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

Comparing Fiberglass Batt Insulation R-Values

gustave_stroes | Posted in Energy Efficiency and Durability on

Hello GBA Folks,

I am grappling with the differences between R19 and R21 fiberglass batt insulation.

R19 is 6.5″ thick whereas R21 is 5.5″ thick. Thinner by one inch and yet 2 points higher in R value. The common explanation as to why this is so, is that the R21 insulation is said to be “more dense”.

But I have always thought of fiberglass batt (any insulation really) as just being a means of trapping air and keeping that air from moving around (due to natural convection). Air is a much better insulator than glass after all.

So how does R21 batt, which is thinner, and has a greater amount of glass per inch, achieve better insulation?

I am either missing something obvious, or this is a subtle thing.

Thank you.

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


    Air on its own isn't a good insulator. If it were you could just leave the wall cavities empty. R-value is just resistance to heat transfer. For batt insulation that is optimized by the best combination of void and material - which is what high density batts are compared to regular ones.

  2. Expert Member
    Akos | | #2

    Also the R19 batts are R19 when installed in a 6.5" cavity. Compressed to 5.5", you only get R18. An R21 HD batt in a 5.5" wall is actually R21.

  3. gustave_stroes | | #3

    Thanks Malcom and Akos,

    Actually air is a pretty good insulator, as long as it is stagnant and not inducing convection heat transfer. Air is a better insulator than glass.

    Air: k = 0.025 W/mK
    Glass: k = 1.05 W/mK

    1" = 0.0254 m

    Air: U = 0.98 W/mK/in R = 1.02 per inch
    Glass: U = 41 W/mK/in R = 0.024 per inch

    R19 at 6.5", then compressed to 5.5" should yield R19 x 5.5/6.5 = R16.1 , not R18 I think. Unless there is something else going on?

    The reason we cannot just leave the wall cavities empty is that the air inside the walls would start moving around. Hot air on against the inside wall would move up. Cold air against the outside wall would flow down, setting up a natural convection current. It's always been my assumption that the fiberglass just stops the air from moving. You could just as well fill the wall with a honeycomb matrix of some sort to achieve the same effect.

    Please feel free to correct me on any of this if I got it wrong.

    1. Expert Member
      Akos | | #4

      All insulation is there mostly to prevent convective air flow. Like you said, air is a pretty good insulator.

      Generally higher density insulation increases R value up to a point. For example, R24 batts are 2x the density of R19 batts.

      The reason the R19 doesn't just scale with width is that as you compress it, density increases, thus higher R value per inch. This is a good source for actual R values:

    2. Expert Member
      MALCOLM TAYLOR | | #5


      No, you have it right. I should have said "Because of it's inherent instability, air isn't a good insulator without being encapsulated, preferably in very small volumes, where it's movement is inhibited. That's what high R batts do better than regular ones".

      We should start a company making very efficient honeycomb insulation.

      1. Expert Member
        BILL WICHERS | | #6

        >"We should start a company making very efficient honeycomb insulation."

        That's pretty much what rigid foam is. If you leave out the blowing agents, look at a micrograph of EPS and it looks like a zillion little bubbles all stuck together. Each little bubble is a tiny air cell. The polystyrene that makes of the "stuff" around the air in EPS is really just acting to greatly restrict convective airflow, so you are maximizing the insulating potential of the air itself. Aerogel is even more effective, but also a lot more expensive.

        About the only thing that works better here is a vacuum panel, but those have structural issues to build. With vacuum, there is no matter to allow for any kind of convective flow, so you end up with only radiated energy transfer between the sides, which is much more limited compared to active transport by convective flow of a gas.


  4. gustave_stroes | | #7

    Thanks all for you replies. So is that it then? That standard fiberglass batt does not limit air movement sufficiently, or optimally. Such that a more dense batt, with smaller air cavities creates better insulation? That is subtle then.

    In a prior life I could have calculated the minimum size air pocket required to eliminate natural convection for a certain delta-T. But I'd have to dust off some books to regain that ability 🤠

    I worked for a while on thermal design for satellites. They employ MLI (multi layer insulation). Space being a vacuum, only conduction and radiation need be considered. The MLI blankets were formed from 20 or so very thin shiny foil layers sandwiched together with spacers to keep them from touching. The very low emissivity of the foil, and the multiple gaps, made it very difficult for heat to move from the outside surface to the inside surface.

    1. Expert Member
      BILL WICHERS | | #8

      Yep, denser batts mean less ability for air to move, so better R value performance, within reason. This is especially noticeable in vertical applications (walls) that have no air barrier on one side (think attic knee walls). Denser insulation, like mineral wool, in those areas has much better performance than a typical R11 batt because it has much less issue with convective flows on open side that has no air barrier. Basically convection in this case makes the batt perform well under it's R11 rating, due to the lack of an air barrier, so the denser batt has a much more pronounced improvement than you'd expect.

      The window guys have already worked out all the optium spacings for minimized convective air flow. The interesting part is that the optimum spacing is somewhat temperature dependent, which is why European U values sometimes look better than US U values -- they use a spacing optimized for a different set of temperature testing parameters. Their windows thus aren't inherently superior, they're just tested differently. It's as the old saying goes "the trouble with standards is that there are so many to choose from".


  5. gustave_stroes | | #9

    "Yep, denser batts mean less ability for air to move, so better R value performance, within reason."
    Yes, it occurred to me as I was mulling this over, that there will be a limit on increasing R-value as fiberglass batt is compressed. Assuming a fixed gap which heat has to traverse, then packing more and more fiberglass batt into that gap will help initially. But at some point the fiberglass will be so densely packed that it forms too many high conductivity "bridges" across the gap, at which point total R-value will start to go down.
    At the limit the wall cavity will simply be filled with solid glass, and the R-value will be that of glass.

    1. Expert Member
      MALCOLM TAYLOR | | #10


      I wonder where that point is? Comfortboard looks a lot like compressed rockwool batts, and must have exponentially more fiber to voids than the latter.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |