0 Helpful?

I-joist cathedral roof / ceiling insulation and moisture control


I'm new here, and I really appreciate all the great information that has been provided by GBA and its users.

I've been reading a lot about cathedral ceiling insulation, and it seems like it is a fine line between success and disaster when it comes to moisture control. So now I have a few questions myself that I hope someone could help me answer.

We just bought a house that is going to get a new roof, and we're thinking about using I-joists for the rafters, and a combination of a combination of XPS rigid foam and closed cell foam spray as insulation. The roof will also be vented (even thought many seem to go with a unvented roof when using closed cell foam spray). See the sketch here: http://oi61.tinypic.com/33z4en6.jpg.

With the I-joists at 24" o.c., our idea is to glue/silicone (?) the XPS rigid board on each of the I-joist flanges, forming a vent channel above it (see sketch). Then using foam spray for the remainder of the I-joists depth, and having another XPS board between the drywall and the I-Joists to minimize thermal bridging.

The vent channel will be 1" to 1.5", XPS board below the vent will be 1" to 1.5", foam layer 8" to 11", and the XPS board above the drywall will be 1.5" to 2".

Does this seem like a good approach (from an insulation and moisture point of view)? Should we put any kind of vapor barriers between the inner XPS board and the foam (or between the drywall and the XPS board)?

Maybe I'm overly concerned about this, but I just like to minimize any risk of having to redo the roof in the near future due to moisture issues.


I-Joist_cathedral_roof_insulation.jpg182.53 KB
Asked by Chris Jensen
Posted Aug 18, 2014 6:21 PM ET
Edited Aug 19, 2014 9:17 AM ET


12 Answers

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Your proposed assembly would probably work, but it has at least three disadvantages.

One: it is expensive, because you want to use a thick layer of closed-cell spray foam.

Two: It isn't particularly environmentally friendly, because the blowing agents used to make closed-cell spray foam have a very high global warming potential.

Three: It includes three layers of vapor-impermeable foam, making it difficult for the assembly to dry out if it ever gets wet.

If I were you, I would design an assembly that can dry in at least one direction. Here's one way to do it: use 1/4-inch plywood for your site-built ventilation baffles. Install a continuous layer of rigid foam on the interior side of your rafters, as you propose, followed by furring strips (16 inches o.c.). And then blow in cellulose insulation to fill the rafter bays. (The cellulose can be installed through holes punched through the rigid foam).

Answered by Martin Holladay
Posted Aug 19, 2014 8:23 AM ET


Martin has it right. That much closed cell polyurethane would never recover in energy-use carbon the total amount of global warming of the HFC245fa blowing agent (which has a global warming potential about 1000x that of CO2), even if you were heating the place with a coal stove.

The HFC134a use for blowing the XPS is worth about 1400x CO2. And as it leaks out over the next 50-75 years it's R-value drops to that of EPS of the same density.

EPS and polyisocyanurate are both blown with pentane (about 7x CO2). If you replaced the interior layer of XPS with polyiso of equal thickness it would outperform the XPS today and every day for the next 200 years (or whenever you gut the place and start over.) What's more, in the event of a fire it won't melt into a puddle and spread, the way polystyrene does.

Foil faced polyiso would also be a very powerful vapor barrier, so you'd be have to set it up for outward drying only, but that's easily done either with Martin's cellulose + 1/4" plywood suggestion, or 1.5" of unfaced Type-II EPS, either of which has a vapor permeance of about 2 perms not super vapor-open, not super tight, the low end of Class-III vapor retardency. If you want a slightly higher R value, using 1.8lb density new-school fiberglass (Optima, Spider, L77) will outperform cellulose ever so slightly, but usually at up-charge. (he thermal mass of the cellulose gives it a modest advantage during the cooling season, the slightly higher R of the high-density fiberglass gives it a modest advantage during the heating season, but it's really "in the noise". Cellulose is more ant & termite resistant due to the borate fire retardents used, and manufactured from predominantly recycled/reclaimed feedstocks, has a lower overall environmental footprint than fiberglass.

Answered by Dana Dorsett
Posted Aug 19, 2014 11:19 AM ET


The vent channel is better off leaky. Just attach it with no caulk.

Answered by aj builder, Upstate NY Zone 6a
Posted Aug 19, 2014 12:35 PM ET


What is the building science explanation for your advice? I think you are confusing air leaks with vapor diffusion.

Answered by Martin Holladay
Posted Aug 19, 2014 1:30 PM ET


No, I am being practical. We use dense glass batts 16" on center with no caulk for vent channels. The key is the rigid foam continuous below sealed well to be the air barrier. I can tell with roof snow melt that the roofs we do have no snow melt. Energy use is low, inside stratified air is less to nill.

And yes just nailing it in place might lead to some useful exchange, but very very little because no air is getting past the rigid.

This site gets carried away with perfect and perfect is not right often, something a bit less is where the curve falls off.

Time to plot some calculus curves? NO.

Build, learn, refine. my roofs are dang good roofs.

Answered by aj builder, Upstate NY Zone 6a
Posted Aug 19, 2014 7:30 PM ET


Wow, a lot of good information, I really appreciate it, thanks!

So if I understand correct with you guys' suggestion; with plywood and cellulose (see attached sketch) instead of xps and spray foam, respectively, any moisture that gets caught in the wall will vent to the outside (at least that is the intention). With that configuration should I try to block any moisture to leak into the wall from the inside by having a vapor barrier between the drywall and the xps or polyisocyanurate (house wrap or maybe tape the seams of the xps/polyisocyanurate plates)?

Also, just to make sure I understand the closed cell foam spray insulation configuration. If I would decide to use closed cell foam spray, would it make sense to allow moisture to vent on both sides of the closed cell foam spray (since the foam spray insulation would act as a vapor barrier)? If so, how do I reduce thermal bridging if I shouldn't use rigid foam boards (as they would act as a vapor barrier) between the drywall and the rafter flanges?

Thanks again! You guys are very helpful!

Answered by Chris Jensen
Posted Aug 20, 2014 10:14 AM ET


Q. "Should I try to block any moisture to leak into the wall [ceiling assembly?] from the inside by having a vapor barrier between the drywall and the xps or polyisocyanurate (house wrap, or maybe tape the seams of the xps/polyisocyanurate plates)?"

A. If you have rigid foam on the interior side of this assembly, then the rigid foam is already a vapor retarder. In any case, your suggestion of using housewrap as a "vapor barrier" is misguided, since housewrap is vapor-permeable. Taping the seams of the rigid foam makes sense, but this measure has nothing to do with vapor diffusion; the point of the tape is to limit air leakage.

Q. "If I would decide to use closed cell foam spray, would it make sense to allow moisture to vent on both sides of the closed cell foam spray (since the foam spray insulation would act as a vapor barrier)?"

A. Closed-cell spray foam is both an air barrier and a vapor barrier. If you use vapor-permeable materials on either side of the closed-cell foam, diffusion will happen whether you want it to or not. You don't have to do anything special.

Q. "How do I reduce thermal bridging if I shouldn't use rigid foam boards (as they would act as a vapor barrier) between the drywall and the rafter flanges?"

A. With this type of vented assembly, the only way to reduce thermal bridging through the rafters is to include a continuous layer of rigid foam on the interior side of the rafters. I don't recommend combining rigid foam with closed-cell foam, however.

Answered by Martin Holladay
Posted Aug 20, 2014 11:14 AM ET


Good thinking here. Use dense packed cellulose and rid yourself of these technical arguments, construction details, expenses and risks. Here's how it works:


Answered by flitch plate
Posted Aug 20, 2014 11:46 AM ET


You need an air barrier. You don't need ten air barriers.

Foam is expensive. Inexpensive insulation works very well when there are proper air barriers.

Continuous insulation used as an air barrier on the inside is good because it does two jobs nicely. It insulates and it is an air barrier. And it is in a location where it is not likely to have moisture and rot issues. Example, put a little rigid foam outside rot loving cheap OSB. I have run into this set up with the OSB rotted enough to think it was garden mulch. The slight gap between the layers sucks in water or has water and then the water stays there not drying out and you have a nice pile of soil instead of a sheet of OSB.

Dreaming up walls and ceilings from reading internet posts is a crap shot in the dark. This site if joined has tab with details. Building Science has details. They have books with details to purchase.

Going to a free site to build a $500,000 home is nuts.

The lowest cost way to produce a set of plans for a superinsulated home is to go with a plan set that is proven and modify the shape to your liking.

Experimenting is great but messing up superinsulating home builds is easy even for Joe L. who has put more effort into understanding this subject than all the rest of us combined.

Answered by aj builder, Upstate NY Zone 6a
Posted Aug 20, 2014 12:11 PM ET
Edited Aug 20, 2014 12:12 PM ET.


I can't fathom why anybody would use 8-10" of closed cell foam in any assembly. Using that much closed cell foam is still going to be a crime against the environment, and would take a long time to install to do it correctly.

Closed cell foam needs to be applied in lifts of no more than 2" at a time to avoid shrinkage & cracking. Spraying in lifts thicker than 2" comes with both higher air leakage and a significant fire risk as it cures (self-igniting several minutes or even several hours after it it applied.) That, and the very low vapor permeance at thicknesses greater than 4" make it a lousy choice for deep cavity fill, independent of its $/R.

Using foil faced polyiso on the interior and taping the seams with FSK tape (2" aluminum duct tape, purpose made for applying to foil facers or shiny-clean galvanized steel) will outperform XPS at any thickness, and it's usually cheaper than XPS at any R-value. And it's manufacture does less than 1% of the climate damage of XPS. Housewrap tapes sort-of work on XPS, but it's not purpose-made for the application, and the long-term tightness in that application isn't fully known. Foil facers are true vapor barriers, Class-I vapor retarders of 0.05 perms or less, which is just fine on the interior side of the assembly, as long as it's vented above the rest of the insulation.

Answered by Dana Dorsett
Posted Aug 20, 2014 3:38 PM ET


Just to be clear; I'm not asking to experiment, I'm just doing some due diligence before I spend my money on having someone to do it. I could of course ask a local contractor and take his words for granted, and hope that it works out ok, or I can do some research first so I'm able to ask the contractor the right questions to make sure he knows what he is doing (and I understand what he is doing). If I knew that all contractors were as knowledgeable as you guys, I probably wouldn't be asking here.

Seems like cellulose + polyiso is a good way to go, but it raises a couple of other questions:

1) For a cathedral ceiling, R-60 equals about 15" of cellulose (based on R3.7/inch) + 1" of polyiso. BUT what if you would want to go higher? Let's say R-90. I know I-joist have a maximum depth of 16" for roofs with a pitch higher than 3:12, so with an 1" vent channel, you would be limited to R-55 with the cellulose (actually it would be a little higher if you include the roof pitch into the calculation). A thicker layer of polyiso could be added, which would get you to maybe R-75 (3.5").

2) What about the drywall screws that penetrates whatever layer of vapor barrier you have, like either the aluminum tape or the polyiso foil, moisture could potential leaks through those holes. Is that something should I should be concerned about? I guess for a vented roof it might not be that big of an issue, but for a unvented roof I could see that being a potential problem...


Answered by Chris Jensen
Posted Aug 21, 2014 7:23 PM ET
Edited Aug 21, 2014 7:30 PM ET.


Keep it simple.

If the roof top is yet to go on, put the XPS or polyiso on top of the I-J's top cords or put a deck there and foam board on top of it. I prefer the rigid foam on top of the I-J's with the deck above that. Rigid foam under the I-J's (interior) is uncomfortable for all kinds of obvious reasons; some of which you mention. Make the thickness of the extruded foam plus the rafter's depth with DP cell sufficient for R 75 or greater. Air-seal the heck out of the bottom side of the deck/foam panels (which ever you put first) and also tape the top side seams.


Use wind-devil screws with caps to attach the rigid foam to the I-J's or the deck. If you use a paneled roof top such as a steel or galvalume, you can lay them on top of the rigid foam. But I like purlins first. All of this takes some mighty long and high quality screws. The purlins can carry the load and save having to penetrate the final roof panels with long screws and accurate chaulk lines. Purlins allows for hidden fastener systems. If you use asphalt, you need to put the deck over the rigid foam; or use a second deck.

This sequence alleviates the need for venting and experimenting and fussing on the in-door’s ceiling.

PS. There is no vapor barrier needed in this system. The DP cell dries to the inside through the drywall. Use low perm paint only; not an airtight drywall system. No pot lights or other unnecessary penetrations in the ceiling; mount everything on the surface.

Applegate Industries' spec sheet for DP cell cathedral ceilings will satisfy the code requirements.

Answered by flitch plate
Posted Aug 21, 2014 9:51 PM ET
Edited Aug 21, 2014 10:00 PM ET.

Other Questions in Energy efficiency and durability

Need some unbiased input on these 2 proposals

In General questions | Asked by milhouse21386 | Jun 20, 18

Building code requirements for stacked window headers, bracing, and stucco over foam

In Building Code Questions | Asked by Zane Bridgers | Jun 17, 18

Carbon monoxide detectors in a fuel-free house?

In Mechanicals | Asked by Trevor Lambert | Jun 20, 18

Do I need an HRV/ERV for a house built in 1947?

In Mechanicals | Asked by Will82 | Jun 14, 18


In Green building techniques | Asked by user-7043474 | Jun 20, 18
Register for a free account and join the conversation

Get a free account and join the conversation!
Become a GBA PRO!