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

Polyiso Exterior Insulation for Vaulted Octagon-Shaped Roof

James Boris | Posted in General Questions on

Hello, thank you all for all you’ve taught me so far! I’m designing a freestanding kitchen/living building in Zone 2A, 50 miles east of Austin.
1) Dimensions: 28’ W octagon, cathedral ceiling (656 sq ft, 8279 cu ft).
2) Walls: ~R-9.4 *overall* (R-20 cavities, U-0.27 windows, 2×6 studs).
3) Windows: Lots of westish-facing glass (SHGC-0.21), which I’ll try to counter with white solar shades, some existing tree cover, and 2′ 4″ overhangs (via “real” rafter tails… aka thermal bridges).
4) Floor: ~R-1.88 (hardwood, plywood, 2×10 joists). ~3’ H fully open crawlspace (think ‘house on stilts’). No floor insulation — I’ve seen too many insects nesting in rigid foam in crawls.
5) Roof/ceiling: R-???. Insulation will solely consist of exterior polyiso – rafters will be exposed on the interior, as the ceiling. 86 SRI white metal roof, zero penetrations… and mostly unshaded, sun-drenched.

The Question: How much polyiso should I put on the roof? Or maybe: When it’s 120 on the roof, how hot is the ceiling, given R-X vs. R-Y? Let’s assume polyiso performs ~R-5.6 in Zone 2A.

Some thoughts:

1) Recommendations for Zone 2A: R-60 (Wilson/GBA), 30-60 (Energy Star), 40-50 (BSC), 38 (Austin). But I’m reticent to just adopt these because…

2) My concern is thermal comfort, not payback. I’ll hit net-zero with even meager polyiso. I’d rather keep heat out of the building in the first place… I don’t like to run/maintain AC and fans all the time. This is forever-ours building, and we may face 100 degree Octobers some day. With my big vaulted ceiling and windows, I feel this is even more important.

3) I’m not sure how to apply [Q = U x A x delta T] to heat gain. Some posts that indicate it can be done:
-Dana Dorsett: “Once you’re at R50 taking it to R100 would make less than a 1F difference in peak & average ceiling temp, and isn’t worth doing for cooling comfort or cooling energy use.”
-Allison Bailes: “[Although the energy savings of adding insulation here may not be substantial,] if we look at the percent reduction in heat transfer… it’s significant. The [assemblies with more insulation] reduce the heat flow through walls by about a quarter.”

4) Costs: I’m the builder and can snag recycled polyiso fairly cheap, so ignore costs, pretty much.

5) Gut feeling: Between 7″ and 10″. Assuming very good air-sealing all around, including on my silly glassy walls, by yours truly.


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
    Dana Dorsett | | #1

    With SRI 86 metal roofing the peak roofing temp is going to be dramatically lower than with SRI 15 asphalt shingles, so even at R38 (any insulation type) the peak ceiling temps are going to be well bounded.

    While used 2lbs roofing polyiso is cheap, anything over 6" gets to be pretty unwieldy to install, requiring extra-long timber screws, and even at a relatively modest peak temp of 120F on the metal roof the average temp through the foam layer would be around 100F, a temp where even brand new roofing polyiso performs closer to R5.0 than R5.6.

    The thermal mass of polyiso isn't very high, and from a peak ceiling temp point of view it matters. Installing R38 of 3lbs density cellulose (modest dense-packing) would do a better job of limiting peak ceiling temps than R50 polyiso or polyurethane despite the higher R. See the comparative charts relative to thermal diffusivity of different materials in this bit of marketing fluff from a German fiberboard insulation manufacturer:

    At roughly half the thermal diffusivity of polyurethane (identical to that of polyiso at a given density) the cellulose would both reduce the peak ceiling temp and delay the peak ceiling temp by hours.

    A reasonable compromise for the exposed rafter look, might be to spec 2x12 rafters and dense-pack only the outer 5.5-6" (R20- R22) and install 3-4" of used roofing polyiso (R15-R20) up top, any of which would beat IRC code min on U-factor basis while providing enough thermal mass to make a difference.

    A higher performance approach would be to lose the exposed rafter look, installing 11.25" (R42-ish) 3lbs cellulose with 2-4" (R10-R20) of used roofing polyiso as a thermal break on the rafters.

    FWIW: The mass effects of cellulose vs. fiberglass for the 2x6 walls also makes a difference in peak wall temp for sun exposed walls, not so much for north facing walls. If dense packing the walls or ceiling prove too expensive there are cellulose batt or cotton(denim) batt options that would work eg:

    If GoLogic's dreams come to fruition even denser R4/inch wood fiber batts with even more favorable diffusivity may become available by 2023:

    1. James Boris | | #2

      Thanks so much for the reply Dana -- you have been a huge help to me in the past year :) Followups:
      1) I'm locked into the polyiso only / exposed rafter look (lots of design work behind it -- 48" oc, lots of bracing, timber frame look). But the info is very helpful and I'll consider it for walls.

      2) I'm not worried about the trickiness of installing foam because (a) I'm doing it and am in no hurry, and (b) my roof decking is (exposed) 2x tongue and groove (running parallel to eaves, with an air-seal detail at the perimeter) so foam screws can get 1 3/8" of bite without hitting a rafter.

      3) "Even at a relatively modest peak temp of 120F on the metal roof the average temp through the foam layer would be around 100F." How did you calculate this? What would the average temp be if it was 9" or 10" of foam? Given my design and concern with thermal comfort, at what point (per your quote in the post) does it become silly to add more polyiso?

      And the Gutex link is fascinating even if a bit fluffy!

    2. James Boris | | #3

      Hi Dana, just wanted to check in on that third question in my reply -- would truly appreciate it! I feel like I have this one shot to get it right, so if adding another few inches will make a significant difference, I'll gladly do it. Just trying to understand the math. Thanks!

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |