GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X 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

What’s the best way to implement polyiso insulation in an old attic about to be converted to living space?

TimGNash | Posted in Energy Efficiency and Durability on

I’m seeing conflicting reports regarding whether moisture barrier should be used for a finished attic roof, or whether the attic should remain vented. I suspect the answer depends on what is done on the exterior side of the roof sheathing (and I don’t plan on touching the outside of the roof myself).

My present plan is to install two 2″ layers of polyiso rigid insulation with seams both staggered and sealed, and then probably another layer on the drop ceiling and poy walls. For the roof itself, is the best method to attach the polyiso to the underside of the rafters and allow the actual rafter space to continue to provide venting? And if I do use that method (assuming my foil side is facing to the interior), should I use a vapor barrier prior to installing the polyiso?

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. TimGNash | | #1

    I probably should have added a few more details:

    * there is existing roof venting
    * the rafters are simple 2x6"
    * the attic roof slopes all the way down to the eaves
    * the plan for the side walls would be to start them at around 4'8" or so (a bit higher in the bathroom, so people don't bump heads when leaning into the sink), and back the framing with 2" of rigid foam. I could also put a bit of fiberglass batt insulation inside the wall framing to up the ante a bit more
    * the roof center line is approx 9'7" above height of the finished floor. If I'm going to retain the top venting, I would do a drop of about 1 foot and then install my first installation layer
    * hoping to still have an 8 foot drop ceiling with a bit of additional insulation. In between those (admittedly close) layers, run my mini-split lineset; probably install a ceiling cassette for the air handler

    Some of this info may or may not be germane, but hopefully it will be more clear what I'm dealing with, and what I'm attempting.

  2. GBA Editor
    Martin Holladay | | #2

    These two articles should answer your questions:

    How to Build an Insulated Cathedral Ceiling

    Two Ways to Insulate Attic Kneewalls

  3. TimGNash | | #3

    Thank you. That's helpful, although I'm still not 100% sure on implementation. The cathedral ceiling article a couple of times appears to equate the principles for spray (closed cell) foam and rigid foam, but it's not explicit enough for me to be sure.

    I'm also concerned that I don't have enough headroom to use venting and still get the necessary amount of insulation value.

    Granted that I don't know what is on the exterior side of the roof (but given how lost-cost everything else was done, I'm guessing it's the bare minimum), it appears to me that if I want to go away from venting, my best option would be to install the rigid foam right up against the interior side of the sheathing, and not use a vapor barrier anywhere. I'm not sure that's the correct interpretation, and I'm also a bit wary because I seem to be remembering that polyiso breaks down under moist conditions compared to other types of rigid foam. Does that sound right?

  4. GBA Editor
    Martin Holladay | | #4

    You're talking about the cut-and-cobble method. The cut-and-cobble method is not recommended for unvented cathedral ceilings -- it can lead to moisture problems -- but it can be used for vented cathedral ceilings.

    For more information, see Cut-and-Cobble Insulation.

  5. Expert Member
    Dana Dorsett | | #5

    Polyiso doesn't break down, but it is slightly hygroscopic, and can wick & retain liquid water if pond-soaked and takes forever to dry. Despite that potential shortcoming it is the most commonly used rigid foam for insulating above the roof deck in commercial construction ( followed by EPS.) When the roof leaks water, parts of the polyiso may become saturated, and may need to be replaced when re-roofing if it has undergone a lot of freeze/thaw cycling while wet from a roof leak, but it has to be substantial &/or chronic roof leak for that to happen. (I've seen a few sheets of frost-damaged polyiso in batches of reclaimed roofing foam over the past half-decade but it's not super-common.)

    The long term air tightness of rigid foam tight to the underside of a roof deck is dubious at best, since it relies on the long term adhesion of whatever materials have been used to seal the edges. Spray-applied foam adheres to 100% of the underside of the roof deck whereever it is applied, and even if the seal breaks in a couple of small locations no paths for substantial convection of humid interior air to a cold roof will open up. Minor faults in the air tightness with the rigid foam layer aren't a problem when the foam is above the roof deck, but can be a BIG problem if it's on the under side.

    With a vented roof minor air leaks from the interior are also not a problem, since the humid air that leaks through mixes with the slowly convecting outdoor air in the vent channel and is purged. Even with some air leaks in the foam cold roof deck is never in contact with undiluted, fully humid indoor air the way air leaks around a unvented cut'n'cobbled foam approach would be.

    [looks like Martin beat me to it...]

  6. TimGNash | | #6

    Thanks again for your thoughts.

    Dana, I suspect the moisture issue is what I was remembering.

    At this point, piecing together the material in the links and the answers here, what I'm thinking is nailing e.g. 1x1 strips 2" inside the interior edge of the rafters and adhering/gap-filling 1 layer of polyiso to that. That will leave 3 1/2" for venting in behind. Then, right over top of the interior face of the rafters, another layer of polyiso. At the peak of the roof, leave about a foot (is that sufficient?) for venting, and have three layers of 2" polyiso in the drop ceiling as well as one layer on the outer side of the knee walls — effectively, an R-value of 36-39 of polyiso alone above and below, not including the insulating value of other wall and roof elements. I'd prefer to do four 2" layers of polyiso for the ceiling, but not sure the $$ will be there, and I'm also going to be running into a headroom issue, at least for the sloped stretch of roof between the drop ceiling and the top of the knee walls.

    Does that venting and insulation situation sound reasonable?

    I'm in middle Tennessee, so climate zone 4, so I know that for the ceiling, that's barely passable, although it's above code for walls. But it will be a big improvement over the current situation. Essentially half the second floor is unfinished attic, and the only insulation in it is a mediocre batting install in the 2x8 floor joists. (To top of it off, I just discovered in the past week that the wall between the old [1950ish] house and the newer [1970ish] addition was not insulated at all.) I suspect that when I'm done, my heating and cooling will be cheaper than it was before I started, despite adding ~450 sq ft of living area.

    The other question is: what direction should the foil face on the exterior-most layer? My instinct is to turn the foil to the outside and try to avoid getting any moisture into the polyiso at all, but now I'm thinking that may make the problem worse if a bit gets in and can't get back out to the ventilated area.

  7. GBA Editor
    Martin Holladay | | #7

    Your ventilation channels don't need to be 3.5 inches deep, although there is no harm if the channels are that deep. Two inches is generous, and 1 inch is legal according to most codes.

    You wrote, "At the peak of the roof, I plan to leave about a foot (is that sufficient?) for venting." I'm not sure I understand you.

    If you are building a vented roof, the vent channels either have to run all the way up to the ridge vent, with a consistent depth (as I said, 2 inches is generous, but 1 inch will work in a pinch), or you need to extend the channels above the insulation layer into a vented attic. In either case, you need a ridge vent. Since you didn't mention the ridge vent, I thought I'd remind you of that fact.

    You don't need a 1-foot high attic. You can have no attic -- just a cathedral ceiling all the way to the ridge vent -- if you want.

  8. TimGNash | | #8

    Martin, thanks for getting back to me. Was sick for a couple days and still had to work, so just getting to see this thread again now.

    The attic is existing and already has vents in the roof as well as some sort of soffit vent on the end. I'll do some research on ridge vents to fully understand what you're saying. (EDIT: okay, that's quite different from what I have. I guess I would need to close up the existing ventilation to use that system.)

    That said, it also seems to me that a full cathedral ceiling isn't going to be as energy efficient as a drop ceiling.

    At this point, I'm beginning to think that if I do go the route of rigid foam (I'm going to get some spray foam quotes to compare cost), I'll put 4" foam from the face of the rafters inward, leaving 1 1/2" for venting. Then a 2" continuous layer over the face of the rafters would give me a really good insulation factor.

  9. TimGNash | | #9

    Okay, on the implementation level:

    1. Would I want to use a vapor barrier? I'm reading conflicting things on this.

    2. It looks like a lot of polyiso is only foil faced on one side, while other products are foil faced on both sides. In the latter case, the question would be redundant, but in a situation where there is only foil facing on one side, would I want that facing toward the exterior or the interior? My initial thought is to face the polyiso that is between the rafters to the exterior, and the continuous polyiso layer that is on the interior side of the rafters would be faced to the interior.

  10. GBA Editor
    Martin Holladay | | #10

    Q. "Would I want to use a vapor barrier?"

    A. No, you don't need a vapor barrier (although it's always important to pay attention to airtightness). If you choose to use foil-faced polyiso, you'll have a vapor barrier whether you want one or not, because foil-faced polyiso is a vapor barrier.

    Q. "In a situation where there is only foil facing on one side [of polyiso insulation], would I want that facing toward the exterior or the interior?"

    A. The foil facing provides no R-value benefit unless it faces an air space. If your assembly includes an air space adjacent to a layer of foil-faced polyiso, then the foil facing should face the air space.

  11. TimGNash | | #11

    Thanks, Martin. Sounds to me like I want to put the foil facing the exterior for the polyiso between the rafters (as that has the air space of the venting channel), and facing the interior for the continuous (as at least part of it will have the air space between the rafters and my knee walls).

  12. TimGNash | | #12

    Well, I'm revisiting this question several months later. It took a while to get my renovation money, but we've been full speed ahead for the past couple weeks.

    I found a good deal on some closed-cell spray foam and some rigid polyiso. Is there anything at all stopping me from spraying an inch layer of the spray foam right against the rafters (no ventilation layer) and then layering the polyiso on the interior side of that? (The polyiso is considerably cheaper per board foot than the spray foam.) I'm presuming that the closed cell will deal with any moisture getting inside, and then I should be good to go.

    If that approach is fine, how should I deal with the existing whirlybird style vents in the roof? Just seal them off with plastic prior to spraying the foam? (I don't really want to do a bunch of roofing just to tear out the vents.)

  13. GBA Editor
    Martin Holladay | | #13

    To avoid the risk of roof rot, your spray foam layer should meet the minimum R-value required for a flash-and-batt job (even if you aren't really doing a flash-and-batt job). Remember, the cut-and-cobble approach in an unvented roof assembly is associated with moisture accumulation and possible rot. You want the spray foam layer to be thick enough to prevent moisture accumulation.

    In Climate Zone 4, a flash-and-batt job requires a spray foam layer with a minimum R-value of R-15 (about 2.5 inches) -- so 1 inch of spray foam is risky. More information here: "Flash-and-Batt Insulation."

  14. TimGNash | | #14

    Thanks, Martin. Guess I should have ordered more spray foam, but it's significantly pricier than the polyiso, although I got a pretty good deal on both.

    Do you think it's still risky if I add moisture-barrier-rated plastic after (or perhaps prior to) spraying 1"?

  15. GBA Editor
    Martin Holladay | | #15

    I don't know what "moisture-barrier-rated plastic" is, but I'm afraid that you have a fundamental misunderstanding of the problem. The problem is that if the cured spray foam is too thin, it will be cold enough to permit condensation. Adding a layer of polyethylene doesn't warm things up, so the condensation problem is still there. You'd just have cold, wet polyethylene.

    The way to warm up the first condensing surface is to make the spray foam layer thick enough that it provides enough R-value to insulate the interior surface of the cured spray foam from the cold exterior. You need R-value, not a moisture barrier -- and the R-value has to come from a material that is a vapor barrier and an air barrier. That material is closed-cell spray foam.

  16. TimGNash | | #16

    Thanks, Martin. A couple further questions and comments.

    1. Assuming your analysis is correct, would the ventilation channel need to be continuous all the way from the eaves up to the ridge vent, or could I insulate at 2 1/2 inches thick spray foam on the lower portion of the wall, and only have the upper part of the rafters ventilated?

    2. While I defer to your expertise, your explanation doesn't quite make sense to me. I well understand what condensation is. While I now live in the Nashville area, most of my life I have lived in northern Alberta, and I have seen plenty of condensation not only on windows but on poorly insulated walls and roofs. In those cases, this is due to warm inside air meeting a surface that is cold from the exterior (of course, the same principle holds if the temperature inside is a great deal cooler than the outside too). This is generally caused either by leakage of warm air through inadequate sealing or by inadequate insulation.

    It's not at all clear to me whether you think we're talking about one of those issues here, or a different sort of cause of condensation. I will grant that if I had one inch of spray foam as my sole insulation in the rafter bays, there would indeed be a dramatic difference of temperature between the surface of the foam and the temperature in the room, resulting in condensation.

    But of course that's not what we are talking about. We're talking about one inch of spray foam, and then five inches of rigid foam (sealed with spray foam crack filler and taped off, of course) to the interior of that. The inside surface of the spray foam, in other words, would never be "seeing" the inside temperature of the room, as there would be ~ R30–35 of insulation value between the spray foam and the interior. It is not at all clear to me how there can be a condensation danger in that situation.

    Perhaps there is a roof rot danger even so, but I don't see how we can attribute it to condensation.

  17. GBA Editor
    Martin Holladay | | #17

    Q. "Would the ventilation channel need to be continuous all the way from the eaves up to the ridge vent, or could I insulate at 2 1/2 inches thick spray foam on the lower portion of the wall, and only have the upper part of the rafters ventilated?"

    A. The only way you could make your plan work is if you could figure out a way to introduce outdoor air into the rafter bays halfway up the roof, where the ventilated part of your roof assembly begins. I can't think of any easy way to do that.

    Q. "We're talking about one inch of spray foam, and then five inches of rigid foam (sealed with spray foam crack filler and taped off, of course) to the interior of that. The inside surface of the spray foam, in other words, would never be "seeing" the inside temperature of the room, as there would be ~ R30–35 of insulation value between the spray foam and the interior."

    A. You're right, of course, that this situation isn't the same as a situation where warm humid air has free access to the cold surface in question. But I'm operating on the assumption that cut-and-cobble insulation is leaky. My assumption is based on field reports of cut-and-cobble failures in roof assemblies. Roof assemblies move, due to snow loads, temperature changes, and humidity fluctuations, and these forces introduce cracks in cut-and-cobble insulation over time. The stack effect ensures that exfiltrating air will find these cracks, bringing moisture to your roof assembly.

  18. TimGNash | | #18

    Okay, thanks. That makes more sense, but I don't see why a cut-and-cobble *has* to be leaky. Obviously, doing it right will take extra effort, but that's always the case.

    I think my planned method will pretty much eliminate cracks, because I'm not just placing rigid foam inside the rafter bays (and sealing the sides with canned spray foam), but also doing a continuous horizontal layer (with taped seams) to the interior face of the rafters.

    Not saying I can't make any mistakes, but I don't really think there's going to be much air leakage with this implementation.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |