If a roof insulation surface is considered "vapor impermeable", can interior vapor pass through such a tightly sealed assembly and condense on the roof and create moisture issues?

What is the vapor impermeability of 12" EPS and how about 6" Polyurethane?

Posted By Lbear on 26 May 2014 07:45 PM
If a roof insulation surface is considered "vapor impermeable", can interior vapor pass through such a tightly sealed assembly and condense on the roof and create moisture issues?

What is the vapor impermeability of 12" EPS and how about 6" Polyurethane?

ASTM E96 methods test the permeance under a standard set of conditions.  The permeance of both EPS and polyurethane vary with both thickness and material density.

For US building code purposes there are three defined orders of magnitude for vapor permeance of building materials.

Class-III vapor retardency is between 1 and 10 perms in an ASTM E96 test, and considered "semi-permeable"

Class-II vapor retardency is anything between 0.1 and 1 perms, and sometimes called "semi-IMpermeable".

Class-I vapor retardency is sub-0.1 perms, and generally regarded as impermeable, though in reality most building materials that meet that spec would allow at least some (if extremely slow) moisture migration via vapor diffusion.  (eg.  6 mil polyethylene sheeting tests at about 0.05 perms, not 0.00000 perms.)

A layup of #30 felt + asphalt shingles runs about 0.1 perms. That's about 2x as permeable as 6 mil poly, but not very permeable.

Type-II EPS (1.5lbs per cubic foot nominal density) is about 2.5-3 perms @ 1", and would run about 0.2-0.25 perms @ 12", which is still more than twice as vapor permeable as the felt + shingles. 

Half pound polyurethane foam ("open cell") runs about 4-6 perms @ 6" which is fairly vapor open but only semi-permeable, - comparable to standard latex paint, whereas most 2lb polyurethane foam would run about 0.15-0.20 perms @ 6"- far more vapor open than 6 mil poly, and still at Class-II vapor retardency range, and only semi-impermeable

Condensation & moisture adsorption are only an "issue" for materials that don't tolerate high moisture content well (such as OSB or plywood, and to a lesser extent,  wood planking). Moisture reaching the #30 felt via vapor diffusion through foam would not be a problem, since it even tolerates immersion in liquid water pretty well.

A layup of #30 felt + asphalt shingles runs about 0.1 perms. That's about 2x as permeable as 6 mil poly, but not very permeable.

Everything I've seen is closer to .6 perms, which puts it well above 12" of EPS. So any moisture making it through the foam (from the interior) won't be trapped against the felt/shingles layer (not even temporarily).

Probably even true when you add OSB or plywood, as long as you consider their permeability at realistic (ie, not ASTM E96 desiccant method) %RH levels (as some flawed studies do).

In some climates you could add poly or smart membrane on the inside to reduce the moisture movement to low levels that can exit through typical outer layers, no matter how you measure it.

So a peel & stick synthetic membrane like Grace has a vapor permeability of 0.05 Perms. Is it possible for exterior water vapor to pass through such a membrane and condense onto the roof OSB?

I was reading an article that claimed exterior water vapor can pass through a synthetic membrane and get trapped and rot out OSB since it has nowhere to dry out.

What is the perm rate of CLOSED-CELL polyurethane?

Posted By Lbear on 27 May 2014 09:02 PM
So a peel & stick synthetic membrane like Grace has a vapor permeability of 0.05 Perms. Is it possible for exterior water vapor to pass through such a membrane and condense onto the roof OSB?

I was reading an article that claimed exterior water vapor can pass through a synthetic membrane and get trapped and rot out OSB since it has nowhere to dry out.

What is the perm rate of CLOSED-CELL polyurethane?

Exterior moisture will not pass through a 0.05 perm membrane at a rapid rate under any circumstances, but there has to be at least some drying path for any moisture that does find it's way in.  Increasing the vapor permeance of the exterior layers to merely the ~0.1-0.3 perm level has no real benefit.  The drying path has to be on the interior side of the OSB, either roof deck venting, or semi-permeable/variable-permeance layers on the underside of the OSB. 

The problem isn't the low permeance of the roofing- all roofing is low permeance, with very low drying rates toward the exterior. The problems come when there is a moisture trap, when there isn't sufficient drying path on the interior side of the OSB.  (The OSB itself runs about 1 perm, 2 perms when it's nearly saturated with moisture.)
 
As stated in my prior posting, 2lb density (closed cell) polyurethane runs about 0.15-0.2 perms @ 6" of thickness.

The average temperature of the OSB is almost exactly the same as the average temperature of the shingles- so there isn't a huge vapor pressure drive across the membrane roofing from the exterior, but there can be from the interior (particularly in winter.)  The more common threat to the roof decking comes from interior side moisture being adsorbed into roof decking over time when the roof decking is stagnating below the dew point of the conditioned space air. This is far more problematic when the transport mechanism is via air leaks from the interior side finish  into air-permeable/fiber insulation, which can easily move orders of magnitude more moisture than vapor diffusion through 3-5 perm latex ceiling paint.  The IRC has developed standards for how much air-impermeable insulation there must be between the fiber insulation & roof deck (at the code-min total R values) to mitigate risk of interior moisture drives reaching rot levels in OSB roof decking, by climate zone.  In that sort of stackup the interior surface of the air-impermeable insulation becomes the condensing surface (and yes, it will condense into liquid water there if there is much air leakage), but the average temperature at that surface in most homes will be no cooler than the wintertime dew point of the conditioned space air at a presumed moisture level. (For US climates that presumptive dew point in the code varies from 35F-45F, depending on climate zone.)

Technically the moisture from vapor drives will never  actually "...condense onto the roof OSB...", but rather, it is taken up by the OSB in the form of adsorb when the temperature of the OSB reaches the dew point temperature of the proximate air volumes.  Some simple science in the home: Put some OSB in the refrigerator overnight next to a full beverage container then pull them both out and set them on a table or counter. Observe what happens over then next 15 minutes.  The beverage container will likely develop condensation on it's surface, whereas the OSB will not.  If you weighed the sample carefully both before & after, you'd find that the water is being taken up inside the OSB rather than forming as liquid on the surface.

Posted By Dana1 on 28 May 2014 12:02 PM
The drying path has to be on the interior side of the OSB, either roof deck venting, or semi-permeable/variable-permeance layers on the underside of the OSB. 

So if I understand this correctly, then a closed spray foam roof or a SIP roof could not have a peel & stick membrane on the exterior roof surface as this would lead to the OSB rotting on the roof surface. The water vapor would collect on the underside of the synthetic membrane between the membrane and the OSB and have no way of drying out, eventually leading to rot and roof failure.

The EPS in a SIP is air-impermeable, but only semi-impermeable to water vapor. The exterior skin can dry very slowly toward the interior. Given the fairly low permeance of the EPS + interior OSB there is effectively no winter time moisture accumulation in the exterior OSB from interior moisture drives.

But any moisture that somehow gets into the exterior OSB takes years, not months to dry. This makes SIP roofs somewhat more susceptible to damage from roof leaks over the long term. It's also a known issue that air leakage at the ridge lines and seams of SIP roof panels will create problems- you can't just squirt some caulk in there and hope. Better SIP builders air & vapor seal the seams & ridgelines with EPDM membrane on both the interior & exterior sides.

Water vapor doesn't "...collect on the underside of the synthetic membrane between the membrane and the OSB...", rather, the moisture collects as adsorb in the OSB and releases into the semi-impermeable EPS as a function of temperature.

The drying rate only needs to be high if the drying season (== warmer temperature season) is very short & shallow relative to the wetting season (== the colder temperature season). If you have 0.05 - 0.1 perm roofing on the exterior side and 0.15 - 0.25 perms of foam on the interior side it isn't really a problem in anything but sub-arctic zone type climates, from a vapor diffusion moisture transport point of view. With just vapor diffusion factors alone the moisture content of the exterior side OSB will see only miniscule seasonal changes, and will oscillate, not climb continuously. Problems only occur when roof leaks or air leaks depositing moisture in the OSB.

The 0.05 perm peel & stick membranes mitigate against roof leaks reaching the OSB, and when applied over seams block continuous air leakage at the seams from drawing air into the seam, depositing moisture. In general they help more than they hurt.

Posted By Dana1 on 28 May 2014 04:30 PM

The 0.05 perm peel & stick membranes mitigate against roof leaks reaching the OSB, and when applied over seams block continuous air leakage at the seams from drawing air into the seam, depositing moisture. In general they help more than they hurt.

Dana,

Thanks for the scientific insight. You really know your building science.

There seems to be so much confusion and contradictory advice out there on how to properly design a roof design based on vapor permeability. I know each climate zone poses its own unique setup.

I was talking to someone and they were claiming that:

We've seen cases where there was liquid water behind the synthetic house wrap high up on a gable wall, above all penetrations and and underneath a reasonable overhang with no sign of air exfiltration. One explanation is that, first thing in the morning, water vapor passed through house wrap from the outside-in and condensed on the relatively cold SIP.

This is where I was confused as it didn't make sense that water vapor can pass through a synthetic house peel & stick wrap with a 0.05 perm rating.

They went on to say:

The rate of water accumulation is relatively slow as it is limited by the cumulative permeance of
the envelope assembly interior of the Ice & Water Shield. The difficulty is caused by the Ice & Water Shield being hydrophobic and having a lower permeance than the total permeance of everything inboard of it and it being located on the cold side of the assembly. The water vapor is stopped by the Ice & Water Shield, condensing on the interior side of it. Slowly, liquid water accumulated to the point where the moisture content of the exterior OSB skin became favorable for mold and fungus growth (wood rot).

Dana’s depth of knowledge about this subject certainly surpasses mine, but I would think we should be able to determine what the indoor/outdoor temp/RH would have to be to in order to cause this to happen. In other words, wouldn’t this ASHRAE based Glaser dew point approach answer this question if we enter the correct parameters for the assembly layers?

http://www.borstengineeringconstruction.com/Building_Assembly_Moisture_Analysis_Calculator.html

This Glaser dew point approach assumes steady state indoor/outdoor temp/RH conditions, or perhaps better stated, it doesn't address transient conditions. So I wonder if it is perhaps the transitory nature of the environmental conditions that may cause this to happen?

Standard Tyvek is north of 50 perms, Typar is about 15 perms, most of their competitors are between those numbers. See:

http://www2.dupont.com/Tyvek_Weathe...K01472.pdf

#15 felt is about 5 perms when dry, north of 50 perms when wet, something of a "smart" class-III vapor retarder.

The OSB and the housewrap are at the same temperature so the housewrap is the condensing surface closest to the exterior skin.  (The OSB can only be a true condensing surface when saturated and cannot take up the moisture as adsorb, but we'll ignore that for now.) The first condensing surface in the AM dew scenario is the siding, then the housewrap- the OSB is third in line. Their explanation doesn't really hold up, even though the housewrap material is fairly vapor permeable.  Chronic wetting from improper flashing, or high vapor drives from sun heating a moisture-reservoir siding such as stucco is a more likely proximate cause of high moisture levels in the exterior skin of a housewrap-clad SIP.

Ice & Water Shield is a vapor barrier, and severely impedes drying toward the side on which it is adhered, and is risky to use on a wall assembly made of air-permeable or highly vapor-permeable construction in a cold climate. But the direction of vapor drive varies with temperature, and in climate with reasonably long seasons where the mean outdoor temp is well above 40F(the nominal indoor air dew point, moisture adsorbed into the outer skin of an Ice & Water Shield clad SIP over a winter will migrate back toward the dryer-cooler conditioned space during the warmer seasons- it doesn't just keep on accumulating from interior moisture drives year-round- the vapor pressure difference changes directions seasonally. (The exception would be high humidity environments such as pool or hot-tub rooms, steam showers running at a high duty-cycle, etc.)

The transitory nature is both why it can be made to work and the challenge in calculating it. As long as the interior side is the source of moisture, things don't dry towards the interior and, without sufficient permeability to the exterior, moisture levels will continue to rise. So in such a case, you count on the situation changing before it gets too bad. And that it changes enough to provide more drying than there was wetting (which may not be the case without AC).

My guess is that solar heating of the roof surface plays a big role in making it work.

> water vapor passed through house wrap from the outside-in and condensed on the relatively cold SIP.

This is where I was confused as it didn't make sense that water vapor can pass through a synthetic house peel & stick

He was talking about a wall, which is going to have a permeable synthetic wrap.

Based on what Dana indicated, it sounds like this is just a seasonal situation. I wouldn’t consider a seasonal situation to be a transient situation. I would consider short duration changes (i.e., perhaps hourly temp/RH changes) to be a transient situation. So, based on what Dana indicated, I would still fully expect the Glaser dew point analysis approach to be reasonable.

However, the Glaser approach does NOT account for solar heating (or wind driven rain, material absorption, or infiltration)…it really only models vapor diffusion. So those unaccounted factors could be significant and could very well be in play here too. However, if these factors are not significant and this is indeed only a seasonal situation, this Glaser analysis should tell a reasonably accurate story (i.e., forecast the mositure accumulation rate and on what assembly layer it will occur).

So with a standing seam metal roof, will condensation occur and get trapped behind the metal roof as stated here?:

Steel will condense on the underside whether the SIPs are properly sealed or not. The condensation will exacerbate the SIP problem described in the above (due to improper sealing you are emphasizing), but even if the SIPs are installed correctly, the steel will still condense. If it has a drain and evaporation plane/space (on purlins for example), drying will occur under the steel. If the steel panels are flat and flush mounted to the deck they will hold condensation in place by capillary action and be slow to dry, so when the sun heats the roof, the vapor drive will push the moisture into the SIP panel faces (passing through any semipermeable bulk water barrier)."

Lbear, I would be happy to run the Glaser analysis for you if you provide the indoor/outdoor temps/RHs you expect and the permeance/R-values values for all the assembly layers. Then maybe Dana could comment on the analysis results.

Okay, here's what I have so far:

Indoor Temp: 70F
Indoor RH: 35%
Outdoor Temp: ?
Outdoor RH: ?
Layer 1, Indoor Air Film Permeance: 160
Layer 1, Indoor Air Film R-value: 0.64
Layer 2, 7/16” OSB Permeance (@ 50% RH): 2
Layer 2, 7/16” OSB R-value: 1.6
Layer 3, 8” Polyurethane Core Permeance: 0.15
Layer 3, 8” Polyurethane Core R-value: 48
Layer 4, 7/16” OSB Permeance (@ 50% RH): 2
Layer 4, 7/16” OSB R-value: 1.6
Layer 5, Roof Membrane Permeance: 0.05
Layer 5, Roof Membrane R-value: 0.5
Layer 6, Metal Roof Permeance: 1
Layer 6, Metal Roof R-value: 0.1
Layer 7, Outdoor Air Film Permeance: 1000
Layer 7, Outdoor Air Film R-value: 0.17

I got most of the values from ASHRAE. Do they seem reasonable Dana? The air films just allow us to have a layer interface to see if there will be condensation on the external sides of the assembly. Are there also sheet rock and paint layers in this assembly? Any internal air gap layers? What shall we use for the outdoor temp/RH? Should we perhaps use a higher temp than normal climate to try to capture the solar heating effect?

Posted By sailawayrb on 30 May 2014 07:29 AM
  Are there also sheet rock and paint layers in this assembly? Any internal air gap layers? What shall we use for the outdoor temp/RH? Should we perhaps use a higher temp than normal climate to try to capture the solar heating effect?

INTERIOR CEILING:
2.5"  space stuffed with 2" Roxul Mineral Batts
1/2" sheet rock taped & finished (some areas will be painted with latex paint while others will have a T&G wood ceiling)

> can interior vapor pass through such a tightly sealed assembly and condense on the roof and create moisture issues?

Here (fig. 13e) is a simulation of what happens when you have an exterior partition that is permeable on the inside and impermeable on the outside. It is well into the mold inducing zone, even in a warm climate (results would be far worse in a cold or mixed climate). Would also be worse with stack effect driven air leaks (the simulation used an optimistic zero air movement). At least in the OP's specific case, the foam is fairly impermeable, so buildup against the Ice & Water Shield is slowed. But the simulation looks relevant to ceilings without thick foam or other vapor barriers and an impermeable exterior barrier.

Nobody would put a polyethylene vapor barrier on the outside of a cold climate exterior wall. But somehow doing an equivalent thing is common for roofs. It is best (but not necessarily required) to build a cold roof and give the moisture a place to exit throughout the cold season.

Okay, here’s what I used as a "first look".

December/January Outdoor Temp & RH: 26F & 81%
July/August Outdoor Temp & RH: 87F & 75%

Indoor Temp: 70F
Indoor RH: 35%

Layer 1, Indoor Air Film R-value: 0.64
Layer 2, 0.5” Latex Painted Gypsum Board R-value: 0.45
Layer 3, 2.5” Roxul Mineral Batts R-value: 10.5
Layer 4, 7/16” OSB R-value: 1.6
Layer 5, 8” Polyurethane Core R-value: 48
Layer 6, 7/16” OSB R-value: 1.6
Layer 7, Roof Membrane R-value: 0.5
Layer 8, Metal Roof R-value: 0.1
Layer 9, Outdoor Air Film R-value: 0.17

Layer 1, Indoor Air Film Permeance: 160
Layer 2, 0.5” Latex Painted Gypsum Board Permeance: 5
Layer 3, 2.5” Roxul Mineral Batts Permeance: 42
Layer 4, 7/16” OSB Permeance (@ 50% RH): 2
Layer 5, 8” Polyurethane Core Permeance: 0.15
Layer 6, 7/16” OSB Permeance (@ 50% RH): 2
Layer 7, Roof Membrane Permeance: 0.05
Layer 8, Metal Roof Permeance: 1
Layer 9, Outdoor Air Film Permeance: 1000

For December/January the Glaser analysis indicates that moisture will accumulate at layer 5/6 interface at rate of 0.0133 grains/sf-hour. For July/August the Glaser analysis indicates that no moisture will accumulate at any of the layer interfaces.

Posted By sailawayrb on 31 May 2014 12:35 PM
Okay, here’s what I used as a "first look".


Layer 5, 8” Polyurethane Core Permeance: 0.15
Layer 6, 7/16” OSB Permeance (@ 50% RH): 2
Layer 7, Roof Membrane Permeance: 0.05
Layer 8, Metal Roof Permeance: 1
Layer 9, Outdoor Air Film Permeance: 1000

For December/January the Glaser analysis indicates that moisture will accumulate at layer 5/6 interface at rate of 0.0133 grains/sf-hour. For July/August the Glaser analysis indicates that no moisture will accumulate at any of the layer interfaces.

Thanks Sailawayrb!

So is the moisture accumulation rate of 0.0133 grains/sf-hour at layer 5/6 during December/January something that would cause concern?

During the swing seasons, I assume the rates would be less than the peak winter rate of December/January, correct?

Hi Peter,

I believe 0.0133 grains/sf-hour is pretty insignificant and would normally just be absorbed by the OSB and then eventually dry if it isn't trapped. This assembly will dry toward the inside. Dana, is the psychrometrics and building assembly expert on this forum and can likely provide you better guidance than I can. I believe I still have your email address from our past discussions about ICF buildings and architecture in general. If so, I will email you the input/output info for the two aforementioned Glaser analyses that I did so you may use the software yourself to play with outdoor conditions or change any of the layer characteristics.

Gayle