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

Open-cell spray foam on roof deck?

wolverine72 | Posted in General Questions on

I could really use some advice on open cell vs. closed cell (or some other option). I have a 100+ year old home in northern NJ (zone 5 border, close to zone 4) with multiple roofs, hips, valleys, etc. Part of the roof has ridge / soffit venting, a lot doesn’t. House is already updated with a new roof (wish I would’ve thought about adding 4″ above before we had it done). Existing insulation ranges from ok to awful (a few inches of fiberglass – maybe r 15) to god awful (r-15 with reflective backing directly against roof, no ventilation, some missing/falling down). Nothing is air sealed. I have 2 hvac units and duct work in the attics. 3rd floor is hot in the summer and cool in the winter (with the units running a lot). After reading endlessly, it sounds like I need to convert to an unvented attic…

So I called a few local Pros to come take a look at it while I researched DIY kits (would rather pay an expert to do the work if I can). And all of them want to do open cell spray foam on the majority, if not all of the roof deck. The reasons vary – some of the comments are ‘tough spots to get into and they are worried about the closed cell heating up’, ‘open cell is fine to use and cheaper as we’ll put 8″ on there to get you to R-30’, ‘moisture on the roof deck is not an issue since the attic will be within ~10 degrees of the house and the extra thick open cell will air seal’.

So this is where I could use advice – closed cell vs. open cell on the attic roof deck to create an unvented attic? I’ve read enough that it sounds like closed cell is the preferred way to go (at least 3 inches, probably more to get higher R-value). But the ‘pros’ are saying they’d use open cell foam.

Any suggestions? Additional questions I should ask them? Something I might be missing? Maybe I just need to go back to looking at DIY options and spray the attic myself with one of the many kits out there (I’ll probably need 10+). I read both of these among the various articles and discussions:

https://www.greenbuildingadvisor.com/blogs/dept/qa-spotlight/pondering-attic-conversion
https://www.greenbuildingadvisor.com/blogs/dept/musings/open-cell-spray-foam-and-damp-roof-sheathing?page=1

Thanks in advance!

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.

Replies

  1. Expert Member
    Dana Dorsett | | #1

    An all open cell solution would be a code violation due to the vapor permeance issues, if that matters to you. Even though it's air-tight, open cell foam is still pretty vapor open at R30. Like thicker layers of HFC blown closed cell foam, at R30 most open cell foam is a fire hazard during the curing period right after installation. To do 8" safely it has to be done in two lifts, with a cooling/curing period between lifts. Many installers will cheat a bit on that and spray 8" in one shot, and while the risk is low, it's non-zero, but it's also becomes a quality issue.

    At 2" all closed cell polyurethane is a class-II vapor retarder, and sufficient to protect the roof deck, which should run ~$2 per square foot if done by a pro. By keeping the ratio of closed-cell to fiber insulation bounded you can do the rest with cheaper fiber. (It may be cheaper to have a pro do the fiber it than as a DIY, so at least get it bid.)

    HFO blown closed cell foams run about R14 @ 2", and don't present the same fire-hazard during curing that DIY kits or HFC blown goods from the pros.

    If you have 8" to deal with, 2" of closed cell foam (R12-R14) on the under side of the roof deck would be sufficient dew point control for another R18-R21 using cheap fluffy stuff if in zone 5, or R28-R32 in zone 4.

    So if it's 2x8 rafters, 2" of ccSPF + R21 "cathedral ceiling" fiberglass or R23 rock wool would be about right for the warm edge of zone 5.

    If they are 2x10s you can go with 2" closed cell + R25 fiberglass at the warm edge of zone 5, or R28-R30 rock wool in zone 4.

  2. GBA Editor
    Martin Holladay | | #2

    Wolverine,
    When you write about installing open-cell spray foam "on" the roof sheathing, I assume that you really mean under the roof sheathing.

    You provided a link to Open-Cell Spray Foam and Damp Roof Sheathing. In addition to that article, I suggest that you also read High Humidity in Unvented Conditioned Attics.

    For the reasons given in those two articles, I don't recommend the installation of open-cell spray foam on the underside of your roof sheathing. Closed-cell spray foam is far safer.

    As Dana noted, if you want to meet minimum code requirements for R-value, it usually makes sense to combine closed-cell spray foam with some type of fluffy insulation. The fluffy insulation needs to be installed under (and in direct contact with) the cured spray foam. For more details, see Creating a Conditioned Attic.

  3. Expert Member
    Dana Dorsett | | #3

    It's unlikely that he'd be able to meet the letter of IRC 2015 for zone 5 with insulation between the rafters unless it was 8" of closed cell foam, which is ridiculously expensive and underperforming on a U-factor basis due to the thermal bridging of the low-R rafters through the high-R foam.

    From a cost effectiveness point of view it's unlikely to make economic sense to do much beyond the minimum amount of closed cell foam to be protective with the remainder of the rafter bay filled with fluff, now that the opportunity for installing rigid insulation above the roof deck has been put off for another 20-25 years. The contractor's statement "...open cell is fine to use and cheaper as we'll put 8..." on there to get you to R-30" implies that it's probably 2 x 8 rafters, maybe 2x10s, which is why I ran suggested stackups for both.

    With ~R12-R14 foam + ~R23-R25 fiber it's still a decent R35+ insulation level at a reasonably low price point. Taking it to R49 at a $8 per square foot won't come close to paying off the difference in reduced energy use over 20 years. When it's time to re-shingle, adding another R10-R15 above the roof deck would be possible, and would raise the performance to IRC 2015 code minimum levels or better.

    With only 2" of closed cell foam the roof deck is protected from peak wintertime moisture drives, and the roof deck still has a reasonable drying rate toward the interior. Any more than that is a waste of high cost high environmental impact foam, since it's potential thermal performance is being robbed by the thermal bridging of the rafters. You only need it for the vapor retardency and dew point control at the foam/fiber boundary.

  4. wolverine72 | | #4

    Dana, Martin - thanks. I did mean under the roof sheathing. Ideally I'd like to get near to code as possible, which I believe is R-38 by me. It will make a world of a difference comfort wise in the entire house, not just the 3rd floor, by stopping the air movement and insulating better. Here's some more info, if that helps. I have 3 distinct 'attic' spaces essentially.

    Section 1 has 2x10 rafters, accessed through a 24"x36" door. Easy enough to spray 2-3" of closed cell foam and add either fiberglass or mineral wool on top of that. Total R value combined could easily top R-30. The ceiling below this section is insulated already (underneath a plywood floor), so that has to stay in place.
    Section 2 has 2x5 rafters (yes 100+ years old), accessed by a small 16"x24" hatch in the ceiling. It's not used for storage so the insulation could extend beyond the rafters. Total R-value combined would likely be ~R30 here given the rafter depth. The existing insulation on the ceiling would be removed.
    Section 3 has the same 2x5 rafters, accessed by a standard door (walk in), is loaded with AC ducts and the main HVAC unit, and has hips and valleys. Some of the ducts are secured to the rafters, but I believe those can be shifted out as needed to accommodate more insulation (ideally mineral wool here for the fire properties). Securing the additional insulation could be a bit of a challenge I'm sure, especially in the valley section. There is some insulation in attic floor, covered up by an old floor and plywood over that.

    So sounds like you both are recommending I stick with closed cell - at least 2". To get the proper R-value (or close to it), and keep the costs down, I should add additional insulation over the foam (fiberglass or mineral wool).

  5. GBA Editor
    Martin Holladay | | #5

    Wolverine,
    You wrote, "to keep the costs down, I should add additional insulation over the foam (fiberglass or mineral wool)."

    I'm a stickler for accuracy: the additional insulation goes under the spray foam, not over the spray foam.

    Dana and I sometimes make different recommendations for the minimum R-value of the spray foam layer in this type of stack-up. The way I read the code, in Climate Zone 5, the spray foam layer must have a minimum R-value of R-20 (meaning at least 3 inches of closed-cell spray foam).

  6. Expert Member
    Dana Dorsett | | #6

    "...stick with closed cell - at least 2"..."

    Not least- MOST. I'm recommending at MOST 2" of closed cell on the underside, for cost, drying rate, and general verditude reasons!

    For 2x5 rafters the rafter bay widths may not fit well and an all foam solution is probably the only realistic quick solution. There's still NO point to installing any more than 2" of closed cell foam between the rafters, followed by 4" of open cell foam (encapsulating the rafter edges.) Competent open cell foam installers shoot acres of 4" foam in 2x4 framing every year, and can gauge it well. You'd be at about R28-R30 at center cavity, but you'd also have an R3-R4-ish thermal break over the rafters, outperforming R30 between rafters.

    A more major re-work of that 5" raftered section,could be to install 2" of closed cell foam, on the roof deck, then 3" of open cell trimmed flush with the rafters, and a continuous layer of 2" rigid rock wool or unfaced EPS strapped in place with 1x4 furring through screwed to the rafters, with gypsum board mounted on the furring. That would come close to meeting code-min performance on a U-factor basis (due to the R8 thermal break and additional air films), but it's a major PITA to assemble, especially while wrestling with a duct-medusa.

    If you're not going for the full R49, there's no need to go for the full R20 of closed cell foam. The temperature at the foam/fiber interface is the what's critical, and that is largely a function of the ratio of R-values. With less total R, you can get away with less closed cell foam R and still keep the average temperature at the interface above the average wintertime indoor dew point.

    If you were to install MORE than code min, say a total of say, R75, a mere R20 of closed cell foam wouldn't be adequate, and the fiber insulation risks becoming wet enough on the cold side to cause problems.

    There are plenty of existence proofs and well vetted WUFI simulated new designs that would demonstrate that R20 isn't necessary at the warm edge of zone 5 even at R49. It's not like the physics of water or the climate changes much from the cold edge of zone 4 where a mere R15 of closed cell meets code, at R49 total-R.

  7. wolverine72 | | #7

    Martin, correct. I should say the additional insulation would go underneath the foam, and the foam goes directly underneath the roof sheathing. I found the section of code you mention which calls for minimum foam R-20 (806.4 5.3) for zone 5. I had seen it when I was looking at cut and cobble as an option (4" of xps).

    Dana - I understand what you are suggesting. I'm right near the border of zone 5, and since I'm not going to get to the full R49, no need to go the full 3" of spray foam.

    Listening to the suggestions, sounds like the best solution is to spray 2", get to R14, and then likely add mineral wool (between R15-R30 additional depending on the space). Is there any value in spraying the rafter itself to encapsulate it and minimize thermal bridging? I have a feeling this is moving towards a DIY solution with all of the Pros in my area only wanting to do open cell.

    Thanks!

  8. GBA Editor
    Martin Holladay | | #8

    Wolverine,
    Q. "Is there any value in spraying the rafter itself to encapsulate it and minimize thermal bridging?"

    A. Yes. Encapsulating the rafters is a good idea, if you have the budget and the patience. (And by the way -- assuming that closed-cell spray foam is R-7 per inch is a little optimistic. R-6.5 per inch is more likely.)

  9. wolverine72 | | #9

    "For 2x5 rafters the rafter bay widths may not fit well and an all foam solution is probably the only realistic quick solution. There's still NO point to installing any more than 2" of closed cell foam between the rafters, followed by 4" of open cell foam (encapsulating the rafter edges.) Competent open cell foam installers shoot acres of 4" foam in 2x4 framing every year, and can gauge it well. You'd be at about R28-R30 at center cavity, but you'd also have an R3-R4-ish thermal break over the rafters, outperforming R30 between rafters."

    Dana - you are spot on. I had the chance to crawl around in "section 3" this weekend. Fortunately (or unfortunately) I found the source of a leak into the BR below. It's coming from a part of the roof and not the HVAC unit condensation as thought. The fiberglass bats with foil backing held the water in, moving it far from the point of origin. So that needs to be fixed, then has to dry out before it gets touched.

    I say you are spot on because the roof rafters are something like 32" on center, some more, some less. The current R-19 foil backed fiberglass bats are held in place by duct tape. Not every joist bay has air baffles in it, and none of the baffles go all the way to the ridge, which doesn't have a vent anyway. And some areas of the roof have no insulation. The only way this is getting addressed properly is via foam. And every pro so far will only do open cell on my roof, so I'm headed towards DIY here if I want closed cell.

    Would it make any sense to use the "peanut brittle" approach that Martin has talked about? I could fairly easily put in 2" of polyiso or xps and spray 1-2" closed cell under that in a large chunk of the area (very little waste on a 4x8 sheet with 30" wide rafter bay). That would completely air seal and get me in the ~R-16 to R-19 range at 1" sprayed under or ~R-23 to R-25 at 2" sprayed under the foam board. If I can, I'll cover some of the joists to address thermal bridging (some of the joists have ducts hanging on them). I'm guessing Dana might say no as that puts too much closed cell in the space? The other concern with that method on this part of the roof is that the sheathing is actually not attached directly to the rafter. it's 100+ years old so there is wood slats across the rafters from the original wood roof. The sheathing is on top of that. Even though it would be sealed from attic air, the foam board/spray foam approach would leave a horizontal air space across the rafters between the foam and the sheathing. Or maybe that's a good thing? Look forward to hearing from you and thanks again for taking the time to share advice!

  10. GBA Editor
    Martin Holladay | | #10

    Wolverine,
    I tend to be conservative when it comes to cathedral ceiling details. There are lots of examples of failed roof assembles with rotten sheathing.

    Because of reports of damp sheathing associated with unvented cut-and-cobble jobs on cathedral ceilings, I don't recommend cut-and-cobble of any flavor -- including the so-called "peanut brittle" method -- for unvented cathedral ceilings.

    If you can manage to establish a real vent space -- with a soffit-to-ridge vent channel between the rigid foam and the underside of the roof sheathing -- the cut-and-cobble approach is worth considering.

  11. wolverine72 | | #11

    Martin, points well taken. In section 3 of the roof I'll stick with spray only. If I can do it, would the preferred method be to put in a soffit to ridge vent channel? I'd lean towards using polyiso if I did this.

    In sections 1 and 2 I should be able to create a 2" soffit-to-ridge vent channel between rigid foam and underside of roof sheathing (in the area where there is a ridge vent). It's a little more work but I'd get to an ~R-23 to R-25, completely air sealed with closed cell. In section 2 where there's only 5" rafters I'd likely be coming over the rafter and addressing thermal bridging as well.
    Thanks!

    1. charlieXu | | #12

      Hi wolverine72

      Wondering what did you finally use for your roof deck insulation. We have almost the exact same situations here and I am concerning that if I use closed cell, any leak from the roof or moisture which makes it underside the shingles would have no way to dry out. I think open cell would make better sense because we are still in zone 4 in which vapor retarder is not required , so moisture can dry out toward interior. But I do see most people are toward closed cell to prevent interior vapor coming in. How has your solution hold up so far?

Log in or create an account to post an answer.

Community

Recent Questions and Replies

  • |
  • |
  • |
  • |