Recording panorama of Zagreb

Roofing Fasteners and How They Affect Insulation Thermal Performance

Print Friendly, PDF & Email

Until now, there has been surprisingly little research into the overall thermal impact of the fasteners that penetrate roofing insulation. A recent study shows that even relatively conservative use of fasteners creates enough thermal bridging between the roof deck and the insulation to substantially reduce the overall thermal performance of the building envelope.

Guest blogger Eric K. Olson, P.E. explains his research (Olson, Saldanha, and Hsu, “Thermal Performance Evaluation of Roofing Details to Improve Thermal Efficiency and Condensation Resistance,” ASTM Roofing Research and Systems and Standards Development, Vol 8, STP 1590, ASTM International, November 2015)


Thermal insulation in roofing systems plays a substantial role in the overall thermal performance of the building envelope.  Energy code requirements for the R-value of the roofing insulation are becoming ever more stringent, requiring increased insulation thickness. Mechanical fasteners are commonly used to secure the insulation and roofing membrane to the structural roof deck.

Each metal fastener creates a thermal bridge that reduces the effectiveness of the insulation.  For a single fastener, the impact would probably be negligible. A typical roof, though, may include thousands of fasteners. The effect of these myriad thermal bridges adds up. That is, the combined impact of the fasteners can substantially reduce thermal performance.

Considering the potential impact involved, there is surprisingly little information in the roofing industry regarding the overall thermal impact of fasteners on roofing insulation. To explore and help quantify these thermal impacts, some colleagues and I decided to perform and publish the results of three-dimensional computer heat flow models of fasteners and other roofing details that penetrate the roofing insulation (Olson, Saldanha, and Hsu, “Thermal Performance Evaluation of Roofing Details to Improve Thermal Efficiency and Condensation Resistance,” ASTM Roofing Research and Systems and Standards Development, Vol 8, STP 1590, ASTM International, November 2015).


Thermal Bridging. Image by GAF.

Modeling and Analysis

We modeled a roofing system with 4 in. of polyisocyanurate insulation and 1/2 in. gypsum cover board with a nominal R-value of R-27.0, over steel deck, with the insulation fastened using steel plates and #14 roofing screws with a diameter of 0.214 in.

Modeling one fastener with plate penetrating a one sq. ft. area of insulation (e.g., sixteen fasteners per 4 ft. by 4 ft. insulation board), we found the following:

  1. Case 1: With the steel plate above the gypsum cover board, the fastener and plate drop the R-value from R-27.0 to R-19.2 (a 29% reduction in R-value).
  2. Case 2: Placing the plate beneath adhered gypsum cover board provides little improvement due to poor thermal resistance of the gypsum, raising the R-value from R-19.2 to R-19.5.

Swapping out the gypsum cover board with 1/2 in. high-density polyisocyanurate cover board raises the nominal R-value of the system from R-27.0 to R-29.0.  Repeating the above analysis, we found the following:

  1. Case 3: With the steel plate above the polyisocyanurate cover board, the fastener and plate drop the R-value from R-29.0 to R-21.2. This is a 27% reduction in R-value as compared to the nominal R-value using polyisocyanurate cover board.
  2. Case 4: Placing the plate beneath adhered high-density polyisocyanurate cover board raises the R-value from R-21.2 to R-23.8. This is a 9% improvement as compared to the case with the plate on top of the polyisocyanurate cover board, but still an 18% reduction as compared to the nominal R-value using polyisocyanurate cover board.

The above cases represent high rates of fastening (one per sq. ft.) that may be encountered at corners or perimeter zones.  In practice, field-of-roof zones require fewer fasteners and have greater area, and thus have a greater influence on thermal performance than corner and perimeter zones.  The figure below graphs the effective R-value versus the number of evenly spaced #12 fasteners and steel plates per 4 ft. x 4 ft. insulation board, using the conditions of Case 1 (fasteners through and plates above gypsum cover board) above.

Change in effective R-value Relative to Number of Fasteners for Case 1

Effective R Value

As can be seen above, the thermal bridging created by even light fastening rates can be significant.  A pattern of five fasteners per board, frequently seen in field areas of a roof, drops the effective R-value to R-24.  This is an 11% reduction in R-value.

Our work to date indicates that mechanically fastening roofing insulation substantially reduces the roof’s thermal performance as compared to a similar system without fasteners. More work remains to be done to quantify thermal bridging through roofing systems. The influence of fastener diameter, the use of less conductive fasteners (like stainless steel), and the use of polymer plates in reducing thermal bridging should be explored.

A better understanding of these thermally bridging elements will help identify options to help mitigate their effect. This, in turn, will help designers to better specify the thermal performance characteristics of their roofing systems. 

There are 11 comments

Add yours
  1. Scott

    I have a few questions. Have you run any calculations using two layers of iso board? Mechanically fasten 2″ layer to steel deck and adhere a second layer of 2″ with the cover-boards as described above?

    Also, Is there an advantage to using high-rise foam in stead of the fasteners and plates in either configuration?

    I have never liked allowing the use of a 4″ board MF and a cover board. I believe you build a better foundation for the roof system by using two layers as I have described. I have had projects where the GC or Architect Value Engineered by changing the specification to allow one layer of 4″ iso mechanically fastened. There was never any discussion of R-Value loss and long-term cost of the VE.

    This is a very informative article. Thanks for your post. I look forward to hearing from you.

    • Eric Olson

      Thanks for your question Scott. I agree – in the real world, two (or more) layers of insulation with staggered joints is beneficial to reduce energy loss at board joints. As far as the modeling goes, for simplicity we just selected 4″ of insulation across the models and used as a constant parameter – we did not intend to endorse going to a single layer of insulation. Since the models did not include board joints anyway, as that was not the scope of our analysis, there wouldn’t me much difference if we did 2 layers of 2″ vs 1 layer of 4″.

      As far as adhering 2″ of insulation over 2″ of mechanically fastened insulation, this would substantially reduce thermal bridging due to insulation fasteners. We have not yet quantified the impact of this configuration by modeling though. Future follow-up will probably include this or similar scenarios.

    • Thomas J Taylor, PhD

      In addition to Eric’s great reply, I’d like to point you to a case study we did on this topic. It looks at the cost of the various attachment methods, especially the cost of lost R-value due to fastener thermal bridging.

    • Thomas J Taylor, PhD

      Scott – just to follow up on one of your comments; yes, we sometimes see two layers of 2 inch polyiso get value engineered down to one layer of 4 inch polyiso. That’s not good practice for a couple of reasons. As you noted, the R-value gets compromised by the joints between boards running all the way from the roof deck up to the membrane. Also, those gaps allow a mechanically attached membrane to draw air up to under the membrane during wind events. In the winter that leads to a moisture condensation risk. When specifying a 4 inch thick polyiso based assembly, its important to consider using a cover board with staggered joints to reduce any air flow and installation of a vapor retarder above the deck and below the insulation.

  2. Ronald Paree

    Thomas, in f.e. The Netherlands, all 35 and 45 degree roofs are insulated which gives you a warm attic. An attic which is in nearly 100% of the cases either directly or in the future converted to one or two bedrooms, study, etc. = with the same footprint of house more m2! Nowadays the slope roofs are coming on site as large prefabricated insulated timber- or steel frame or SIP elements, all with with hinges. Installation of a complete roof directly from trailer to top of the building for a one-family house takes about 20 minutes. Often roof windows including thermophane glass are off-site installed in the prefab elements. After installation of the prefab roof elements the building is waterproof. Prefab elements meet the requirments of the European Union Directive 2030 for zero-energy constructuction. Off site chimneys and dormers (incl. glazing) are installed accordingly

    • Thomas J Taylor, PhD

      Ronald – thanks for the European perspective. As you note, many attics in Europe are designed as living spaces and the insulation is therefore under the roof as opposed to the US where it is normally above the ceiling of the top floor. There are some interesting contrasts between the two approaches – but the common thread is the need to avoid condensation. All good roofs keep out the weather, but systems need to be designed well to avoid interior issues such as condensation. In the US, that means sufficient attic ventilation plus minimizing air leakage from the living space up into the attic. In Europe, per your comment, even the materials can be different to avoid air and moisture penetration to the outer cold surface. Thanks for checking in with us.

  3. Marcus Evans

    Wonderful article. I often thought about the effect of the screws and pressure plates on the insulation’s r-value. Now I have a document to back out my claims.

    I’m assuming using low density foam to adhere the insulation will take care of this problem?

    • Thomas J Taylor, PhD

      Marcus – thank you. I had seen Eric present this work at a conference in a lot more detail, and thought it deserved to get a wider audience. Yes, using low density foam to adhere the insulation would take care of this issue. As you may have seen, we have done a lot of work on topics such as TPO performance but are starting to address topics related to overall system design. Look for more of these in the future. Thanks for checking in with us.

  4. Andrew

    Most roof decks are not insulated. You want your attic to be as close to outside temperature as possible. That’s why there ventilated and not insulated.

    • Thomas J Taylor, PhD

      Andrew – great comment; the blog covers low slope commercial roofs that are based on a steel deck with insulation above it. To your point, residential and light commercial buildings with sloped roofs normally have attics. Attics normally have insulated floors but not insulated decks for the reason you mention. We will be blogging about this in the future and discuss the best ways of avoiding both moisture build-up in a residential system as well as ice dam avoidance in the winter.

Post a new comment