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
Musings of an Energy Nerd

Using Interior Poly As an Air Barrier

Alaskan builders still use a technique developed in the 1980s by Canadian superinsulation pioneers: the use of interior polyethylene as an air barrier

The Cold Climate Housing Research Center has produced a new video showing Alaskan builders how to use interior polyethylene as an air barrier. The video is narrated by Ilya Benesch, a building educator for the CCHRC.
Image Credit: Colc Climate Housing Research Center

Back in the 1980s, Canadian energy experts urged builders to use interior polyethylene as an air barrier material. If the poly was installed conscientiously, and all seams were sealed with Tremco acoustical sealant, the approach worked well — at least in cold climates.

When the same techniques were later adopted by American builders in places like Ohio and North Carolina, builders learned an important lesson: “Climate matters.” In hot climates or mixed climates where air conditioning is common, interior poly can become a condensing surface for exterior moisture during the summer. In many cases, the result was mold and rot.

These days, interior polyethylene is not recommended in most U.S. climates, especially in any home that may be air conditioned during the summer.

Alaskan builders still use interior polyethylene

But what if you are building in Alaska or northern Canada? Well, in those climates, interior polyethylene still makes sense. That’s why the Cold Climate Housing Research Center (CCHRC) in Fairbanks, Alaska, recommends that Alaskan builders use interior polyethylene as an air barrier.

The CCHRC has produced an excellent video on this topic. In the video, Ilya Benesch, a building educator at the CCHRC, explains how to do a good job of air sealing with interior poly. If you are building in northern Canada or Alaska, the techniques shown in this video (below) are worth emulating.

For more excellent advice on building techniques for very cold climates, visit the CCHRC website.

[Author’s postscript: in July 2012, when I was vacationing with my family in Alaska, we all showed up unannounced at the Cold Climate Housing Research Center in Fairbanks, where we were warmly greeted by Jack Hébert and Ilya Benesch. Not only did Jack and Ilya give us a tour of the CCHCR facilities, they let us use the building’s…

GBA Prime

This article is only available to GBA Prime Members

Sign up for a free trial and get instant access to this article as well as GBA’s complete library of premium articles and construction details.

Start Free Trial


  1. Expert Member
    Dana Dorsett | | #1

    Not just "Back in the 1980s..."
    Poly sheeting is still the most commonly used sheeting for air & vapor barriers in Canada, and in some Canadian climates that makes sense. It can still be a bit problematic in brick-clad air conditioned buildings in southern Ontario though.

    BTW: I wish they'd done as good a job in that video on the batt installation as they did with the vapor barrier. It doesn't take eagle-eyes to spot scores of compressions and some significant voids in the insulation job.

  2. Cjb80 | | #2

    I am very interested in the framing technique that is used there. I've never the top plates angled to match the pitch of the roof like that before. Does anyone have any thoughts on where I could find more information about such a technique? It seems like the top outside corner of the top plate would not be flush with the exterior sheathing.... (maybe this doesn't matter?)

  3. nvman | | #3

    I guess because it is a demonstration building for the poly video, they did not pay as much attention to the insulating. It also appears that some of the insulation is well used.

  4. IlyaBenesch | | #4

    Insulation and Framing
    I agree with you guys in that we could have done a better job with the insulation and yes, the materials were left overs from other jobs so the color scheme looks like neopolitan ice cream. When the opportunity came to film I was focused on the vapor retarder and not the insulation which had been installed and removed several times at that point due to weather. I reckon we should have gone back and tuned up a few spots here and there. Hopefully the audience can look past this oversight. Regarding cutting the plates to the roof pitch, we used 2x8's and ripped a 6/12 onto both edges to create a parallelogram so the sheathing could be nailed into the double top plate. I feel framing on a pitch like this has quite a few advantages but it costs a bit more and does take more time so I don't know that it would be an easy sell. In any case, I do prefer it over cutting birdsmouths into the I-joists, which lessens the r-value over the wall - and you don't have to block between the webs over the bearing wall as the I-joist isn't compromised by the birdsmouth cut. A little more on the subject of cathedral roofs can be found in this video we put together using the same structure:
    Martin was gracious in hosting the vapor retarder video and I welcome any feedback people might have. Having the GBA forum as an audience is good for us as I would hope that the feedback we get here can help improve on any videos we do in the future.

  5. Expert Member

    Reply to Chris
    It's fairly common when framing with exposed Glu-lam or solid wood beams to order them with a sloped top to match the roof pitch. Less common to also slope the outside walls.

    You should check with your building inspector before deciding to forgo a flat bearing surface and using birds mouths on your roof joists. A roof structure with a supported ridge exerts only vertical loads on the beam and exterior walls. Sloped bearing surfaces introduce a small horizontal load which must be countered by fasteners.

  6. Cjb80 | | #6

    Excellent video Ilya! I am
    Excellent video Ilya! I am now considering changing our approach to using I-beams; I will have to check with my structural engineer to get his commentary. I was planning on using the Simpson SDWC screw to fasten the studs to the I-beams, which I understand is stronger than the typical truss clips.

    In the second video that was posted, I didn't understand the comment made about "putting a spreader on layout"; I am not familiar with these terms. Does this mean to fasten a 2x4 across the rafter to use as a guide to ensure that the Tyvek is straight and square?

    Also, why did you not install Tyvek over the overhang lookouts? Insulating that area is referenced later in the video, the Tyvek was never shown to be installed over that portion of the rafters.

    Thanks for the videos, they're excellent!


  7. tenbob | | #7

    Well done!
    I'm disappointed; I didn't see any reference to using some exotic European materials! I thought for sure that to achieve any kind of an adequate assembly something had to be used that I couldn't get at my local lumber yard (although the Tremco can be a little dicey, sometimes!). Irony aside, with the addition of the foam on the inside of the assembly I believe you've even mitigated any vapor drive issues in a mixed-climate region where airconditioning would be prevalent; the poly is kept at a high enough temperature within the assembly make up so that condensation shouldn't be as issue. Although without performing some kind of hygrothermal analysis, I'd be cautious in making those kind of claims. I also applaud your achieving a thermally broken assembly while maximizing the use of lower-cost/R value insulation materials as opposed to just piling up expensive foam on the outside of the assembly (a real cost driver in an Alaskan environment, I'm sure).

    I do have a question: for a metal roof on this assembly, would it be advisable to just install the purlins across the 2x4-based ventilation channel and depend on the Tyvek for any moisture mitigation (I'm thinking of the numerous mornings when dew is dripping off the roof; I'm sure the bottom of the panels are wet, also). Or, just install the plywood/OSB roof sheathing faced with tar paper.

    Anyway, great job! the second video was definitely required to put the first video into context.

  8. Cjb80 | | #8

    I imagine that it would work
    I imagine that it would work fine to not have OSB on the roof sheathing in regard to dripping water; however, the sheathing is a structural element, so I don't think that it wouldn't be sufficient to only have Tyvek up there with a steel roof.

  9. Northernbuilder | | #9

    Better detailing
    I know this is some sort of shed, but in a residential building I would expect more.
    Putting the poly joints in outside corners tends to build up and tends to get ripped when installing interior strapping / curtain walls / drywall or creates more work for tapers and boarder = more $$ or a sub-par finished product (crooked corner beads). Leaving any joints a few feet off the corner, lap min 2ft with sealant and tape will reduce buildup in the corners and provide a better seal with less chances of tearing.
    Also the addition of a bead of sealant in all outside corners (on wood, exterior side of poly) will reduce cold air from penetrating through the ‘L’ wall joints causing frost to form on the interior side of the poly which can wet the back of the drywall or if cold enough penetrate through and form frost on the interior wall surface. This can be reduced with a multi-wall construction practices and the placement of the poly within different wall assemblies.
    I prefer to see larger pieces or one continuous piece rather than the ‘one piece per wall’ approach which in turn means fewer joints, less tape and less sealant. Single continuous pieces are very simple to achieve.
    Also more detail could be put into the unwanted but sometimes unavoidable un-backed joint, by providing more lap (2ft), a double bead of sealant finished with tape. Just my 2 cents.

  10. IlyaBenesch | | #10

    In response to questions...
    Chris, to answer your questions on the second (linked) video, the spreader shown at 6:58 is there to hold the rafters straight and on layout before you staple down the first run of house wrap. Once the house wrap is stapled on, there will be no more adjusting the rafters left or right if they aren’t running straight . Among other things, off layout/crooked rafters can affect whether your drywall edges land properly centered on the rafter edges on the ceiling underneath. I’m less concerned about the top edges as those get the 2x4 nailer laid flat on them to create the air space so you end up with a wider nailing surface for the roof sheathing that follows. You can also run that spreader on the underside of the rafters instead of the top if that works out better. As far as the house wrap goes… it can be treacherous, and on a big roof I’d put down temporary 2x4 purlins as needed to stand on when working my way up the roof. Trying to walk up the 2x4’s nailed on the flat on the top edge of the rafters isn’t something I can advise. And, as I mentioned in the video, avoid windy days for laying the tyvek, and it’s easier and safer if you have several people on hand to help when rolling it out and getting it tight.
    We used the lookouts to create a 4 foot roof overhang and they do travel the same distance (4 feet) back into the roof. If you look at 13:10 there’s a pretty good shot of the lookouts where they travel over the heated portion of the roof assembly. The entire heated portion of the roof is covered in house wrap. The eaves/overhangs around the perimeter stay uncovered. It’s not needed there and will be a safer place to work from if not obstructed by house wrap. Since this particular roof is framed with 11 7/8” I - joists, when you subtract the depth of the lookout (2x6) you are left with a space of 6 3/8”. So, before we put the lookouts in, we laid down the gable end rafter and the second rafter ripped out of dimensional lumber to the 6 3/8” width, which then brought the underside of the roof into plane with the other rafters and maintained ceiling layout.
    Malcolm brings up a good point – and we try to cover this consistently: Any techniques involving structural framing and attachments that deviate from accepted practice and/or prescriptive measures should be reviewed by a structural engineer to make sure that the construction method being employed meets local codes – including snow loads, and seismic or wind requirements. We made statements of that nature a couple of times in the video – once while discussing the screws used to fasten the underside strapping on at 27:48 and also 7:35 while discussing the 2x4 sleepers used to create the air space. I think we also mentioned it when talking about the rollover blocking too. In any case, I can’t stress that aspect enough.
    Also, if you plan on cutting your studs and plates on angles to match the roof pitch as we did in the video, take note of the total height as measured from the subfloor to the bottom of the rafter on the warm (long) side. If you are building a wall that is expected to cover with of two sheets of drywall, then you want to avoid having anything too tall (including of course allowances for the thickness of the ceiling drywall angled to the roof pitch and the approximate ½” space at the bottom). The last thing you want is to have a wall that’s too tall for 2 sheets of drywall to cover it adequately - if that’s what you were planning on. If you will be putting foam board and strapping on the ceiling, the same rules apply and you’ll still want it to work out so that two sheets of drywall can go on the wall without a lot of left over space.
    In regards to Bob’s question, the roof on that shed will get metal on it after it gets either a protective layer of 30lb felt or a self adhering membrane like Grace Ice & Water shield - either of which will serve to cover the OSB sheathing and protect the structure underneath. We are counting on the sheathing for structural reasons and to support the tin and weather barrier above. Although many folks just nail 2x4 purlins on 2 foot centers across the rafters and then put the metal down, I can’t recommend it. Also, I wouldn’t want to count on the tyvek alone to stop any condensation or roof leaks from reaching the insulation underneath. I’m sure it will help, but I think you’ll do best with a layer of sheathing followed by a weather barrier approved to go under the tin in your climate. Maintaining an unobstructed air space at 1” minimum up each rafter bay is critical. In my opinion, the 1 ½” air space we created by nailing the 2x4’s down flat is a case where more is better. If instead, you put down tyvek over the rafters and then perpendicular nailers/purlins followed by tin, you lose that continuous vent space. I don’t think most inspectors would qualify any vertical channels in the tin as vented air space.
    Hot climates have other considerations for what underlayment can go under the tin to withstand high heat. I am assuming any condensation forming is atmospheric and related to outdoor humidity and temperature drops rather than indoor air leakage from the heated enclosure underneath. Aside from the condensation issue, that weather barrier over sheathing will protect the rest of the structure from any minor leaks that may occur over the life of the metal roof. I think the roof system we filmed should work well in mixed climates, but I need to be clear that I don’t work in places with extended hot/humid seasons so someone else may wish to chime in on the condensation on the underside of the tin issue. Frankly, I prefer attaching metal roofing to plywood. The laminations in the plywood seem to do the best job in keeping the fasteners in place in spite of expansion and contraction of the metal with temperature swings. I’ve seen many cases up here (including my shop) where the gasketed screws in an exposed fastener type roof have started to back out when they were screwed into purlins – particularly if those purlins were milled out of fresh cut spruce.
    To a point, I agree with Rob’s comment on the build up of the poly in the corners being a concern. You want someone detail oriented installing the poly – not necessarily the laborer who just got hired, but that’s often how it goes. Big blobs of acoustical sealant and build up in the poly, no matter where you put them, can cause grief behind drywall – but the corners may be more vulnerable. Ditto for wrapping the poly around several walls or you’ll get bound up – someone who’s done it before, and knows where to leave a little slack as needed, will do fine -and less seams are certainly more desirable. Crappy looking finished drywall can result from anything that interferes with flatness on the back side, including the flanges on “vapor tight” electrical boxes, mud rings, nail plates over wiring or plumbing, misaligned studs, framing hold downs/plates, and studs not all crowned in the same direction.
    In our area, the city inspectors look at the vapor retarder sealing job and most houses are beholden to a blower door test when completed, so the subs are usually pretty aware that the vapor retarder is sacred. It didn’t used to be that way here and even many of the better built older houses (into the 1980’s) are heinous leakers.
    All that said, I feel horizontal strapping is a better way to go than applying drywall directly over the vapor retarder. Yes, you do lose a little more real estate on the inside but you can get a better product for it. Strapping the interior of the exterior walls is relatively common in this region and is one of the easier ways to get a little extra performance out of the standard 2x6 R-21 framed wall. It solves many problems, including the ones I mentioned above and those that Rob mentioned regarding cold spots. Once the poly is installed, then 2x thickness horizontal furring is nailed on 16” centers up the wall. At least in my experience, two carpenters with two framing guns should be able to rip the material on a jobsite table saw and strap a basic 1600 sqft ranch style house in two days or less. The spaces between the strapping then become a wiring chase for 1 ½” shallow electrical boxes – similar to what we showed with the ceiling strapping in the vapor retarder video. You still need to take your time in the corners and make sure everything lines up in plane. You can also add additional insulation into the spaces between the furring strips once the wiring has been roughed in -and it’s easiest if you can contain the bulk of the wiring to one bay. We hope to have a video out on strapping walls in a couple of months to cover some of these details. Don’t get me wrong, I agree with Rob that there are more thermally efficient ways to build walls and deal with the cold spots!! R-40 plus walls are easy to find in Alaska’s Interior - including our research center, but for someone who wants a little more performance out of a generic 2x6 wall assembly (perhaps in a milder climate?), the strapping is at least worth considering.

  11. Northernbuilder | | #11

    I didn’t even see the link to the other video, I’ll try and watch it when I have some time!
    It sounds like you pay attention to the same details I do when it comes to framing. Although I wasn’t paying attention as to whether you built the gables as rake walls?
    I agree interior strapping is the way to go and will limit the paly in corners; I don’t get why you would rip 2x4?6? down? when we did 2x6 with interior strapping (went to an R52 wall) we just used 2x3 adding an extra R8 FG batt or R10 Roxul to the mix; no shallow boxes, no ripping boards ,and is available at most lumber yards. One guy one day for an equivalent house.
    When doing continuous poly always do your entire top line first then finish the corner to the floor DO NOT CONTINUE ON THE NEXT WALL. Go back to the middle take that to the floor finish the rest. Do the next wall the same. We were regularly hitting .3-.5ACH.
    Although I have my own thoughts on the roof especially regarding tin and air space we’re on the same page just a few different paragraphs. By the way I absolutely love the rigid styro for wall penetrations, that is brilliant and I’m totally stealing that!

  12. user-1105327 | | #12

    how do you seal the poly when
    how do you seal the poly when penetrating the rigid foam for services after the gwb is installed? also, my grandfather had a stapler like that...

  13. user-1140531 | | #13

    Warm Side Vapor Barrier
    I prefer warm-side poly as a vapor and air barrier for Minnesota. My general objective is to keep water from entering from the exterior, avoid reservoir cladding, and completely seal against outward air leaks and vapor diffusion. The wall then has the ability to dry only to the outside, but no reason for a need to dry is expected.

    I like the strapping approach to over the poly that is mentioned above. I would use a furring out of the studs to achieve a service cavity and sandwich the poly between the stud and the 2 X 2 furring strip. So the poly would be tightly clamped at each stud and corner by the furring and not need to achieve a clamping effect between the drywall and the studs.

    In Minnesota, I do not believe the summertime air conditioning will pose a problem of inward vapor drive condensing on the vapor barrier because I do not expect the vapor barrier temperature to go lower than the outside dewpoint temperature, except on extremely rare occasions. The R-value of the service cavity and the three air films, though small, will further assure that the poly will stay above the outdoor dewpoint temperature. Also working in favor of this outcome is my preference for keeping the air conditioning above 75 degrees for the living space.

    It would be interesting to install temperature sensors on the exterior side of the poly just to monitor the temperature during periods of high dewpoint and air conditioning. If necessary, the air conditioning could be control limited by that temperature.

  14. synergytodd | | #14

    Back in the early 1990's, I helped my family's cellulose manufacturing business develop wet-spray cellulose installations in the Southeastern United States. I read everything I could get my hands on (mainly Energy Design Update) to learn more about how to merge what the code was saying with what we were learning from field installations. We learned almost immediately that an interior vapor barrier of poly trapped moisture inside of the wall and prevented the wall from drying out. For a long time, a lot of building inspectors, builders, competing insulation contractors, and others called me a "crook" or "'crazy" for insisting on not using poly as an interior vapor barrier for both wet spray cellulose as well as unfaced fiberglass.

    1. GBA Editor
      Martin Holladay | | #15

      Yes, even in the early 1990s, most builders and code officials misunderstood the science behind vapor barriers and vapor permeance. Even today, the confusions persist.

Log in or become a member to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |