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

Community and Q&A

Why does insulation have to contact six sides?

JeffSch1 | Posted in GBA Pro Help on

On the internet I’ve read about 15 years worth of claims that insulation must contact all six sides for it to be effective, but haven’t seen an explanation why. Yes, I’ve read many times that a building’s insulation and air barrier must completely surround the conditioned space and be in contact with each other, but can’t find the science behind it explained. I would sure appreciate a scientific explanation from anybody. Thanks very sincerely.

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. Expert Member
    Dana Dorsett | | #1

    If it's only 5 sides of the cube there is going to be a thermal bypass path around the insulation. Soil isn't the greatest insulator, nor is it air tight.

    If there's a discontinuity in the primary air barrier anywhere, the potential for leakage at the discontinuity is high.

  2. JeffSch1 | | #2

    I'm grateful for your reply, Dana, but I still don't understand. Your first explanation re: thermal bypass--I'm picturing a 2x6 stud bay with R-13 in it against the exterior sheathing, tight to the sides of the wall members. Where is the thermal bypass?

    Re: your 2nd sentence--discontinuity in the primary air barrier: If I've sealed a sheet of interior drywall at its perimeter, where's the discontinuity? Air goes between to studs tightly nailed together, hits the back of the drywall at that point, but can 't get into the conditioned space. Or, are you saying that caulking between the two studs in my example just provides redundancy in case the perimeter sealing fails?

    Again, I'm really trying to understand the science behind this concept, and for as often as it is cited, it is a bit shocking that my searches haven't yielded any scientific answers.

  3. _matt_p | | #3

    Imagine your insulation is in a completely airtight box, but does not make full contact on all 6 sides with the box. Then you have a potential for convective loops. Aka, airflow that stays within your box but moves cold air from the inner side of the cold sheathing in the box to the inner side of the inward facing warm sheathing wall and then back. This sort of bypasses the insulation, without requiring any air leakages, all going on inside your airtight box. That is not a big deal if your insulation itself is air impermeable, but with really fluffy stuff there is the potential for that convective loop. How big a deal this is, I have no idea. My guess is that if its truly fully airtight, this is mostly minor. My guess is convection in a full airtight box can reduce an R13 by an R2 or so if it is low density fiberglass. Obviously, it depends on whether you are not making contact on just one of the 6 sides or on say 5 out of the 6 sides. I am making this R2 number up. - Matthias

  4. Expert Member
    Dana Dorsett | | #4

    Jeff: It wasn't clear in your initial post that you were talking about studwall cavities. The statement:

    "..a building's insulation and air barrier must completely surround the conditioned space and be in contact with each other..."

    ...reads like you're talking about the whole house, not a stud bay, since you're talking about surrounding the conditioned space with a continuous and connected thermal & air barrier rather than an air barrier surrounding a batt.

    Matthias has it right: In the 2x6 w/ R13 scenario air can convect into and out of the batt into the 2" air gap at any level. Even batt with facers aren't air barriers- they invariably leak at the edges.

    If there are any gaps between the batt & framing at the sides, bottom, or top, the 2" air gap provides an essentially unimpeded convection path around the batt. If the sixth side of the batt the convection paths from those gaps all must flow through the batt to connect to one another, which is far more air-retardent than a 2" air gap. Without a compression fit there is also a narrower but still real gap between the batt and the sheathing too. Batts are intentionally designed with enough loft to be a compression fit into standard framing dimensions, 1/4- 3/4" thicker than the framing. An unfaced R13 will never be tight enough to guarantee there is no gap to the sheathing unless a means of providing some mechanical pressure is provided. A faced R13 with the tabs stapled 3.5" or less from the sheathing would be about as good as can be done, but still not good enough, since there's no air seal between the tab and 2x6, and even less of a seal at the top & bottom to prevent convection around the facer.

    With perfectly installed high density highly air retardent R15 batt under some amount of compression the difference would be small. But in construction (particularly batt insulation fitting) there is only good, better, and best, and no such thing as "perfect". With 3.5" of open cell foam it'll be air tight and gap-free enough with a less than full cavity fill, but not with batt insulation.

  5. JeffSch1 | | #5

    Thank you Matthias and Dana! I'm surprised I haven't encountered the term "convective loop" yet because I've read about contact on all six sides about 5 million times. Now I've I found a number of articles discussing it. So, is convection the chief reason to completely fill a cavity? (Convection losses must be particularly bad then in insulated floor framing over unconditioned crawl spaces and likewise in ceilings under unconditioned attics.)

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |