GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Audio Play Icon Headphones Icon Plus Icon Minus Icon Check Icon Print Icon Picture icon Single Arrow Icon Double Arrow Icon Hamburger Icon TV Icon Close Icon Sorted Hamburger/Search Icon
Q&A Spotlight

How to Make a SIP Roof Better

Will a layer of rigid foam over the panel help or hurt?

Rot in a SIP roof:

Unchecked air leaks in a structural insulated panel roof can lead to decay. That was the conclusion reached by Joseph Lstiburek, a principal at the Building Science Corporation, after investigating roof failures in Alaska. The question is whether a layer of rigid foam on top of the panel has the potential to trap moisture and create a similar problem.
Image Credit: Building Science Corporation

Roger Lin’s Washington, D.C., house will have a roof of 12-inch-thick structural insulated panels (SIPs). By most standards, that’s a well-insulated roof. But Lin wants to add 2 inches of rigid foam on top of the panels to reduce thermal bridging.

He’s uncertain about the details. He has already installed roofing underlayment over the panels. Can he put expanded polystyrene foam on top of the underlayment and cap it with metal roofing? Or does he need a layer of plywood or furring strips over the foam before the metal roofing is installed?

His question, posted in GreenBuildingAdvisor’s Q&A forum, is the topic for this week’s Q&A Spotlight.

The discussion quickly turns to the risk of rot in the oriented strand board (OSB) skins of the SIPs, and two competing points of view emerge.

The foam will help protect the roof

GBA senior editor Martin Holladay likes the idea of using foam over the roof panels. “Installing a layer of rigid foam on top of your SIP roof to address thermal bridging is an excellent idea,” Holladay writes.

The foam layer should be airtight, he adds, so seams should be sealed with high-quality tape. Metal roofing should not be installed directly over the foam but on furring strips or a continuous layer of plywood. And if plywood is used, then Holladay recommends that Lin install 2x4s from soffit to ridge between the foam and the plywood to create ventilation channels.

“These ventilation channels will give you a cold roof — cheap insurance if you are going to the trouble of getting your details right,” he says.

The bottom line: the extra layer of exterior foam will keep the OSB warmer and dryer,…

GBA Prime

This article is only available to GBA Prime Members

Sign up for a free trial and get instant access to this article as well as GBA’s complete library of premium articles and construction details.

Start Free Trial

27 Comments

  1. Expert Member
    Dana Dorsett | | #1

    EPS yes, vented nailer deck yes, mineral wool, not so much
    The OSB skin of the SIP can dry plenty well through 2" of EPS, but only if it has somewhere to dry TO. Roofing felt and composite shingles are relatively un-ventilated and the shingles are very low perm. Even 1x furring between the foam & nailer deck provides a huge improvement in drying capacity if vented both top & bottom, but 2x furring cut in sections with 1' gaps to allow cross ventilation is an even larger improvement. It's the moral equivalent of a rainscreen on a wall, but needs a bigger cross section to provide the same ventilation due to the lower convective drive of a sloped roof.

    Mineral wool has the disadvantage of gravity pulling any moisture leakage through to the bottom. If vented top & bottom it becomes something of a high-impedance mesh-type rainscreen, but loses a fraction of it's insulating capacity to the very convection that's doing the drying. EPS is waterproof and an air-barrier, definitely a better way to go.

  2. Philipp Gross | | #2

    Can diffusion be a problem or not!?
    The note of Peter Yost: "Really good mechanical, spot ventilation, and a hygrometer installed for homeowners to KNOW their interior relative humidity." makes me believe that someone else is worried about diffusion. I stand with my point that 2" foam increase that problem and contribute little to the overall airthighness, To my knowledge 2" EPS has a perm rating of approx. 1. Yes there is trying potential but after all the foam causes more water (given all the details are done correctly) and decreases the trying potential. I also stand with my point that thermal bridges caused by wood members should be considered by calculating with the "true R-values". If you do this they are not evil and you are better of increasing the SIP thickness if you want a better thermal performance.

  3. GBA Editor
    Martin Holladay | | #3

    Response to Philipp Gross
    I can't imagine any scenario in which vapor diffusion would cause OSB rot in SIPs. How much vapor makes it through 6 inches or 8 inches of 12 inches of EPS? Not much, if any.

    The moisture transport mechanism is air exfiltration, not diffusion. Dana Dorsett notes that vapor diffusion through EPS can be your friend rather than your enemy, by allowing some outward drying through a layer of exterior foam on top of the SIPs. Like you, I'm skeptical that much diffusion can occur through the EPS -- a tiny amount, which is certainly good, because it will help remove any moisture -- but not much.

    However, as long as the EPS is installed on a sunny day when the OSB is dry, I remain convinced that an exterior layer of EPS lowers the risk of OSB rot rather than increasing it, for the same reason that exterior foam sheathing on a wall lowers the risk of OSB rot. It keeps the OSB warm and dry, so it won't accumulate moisture in the winter.

  4. John Harris | | #4

    Seal it tight
    I plan to use SIPs in my next house, wall and roof. Proper sealing of the joints is key for a tight structure, no air movement in or out which removes moisture problems inside the panel. I had the idea that once the shell is up that the whole structure is sprayed inside and out with a sealer that fills any joints and OSB edges and flats, any sills, openings, penetrations, etc., so the shell unit is water tight and air movement proof.
    Finding panels greater than 12" thick is not easy, probably a custom panel. Adding a couple inches of foam to the roof to reach R60 is a good idea, in my opinion, removing any thermal bridging from the internal I-joists I would need in the sip roof panel joints. If you have sealed the panel by spray or roll on membrane and add foam then 2"x4" furring, covering with plywood or sealed OSB venting from facia to ridge for a cold roof and air movement then there should never be a problem with rot. I plan to do something similar with siding, a 1x2 furring up the side and siding over that for an air channel. As Martin says: "it's all in the proper detailing", no matter what you are building with.

  5. Philipp Gross | | #5

    Confusion about diffusion
    Martin, You could probably convince me that (depending on the climate) diffusion is not a problem in this and many other cases. I also agree with you that air exfiltration is a much bigger problem with SIP`s and in general. However your last point ("It keeps the OSB warm and dry, so it won't accumulate moisture in the winter") I do not agree with and maybe we just have to agree to disagree. But I will ones more try to make my point: Your outside temperature and relative humidity (or just the absolute humidity) gives you a water vapor pressure and so do your inside conditions. In the winter the inside pressure is much higher causing water vapor to move outward. The perm rating tells you how easy or hard it is for the water vapor to pass (resistance) but it will pass no matter what. In Europe we use the sd-value instead (pretty linear to perm) which gives you an equivalent thickness of air that would have the same resistance as the material you are looking at. The advantage with the sd-value is that the water vapor pressure profile inside to outside can be shown in a linear trend. You can than calculate the dew point at every material transition point. and find out your condensation potential or if there is any at all. I guess all I am saying is that it is a little more complicated than just keeping the sheathing warm. In general that sure helps but in this case it would actually increase the condensation potential because the absolute humidity at the OSB surface is higher than without the EPS.

    It is true that condensation happens on cold surfaces but cold is relative. I moved from Germany to Minnesota and I was quite impressed when my glasses fogged on a hot humid day leaving work....

  6. GBA Editor
    Martin Holladay | | #6

    Response to Philipp
    Philipp,
    There are two ways to resolve this question. One way is with WUFI. The only problem with WUFI is that it requires a thorough understanding of the program and the underlying physics, or else you can fall victim to the garbage in, garbage out problem.

    The other way is to apply basic principles. Do you really think that the permeance of 6 inches of EPS is high enough to give rise to significant water vapor transfer through the SIPs? I don't. Vapor diffusion pales in significance compared to the amount of moisture transport that can accompany exfiltration.

  7. Philipp Gross | | #7

    Martin,
    1) I agree that WUFI

    Martin,

    1) I agree that WUFI could give some answers to this. I was trained in the program and had to go through a lot of the underlying physics during my studies and claim to understand them pretty well. I would never choose this assembly and therefore not do the calcs. Also this is so climate dependent as you know, which was never part of the discussion.

    2) I don`t think water vapor transfer is a good name. There is a absolute humidity outside and a different one inside. Within the assembly the absolute humidity is in between the two and the permeance defines the profile. The question is: Could any surface temperature be "cold" enough to cause condensation given the absolute humidity at that point? I think the answer is yes. Which brings us/me back to number 1) and look at it more closely.

  8. GBA Editor
    Martin Holladay | | #8

    Response to Philipp
    Philipp,
    You wrote, "The question is: Could any surface temperature be 'cold' enough to cause condensation given the absolute humidity at that point?" I agree -- that is the question.

    Case 1: A SIP roof without any additional exterior foam. The exterior OSB is cold in winter.

    Case 2: A SIP roof with an additional layer of exterior foam. The exterior OSB is warmer than Case 1.

  9. Aaron Gatzke | | #9

    How to install mineral wool?
    Not to sound stupid but is the reference to mineral wool in regards to a mineral wool panel? Don't mineral wool panels have an impermeable covering? Once taped, wouldn't the result be the same as foam panels?

  10. Philipp Gross | | #10

    Martin, what you miss is the
    Martin, what you miss is the second parameter, the absolute humidity at that location.

    Case 1: exterior OSB cold and little absolute humidity

    Case 2: exterior OSB warmer but more absolute humidity

  11. Expert Member
    Dana Dorsett | | #11

    It's the air leaks at, not the diffusion...
    I'm with Martin- vapor diffusion through even a 6" EPS core and the interior OSB skin will be under 1-perm, maybe even less than that through an asphalted kraft facer on batt insulation (~0.4 perms). Moisture accumulation on the exterior skins of a roof SIP from interior moisture drives will be due to air leakage (usually at the ridge line first.)

    SIP roofs tend to rot from the ridge down and along section seams, due to incomplete air sealing or failure of the air seals. Air seals on SIP roofs are more likely to fail than those of SIP walls because roofs are more dynamic than walls- they see higher temperature peaks & valleys, and dry more slowly, and in some climate there are significant static loading of snow, etc. All this leads to greater seasonal dimensional shifts and separation at the butted up OSB skins, but it's the ridge seam that sees the biggest shift, and is the hardest seam to seal well in the first place. Study where the rot is occuring in that first picture. It's pretty obvious that the greatest damp is at the seams. Were it a vapor diffusion problem it would be everywhere and fairly even. This is air leakage.

    Under moisture bearing stucco or brick cladding on OSB sheathed walls it's customary to leave 1/8" of gap between OSB sheathing sections to accommodate the dimensional changes from seasonal moisture, but there's no really great way to do something similar on a SIP roof and still get a good air seal. A little venting on the exterior goes a long way, even it has to diffuse through 2" of exterior EPS.

  12. Philipp Gross | | #12

    Dana, The funny thing is that
    Dana, The funny thing is that I agree with all you say and yet argue that the 2" EPS on top worsen a potential condensation problem. I am not saying that the damage shown in the photo has anything to do with condensation it is very obvious that this as most problems with SIPs are caused by air exfiltration, not condensation. What I am concerned about is that because we build for 100 years (or at least should be) we should still consider potential condensation especially when we create assemblies that are out of the ordinary. Potential problems may show after decades because as we all agree the amount of condensation may not be much.

  13. Albert Rooks | | #13

    Back to Mineral Wool
    Dana,

    This confuses me: 

    "Mineral wool has the disadvantage of gravity pulling any moisture leakage through to the bottom. If vented top & bottom it becomes something of a high-impedance mesh-type rainscreen, but loses a fraction of it's insulating capacity to the very convection that's doing the drying. EPS is waterproof and an air-barrier, definitely a better way to go."

    Are you taking about moisture as a liquid or a gas? In either case I don't see the issue. It's too dense for liquid to be a serious issue under a vented roof (which was the original application that we were talking about). A gas will vent up, not down since the sheathing will be warmer under the MW than above it.

    I don't see where a high density / highly permeable panel is anything but an improvement over the moisture trapping scenarios that Martin and Phillip are discussing.

  14. Albert Rooks | | #14

    Response to Aaron
    Aaron,

    Thats Not a dumb question. Mineral wool is available with or without applied membranes. The typical "outsulation" application is to use MW because it is an excellent insulator, nearly indestructible to time water and pests, and yet still vapor open so that it does not trap moisture in the layers below it.

    When applied, it's usually 6lb density or better. At that density it is not susceptible to convective looping and therefore does not need additional  air sealing by tape or other means.

    I like MW because with it's high permeability effectively removes all of the concerns of trapped moisture that are being debated by Martin and Phillip while still performing the same insulation functions.

  15. Craig Payne | | #15

    Not all SIP Panels are EPS
    I thought it only fair to make sure others reading this blog and stream of comments understands that not all SIPS are made with EPS. A few companies use high density PU and can achieve twice the R-Value per inch while having a lower perm rating as compared with EPS...even though I agree that is not the main area of moisture concern. Also, one of these PU SIP companies can build roof panels with integrated cam locks (no splines or other structural connection needed) and achieve R-60 in an 8 1/2" panel.
    I'd like to get some opinions about one option: How about using the ZIP roof sheathing as the exterior skin on the SIP and then ZIP taping the seams after the residual minimally expanding foam that was put inbetween the panels was scraped off? Would any of you be hesitant to use this application and or want to add another layer of underlayment over the ZIP sheathing before the final roofing material is installed?

  16. Expert Member
    Dana Dorsett | | #16

    Maybe I wasn't being so clear
    Philip: "Dana, The funny thing is that I agree with all you say and yet argue that the 2" EPS on top worsen a potential condensation problem. I am not saying that the damage shown in the photo has anything to do with condensation it is very obvious that this as most problems with SIPs are caused by air exfiltration, not condensation."

    The moisture damage from exfiltration is very much a condensation issue, just not a vapor-diffusion issue. With air leaks a lot more moisture is moved up to where can condense and be trapped than would ever happen with vapor diffusion through the SIP.
    Adding EPS alone doesn't worsen the condensation, but it doesn't enhance drying unless there's a vented cavity to dry into. With enough exterior R value as foam above it, the OSB can stay warm enough to be above the dew point of the interior air preventing ANY condensation at the OSB layer, even at a leaky joint, but even at R values short of that it will still reduce the number of hours where the OSB is below the dew point substantially, allowing it to dry quickly enough after it warms up. (This is climate & R value dependent, of course.) Staggering the seams of the exterior foam with that of the SIP provides a secondary air barrier over most of the seams blocking exfiltration paths there, but not over the more susceptible ridge joint. Adding a vent cavity to dry into above the foam makes it far more resiliant to moisture from both air leaks and roofing leaks.

    albert: "I don't see where a high density / highly permeable panel is anything but an improvement over the moisture trapping scenarios that Martin and Phillip are discussing."

    It allows liquid moisture from leaks to find it's way to the OSB. Without a vent space above it the high air-retardency of the high density fiber that is it's virtue becomes a problem. It still has the impermeable shingles above, and it needs air flow to dry. To be sure WITH a vented nailer deck above it's drying capacity will be orders of magnitude higher than with EPS, but it also allows orders of magnitude more water in from exterior moisture from leaks. With the vent space above it'll work great, without the vent, maybe not. Martin's approach of creating a vent gap is nearly foolproof, as it adds a secondary air barrier for when the air-seal of the SIP (almost inevitably) fails, and provides a substantial drying path toward the exterior. Whether insulating above the SIP with semi-permeable foam or highly permeable high-density fiber, the vent cavity is critical.

  17. Philipp Gross | | #17

    Maybe I wasn`t eloquent
    Dana, Sorry for my mistake of saying condensation instead of condensation caused by diffusion. English is my second language and I sometimes have a problem with expressing myself clearly. Obviously exfiltration causes condensation and huge issues and so can diffusion just not as fast. I think we are on the same page here. Now for my argument let take exfiltration out of the equation and just look at diffusion. Assuming the SIP`s are connected right and tight. I argue that the condensation caused by diffusion gets worse with 2" EPS. The reason is as mentioned above that the absolute humidity at the OSB surface is higher than without the EPS and despite the OSB being "warmer" the relative humidity and the times it exceeds 100% increases. More foam on the other hand changes the equation. I posted the attachment (glaser diagram) already in the Q&A section but I`ll post it here ones more. The top graph shows the saturated water vapor pressure (dew point) and the bottom graph the actual water vapor pressure at the location of the assembly. When the two meet there might be condensation and the more the bottom graph is interrupted the more condensation potential there is. This is for Minneapolis (climate zone 6) where we might have to actually care about diffusion. The glaser analysis has its weaknesses but it definitely can give you an idea of whats going on. In this case it tells us 2" of EPS make things worse not better. I 100% agree with you on the importance of a vent-space for either solution but the EPS slows the drying potential down.

  18. Leif Dressler | | #18

    Philipp is right
    A bit of a late commentary, but I think Philipp is right. Sure the added EPS lowers the OSB temperature, but not enough to offset the additional moisture it traps. I think Martin has elsewhere written about the exterior 2/3 insulation 'rule of thumb' that applies with systems like the REMOTE wall to adequately lower condensation levels at the sheathing: it seems to me that this same logic should apply here.

  19. GBA Editor
    Martin Holladay | | #19

    Response to Leif Dressler
    Leif,
    You wrote, "the added [exterior] EPS lowers the OSB temperature." That is only likely to be true during the summer.

    During the winter, on the other hand, the addition of a layer of exterior EPS should raise the OSB temperature, not lower it.

  20. Nicolaas Wilkens | | #20

    Rigid Insulation over SIPS Walls
    I am following up on this conversation and curious if the same concerns are applied to SIPS walls. Specifically, how applying a layer of rigid insulation (probably XPS) over a 4 1/2" SIPS panel wall would run into the same performance issues if minimum required thickness foam is installed for a climate region 5 (near the border of 6). The installation of foam would allow uniform extension of a continuous plane extending down over floor support beams and wrapping under to the under into under floor insulation. 1x4 strapping for rainscreen exterior wood siding would be applied over a WRB over the foam.

  21. GBA Editor
    Martin Holladay | | #21

    Reply to Nicolaas Wilkens
    Nicolaas,
    If the exterior OSB facing of the SIPs is in good shape, and if the SIP seams are well sealed, I see no problem with your plan.

    Attention to details is important, however. You want excellent flashing to keep wind-driven rain out of your walls. Since you are planning a rainscreen gap behind your siding, I think you will be fine.

  22. Nicolaas Wilkens | | #22

    Thank you!
    Thank you for the feedback.
    N

  23. William Sherman | | #23

    Existing PU roof SIPs -- only 3 1/2" PU!
    I live in a timber frame home built in 1991. It has SIP walls and roof. It appears as though the PU insulation in the roof panels is only 3 1/2" thick; woefully inadequate for central Massachusetts. With all the issues mentioned in the article and posts above, would I be better off removing the old SIP roof and replacing with SIPs of sufficient thickness to get closer to R-50? The inside (underside) of the all SIP roof panels can be seen from the interior top floor as all rooms on the top floor have cathedral ceilings. I'm seeing telegraphing of the SIP panel edges on the roof top when there's snow up there.

  24. GBA Editor
    Martin Holladay | | #24

    Response to William Sherman
    William,
    If I were you, I would wait for good weather, and would assess the condition of your existing SIPs during your next re-roofing project. Strip the existing roofing down to the SIPs. If the outer face of the SIPs are dry and in good shape, there is no reason that you can't install one or more layers of rigid foam on the exterior side of the SIPs, followed by new roof sheathing (OSB or plywood) and new roofing.

    In your climate, I strongly recommend that you include a ventilated air gap between the outermost layer of rigid foam and the top layer of roof sheathing. You can create this ventilation space by installing 2x4s on the flat, soffit to ridge, 16 inches on center.

    Here is a link to an article with more information: How to Install Rigid Foam On Top of Roof Sheathing.

    One last point: You should consult an engineer to make sure that your plan for fastening your upper layers of material (especially the 2x4s that create the ventilation channels and the new roof sheathing) to the lower layers (the existing SIPs or the timber-frame below the SIPs) is adequate. You need enough fasteners, and the fasteners need to be long enough, to hold everything together.

  25. Cliff Walker | | #25

    I have used sip panels since 1982. I build log and timber frame homes. My log homes usually have timber frame roofs so we use sips here also.
    Since the interior timber frame is the structure the sips are curtain walls with little thermal bridging, only at openings with only a 2x set in to allow attachment of fenestration. This is a very good system and I have achieved ACH of .57 and consistently below .8 in my net zero ready timber frame homes., for the past 10 years. I built my first R 2000 home in 1986. We of course pay great attention to joint sealing techniques on the interior side of the sip. We rain clad vent all exterior surfaces of the sips using 3/4” venting cavity’s bottom to top. Often double strapping with 1 x 4 to allow for attachment of siding and roofing. We use only breathable membrane on the exterior of sips, except perimeter of fenestration, roof valleys, roof perimeter to within 3’ of interior space. We never use asphalt shingles and almost exclusively steel roofing.
    We close a 2500 sq ft home in 5 days plus or minus, with interior finish of exterior walls and ceilings complete. The exterior is membrane wrapped and thus once fenestration is installed including doors the home is weather tight and protected from the elements. We commonly use R 40 walls and R 50 roofs and triple glazing in heating zones 4, 5, 6.
    In 1982 we used Atlas sips out of Ayer, mass., we have used Insulspan, Frank Baker founder, professional engineer, sips since 1985 plus or minus.
    Been a fantastic relationship and we have plus 200 successful sip installs.
    In my opinion there would be no robust timber frame industry without sips.
    In addition during this time I have had 3 clients, 2 of them contractors, opt to conventional frame walls and roof, against my objections. Each one, candidly admitted after their experience that sips would have worked better.
    In my experience the expense up front of this system is well worth the integrity of it’s longevity.

  26. Ed_WI | | #26

    In 2005 I built a timber frame home in Wisconsin with SIP walls and roof. I taped the interior SIP seams except where timbers interfered. During the first year or two I occasionally heard what I believe was the sound of foamed joints failing. It sounded like a squirrel running across my roof as the foam split traveled down my seams. Within a few years I noticed ghosting along many of the seams through my hot roof asphalt shingles. Two springs ago I had some dripping inside my house that I presume was melting frost within SIP seams. I think I've read everything in this forum on this topic. I have my doubts whether a permanently perfectly sealed SIP house is achievable. Regardless, that ship has sailed for my house and I am searching for an exterior solution. I plan to take action this spring and am working on the assumption that the exterior SIP OSB faces are not rotten. I'm leaning toward tearing off the shingles, putting down mineral wool in the hope that escaping moisture will condense above the SIP seams, and creating a cold and ventilated roof above the mineral wool. Any advice will be appreciated.

    1. BlackHawkDesigns | | #27

      Ed it looks like you're experiencing the classic ridging issue for a sips roof. I'd definitely take off the shingles as soon as you can and do a cold roof and then maintain vapor open with a new or updated underlayment to dry out the OSB now and for the future.

Log in or become a member to post a comment.

Related

Community

Recent Questions and Replies

  • |
  • |
  • |
  • |