Roof Air Leakage

10/14/2009 | By Richard L. Fricklas

Many readers responded with helpful and valid comments – some of which are featured here. First of all, there’s considerable disagreement on whether air barriers are of benefit in low-slope roofing systems.

    “I think there’s great ‘practical and economic support for installing air barriers’ in almost all buildings in North America. In some of these buildings, we do that already by virtue of the assemblies we use (e.g. fully adhered compact membrane roofing systems), but in others, such as framed assemblies (roofs and walls), we need to significantly improve air tightness by installing deliberate air barrier systems.”

    “Air Leakage: More roofs are subjected to damage by air exfiltration than air infiltration. Since most moisture problems in roofs are still due to waterproofing deficiencies, not condensation, vapor retarders and air barrier systems are not ‘at least as important’ as attending to waterproofing issues. I suggest you rephrase to acknowledge that ‘another important consideration is air leakage’.”

    “While everyone acknowledges that there is some air leakage into a building, there’s no practical or economic support for installing an air barrier … I would think, if anything, the air leakage through a roof would be a minor problem at best.”

    “In terms of condensation control, I’ve seen many examples of serious air-leakage-fueled condensation in roof assemblies where the convention rules of thumb and hygrothermal modeling predicted little or non-condensation accumulation. (To be fair, most of these involved below-deck insulation with readily available access for air leakage into spaces with cold surfaces.)”
Different roof systems will have different requirements with respect to air leakage. In “A Guide for the Wind Design of Mechanically Attached Flexible Membrane Roofs,” National Research Council of Canada, 2005, the authors point out:

The speed of wind flow over a roof isn’t constant. The variation: speed causes rapidly changing fluctuations in the negative (suction) pressure induced on the roof assembly at a given time and location on the roof. Mechanically attached systems respond to wind loading differently than other roofing systems (ballasted, fully adhered, or protected membrane).

Because mechanically attached systems are attached at discrete points or rows, moderate to strong wind will cause the membrane to lift and billow between the attachment points or rows. The billow height is primarily a function of the wind load, the fastener row spacing, the modulus of elasticity of the membrane, and the presence of air/vapor barrier (retarder).

The insulation wetting studies conducted at CRREL (see “New Wetting Curves for Common Roof Insulations” in the Proceedings of the 1991 International Symposium on Roofing Technology) show that moisture can decrease the thermal resistance of many common roof insulations to less than 25 percent of their store-bought “dry” value. That represents an increase in the energy loss through them of 300 percent (4 times as much energy is passing through them, which is an increase of 300 percent).Since there’s so much attention focused on energy conservation, we need to emphasize the consequences of this insulation getting wet.

In Vapor Retarders for Membrane Roofing Systems, the author points out:

“Membrane roofing systems suffer more than their share of moisture problems, but most of those problems are due to flaws in the exterior waterproofing system, not to improper control of condensation. Flaws at flashing and penetrations are the primary cause of moisture problems for low-slope membrane roofs.”

The primary reason membrane roofing systems suffer few condensation problems is that most are built as tightly sandwiched compact systems that are quite resistant to air leakage.

Compact roofing systems, with their membrane fully attached with hot bitumen and with low permeability insulations adhered with hot bitumen, are remarkably resistant to air leakage even if no deliberate vapor retarder is present.

Loose-laid and mechanically attached single-ply roof membrane have some advantages since they’re not as sensitive to substrate movements as fully adhered membranes; however, their lack of complete attachment increases the potential for air leakage and condensation problems.

There are a couple of things to add to this. There is the Oak Ridge software, WUFI, which will calculate moisture movement in systems and whether there’s accumulation of moisture in components. Penn State University (PSU) has a group working in the sustainable building area that’s focused on internal air quality. The PSU position is that moisture is, by far, the dominate issue with air quality, and they’re working on system designs, including the use of barriers, to control this problem. This could be something the industry could support to develop better understanding on how these systems work and the best design aspects for their installation.

If it’s possible to form some conclusions from last month’s column and the information provided in this month’s column, it’s this:

  1. For low-slope roofing systems, air leakage is far more likely to occur at terminations and penetrations than through a fully adhered roof membrane system.
  2. For loose-laid and mechanically fastened roof systems, an air barrier primarily serves to reduce bulging and fluttering of the membrane, both of which can pump moisture-containing interior air into the roof system.
  3. For top-side vented systems to provide wind-uplift resistance, it’s necessary to have a totally sealed air barrier installed.
  4. While much attention has been focused on the vapor retarder vs. air barrier debate, at least for low-slope roofs, the primary concern is to keep the insulation dry by maintaining the installed membrane and flashings.

Related Coverage