Planning a new house in Climate Zone 6, Chad Kotlarz is reviewing his architect’s plans for the roof — and discovers he has a few misgivings.
The unvented roof will be framed with 2×12 rafters, sheathed with plywood and capped with standing-seam metal roofing. Closed-cell spray foam will insulate the rafter bays, and the interior of the cathedral ceiling will be finished with gypsum drywall. An exposed truss with a collar tie provides structural support.
“The main concern is with the thermal bridging from the 2x12s,” Kotlarz writes in Q&A post at Green Building Advisor. “The roof is a 10/12 pitch, with a dormer and two valleys. I was told that the thermal bridging created from the 2x12s would cause hot/cold spots in the roof and would lead to problems with the metal roof. I’m assuming this is from expansion/contraction or condensation, but I’m not certain.”
For reasons he doesn’t explain, Kotlarz doesn’t want to use rigid foam on top of the roof sheathing or on the interior, so those approaches to thermal bridge reduction won’t work.
Will thermal bridging through the rafters cause problems over time, such as rippling in the metal roofing or condensation in the roof framing?
Those are the questions for this Q&A Spotlight.
Consider using scissor trusses
Stephen Sheehy, who built a well-insulated house with a standing-seam roof in Maine, suggests that Kotlarz look at a different way of framing the roof. Sheehy used raised-heel scissor trusses that were deep enough to allow about 20 inches of cellulose plus a ventilation space.
Raised-heel, or energy, trusses are extra deep at the building perimeter to provide more room for insulation than a conventional truss. “A nice benefit to trusses,” Sheehy adds, “is the ability to put interior partitions anywhere you want, since the trusses carry the entire roof load.”
Without foam above the roof sheathing, Sheehy adds, Kotlarz’s roof will see some thermal bridging.
Scissor trusses aren’t an option in this case, but Kotlarz’s research has uncovered a way of creating what amounts of a parallel chord truss by gusseting 2×3 or 2x4s to the bottom of the rafters with strips of plywood. That would create a layer of insulation between the rafters and the interior.
That plan seems similar to an approach used by Dan Kolbert, replies Malcolm Taylor, explored in a 2009 article in The Journal of Light Construction.
Another option that Kotlarz suggests would be to use 2×10 rafters in a staggered pattern to allow some insulation above and below every other rafter.
Spray foam will not affect the roofing, but has a drawback
How a roof is insulated has no effect on whether the metal roof will have rippling (also called “oil canning”), says Chris M. Factors that are important, he adds, include the gauge of metal used in the roofing, whether the panels have been clipped together in such a way to allow movement from thermal expansion and contraction, the length of the panels, and the experience of the installer.
But according to Dana Dorsett, Kotlarz’s plans to use about 10 inches of spray foam is the “opposite of green,” because of the high polymer content of the insulation and the HFC 245fa blowing agent.
“In a Zone 6 climate, with closed-cell foam under the roof deck and fiber insulation under the foam, a minimum of half the total R-value would have to be closed-cell foam to avoid moisture accumulation at the foam/fiber boundary over a winter,” Dorsett says, referring Kotlarz to a section of the 2012 International Residential Code. (For more information on the required ratios between the foam insulation later and the fibrous insulation layer in this type of stack-up, see “Combining Exterior Rigid Foam With Fluffy Insulation.”)
“If you install 2x3s perpendicular to the rafters, ‘Mooney wall’ style, that would yield a total cavity depth of 14 inches, and there would be an R-10-ish thermal break over the rafters,” Dorsett continues. “If you installed 5 inches of closed-cell foam and 9 inches of cellulose, you would be at about R-60 to R-65, and would have sufficient dew point control for the foam/fiber boundary to be fine with a Class-III vapor retarder (standard latex paint on gypsum board) on the interior. Cellulose is able to store and redistribute whatever moisture accumulates without damage or loss of function, even if the ratio is slightly off.”
Kotlarz could get comparable thermal performance in that assembly with 5 inches of closed-cell foam and 9 inches of half-pound density open-cell foam. Although that combination would still mean a lot of foam and blowing agent, it would be cheaper and work better than a full 10 inches of closed-cell foam.
Installing rigid foam above the roof sheathing is always the best solution, says GBA senior editor Martin Holladay, but if Kotlarz is totally against that, the next best plan would be to use a combination of spray foam against the roof sheathing supplemented by “fluffy insulation” — cellulose, mineral wool, or dense fiberglass batts.
Is thermal bridging also a rot issue?
Although thermal bridging is a bona-fide issue with the proposed roof design, rot won’t be a concern, says Charlie Sullivan.
“Will 2×12 bridging lead to mold or rot? I don’t think so,” he writes. “The thermal bridging in this case is mostly a heat loss issue, not a durability issue. You might be worried about the problem occurring on the surface of the drywall inside, if the thermal bridging was bad enough to make that surface cold. That is not a concern — the wood is about R-1 per inch, so 12 inches of it is about R-12. You’d need to be a lot worse than that to have condensation there.”
The other spot where condensation is a possible problem, Sullivan adds, is at the top of the cellulose-filled rafter bay where rafters poke out of the foam. “There,” he says, “you have a shorter bridge through the foam to the cold outside, so you might worry about it, but there are two things that mitigate that. One is that you also have a thermal bridge to the inside, helping to keep that spot warm. Also, wood is vapor-permeable, so it can dry to the outside a bit.”
Sullivan goes on to outline one more option, a combination of open- and closed-cell spray foam.
“An idea that I haven’t seen done before might be good for an application like this — and I’m curious to see what Dana and others think of this,” he says. “Instead of 5 inches of closed-cell spray polyurethane foam (ccSPF), how about 4 inches of open-cell spray foam up against the roof decking, followed by 2 inches of ccSPF sprayed onto the open-cell? That reduces cost, global warming impact, and polymer content, while still providing the vapor barrier at the point where we transition from fluff to foam.”
Our expert’s opinion
Peter Yost, GBA’s technical director, had this to add:
I have never seen or heard of standing-seam metal roof buckling because of thermal bridging. There is plenty of room for thermal contraction and expansion built into the standing-seam panel assemblies to deal with much more significant differential heating from the sun than what is conducted by 2x12s.
Some other points:
Type of closed-cell spray foam. Several builders and insulators from around the country have told me LaPolla Foamlok 4000 spray foam with the new Honeywell HFO, low-global warming potential blowing agent is working out just fine. However, there still is the issue of brominated flame retardants.
Lack of drying potential. With a borderline Class I vapor retarder insulation sprayed against the underside of the structural roof sheathing and standing seam metal roof topside, this assembly has little to no drying potential should it get wet. (We’re not concerned here with condensation, but with bulk water leaks over time.)
A best practice for increasing the drying potential is one of these two approaches:
Create a vent space from the soffit to the ridge below the roof sheathing of at least 1 inch in depth. Make up this loss of R-value by installing continuous rigid insulation on the bottom (interior ) of the rafters. This also will reduce thermal bridging, although I have never seen significant thermal bridging across the depth of 2x12s when each is sandwiched in with airtight cavity insulation.
Or, create a vent space (soffit to ridge) on the topside of the roof sheathing by installing 2x4s and another layer of sheathing over the structural sheathing. Run the 2x4s in the same direction as the roofing panels will go, then add a second layer of sheathing as a nailbase. Space the 2x4s so the seams of the roofing panels land on them.
Of course, the second approach is quite a bit more expensive than the first, and involves quite a bit of reconfiguring of eave and gable trim. But frankly, if you can afford standing seam and closed-cell spray foam you should be able to afford the best practice for long-term durability and drying potential.
Thanks to my good friend Brian Knowles at Jancewicz & Son for talking through these options with me.
Reversing the open- and closed-cell spray foam. I also wanted to check this issue with a field expert, in this case Henri Fennell of H C Fennell Consulting, LLC. Henri agreed the main problem with trapping the open-cell foam between the structural sheathing and the closed-cell spray foam is the hydric capacity of the open cell foam. Should there be a bulk water leak, the open-cell foam will absorb quite a bit of moisture, aggravating the lack of drying potential of the overall assembly. Henri did not mention any chemical compatibility or drying/setting time requirements there might be when sequencing two different spray foam applications but I am sure he would agree that checking in with both product manufacturers would be the best practice, regardless of how the insulation is layered.