiling(to have a peaceful marriage.. :P .. my preference is 9'(2X8) or 10'(2X8) ceiling with R-60 blown-in cellulose), ofcourse will seal the 2nd floor ceiling for air tightness. Since its going to be cathedral i was thinking of 14" truss and have 12" CC-SPF and will have Zip sheathing and have a Ventilation above Zip Sheathing with 1X2 purlins and install fibre based shingles. I am in Zone4a.
    The more i read and listen to Joe i think i get it, however when i read other papers and articles, i get worried. In short given that there will be roofers failing's... does the above roof system have any chance of surviving over a long period? in other words what is wrong here?

  10. GBA Editor
    Martin Holladay | | #23

    Please describe your roof assembly in more detail. Will your insulation consist of R-60 cellulose, 12 inches of closed-cell spray foam, or both?

    1. yavaid | | #26

      ok, Top to bottom this is the sequence of layer
      1. fibreglass shingles
      2. 1x2 purlins
      3. Zip sheathing
      4. 12" CloseCell spray Foam
      5. Drywall


      1. GBA Editor
        Martin Holladay | | #27

        I'm guessing that you forgot to mention a layer. You can't install fiberglass shingles on 1x2 purlins, so I'm guessing that there will be a layer of plywood or OSB between the purlins and the shingles.

        The type of assembly you are talking about is the same assembly mentioned by Mark Walker in Comment #2, above. Directly underneath Mark Walker's comment is my answer (Comment #4). If you read my Comment #4, I think you'll find the answer to your questions.

        Installing a vent channel above roof sheathing only works well if (a) the roof sheathing is fairly vapor-permeable (Zip sheathing isn't), and (b) the roofing underlayment is fairly vapor-permeable.

        1. yavaid | | #28

          So i would use fiberboard instead of Zip sheathing and then use OSB/plywood on top of 1X2 purlins to place the shingles? or should i go the route shown in the picture attached(seems less expensive)

          i am trying to build a house as tight as possible without going Passive house route, so i am kind of restricted in terms of checking truss regularly(to top it going for cathedral ceiling for family room alone. What would i have to be cautious about?


          1. GBA Editor
            Martin Holladay | | #29

            There are lots of ways to build an insulated cathedral ceiling. The way you have chosen is neither the least expensive, nor the easiest, nor the best performing, nor the greenest, way to build.

            For lots of information on all the different ways to create a cathedral ceiling, you might want to read this article: "How to Build an Insulated Cathedral Ceiling."

  11. yavaid | | #31

    If i am not confused and clear after having read that and few other articles from JLC i wouldnt be asking the questions i have asked earlier.
    Question is what is the simple way to build unvented cathedral ceiling( i have a dormer for a 15' section of 45' ridge.), which is preferably green too.

    1. GBA Editor
      Martin Holladay | | #32

      Q. "What is the simple way to build unvented cathedral ceiling?"

      A. I think the simplest way is to install an adequately thick layer of rigid foam on the exterior side of the roof sheathing, supplemented by fluffy insulation (fiberglass, cellulose, or mineral wool) between the rafters. Ideally -- especially if you are a green builder -- you will use recycled (reclaimed) rigid foam. For more information, see "How to Install Rigid Foam On Top of Roof Sheathing."

  12. Hutlife | | #33

    I have been reading many related articles and I am in a unique situation as I am converting a wood framed dirt floor Quonset Hut in climate zone 5 into a Workshop/small apartment. My plan was to tighten up the roof to prevent leaking and spray 4" closed cell spray foam the underside. I am starting to think it may be better to take the metal roofing off as i have to replace the nails with screws and butyl tape the seams and install a plywood sheathing on top of the pearlins. I feel like this might add a layer of protection between the roof and the wood frame of the building.
    At this point I was hoping to just add an underlayment on top of sheathing and reinstall metal roofing. I would at some point in the future like to replace the roofing with a nicer metal roofing but would like to get by with this for a while. I'm hoping this approach would work just fine with the right underlayment. Any suggestions?

    1. GBA Editor
      Martin Holladay | | #34

      It would be easier to give you advice if you asked a specific question.

      Most metal roofing can't conform to a curve, so that is problem #1. Curved metal roofs exist, but most roofers have never encountered one -- so you'll need to get advice from someone with relevant experience.

      Installing sheathing on a curved surface is also challenging. In some cases, you can't install 1/2-inch sheathing -- so you may be forced to install two layers of 1/4-inch sheathing with adhesive between the layers. This approach may raise engineering questions.

      1. Hutlife | | #35


        Thank You for your response. I am not concerned with the ability to bend even 1/2 sheathing to a 19' radius so thats not really my question. My question is with 4" closed cell foam under the sheathing creating a vapor barrier what type of underlayment would you use to then install the metal roofing on top. Should the underlayment allow the sheathing to breath and if so would that still be good to repel water from the outside? Or should I use something that would seal the whole thing better protecting it from outside water intrusion Like adhesive backed water and ice type underlayment?

        1. Expert Member
          MALCOLM TAYLOR | | #36

          "I am not concerned with the ability to bend even 1/2 sheathing to a 19' radius"

          You should be. The structure of your building is an arch. The forces it is resisting are primarily outward. Adding the plywood, which wants to remain straight, significantly increases these forces.

          Most quonset huts rely on the metal structure itself as the structure. Before committing too far on the interior and insulation, I'd see if the existing roofing can be effectively made weatherproof. My guess is the difference in having to replace the existing roofing with new custom bended stock, and replacing the whole thing with a load-bearing metal hut wouldn't be that significant.

          1. Hutlife | | #39

            Hello Malcolm,

            Thats a aspect I have not yet considered. I do however have a structural engineer that would be happy to go over that issue with me. I am thinking I may not need 1/2" for sheathing due to the width of the existing purlins the space between them is around 12". i am thinking of using 3/8" sheathing oriented horizontally in the direction it tends to bend anyhow. I have sheets of this sheathing on hand and find that it conforms to the arch of the building with little or no effort. If it is still a concern I would be fine with possibly adding some rafter ties maybe cables around 8' high from side to side to counter that. Any thoughts?

          2. Expert Member
            MALCOLM TAYLOR | | #40


            3/8" will work like a charm. I wonder though - if you are going to add 4" of closed cell foam, what's the advantage of adding the plywood? A synthetic underlayment will both protect the roof and keep the foam from adhering to the metal. That might be all that is necessary.

            There is a good chance the building is fine as it is structurally. Monkeying with arches by adding ties might cause more problems that they solve. It's definitely worth a quick chat with an engineer.

          3. Hutlife | | #41


            Now thats interesting! The building has been as is since 1942 when it was built. One of the issues I have is that they installed the roofing in opposite directions so that the ribs on either side didnt line up at the seems. The top is overlapping only by about 3" then they beat it down with a hammer to make it meet. I plan on addressing this by removing the roofing I think the panels are a little under 12' meaning its takes three pieces to get to the bottom of one side. I will take a section cut it in half for the bottoms and end with a section centered on top to eliminate the seam at the peak. Then butyl tape and screws. One of the reasons I wanted to use roof sheathing is to give the foam something to stick to besides the metal so it would be replaceable down the line. the other part is it would make the building more resistant to sway which it has a very slight lean to it and i'm not sure if thats been there since the beginning or happened over time. I guess I would question if just the spray foam alone would add enough rigidity to stablize it if it is starting to lean. And would the sheathing be worth it as an extra layer of protection in the sandwich?

          4. Expert Member
            MALCOLM TAYLOR | | #42


            Thinking about it more I'm going to walk back on my comments about the advantages of a load-bearing metal structure. The wood arches and purlins provide a good base to support whatever insulation you choose, and as you say, filling in the voids with foam will add quite a bit of rigidity to the structure. Unless you are in a high seismic or wind area I'd bet it is enough on it's own.

        2. GBA Editor
          Martin Holladay | | #37

          Leaving aside (for the moment) the structural issues and the issues surrounding the need to bend the plywood and the metal roofing, the answers to your questions about "does the plywood sheathing need to breathe?" can be found in the article on this page.

          It's certainly better if the plywood can dry out in at least one direction, but that won't be possible if you install closed-cell spray foam on the underside of the roof sheathing. That said, people do it all the time, and the plywood is going to be fine as long as the plywood is dry on the day that the spray foam is installed.

          The vapor permeance of your roofing underlayment is irrelevant, because there won't be any outward drying. You can select any type of roofing underlayment you want.

          1. Hutlife | | #38

            Thank you Martin that is helpful.

  13. AdamJ_7685612 | | #43

    I'm replacing my shingles, adding 5" ccSPF in my rafter bays, and finishing with an additional 2" polyiso foam board to provide continuous insulation and additional vapor barrier over the 2x6 rafters. (Taking framing into account, that gets me just over R-40 - good for Zone 5?). Since I'm already in 'sandwich' territory due to the ccSPF and shingles, would it be advisable to lay full I&W beneath the underlayment to provide additional leak protection? I don't see any down-side aside from cost, if the asphalt shingles will already prevent my deck planks from drying outward.

    1. GBA Editor
      Martin Holladay | | #44

      As you undoubtedly know from reading the article, I'm not a fan of your approach (sandwiching your roof sheathing between two vapor-impermeable layers). But if you're already committed to ignoring my advice, the Ice & Water Shield won't make things any worse.

      1. AdamJ_7685612 | | #45

        I understand your perspective, and I'm not ignoring it. Although it's not your preference, I feel it's safe to move ahead based on the work of Grin and colleagues (2013) and other helpful resources you cite:

        "Based on this modeling there are no known risks with using SPF insulation under plywood and OSB roof decks if the following requirements are met:
        • The installation complies with the 2012 IRC.
        • A fully adhered leak-free roof membrane is installed.
        • The roof sheathing is and framing dry below 18% before SPF installation.
        • And when using ocSPF a low-perm Class II vapor retarder is installed where required."

        But that study notes that drying performance requires that "fully adhered membrane is properly designed, detailed, and installed" to resist expected rainwater leakage. Will breathable synthetic underlayments perform that role, or does 'membrane' imply a full vapor barrier?

        Thanks for your time and detailed articles!

        1. GBA Editor
          Martin Holladay | | #46

          If you want to follow the advice of Grin, Smegal, and Lstiburek -- and there is no reason you shouldn't, since they are eminent building scientists -- the answer is clear: "fully adhered" means "fully adhered." They're referring to a product like Grace Ice & Water Shield, not a synthetic roofing underlayment.

          1. AdamJ_7685612 | | #47

            Thank you, Martin!

  14. austintechsmith | | #48

    I have been taking the task on of building my own tiny house on wheels. Glad to find your article because I thought all hope was lost when I realized I didn't have enough knowledge about thermal transference with CFS and had insulated the inside with closed cell foam. It is essentially a solar powered easy bake oven in this texas weather right now... Ooops.

    Do you think I can save the thermal performance of this structure by sheathing the roof and walls with foil faced foam board and air spaces between foam board and cladding/metal roof?

    The Closed cell was seemingly installed well, but Is now shrinking in the areas they applied later in the day when it got hotter. There is about a 1/8th-1/4th" gap in some of the stud bays about a month after closed cell was installed. In this case that may be a good thing as it would allow some drying to the inside?

    1. GBA Editor
      Martin Holladay | | #49

      Your questions were confusing to me until I clicked the photos. I looks like you installed steel studs, steel joists, and maybe steel rafters. Is that the case? If so, your problem has nothing to do with the spray foam -- it has to do with thermal bridging through the framing.

      If this building has steel framing, then it certainly needs a continuous layer of exterior rigid foam -- the thicker, the better. The foam doesn't have to be foil-faced.

  15. michaelbluejay | | #50

    I have a hot garage with ice+watershield over the sheathing, and no insulation above the sheathing. I have a lot of pieces of 1/2", 3/4", and 1" rigid insulation left over from other projects, and I'm tempted to cut & cobble them between the rafter bays, so I can get some use out of them rather than landfilling them. (I wouldn't put them across the rafters, for aesthetics.) While the article didn't seem to address this particular situation (it addressed ice+water shield on top, and spray foam on the bottom), I still presume that this would make an unacceptable vapor sandwich? I assume that not much drying can take place through the uncovered rafters. I'm in Climate Zone 2.

    1. GBA Editor
      Martin Holladay | | #51

      Most unvented roof assemblies don't dry to the exterior, whether or not they have Ice & Water Shield above the roof sheathing, since neither asphalt shingles nor metal roofing allow for outward drying. So the presence of Ice & Water Shield isn't really the problem.

      The safest way for you to proceed would be to create a vented assembly -- that is, to create a 1-inch-deep air channel between the top of the insulation and the underside of the roof sheathing. This can be accomplished by installing 1" x 1" sticks in the corners of your rafter bays, tight to the underside of the roof sheathing, before installing the rigid foam. Of course, this approach will require soffit vents and a ridge vent.

      If you read my article on the cut-and-cobble approach ("Cut-and-Cobble Insulation") you'll read about roof failures that have occurred when homeowners have tried to use the cut-and-cobble approach for unvented roof assemblies. Your suggested method isn't recommended. (In my article, see the third bullet point in the list that begins, "Cut-and-cobble has some disadvantages.")

    2. Expert Member
      Dana Dorsett | | #52

      >"I have a hot garage with ice+watershield over the sheathing, and no insulation above the sheathing."

      Is this garage actively cooled & heated?

      How air-leaky or ventilated is the garage?

      What is the ultimate goal here?

      >I have a lot of pieces of 1/2", 3/4", and 1" rigid insulation left over from other projects..."

      What type of foam, and are there any facers?

      Unfaced polystyrene (EPS or XPS) is somewhat vapor permeable, and does not create a moisture trap. Foil or vinyl FACERS (sometimes found on low-density EPS) CAN create a moisture trap.

      How deep are the rafters? Do you intend to fill the rafters with the cut'n'cobble?

      At 1.5" most XPS becomes a Class-II vapor retarder, but that isn't a particularly big deal- it still dries to the interior but it takes some time. At 5" (0.2-0.25 perms) + it has the potential to create a moisture trap, but without a significant interior side moisture drive (are you boiling 10 gallon pots of chili in there on a regular basis with the pot lids off, doors closed?) even that isn't a big deal.

      Type-II EPS is about 3x as vapor permeable as XPS at any given thickness. At 1" it is no more vapor tight than a couple coats of standard interior latex paint, with very low moisture trapping potential.

      >"I'm in Climate Zone 2."

      In Climate zone as little as R5 of insulation provides enough dew point control a the foam/fiber insulation boundary for up to R35 of fiber insulation snugged up to the underside of the foam.

      If the whole point is to limit the amount of heat blasting at you near the floor in an uncooled garage from a 150F roof deck you may be better off installing a perforated aluminized fabric type radiant barrier on the underside of the rafters than installing a couple inches of cut'n'cobbled foam, with near zero moisture trapping potential. Most perforated RB runs a bout 5 perms, about the same vapor retardency of a single coat of interior latex paint. Going one better would be to use "contractor roll" R13 or R19 KRAFT faced batts snugged up to the roof deck (facer side down) and installing the perforated RB an inch or so below the batts, using 1-1.5" wide strips of your scrap foam board as spacers. That wouldn't necessarily be a good idea in fully conditioned living space (though not super-risky in zone 2, but is fine for a garage since there is little to no moisture being put in the air by life occupants, bathing, cooking, etc. The RB on it's own provides quite a bit of heat rejection from a superheated roof deck but does very little toward keeping the place warm during heating season. The fiber insulation helps year-round.

  16. kat_h | | #53

    To preserve storage space and avoid a complicated air sealing job between the 2nd story and the attic, we want to add closed cell spray foam to the underside of our board-sheathed roof.

    We are also about to have steel shingles installed onto furring strips. The current plan from the roofer is that they will install the strips over the top of the (deteriorating) asphalt shingles that are already there. We have no idea what kind of underlayment was used when the shingles were installed.

    Should we create the above-sheathing ventilation channels mentioned in the article by having the roofer remove the current shingles and underlayment and install a permeable underlayment underneath the furring strips? Any suggestions for that underlayment?

    I'm new to all of this, so I'm just trying to make sure I'm not missing something.

    Oh, we're in CZ6.

    1. GBA Editor
      Martin Holladay | | #54

      It sounds as if you are creating an unvented insulated roof assembly. For more information on this type of roof, see this article: "How to Build an Insulated Cathedral Ceiling."

      With this type of roof, there is no need for above-sheathing ventilation channels. That said, if your roofer proceeds as planned, you'll have an air space between the existing asphalt shingles and the new roofing -- and that air space will reduce the chance of ice damming.

      Your roof assembly, as planned, will not allow the existing roof sheathing to dry to the exterior. That's not necessarily a problem. That said, if you want your existing roof sheathing to be able to dry to the exterior, you'll need to strip the existing roof shingles and install new vapor-permeable roofing underlayment before your roofer installs the new steel shingles.

      1. kat_h | | #55

        Thanks, Martin.

        I guess the question is *should* I want the sheathing to be able to dry to the exterior, or is it okay to to go ahead with the roof as planned.

        And if we go ahead as planned, I assume it's essential to convince the insulation contractor to test the sheathing for moisture before applying the spray foam? How common is it for moisture content to be higher than 18%?

        1. GBA Editor
          Martin Holladay | | #56

          My article, and the comments on this page, discuss your question at length.

          I don't like roof assemblies that sandwich roof sheathing between two impermeable layers. But contractors do it all the time.

          My advice: Don't do it unless you're backed into a corner. In your case, adding a sufficiently thick layer of rigid foam on the exterior side of the roof sheathing would certainly be preferable to your plan -- especially since you are installing new roofing. For more information, see "How to Install Rigid Foam On Top of Roof Sheathing."

  17. noconflictsplease | | #57

    Howdy! I am about to re-roof a house on occasionally hurricane ravaged Galveston Island (10 feet from salt water in climate zone 2). The upper floor has essentially a hip roof that rises on 4 sides to a central point, no ridges. The enclosed space is a master bed/bath, with a cathedral ceiling (drywall on bottom of 2x6 rafters) on the perimeter and a central tray ceiling. There are 14 5"x10" soffit vents on the sides and 2 static vents near the apex. The fiberglass-choked rafter bays (no baffles) terminate into the soffit, without a common eave, so at most 14 of the bays are open to air intake. The centrally located of these bays (that do not terminate in a "Hip" behind the cathedral section) would allow air to enter the modest attic space above the "tray" part of the ceiling, and exhaust thru the air hawks. Current roof is plywood+#15 felt+asphalt shingle. THE PLAN: We will tear down to the deck and install a peel and stick I&WS equivalent. I want the extra protection in case of shingle blowoff in a hurricane. New wind-rated "cool roof" asphalt shingles will go on top of that. QUESTION: Should I install a continuous soffit and baffles for the rafter bays that are open to the attic above the "tray" ceiling? Since I currently have what approximates an unvented attic. I worry that given such changes, our AC cooled sheetrock could invite condensation on the outer surface in contact with additional supplies of warm moist air.

    1. GBA Editor
      Martin Holladay | | #58

      You can't vent the roof you describe. You need to create an unvented roof assembly. Above the Ice & Water Shield, you need an adequately thick layer of continuous rigid foam insulation, followed by another layer of sheathing and roofing.

      1. noconflictsplease | | #59

        Thank you!

  18. AaronAho06 | | #60

    Hello Martin,
    I have a house I bought last year that needs a new roof put on it. It's in climate zone 4a. It has a gambrel style roof with living space under the roof. The roof rafters are 2x6's. Upon investigation it seems the previous owner insulated and finished this space off by using 3-3.5" of closed cell spray foam directly onto the roof sheathing and did not install any batts due to the limited space left in the joist cavity. So it has approximately 2" of air space. Then they applied drywall. The drywall is not air tight since it has can lights installed throughout. There is no venting along the eaves but from what I can tell the foam sealed off the joist cavity to the top plate of the wall so there should not be any exterior air flow.

    I am now looking at installing a new roof and am trying to figure out what's the best way to improve the situation. I've been reading your articles and think I definitely need to install at least R15 of rigid insulation above the sheathing. However, what would you recommend would be the best course of action in this situation?

    1. MartinHolladay | | #61

      Your roof has about R-19 or R-20 insulation, so it would be great if you could add R-38 or R-40 more. That's difficult, however -- it would require about 10 inches of EPS. If it were my house, I'd install about 6 inches of EPS (that's easier to install than 10 inches) above the roof sheathing. But if your local roofers balk at that plan, you may have to make do with a little less rigid insulation, as you propose.

      It's important to emphasize to the roofers that you need perfect sunny weather (ideally for several days) when the old sheathing is exposed. That sheathing must be dry when the new rigid foam is installed.

Log in or become a member to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |