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

Closed- or Open-Cell Spray Foam for Unvented Cathedral Ceiling

kramttocs | Posted in Energy Efficiency and Durability on

I know this topic gets brought up a lot and I have found a ton of information on this site but at this point my head is spinning.

Remodeling a 1979 Cape Code in Zone 4 (SW MO).
Garage was converted to living space unknown years ago with a cathedral ceiling bonus room above.
Recently had a roof replacement due to asphalt shingle age and decided to redo that whole room as it was uncomfortable in summer and winter.
No soffits and no ridge vent. Removed existing R-11 fiberglass insulation from 2×8 rafters.

Originally had planned to do the flash and batt with 2.5 inches closed cell (to get R-15) and then compressed fiberglass.
All of the spray foam installers I have talked to locally say they always do open cell in a situation like this but that’s contradicted most of what I have read on here so was still thinking the closed cell.
Then yesterday I ran across an article mentioning roof underlayments and spray foam that has sent me back to the drawing board.
The recent roof was installed with synthetic underlayment (RoofRunner).
Talked to a local installer again and they once again recommended open cell so it could breath to the interior.

If I go with closed cell I am now worried about creating the sheathing sandwich. (Would asphalt shingles not already do that though?)
I’ve read an article from Martin that talks about this situation due to poor planning which I admit to and that it is ok if the sheathing is dry <18% during install.

Looking for some direction.

In addition to the cathedral area (687sq ft roof area) there is a smaller attic connected to it (315 sq ft roof area) that currently has fiberglass insulation in the floor.  Also no soffits. I’d like to use that space for storage.
This has 2×6 rafters.
Currently it has no hvac present and was accessible by crawling through a kneewall (now removed) but I can condition it.

Thanks again for all the great information here.

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. kramttocs | | #1

    This is what is said about RoofRunner: "strongly recommends that it be installed over adequately ventilated attic spaces"

    Which leads me to thinking that full cavity open cell (won't hit minimum R-values) and conditioned room/attic may be the safe bet? That way it can dry to the interior.

    Give me an hour and I'll change my thinking again...

  2. Expert Member
    BILL WICHERS | | #2

    Lots of roofing manufacturers want a vented roof, but all kinds of roofing get installed over unvented roofs without problems.

    I would use closed cell spray foam here against the underside of the sheathing as it’s the safest option. You need to use enough for your climate zone to get a good vapor barrier against the sheathing. Since it’s fully adhered, there is no way for moisture to get in and you don’t have a “sandwich” so no concerns there. I would then fill the remainder of the rafter bays with open cell spray foam which is cheaper and will be trimmed flush with the underside of the rafters. You could also use compressed batts, but open cell spray foam will be faster and easier, and might not really cost much different since you’ll already have the spray foam crew on site anyway.


  3. GBA Editor
    Kiley Jacques | | #3

    There are some informative comments in this thread that I think would be worth considering: Closed and open cell foams on unvented cathedral ceiling .

  4. kramttocs | | #4

    Thanks Bill and Kiley. Since I posted this I have switched back to going the closed cell route.
    Curious: so you are saying it's not a sandwich because the closed cell and the decking would be one unit essentially?
    I can't count the number of times I've read that link so at this point I really should settle on something :)

    My thoughts on that were:

    2x8 rafters (7.25)
    2.5" closed cell = R16.5
    Remaining 4.5" = unfaced R-21 compressed which I believe is around R-18
    Total R34.5 which while may not be ideal is a big improvement over the R11

    2x6 rafters (5.5)
    2.5" closed cell = R16.5
    Remaining 3"* = R15 Roxul
    *In this room I can add to the rafters to get the extra half inch
    Total R31.5

    2x8: Unfaced fiberglass and then just drywall is correct, right? No facing or anything else needed?
    2x6: Any issues with just leaving the Roxul open without drywall over it? I'd add some 1x horizontal to keep it all held up.

    I could check into the open cell on top of the closed also

  5. Expert Member
    BILL WICHERS | | #5

    The ratio of closed cell spray foam to "other stuff" in a roof assembly is different (more conservative, so more ccSPF compared to "other stuff") than for a wall. I had forgotten the exact ratio, but Dana mentions ~50% is "prudent" in the thread Kylie linked to, so I'd shoot for that. That means 50% of the total R value in the roof should be closed cell spray foam.

    I would used open cell spray foam for the rest. That will give you a good insulation job, quickly, and save all the labor of dealing with batts. If you do go with batts, a facer (which acts as a vapor retarder) isn't really necassary, but it doesn't hurt. In this situation, the vapor retarder would be extra insurance. I've found that faced batts are often cheaper than unfaced batts in some sizes, so be sure to check that.


  6. kramttocs | | #6

    Thanks Bill.
    I could be wrong but I believe the 50% is in reference to what that user would have.
    I was basing the 2.5" off a table I found that shows for zone 4A that R15 is required for the closed cell and at R38 that is 39% (31% at R49).
    Since I won't make it to R38 my closed cell will be a higher ratio than those.
    Again, please correct me if wrong. I'd provide the table but not sure if allowed. Here is the link though:

  7. kramttocs | | #7

    This is the article that has me worried about using closed cell with the synthetic RoofRunner and asphalt shingles on top:

    1. Expert Member
      BILL WICHERS | | #8

      It’s done all the time without problems. Closed cell spray foam on the underside and ice and water shield on the top is pretty common. Just make sure you don’t have any leaky spots on top and you should be fine.


      1. kramttocs | | #9

        Thanks Bill!

  8. Deleted | | #10


  9. kramttocs | | #11

    Deleted my post from last night since I needed to gather my thoughts a bit.

    Had the closed cell installed on Friday. Never had spray foam done before but coverage seems good and I spot checked with them and everything is >2.75". In some bays it's very uniform and would take a batt well whereas in others they did a couple passes (shy on depth the first time) so in those (and there are quite a few) the depth is quite a bit thicker non-uniformly and would require more compression of the batts.

    My question is this:

    Originally I'd thought to do fiberglass batts (unfaced) to make up the difference.
    A coworker had a sound booth they were getting rid of so I got a lot of used Roxul Safe-n-Sound for free.
    It's 3" with no technical R-value though most seem to put it at R-13.
    It would be nice to use that here since it's available. I could mix and match it with regular 3.5" Roxul when the depth allows it or I run out of Safe-n-Sound.
    If going this route, in the places where the depth is a little greater than 3.5" should the air gap be between the roxul and the ccspf or between the roxul and the future drywall?

    I am guessing the roxul should be in contact with the ccspf but seems like air gaps also help R-Value in some articles I read over the years.


  10. kramttocs | | #12

    Obviously used some better keywords in my search after posting that as I came across:
    Where Dana states that the roxul would need to contact the spray foam.

    Even though the Safe-n-Sound was free, should I still go back to the original plan of getting 'too thick' fiberglass batts and filling the entire cavity?

  11. Expert Member
    BILL WICHERS | | #13

    The Safe'n'Sound product is pretty much the same thing as the R15 product for 2x4 walls, just reduced down to about a 3" depth. The R value for the Safe'n'Sound product is a bit shy of R13 as a result, and will work just fine for thermal insulation. If you need to be inspected, you can write to the manufacturer and they will send you a letter with an R value for the Safe'n'Sound product (there is a note about this in the documentation for the product).

    Dana is correct that you don't want any air gaps between layers of insulation. If you have to have a gap, it's better to have the air gap between the outmost insulation and the finished wall material (drywall, etc.). Ideally you don't want any air gaps at all though. I don't see a problem using the material you have on hand, but it won't be ideal if it doesn't fill the cavity completely. Note also that low density fiberglass batts will be easier to squish into irregularly filled cavities such as you get after a layer of closed cell spray foam has been installed. Low density fiberglass batts are much squishier than mineral wool batts, and are also squishier than high density fiberglass batts.


  12. kramttocs | | #14

    Thanks Bill. I won't rock the boat and will just go with the R-19 since it will fill in the gaps easier.
    I can throw the Safe-n-Sound between floors somewhere or whatnot.

    Did some checking and you are right - faced is cheaper and easier to come by so will go that route.

    1. Expert Member
      BILL WICHERS | | #15

      The facer will also help to hold the batts in place as you work to install your finished ceiling. Remember to staple the facer on the outside edges of the rafters/joists, NOT on the inside edges alongside the batts. You don't want any voids or gaps in the insulation. You also want to stretch the facer tight as you go on both sides, don't try to conserve on staples one one side so that you can staple the facers of two adjoining rafter bays at the same time. Been there, done that, you end up with a less optimal installation since the batts won't be in as evenly.

      I highly recommend pneumatic staple guns that shoot the regular Arrow staples too -- these staple guns will make your job much easier. I highly recommend this one:
      Which isn't even $40. Just remember to give it a drop of oil every time you use it. I installed a swivel fitting on the air port for mine, which makes the air hose fight you a lot less as you maneuver the staple gun into position. Using 1/4" air hose also helps.


  13. kramttocs | | #16

    Great info! I hear people say that insulation installers and sheetrockers have different points of view on the stapling but to the face makes sense and I'll be installing the drywall myself anyways.

    Just added the stapler to my cart. Have a swing one but seems like one is never at a good angle for those.

    Installing some roxul on the walls as we speak so will move to the cathedral ceiling next.

    I know being compared to poorly installed R-11 it's not saying much but the ccspf alone has made a huge difference in how often the heat pump kicks on. And that's with a bare uninsulated concrete floor and nothing in these two rooms. That said, thermal bridging is no joke. Went outside just now and can see the rafter lines in the frost. But the same can be seen on the rest of this Cape Cod that hasn't been touched since '79 so I guess it is what it is (until I formerly meet my rich Nigerian uncle that is)...

    1. Expert Member
      BILL WICHERS | | #17

      If you staple the facer on the inside face of the rafter or joist, you have to push up the sides of the batt, which results in an incomplete fill and compressed insulation. That is less than idea from an insulating standpoint. If you staple the facer onto the outside edge of the rafter or joist, the batt can completely fill the cavity space between rafters/joists, which is much better. This is why the kraft facer has those fold out edges anyway -- to give you enough reach to span the framing.

      BTW, what I normally do is wrap one side over the edge of the far rafter from me, and staple on the inside face OPPOSITE of the bay where the batt is. I then pull the facer on the near rafter and staple that one on the outside edge as I go along. I find that this results in less tearing of the far edge of the facer, so it makes for an easier installation. You then just continue down the ceiling as you go, with each new batt's finished edge overlapping the one you just completed. Hopefully that makes sense -- it's really easy to see in person.

      Drywall guys can just go over the top of the staples. It's not like all the edges of the framing are going to line up perfectly anyway, so the <<1/16" or so from the staples won't make any noticeable difference in the finished surface.


      1. kramttocs | | #18

        I think I follow.
        So is your process below?

        Installing from left to right for this example.
        1. Install batt 1 in bay A
        2. Wrap the flap 1/4" over the left rafter and staple
        3. Repeat for batt 2 in bay B while the right flap of batt 1 is still free
        4. Pull the right flap for batt 1 over batt 2's flap and then face staple. It won't be centered in the right rafter due to the 1/4" from step 2

        1. Expert Member
          BILL WICHERS | | #19

          Yep, pretty much. The Knauf brand batts I've used as of late have a pretty long flap with a crease in it, and the crease is about 1.5" from the edge of the actual fiberglass of the bat. This means you pull the flap over, line the crease up with the corner of the rafter/joist, then tack the rest of the flap up (the "rest" is about an inch of so). It actually goes pretty quickly once you get into the project.

          As you put up the first flap that is wrapped over along the crease, pull the facer taught linearly along the rafter/joist as you go. Only adjust the other side enough that the batt doesn't try to crease up or otherwise get pulled at an angle, but don't try to get into the fine details much yet. When you do the "other side", then you're pulling across (perpindicular to) the rafters/joists instead of along them (parallel to) like you were with the side with the crease. This time you want to get the facer to be flat.

          When you're done, it should look similar to the FAR SIDE of the attached pic, where the batts are pretty much straight and even without too many wrinkles. The near side of that pic is where I had to pull several batts down to correct something else, and it was difficult to get them back in as well since the facer tends to tear with a pull out and reinstall cycle.

          BTW, remember to fluff the batts too so that they fully fill the cavity without squished edges.


          1. kramttocs | | #20

            Excellent. I'll follow this approach. Much appreciated Bill.

  14. ProductionManager | | #21

    So, I have been following along on your chat back and forth about insulating a raftered vault ceiling. Two things come to mind, when I read through it all. I took a building science class and learned how condensation is a more intrusive problem than outside water getting in. We know what causes condensation, warm moisture laden air coming into contact with a cooler surface. When the warm vapor turns to water that's the dew point. So, what we need to calculate is how much (thickness) insulation is needed to ensure that the warm vapor migrating thru the insulation doesn't cool to the dew on the cold side of the insulation. Also having the rafter space completely full of foam insulation will prevent any air transferring heat or cold to the middle area of the insulation. Because the space between the rafter framing is completely filled therefore no air movement, we need to use a closed cell foam to prevent any vapor from entering the insulation.

    1. kramttocs | | #22

      Hey. Do a site search for flash and batt. That's the name of what I've got going on. You are right in that there is a specific thickness that is needed for the flash portion. For me in zone 4A that is R15. With the 2.75-3"+ that I have I am looking at [conservatively] R-18. The ratio set forth in the IRC for r-38 in zone 4 is 31-39% of the total R value being the ccspf. For me, it will be a higher percentage since I won't make it to r-38 with the batts unless I add to the rafters and that would make the room more difficult to use. Basically, the fiberglass won't be efficient enough to drop the surface temp of the ccspf enough to cause condensation.

      1. Expert Member
        Michael Maines | | #23

        Conservatively, your foam layer will eventually be more like R-15-16, once air displaces the blowing agent inside the foam cells. Still enough to meet the reasonably safe IRC ratios, but less than spray foam marketers would have you believe. From an energy loss point of view it's not a big difference, but because the IRC ratios are not overly safe, I use the long-term R-values, not the initial R-values.

  15. kramttocs | | #24

    Installed the r-19 last night and installing furring strips now to level the rafters. Since there will now be an air gap between the drywall (or planks..not sure yet) and the insulation due to the 1x3 furring strips, would it make sense to rip some .55" thick xps foam strips (around 13" wide) to fit parallel to the furring strips and take up that space?
    The R19 is faced so not sure if having the .55" foam over that would be an issue though or would just be essentially a thicker vapor retarder.

  16. kramttocs | | #25

    I was planning to installing vertical pieces covering the rafters between the strips to do something, not much I know, for thermal bridging like the photo (but cut down to actually fit between)but figure it'd be just as easy to rip wider pieces on the table saw and fill everything in. R2.5 is R2.5 and I'll take what I can get.

  17. Expert Member
    BILL WICHERS | | #26

    I wouldn't bother with the XPS here. I think you'll find the sag of those batts will probably fill in around half of the gap anyway, maybe more. XPS also risks creating a bit of a vapor trap. EPS would be better if you really want to do this.

    You'll be hating life if you rip that foam on a table saw. The saw will create about a bazillion little static charged foam bits that will fly everywhere, cling to everything, and make a mess that you'll probably never quite manage to fully clean up -- you'll keep finding missed foam bits stuck behind stuff. Use a knife and score/snap the panels, or use a hot wire cutter.


  18. kramttocs | | #27

    Thanks Bill. I haven't messed with EPS and have only ever used the XPS for my crawlspace walls and under my siding.
    Since I am going to cover the rafter ends anyways, aside from cost, is there any downside to using 3/4 non-foil-faced EPS and filling the entire space? You are right that in the center of the span, the batts bow out to be even with the furring strips.
    Going back to the hybrid ratio percentage of 39 (at R-38) I should still be safe. If I use R-15 to be long term conservative as Michael mentioned above for the ccspf, then the compressed R-19 + 3/4 EPS at R-20 that would still have the ccsfp at 43%.
    Probably higher as the R-20 is a gracious guess since it's hard to factor in the varying compression.
    It's one of those things where I'd rather spend the $150 now if there is even the slightest benefit but I don't want to be doing anything detrimental just because it seems like a good idea in my mind.

    You are spot on with the static foam bits! I did that a few years ago when adding foam to the walls of my crawlspace and now I remember all the time spent vacuuming out all the nooks and crannies of the table saw. So will stick with the 8' straightedge and knife approach.

    1. Expert Member
      BILL WICHERS | | #28

      Adding EPS would gain you around R3 or so here. You'd want to make sure to use UNfaced EPS (no foil facer, no poly facer), since you want it to be vapor permeable. You're going to have a lot of leaks around the furring strips anyway though, so now that I think about it, it's probably not super critical.

      I personally would probably not bother, but it won't hurt you to put it in.


      1. kramttocs | | #29

        You are probably right - not worth it. I'll just stick with the xps cut in 3" wide strips to go over the rafter ends and call it good.

  19. kramttocs | | #30

    Got to thinking and doing some more reading today (which isn't always a good thing).
    Regarding the kraft faced batts. You mentioned that it was just added insurance which makes sense as it will limit moisture moving into the wall and running the risk of condensing on the foam.
    A lot of what I am reading (and what had me thinking unfaced originally) is saying that it will prevent any moisture that does make its way into the wall from easily drying out. I've found comments going both ways with the ones not against it saying it won't hurt.
    I guess my question is why won't it hurt?

    What got me thinking about it was that the xps covering the ends of the rafter will further seal the kraft facing joints.

    I guess my question is should I purposefully not seal the batt layer or even go so far as to cut slits in the kraft paper? But then what the drywall + paint as that would form about layer preventing interior drying?

    I felt like I had all the answers to this before going this route but second guessing it all at the moment.

    1. Expert Member
      BILL WICHERS | | #31

      The kraft paper acts as a sort of old-school vapor RETARDER. It is not a vapor BARRIER. A lot of people mix these two terms up. A sheet of polyethylene is a vapor barrier, and won't let moisture move through. Moisture tends to be sneaky, and finds ways to get into things better than it can find ways to get out. Kraft paper will let it out, but slow it down getting in. If it gets more humid in the rafter bays, the kraft paper gets more vapor open, so it allows for faster drying. The kraft paper helps you by slowing moisture ingress into the ceiling here, but it doesn't hurt you because it will still let moisture get out.


      1. kramttocs | | #32

        As always, thanks Bill!

        Just for confirmation - I should install the kraft faced R-19 in a professional manner but since I do want it to dry to the interior I shouldn't tape the seams or go crazy trying to airseal it, right?

        1. Expert Member
          BILL WICHERS | | #33

          I would say to try to install everything in a professional manner :-)

          You don't need to tape it. I would rely on the kraft facer as an air barrier regardless. Just tack it up with a stapler in the usual way and call it good. The vapor can move through the kraft facer, it doesn't need gaps or seams that are open to work.


          1. kramttocs | | #34

            Words to live by :)

  20. PBP1 | | #35

    I have a flash and batt shed roof lid with tongue n groove inside and standing seam metal outside and agree with open cell foam over batt as a good solution - however, with batts it might be easier to install/service electrical, Ethernet, etc. I have 15 puck LEDs spring loaded against the tng along with Ethernet cable running in the non-closed cell foam space.

    1. kramttocs | | #36

      If I ever do another part of the house I will likely go that route as well. Live and learn :)

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |