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

Thermal Conductivity of Structural Screws

M Gillespie | Posted in Green Building Techniques on

Climate Zone 5 deep energy remodel on A Frame structure.

There is no attic, structural tongue and groove (roof deck) is attached to the exterior of the internal A-Frame timber structure.

There is no internal insulation “under” the tongue and groove, as the tongue and groove presents as the decorative ceiling within the A frame.

Roof will have air/water barrier at structural deck (T&G), rigid insulation, then insulated and vented Hunter panel (or equivalent).  Metal roof underlayment on top of the Hunter panel and then standing seam, hidden fastener, metal roofing as the final layer.  When attaching through the Hunter panel and rigid foam (all layers) into the T&G roof structure, long and appropriately rated screws will be used in a resilient roof screw pattern (opposing angles).

There are two (2) concerns:
1.  During extended cold periods and/or snow on roof, the top of the screw head under the metal roofing getting cold and transferring the cold into the roof assembly/structure.

2.  During extended heating days, the top of screw heads becoming “hot” and transferring that heat into the roof assembly/structure.


1.  Is this even something to be concerned about?

2.  If so, do the screws needs to be thermally broken?

3.  If screws need to be thermally broken, how? Technique or Product recommendations?

1.  Counter sinking of the screw heads on the Hunter panel would be difficult.
2.  Any thermal break material (fiberglass or foam blocking) on top of the screw heads could show up in as “bumps” under the metal roofing.

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. Kyle Bentley | | #2

    You have to attach the panels somehow, and while there is some thermal bridging, it's small compared to the energy saved by the additional insulation. There's no way to practically isolate a screw from it's substrate.

    You could use stainless steel screws if you really wanted, as they have a much smaller thermal conductivity compared to mild steel, but you'd never recoup the costs in energy savings.

    If you could aim to terminate the screws into the rafters instead of the 1.5" decking, they'll be somewhat insulated from the interior environment, without any chance of condensation on the screw tips in the interior in the winter.

    If you can't countersink, make sure to start the snaplock panels in a way that a raised component will land over the heads. With polyiso under those blocks in the hunter panels, you can probably draw the head of a timberlocks or powerlag pretty much flush as it is.

  2. Expert Member
    Zephyr7 | | #3

    This question comes up here every now and then. In a word, no -- it's not necassary to thermally break long screws. The total energy lost through the screws is very low. I actually calculated it out as BTUs per unit of 4x8 sheathing, you can probably find that Q+A thread on here with my calculations if you do some digging.

    Note that the screws are already thermally broken a little anyway, because the wood they screw into insulates them a little -- the screw isn't actually a piece of metal going tall the way through the wall as a result. If you're worried about moisture rotting out the screw head, stick a piece of 20mil tape on the top of every screw head as extra protection, or give it a smear of roofing sealant. I don't really think either is necassary though.


  3. Roger Berry | | #4

    Do read the link posted by John, look to comments 3,5,&11. The roof in picture has 8" carbon steel structural screws going through 4" nailbase, 2" polyiso into 2x8 rafters sprayed with flash of spray foam. My climate is 6B. I also can see ghosting of screws under my standing seam which is attached to 8.25" nailbase screwed with 10" carbon steel structural screws into 2x12 rafters filled with batts. You may draw your own conclusions.

  4. Kyle Bentley | | #5


    Your picture made me google a few things. I think this report that I found from the NREL (2010) models precisely what you're seeing.

  5. Expert Member
    Zephyr7 | | #6

    It doesn't take a lot of energy to melt small dots of snow. I see this myself in my dog's doghouse, that I built using simple 2x2s for framing and 1.5" EPS between the inner and outer wall surfaces (~R6 or so). I assembled things using regular ol' deck screws, so those screw through the outer 1/2" or so sheathing and into the 2x2 framing. I can actually look out my office window right now and clearly see depressions on the roof from thermal briding of the 2x2 framing, although the screws are not visible since we had a bunch of fresh snow last night. In time, I know from past experience that the screws will eventually thaw little dimples that will be visible.

    The overall energy loss is low. Over time, you gradually lose enough energy to melt the snow over the screw heads. How long it takes depends on the outdoor temperature and the amount of snow or ice that is present. Screws DO add some additional thermal loss, it's just SMALL compared to other things. In my other post some years ago in another thread that I mentioned earlier, I calculated this total loss and it's small, a few tens of BTUs per hour or so if I remember correctly. I think I compared it to the energy lost by an air leak through a keyhole in a door. You would be better off do some extra air sealing work than worrying about thermal bridging of fasteners in rigid foam.

    You can minimize thermal bridging of fasteners be using long fasteners of small diameter to minimize their cross sectional area. You can be careful to use only enough to do the job and not go over board with the total fastener count too, for the same reason. Anything to really cut down on that thermal bridging will significantly complicate things. I can think of two relatively simple ways to do it though:

    1- Glue up a layer of 1/2" rigid foam as the last step of your build, using it to cover all of the fastener heads. You don't need a thick layer here, even a little bit will significantly reduce thermal bridging. This will complicate the installation of siding though if you want to maintain your thermal break.
    2- Glue up a second furring strip over the first (assuming you have a rain screen sort of arrangement going up), or screw them together, using the outer furring strip as "insulation" to cover the fastener heads. This won't be as effective as a layer of insulation, but it will help, and it doesn't add too much complexity to the siding install.

    I can think of a third more complex reason:
    3- put up furring strips over your first layer of insulation horizontally. Add rigid foam between these. Put up your final, outer furring strips up in the usual vertical way, but centered on the first furring strips instead of lined up with the studs. This creates a sort of zig-zag thermal path from the first fastener to the middle of the inner furring strip, then horizontally along the length of the inner furring strip to the fastener securing that inner furring strip to the studs. This is a lot of extra work, but will result in very little thermal bridging of the outer fasteners to the inner structure of the home. You'd need to use 2x4s for the inner "furring strips" here though, since they have to carry all of the load of the outer furring strips and siding.

    I just noticed that I gave an example for a wall, when the OP asked about a roof. Please tilt my "wall" example on it's side when reading and pretend it's a roof instead to save me from retyping everything :-)


  6. Expert Member
    Akos | | #7

    Everything seems to be better when it comes to efficiency across the pond:

    This puts the screw inside the insulation plus it also means you don't need extra long specialty screws.

    This won't work with Hunter panels, but would be great option for a rigid roof install to squeeze that extra 5% to 10% of R value out.

    1. DCContrarian | | #8

      After a cursory scanning of the specifications it seems like those are only rated for attaching insulation, they wouldn't be rated for attaching the roof substrate to the framing. I think a similar product available here is the Plasti-Grip fastener:

      Although not as long.

      If you want to break the thermal bridging I think the way to do it is to have a multilayer roof where you have a layer of strapping in the middle which is attached to the framing, and then the roof substrate is attached with the attachment points offset. It might be cheaper to have twice as many fasteners half as long.

  7. M Gillespie | | #9

    All: Appreciate the feedback.

  8. RussEllisCQB | | #10

    There are plenty of ways to attach the roof substrate to the framing. I think here is better to use some plastic fasteners to secure the substrate from moving. And for the roof itself, you can use roofing screws timber T17. They are self-sealing and protected from corrosion for a long time. This might be the best decision for the roof and for the insulation. However, for the insulation, you can also use plasterboard fixing screws that are able to carry heavy loads in sheet materials. They are also simple to use and very reliable. Good luck!

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |