GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter 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

Unvented cathedral ceiling with interior rigid foam

Jtrhodes | Posted in General Questions on

I have a Cape cod style house. The upstairs is completely gutted and will be finished as unvented cathederal ceilings. The roof rafters are 2x6s with some 2x8s on a big dormer. I was planning to open cell foam all cavities and add 2 layers of rigid foam under the rafters with offset taped seams. ( 4″ total under 2x6s and 2″ total under 2x8s) Spray foaming any gaps. Then use furring strips and drywall. Does this seem ok? Would untrimmed closed cell foam be better? But it would leave air pockets.

I was planning the same thing for the 2×4 walls with 1″ of interior rigid foam.

John R.

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
    Michael Maines | | #1

    John, a few questions: what climate zone are you in? Why do you want an unvented roof? Why foam in the rafter cavities instead of an insulation with lower embodied energy?

  2. Jtrhodes | | #2

    I am in Philadelphia. The house has no eaves to have effective roof venting. The shingles are only 5 years old and in good shape. My thought on foam was that it would seal air out well.

    John R.

  3. walta100 | | #3

    In my opinion unvented cathedral ceiling is a risky assembly. The more insulation you add the more dangerous it becomes.

    You may want to take a look this on Martins blog


  4. Jtrhodes | | #4

    Thanks. The bottom 2/3's of the roof sheathing has been an unvented roof using fiberglass insulation since 1951. Luckly without rot. I have no ability to vent it from below the sheathing (no eaves). I am trying to make things better. The roof is just 5 years old so venting from the top is not on the table either. Yes I plan to eliminate the venting on the top third if the roof as well.

    So the question is untrimmed closed cell foam, air gap & polyiso sheets inside or trimmed open cell foam, no airgap & polyiso sheets inside?

    Closed cell is more traditional for this but what about the airgap and polyiso? Also should the polyiso not be foil faced so some drying could slowly take place?

  5. Expert Member
    Michael Maines | | #5

    Jtrhodes, having the attic open (and likely vented at the gables?) makes a big difference on how roofs like yours perform. That said, while open cell foam meets IRC code language in this application, because it's air-impermeable, it is open to water vapor movement, so there is some risk of moisture accumulating at the sheathing. Closed cell foam would be a safer option. It would also be smart to allow drying to the interior; in fac, using your proposed assembly, I would not use polyiso at all, but either EPS or mineral wool board stock at the interior.

    My approach would be to furr the rafters down, hanging 2x2s or 2x3s from plywood gussets, to create a thermally-broken assembly, and install a deeper layer of closed-cell foam. If you use an HFO-blown foam, which you should because the blowing agent is much less environmentally damaging than that in regular foam, you can get 8" (R-49, required by code in your 4a zone for new construction) in two passes.

    1. Jtrhodes | | #9

      Why no polyiso? Based on space I would loose about r12.

      John R.

      1. Expert Member
        Michael Maines | | #11

        Because most polyiso is not vapor permeable, you'd have a void that could collect moisture, and screwing through multiple layers of foam into framing is not always easy. But if you get fiber-faced polyiso, you would have some drying to the interior, so it would probably be fine.

        1. Jtrhodes | | #12

          Got it no foil faced polyiso.

  6. Jon_R | | #6

    Consider some closed cell spray foam and fiberglass fill.

    1. Expert Member
      Michael Maines | | #8

      Jon, that's a great idea, and one I should have thought to mention as well. In zone 4A he would only need 31% of the R-value to be in the closed cell foam layer, with a minimum of R-15 to meet IRC codes. Cellulose or open-cell foam could also be used on the interior portion.

  7. walta100 | | #7

    John did you read Martin’s blog?
    Note every valid email address gets 1 free 10 day membership.

    It is the stories of 6 people that had working unvented cathedral ceilings who made well meaning changes. All 6 homes are now moldy messes that drip water with no clear solutions.


    1. Jtrhodes | | #10

      Yes I just read it. Air leaks and not having stratified indoor air seem important.

  8. GBA Editor
    Martin Holladay | | #13

    Walter Ahlgrim,
    This is the second blog where you have posted comments that imply that my article about summertime condensation is some type of warning against unvented roof assemblies. It isn't. You are misrepresenting the content of my most recent article.

    Unvented roof assemblies can be detailed well, and often are. A well-detailed unvented roof assembly will be problem-free.

    My most recent article ("Summertime Condensation Near the Peak of a Cathedral Ceiling") has nothing to do with the performance of unvented roof assemblies. It has to do with a building science mystery reported by 6 GBA readers.

    For more information on good details for unvented roof assemblies, see these two articles:

    "How to Build an Insulated Cathedral Ceiling"

    "How to Install Rigid Foam On Top of Roof Sheathing"

  9. Expert Member
    Dana Dorsett | | #14

    In zone 4A it only takes 1-2" of closed cell foam to protect the roof deck, but you don't want to create a moisture trap. With ~0.1 perm asphalt shingle layup and 2" of interior side foam board you'd have to use unfaced EPS (~1-1.5 perms @ 2") to have a reasonable drying path. Even fiber-faced polyiso would be under 1 perm.

    With 2" of HFO blown closed cell foam in the 2x6 cavities you'd have R14 foam, and 3.5" of space for R15 rock wool (recommended), and with 2" of unfaced Type-II EPS it would come to R8.4 for the foam layer for R37-38 t0tal, and more that 35% of that total as closed cell foam.

    With 2" of HFC blown closed cell foam (not recommended for environmental reasons) it would come to R35-R36 total, with about 33% being the closed cell foam.

    On the 2x8s use 2" of closed cell foam and compressed R20-R21 fiberglass (which performs at R19-R20 when compressed to 5-1/4"). With HFC blown closed cell would have R40-R41 with a 30% closed cell foam ratio, with HFO blown it would be R42-R43 and a 32% ratio.

    1. Jtrhodes | | #15

      I had planned to shoot for r45-r60. Either 5" or 7" of closed cell in the rafter bays 2x6 or 2x8s leaving only a small air gap. Then 2" of reclaimed foam on the 2x8 and 4" on the 2x6s. Trimming closed cell flush is difficult but the air gap should only be 1/4" to 1/2" thick. At that thickness not sure if any foam is permeable. Should I not shoot for that r value?

      1. Expert Member
        Dana Dorsett | | #16

        Any air gap is a potential thermal bypass and moisture transfer path, even 1/4". It's better to install less closed cell to be able to get a complete fill using fiber insulation or trimmable open cell foam.

        What's the purpose of such a high center-cavity R-value? (Bragging rights?) At 7" HFO blown foam is already R49, at 5" it's R35, but when installed between framing it's high R/inch performance is severely undercut by the substantially lower R/inch framing. The 2" of continuous rigid foam does more for the net thermal performance than the difference between going with just 2" vs. a complete cavity fill of closed cell foam. At R-values ranges this high the difference in energy use is somewhat "in the noise", and it's not necessarily "worth it" on a lifecycle energy cost basis to go much higher than the code-min U0.026 , which is R38.5 "whole assembly-R". There's an easy financial rational for going that high or a bit higher with fiber insulation at less than 5 cents per R per square foot, not so much when it's closed cell foam at 18-22 cents/R-foot^2.

        Even at 7" closed cell foam is still a class-II vapor retarder, but it's at the very vapor tight end of the Class-II range. At 2" it's typically between 0.5-1 perm (depending on vendor), which is still an adequate drying path. Even if it needs to bump to 3" for dew point control it's still a seasonal drying path, not a moisture trap.

        Closed cell foam is not a very "green" insulation material, even when blown with relatively benign HFO blowing agents. An inch of closed cell foam uses 4x as much polymer per inch of depth, and about 2x as much polymer per R, and it's quite expensive. Anything more than the minimum needed for dew point control is really something of a waste, since it's effect on the "whole-assembly-R" performance is small compared to just adding another half-inch to inch of continuous foam on the interior.

        Using reclaimed foam on the interior is certainly greener. But stick with reclaimed EPS, not polyiso, not XPS. Most reclaimed roofing EPS is Type-VIII (1.25lbs per cubic foot nominal density), and would still be ~1-perm or a bit more even 3", which is enough to keep fiber insulation or open cell foam from accumulating moisture at/near the boundary with the sprayed closed cell foam. At 2" 2lb density roofing polyiso could come in a bit under 1 perm even if you tediously stripped the facers (though that won't show up in a product specification sheet.) It still might be fine though, but only if there is enough closed cell spray foam for dew point control at the higher R.

  10. Jtrhodes | | #17

    Thanks. I was reading on new construction r60. Maybe I misunderstood. So how about this:
    3" of closed cell foam (45/48% of r value)
    Fill rest with fiber glass r8 to r12 (17/32% of r value)
    2 or 4" of eps r8 or r16 (20/37% of r value)
    So r40 to r43 for roof.
    Some areas I can't get foam board in will get only spray foam.

  11. Expert Member
    Dana Dorsett | | #18

    With 3" of HFO blown 2lb foam you're at ~R21.

    Then 2.5" of 1.8lb fiberglass or a compressed R15 (the 2 x 6 case) would be another R11...

    ...or 4.25" (the 2x8 case) would be R18.

    Add to that 2" of Type VIII EPS for R8.3.

    That makes the 2x6 case ~R40-R41 at center cavity, 51% of which is the closed cell spray foam...

    ...or R48-R49 (= IRC code-min for zone 4A) for the 2x8 case, 44% of which is the closed cell foam.

    Using HFC blown foam it knocks R2-R3 off the totals, but there is still a lot of dew point margin.

    Looks like a very moisture safe (if expensive) stackup.

    From a financial rationality point of view, take a look at the "Compact Roof" column of the zone 4 row of Table 2 on p.10 of this document:

    Note, those are "whole assembly R" numbers not center-cavity R, but it's also just the starting point as a cost-effectiveness point for new construction, not as a retrofit.

    The R45 whole-assembly starting point is more expensive to do as a retrofit than new construcion, since in new construction there is no need to use closed cell spray foam (at twice the cost per R-foot^2 of EPS or polyiso). Just 3" of HFO blown foam would run about 22 cents per R-foot^2, so that R21 is cost at least $4 per square foot, whereas even an R21 polyiso nailbase panel would be both somewhat cheaper & higher performance, since it would thermally break the rafters. Combined with a 2x10 rafters w/ R30 batts it would come in north of the recommended R45 whole-assembly-R for less money than what you will be spending doing a retrofit.

    1. Jtrhodes | | #19

      Thanks for all the help. I appreciate it.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |