October 8, 2015

Yes, Unvented Roof Assemblies Can Be Insulated With Fiberglass – A WUFI Post

And Not Just In Climate Zones 2B & 3B: Cold Climates Too
cathedral ceiling WUFI pro foam free

A post on August 21st, 2015 on GreenBuildingAdvisor.com, Martin Holladay asked the question, Can Unvented Roof Assemlies Be Insulated With Fiberglass?  Martin’s post focuses on the marketing of Owens Corning’s ProPink Unvented Attic Insulation System.   The Owens Corning system is only allowed by code in zones 2B & 3B with tile roofs.

And even still, Joe Lstiburek is quoted in the post saying “I don’t believe it will work without a vapor diffusion vent,” …“It’s too risky…”  Martin, then quotes the response from Achilles Karagiozis, Owens Corning’s director of building science: “Absolutely,”…“We are stuck with the code as written. It was written a long time ago, and it was written for one particular industry: the spray foam industry. Joe [Lstiburek] worked on that.”

Later Martin Holladay notes:

Owens Corning says that this system works with asphalt shingle roofing in Climate Zones 2B and 3B — but using asphalt shingle roofing on this type of roof is a code violation. “We all have to follow code, but at the same time, we can provide code officials with the engineering behind our recommendations,” Karagiozis said. “If we feel it makes sense, we will help the builder to provide information to the code official to get approval. We take these WUFI files.”

At 475 we agree with Joe Lstiburek, that we must make safe and robust enclosures.   And 475 agrees with Owens Corning, and Achilles Karagiozis that the current code was written for the spray foam industry and we can make safe unvented roofs with fiberglass insulation given the right approach – and that educating code officials for approvals of such systems is essential.

So building inspectors and plan examiners please take notice: we want to work with you and show we need not rely on toxic and unreliable spray foam but can instead safely and robustly use fiberglass insulation.   But must we be stuck in climate zones 2B & 3B?  What about the zones where most of you live and work: 4, 5 and 6?

In this post we will show, that while the ProPink Unvented Insulation System may provide a safe roof in zones 2B &3B, if we instead combine dense fiberglass insulation (batts or densepacked) with INTELLO smart vapor retarder and airtight membrane we can make safe and robust unvented roofs in climate zone 6!

Code pushes the envelope – IECC vs IRC/IBC

Wedlick taped sip on interiorCodes should provide the minimum parameters for performance, safety and durability. For instance IBC# does indicate that wood moisture content should be below 19% when installed, a dry roof (no condensation or leaks) is a healthy, structurally sound roof.  However since the Energy Code continues to increase R-value minimums, the risk of condensation/moisture accumulation in roof assemblies in cold and mixed climates has increased too. The only solution the code provides to this moisture drive issue is to add “sufficient” air-impermeable insulation below the deck to theoretically keep protect the sheathing from condensation/moist air-currents. None-the-less, it is known that an air-impermeable insulation layer will only work as an air barrier/vapor control layer, if it is also installed as one (board seams taped, spray foam mixed/applied correctly). This work not only should be performed meticulously, it should be verified by a blower door. However, if the installation is sloppy or not durable (foam shrinks – cheap tape is used or worse – caulk), the “airtight” insulation will not be airtight for very long. If the airtightness fails, then the risk of damages/rot or mold in the assembly increases substantially. You might get lucky, but if things go wrong – the consequences are costly (repairs, healthy problems, etc).

IECC 2012 took a first step to acknowledge these issues, and is mandating a verification of the airtight layer installation at 3.0ACH50 (Unfortunately, a residential code mandate only, so far, that a number of states have adopted). And we do see good designers, even when using SIPs, acknowledge the importance of sealing the interior seams of this airtight insulation material with airsealing tape (see photo on right).

We have gotten assemblies approved in several jurisdictions while working with the authorities having jurisdiction (AHJ), such as the building inspector, to show that one can design and build a roof that has air-premeable insulation (fiberglass, mineral wool, sheepswool, cellulose, etc) that can be declared  “air-impermeable insulation” with an air barrier which is also a (very) smart vapor retarder: INTELLO Plus as an integral part of the insulation installation. From experience, 475 knows that with the Pro Clima system, one can quickly make a very durable airbarrier. Every month we have dozens of customers that blow-by 3.0ACH50 and many reach 0.6 or better for Passive House certification – with more and more hitting 0.3 and lower. Since this membrane is part of a complete airtightness system from Pro Clima with TESCON tapes, gaskets, and adhesives – we know that the contractors can reliably build and roof assembly that is durably airtight (100 year lab tested tape) on the interior and can verify this with a blowerdoor test. Giving the AHJ confidence that the WUFI’s as shown below, can be relied upon – because without a certain level of airtightness, no reliable performance (not with INTELLO, not with foams).

If Owens Corning ProPink Boxed Netting is a Step Forward, Then INTELLO is a Leap Forward

Joe Lstiburek called the ProPink Boxed Netting in the GBA article “brilliant” and “fabulous”, and Martin Holladay notes the netting is well designed.  Both may be the case.  But it should be noted that it is this netting material that is the cause for the system’s narrow reach.  The netting is vapor open and not airtight.  Airightness is to be provided outboard of the insulation at the sheathing with canned spray foam, as GBA notes.   Consequently, if the ProPink system is installed in a cold climate, the conditioned air can easily reach condensing surfaces and cause moisture damage.  (And of course, airsealing is so important because air can carry 100x as much moisture into the structure then diffusion will – leading to damages much faster.)

But if the fiberglass is held in place with INTELLO membrane – the insulation has vapor control and airtightness inboard, where it should be – the conditioned air stays in the conditioned space and away from condensing surfaces.

WUFI The Challenge

To extend the use of non-foam insulated attic and cathedral ceilings with asphalt shingles, we have WUFI-ed a challenging roof in climate zone 6. The assembly has vapor closed, light colored asphalt shingles on the exterior over OSB sheathing. It avoids vent baffles, ridge vents etc – so considerably faster to construct both for new builds as well as renovations (details are available here). The required R-49 in Burlington VT (climate zone 6) was made with dense fiberglass.

We then assessed the performance of such roofs with different interior airbarriers, each with different vapor permeability profiles:  ProClima INTELLO Plus as the vapor variable retarder (0.17 to >13 perms), 6 mill polyethylene (fixed 0.05 perms), painted ‘airtight’ drywall (15 perms) and siga majpell (fixed 0.7 perms).   We want to know if air-sealing the insulation with these materials on the interior will lead to safe assemblies (ie once with some reserves to deal with unforeseen moisture) , that are easy to build or to retrofit.

Since we modeled the worst case roof, steep north facing, 10:12 pitch and light grey shingles – the solutions shown is a foam free cathedral ceiling that can be used throughout climate zone 6*, that use a limited number of readily available materials in a cost effective, durable way, with repeatable, easily verifiable methods that work for any orientation or shingle color. We want a roof with dependable performance: one that has reserves so it can deal with some small airleaks (at 3.0ACH50 small leaks can be expected).  Another objective is to have an assembly that does not require foam and is easy to install – applying batts in between rafters can be done fast and cost effectively – the cost of INTELLO Plus and TESCON VANA is less then the minimum required layer sprayfoam – and does not require biohazard suits to enter the enclosure, nor do homeowners have to evacuate the house.

In this WUFI study, we did assume a airtight construction with only minor airleakage into the insulation layer.  (Contractors must blower door the building and confirm airtightness.) The first graph below show total moisture content over 5 years from different approaches.  The assemblies sealed with 6 mill polyethylene and siga majpell on the interior are getting more damp every year – this is an indication one can expect issues in those assemblies over time (rot/mold).  The moisture content in the INTELLO Plus roof decreases and stabilizes at a safe level in year one.

WUFI INTELLO burlington VT

Do note that the drywall roof goes off the chart every winter. This should be a red flag as well. If we look at the graph below, we see that there is condensation on the OSB – it’s RH maxes out in November and it stays wet the end of May – so even though this roof dries out every summer, there are concerns for structural degradation and mold in this roof.

We should then also look at the relative humidity thresholds that ASHRAE 160P has defined. It states roof assemblies will be safe and reliable if 30 day averages of relative humidity do not exceed 80%. And if it does exceed 80%RH,  then the temp should be below 40F – which is too cold for rot and mold to occur. The graphs below show that with airtight drywall the sheathing is saturated in the winter.   But with the INTELLO roof, only when the sheathing is extremely cold – temps in low single digits – that the OSB sheathing goes above 80%RH.  This spike in RH can be explained, because when it is that cold, given laws of physics, materials hold less moisture, hence the RH briefly increases during those cold snaps.

RH and temp wufi fiburous insulation

A second safety check can be done in this case. It is generally accepted that if equivalent moisture content of wood stays below 18M% it will not degrade. For OSB we like to stay a bit lower – 15M% is a more conservative threshold, as it allows for some reserves in the structure. Below is the M% shown for the OSB sheathing. Except for winter #1 (when there is some residual construction moisture being pushed into the OSB) – the OSB stays right below 15M%.

M% of INTELLO covered fiberglass unvented roof

Drying reserves – sufficient at R49, convincing AHJ and can you do more?

This WUFI Pro modelling shows, that in climate zone 6 – unvented R49 asphalt shingle covered roof can be build without foam, and have a sufficient amount of reserves that enable it to cope with some unforeseen moisture ingress. As shown, it does need a vapor variable membrane that modulates from very vapor closed (0.17 perms) to really permeable (over 13) -> INTELLO Plus, that is installed in an airtight, verified manner (TESCON VANA tapes, gaskets, blowerdoor), with fibrous insulation that completely fills the bays (class I install for batts, density checked for blow-in insulation). For any cases that go beyond R49 or require assistance convincing the AHJ in your specific location/climate zone – please contact us as we are happy to assist with additional WUFI models.

We applaud Owens Corning’s efforts to to develop new and inovative solutions that reduce our dependence on spray foam.  INTELLO is a great partner with fiberglass to move this approach across most of our climate zones.

If you are an architect,an insulation installers and contractor can design and construct high performance roofs – that make plan examiners comfortable.  If you need additional support or WUFI files for your location – please contact us.



# Moisture content max per JM Spider page 2  and IBC 2304.7 Floor and roof sheathing.

* exempted are extra challenging constructions: greenroofs, high altitude locations or heavily shaded locations – please contact us for specific recommendations for such assemblies.

Note: Proclima INTELLO is included in Fraunhofer material menu in WUFI Pro, including it’s optimized vapor variable properties (see graph below)ProClima INTELLO included in WUFI Pro

, , , , , ,

Tweet about this on TwitterShare on FacebookShare on LinkedInPin on PinterestEmail to someone

15 Responses to Yes, Unvented Roof Assemblies Can Be Insulated With Fiberglass – A WUFI Post

  1. Sam Glass October 13, 2015 at 2:03 pm #

    Thank you for posting this interesting modeling exercise. I’m curious about a few things. You mentioned, “In this WUFI study, we did assume a airtight construction with only minor air leakage into the insulation layer.” Would you be willing to state how much air leakage and describe your method of modeling air leakage? Did you perhaps use the IBP model for air leakage that is built into WUFI? (See http://www.nibs.org/resource/resmgr/BEST/best3_kuenzel.2.14.pdf) What did you assume for solar absorptivity of the shingles?

    For the temperature and RH graphs, are you looking at the interior OSB surface (interface between OSB and fiberglass)? I noticed that the legend in the RH figure slightly obscures the RH peak for the assembly with INTELLO (blue curve) during the first winter. I also wanted to clarify the reason that the RH peaks during cold weather. It is not because “when it is that cold, given laws of physics, materials hold less moisture.” In reality, equilibrium moisture content of wood at a given RH is slightly higher at lower temperature (opposite of what you’re saying). WUFI, however, does not include temperature-dependence in the moisture storage function (it’s a fairly small effect). What’s going on in the simulation has to do with diffusion rates. When the OSB is cold, water vapor diffuses from the interior to the surface of the OSB and adsorbs (changes phase from vapor to bound water). Diffusion of moisture within the OSB, away from the surface, is relatively slow because it is cold, so the OSB moisture content increases and the RH increases (because the RH and moisture content are linked through the moisture storage function). Another phenomenon that would be occurring if you are using the IBP model for air leakage is that when it is cold, this model would deposit moisture into the OSB, which would increase its moisture content and the RH at the surface.

  2. Floris October 14, 2015 at 12:23 pm #

    Sam, thanks for your comment and question.
    We did include and airleakage per IBP model at airtightness class 3 in the assembly at OSB/fiberglass interface – with a stack height of 5m (17′), but no mechanical over pressure. This so we would be pretty conservative regarding the mistakes in the air-tightness of construction, but is not equivalent to one big airleak in the roof.
    If one turns off the airleakage, the assembly with INTELLO performs even better and the RH stays below 80%RH and the EMC completely below 15M% after year 2.

    The fact that the air-leakage would deposit vapor at the OSB interface and the diffusion of vapor into the OSB is slow at such low temperatures, as you point out is accurate and explains the spike marginally.
    However the RH and the temperature swings correlate too much, for the laws of physics to be ignored and WUFI does include it. I am double checking with the WUFI developers to confirm my understanding is correct.

    You are correct that the vapor diffusion away from the face the OSB is slow at low temps, however it doesn’t explain the short time dip in RH between the two spikes above 80%. In that period, vapor would still be deposited (temperatures in the 10′.20’s) and with below freezing temperatures the diffusion in OSB would still be rather slow as well. So a spike can logically only be the result of relative humidity increasing in the material when it becomes (much) colder quickly.

  3. Tim Johnson February 5, 2016 at 3:50 pm #

    Good article! My concern is with sun driven moisture from above. Under the shingles did you use an impermeable layer as in ice and water or what was the perm rating for the underlayment? I am in Zone 7 and often thought if you put a steel roof on with a vapor open (to some degree) underlayment that this would be safe. There would be some vapor release because of the air channels in the steel yet the sun would not push in moisture.

  4. myrooff October 6, 2016 at 8:04 am #

    Thanks, useful.

  5. John M February 17, 2017 at 4:04 pm #

    Very helpful article – I’m not sold on the effectiveness of spray foam, and it seems like everything I read spray foam is the only solution available to insulate an unvented (cathedral) roof.

    Question, do you think an application of fiberglass (or more specifically stone wool is my preference) combined with intello under a standing steam metal roof would prevent moisture issues? Our roof has an ice & water shield on top of the OSB as well, but is completely unvented. I’m in VT as well (zone 6).

    • Floris February 17, 2017 at 4:33 pm #

      JOhn, you certainly can use fiburous insulation (fiberglass, stone/mineral wool, etc) in such a roof if covered on the interior with airtightly installed INTELLO Plus in residential or similar uses (ie normal interior RH).
      We have a growing number of projects, with vapor closed to the exterior (like your ice&water) that have used this approach, and which have no moisture (condensation) issues – which we have confirmed by both computer/WUFI Pro modelling as well as on site/real condition monitoring.

  6. Chad March 13, 2017 at 4:44 pm #

    I have a tar and gravel roof on my 1950’s ranch house in Zone 5. Everything I read points to using only closed-cell foam for my unvented roof but the cost is prohibitive.

    I would like to fill the cavity with Roxul instead and use INTELLO Plus. Does the tar and gravel roof change any dynamics with moisture control?

    • John March 14, 2017 at 2:47 pm #

      Hello Chad,

      With any retrofit, there are a lot of variables – but if you’re following the rules from our Ten Golden Rules for Foam-Free Flat Roofs there’s a good chance that Roxul + INTELLO will be the right fit. Give us a call when you have a chance and let’s discuss further. If you have a sketch of the assembly or any photos you can share that would make it even better.

  7. Keith H April 10, 2017 at 3:46 pm #

    Does plywood vs osb change the equation at all? With a flat roof, does steel help or hurt (I could believe without exterior insulation that it would hurt)?

  8. Keith H April 10, 2017 at 3:48 pm #

    I’m also wondering about the assumptions regarding interior heating controls. If you have an owner/occupant who sets a very low thermostat setting, does the equation change, presumably via increased interior RH?

  9. Floris April 12, 2017 at 2:56 pm #

    Keith, thanks for your questions. Any changes in setup/enclosure will lead to different results. Would think that changing to plywood probably doesn’t make a large difference, since there is no outward drying in vapor closed roofs. However, plywood would be able to cope with a bit more humidity before getting damaged.

    Regarding the interior heating controls, of course this changes things -however the absolute vapor pressure most likely will be the same, so it is probably safe to theorize that the exterior vapor drive wouldn’t be changed much even with a lower interior temperature.

  10. stephan jacobs September 2, 2017 at 2:17 pm #

    great thread. I have a detached garage unvented pyramid hip roof. Circa 1924. T&G fir decking. 2×6 rafters. New GAF membrane and Ice & Water, with asphalt shingles. Zone 5 / Boston. I do not want to use closed cell foam. Are there installation instructions for an INTELLO Plus / Fiberglass assembly on your site?

  11. Marianne Castillo April 28, 2018 at 3:30 pm #

    I’m insulating the interior side of an asphalt roof w/no insulation and wonder if your product might work. The walls of the building are brick, an historic property, in Kingston NY (don’t know the climate zone). Would the Intello Plus batt insulation product work. Looking forward to your suggestions.

    • John April 30, 2018 at 2:03 pm #

      Hi Marianne,
      Have you downloaded our Historic Masonry details? https://foursevenfive.com/product/high-performance-historic-masonry-ebook/ It’s where we’ve consolidated all our knowledge resources regarding brick retrofits like yours. And we developed these based on our experience in New York City, so we’re talking about the same climate.

      The short answer to your question is: yes, INTELLO Plus is going to help you during the retrofit process. For more on how, use the details found in the ebook, in addition there are loads of examples on our videos page: https://foursevenfive.com/knowledge-resources/videos/

      Hope that helps!

Leave a Reply