More Building Science
For the last ten years, most of my work has been designing renovations and new homes, and now I am building again as well. But I also sometimes help clients fine-tune the details on homes designed by others.
In early 2017, I was hired to consult on details for a small, high-performance home. Nick, the owner, and his mother, for whom he was building the house, care deeply about our environment and wanted to minimize both embodied carbon and operating carbon, which are similar to embodied energy and operating energy, through the lens of global-warming emissions.
Although exposed, sealed concrete slabs are popular for finished floors in high-performance homes, concrete comes with a large carbon footprint, responsible for 10% of man-made greenhouse gas emissions. Nick and his mom were adamant about having warm-looking, locally sawn white-pine floors. They did not want a basement or a crawlspace, the conventional options in the northeast, because they require extra labor and material to build, and they can be hard to detail so that they remain free of moisture (and resulting mold). Plus, there was ledge (the regional term for bedrock) not far below grade, and we weren’t sure if we could even get code-compliant clearance in a crawlspace. They were committed to the concept of a slab, just not the concrete material.
Nick asked if we really needed concrete under the flooring. Conventional wisdom says yes: I’ve installed and/or specified wood floors over concrete many times, and while the assemblies can be handled a number of ways, they always include a layer of concrete somewhere. But then I recalled a project by builder Stephen DeMetrick and architect Steve Baczek including a concrete-free basement floor. They floated two overlapping layers of subfloor over foam and a vapor retarder, without concrete. But that was a basement…
Weekly Newsletter
Get building science and energy efficiency advice, plus special offers, in your inbox.
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 TrialAlready a member? Log in
49 Comments
One minor clarification is that frost wings are not actually required by code where this house was built, as long as the vertical exterior insulation extends at least 12" below grade. But when you go beyond code minimum insulation below the slab, you are reducing the amount of heat pushing down from the conditioned space. The ASCE 32-01 design guidelines for frost protected shallow foundations doesn't really address this issue, so I split the difference with their requirements for unconditioned buildings, which in my zone require 4' to 5' frost wings to retain natural geothermal energy to keep the ground at the footings from freezing.
Great detail and great to hear that after a year it's performing flawlessly. One of my biggest hesitations would have been with squeaks. I've done a number of FPSF high-performance houses (in Montana) with monoslabs & shallow crawlspaces. I always design them per the Revised Builders Guide mentioned above as unheated buildings since I go to such great effort to try not to inject any of the heat from the building into the ground. The cost to do the full wings required with EPS foam around the entire perimeter always seemed like pretty cheap insurance to make sure there weren't any future issues with frost heaving or other structural issues associated with no digging to frost depth. No issues so far on any of my projects either. Thanks again for sharing. I will definitely keep this method in mind on future projects to try to eliminate some concrete as well.
Do you have any insight into how the cost compared to an insulated raft-slab or a more conventionally insulated (fully wrapped) mono-slab?
Thanks James. While the EPS frost wings are good insurance, they do represent embodied carbon, so I try not to overdo it. You could use recycled foam but it does not have the insect-resistant borate treatment that I prefer, only available in new EPS (and even then it's a bit hard to find).
It's good to hear that you haven't had any trouble with FPSFs. Mike Guertin just posted on Instagram that he is about to try a similar method to what I wrote about, except with a frost-protected permanent wood foundation. On the project I wrote about we considered several different foundation types before settling on the ICFs as a DIY-friendly approach; if working with a concrete contractor, it might be simpler to pour a solid wall.
I don't have specific cost information but when I present this idea to people, I typically say that if you want the least-expensive system, go with an exposed concrete floor--just use pozzolans to replace some of the Portland cement. But if you require a wood or tiled floor, consider skipping the concrete, as it's just not necessary.
Only though that comes to mind is twisting 2x4 under that advantech becoming a nightmare.
I suppose a couple months drying on stickers would fix that culling issue
That did come up during discussions, and we considered both the overlapping Advantech approach and using LVL or PSL for sleepers instead of spruce, the standard framing material here in New England. The owner/builder said he was careful to only use straight and straight-grained stock, and that he thinks the subfloor and finish flooring keeps enough pressure on the sleepers that they can't twist enough to matter. Letting the lumber fully dry would also work.
Hi Michael,
Great article. I'll take your concrete free slab foundation and raise it one. About 3 years back I had an 'aha moment' sitting in a bath pondering how one could remove concrete altogether from a foundation while making it foam free and high performing. I came up with the attached. We have since run a THERM model on it without any issues and have run it through a structural engineer and it works with the seismic requirements of the PNW where our firm works. I've also contacted a CLT manufacturer and have gotten some initial pricing feedback which is still higher than a typical excavation/insulation/footing/slab, but within about 15%.
We haven't built it yet in the wild, and there are some details to hash out in terms of constructibility timelines and MEP but in theory it seems to work. Just food for thought...
Joshua, that's a creative design, and looks like it would work. Nice job. Learning more about CLT is high on my to-do list.
INteresting thought. The CLT will certainly provide a stiff floor. I'm a bit concerned that the CLT is nto strong enough to act as a beam between the pile caps. Since the piles are so much stiffer than the Comfortboard, most of the weight of the house is going to be bearing on the piles and the CLT will be spanning between the piles. You could beef it up by bonding an engineered beam to the CLT at the perimeter. A beam the width of the wall would work find, and properly fastened and bonded to the CLT floor, would become part of the CLT. Might let you drop back on the overall stiffness of the floor itself. Essentially a stiffened CLT slab. You could probably even allow for interior wall loads this way. Just thinkin.....
@ Peter, I suppose it depends on how many piles one has, how they are attached and how thick the CLT is. Our engineer didn't see this as an issue, but I am certainly not an engineer and your point is taken.
Also, we intentionally used Comfo bats in lieu of Comfortboard as we figured we would set them a bit higher than the bottom of the CLT so they would compress and provide complete contact to the bottom side of the floor. We didn't intend for it to have any bearing capacity at this condition.
Joshua, this is a Passive House on helical piers: https://www.ecocor.us/our-work/ash-point-passive-house.
I worked for Ecocor when we built this house and I helped develop the framing details, so I'm speaking from experience to some degree. Helical piers are inexpensive but not free, probably around $350/pier but it varies depending on several factors.
One thing you might consider in your floor framing is something they do in parking garages and other large structures, with a relatively thin concrete slab (comparable to your CLT) with integral beams. Search "T beams" to see the concept. The CLT manufacturer might be able to do something similar for you, if it helps with the economics--the fewer piers you need, the less expensive and easier to build it will be.
@Michael That is a beautiful project, nice work! I'd love to see the details on helical pier foundation. I'm guessing it is crafted of piers and beams as the finished floor looks to be a few steps up from grade. Is that the case?
Thanks for the info on the T beams. I agree the less piers the better
Response to #30:
Joshua, yes because this was a Passive House we used deep I-joists filled with insulation, with code-required 18" clearance above grade. The joists rest on girders.
Joshua,
That's a really interesting approach. Conceptually it's a bit of a different animal than Michael's slab-on-grade without the slab. More a floor on piers with the area under the skirt filled with insulation. Thinking of it that way, is there a big advantage to using CLT over flush beams and floor joists?
I'd agree with Malcolm's interpretation of Joshua's design as structurally having more in common with pier construction rather than a slab on grade, but it eliminates some of the major issues with pier construction--providing an 18" clear space, air sealing, providing for mechanicals, etc.. From an engineering perspective it would likely be more resource-efficient to frame the floor more conventionally, rather than using CLTs--you would need less total lumber and fewer helical piers. But I like the out-of-the-box thinking.
It is similar to piers, but even those use concrete (admittedly, not nearly as much) and my thought experiment was to not use concrete, so being true to my self imposed terms- this was the solution.
One of the other reasons we thought to use CLT's is that one can order them with a different species of wood on the top layer. In our case, we would order them with oak and use the top layer as the finished floor. This was a cost and time savings idea since we could cut out the subfloor and flooring install. Another reason for CLT use is the labor savings since (in theory) we could set the floor system much quicker. We only need a 4.5" thick CLT which would reduce the finish floor height of the foundation relative to grade vs. using beams and joists. This would also make the insulation easier since one would have to insulate between the joist cavities, whereas with the CLT's this isn't the case.
Granted, some of these ideas may or may not be worth the hassle of the CLT (long lead times, engineering, etc.). It is also true that if the intention is to avoid excavation the finish floor height will be higher than a slab on grade foundation making ADA/accessibility a potential issue.
The idea of using flush beams and floor joists is a perfectly good idea and would also avoid the use of concrete. I could see this being a more traditional and approachable way of crafting a concrete free foundation and worth exploring further.
With either approach, the thinking was one could show up at a flat lot, screw in the piles, dump and spread gravel, set the rock wool bats and then set the floor system. This would save a ton of costs- I know in the PNW excavation and hauling pricing is pretty dynamic (and mostly in the more expensive direction) and so it concrete. Avoiding the CLT's in lieu of a framed system could save even more costs and make this significantly cheaper than a traditional footing/slab on grade system.
Thanks for entertaining this idea and adding to the conversation! Its one of the great things about this site.
Jonathan,
If I seem to be looking for weaknesses in the approach, it's because it's really interesting one that it's worth thinking through the implications of.
The sequencing is a bit more involved, in that the plumbing has to be roughed in and located precisely before the insulation and CLT floor panels are placed.
The two other areas that look a bit tricky, are attaching and sealing the continuous vapour-barrier, and completing the skirt, which needs something to back it, and as the consequences of pests entering the inaccessible area below the floor are pretty serious, a more robust material extending deeper below grade.
@Malcolm
You zero'd in on exactly the things we have been discussing in regards to sequencing. The plumbing is a tough one, but not overcome-able. We thought it could be easy to route out a recess in the CLT for a threshold free shower pan and had some other ideas for wet areas. We also had some discussions about the vapor barrier, and one idea was to use pond liner. It is robust available and affordable. We thought maybe we could laminate it to the bottom of the CLT panel prior to install (exactly how is TBD) and then come up with a solution for the seams. We had a rabbited detail that could have a vapor barrier wrap into it and then attach with sealant during install. You are totally right about the skirt- we are still fishing for ideas to make this a durable yet efficient install. We've had a few, but nothing that has stuck. The pest issue is a real one and also needs to be thought out.
Thanks for the feedback!
If the main goal is to eliminate concrete, could the helical piers be used at the perimeter only to replace the ICF stem wall. Then span them with a beam the same thickness as the sub-slab (floor) foam, attach a skirt and EPS to the outside of that. Then frost protect the whole thing with wing foam and continue as Michael did above. The exterior bearing on pier and beam surrounding a frost protected floor inside. The beam would have to be treated and which point cutting right to permanent wood might make more sense. For that matter would helical piers (being metal) have less carbon footprint than concrete here?
Lastly, the article, the concept, and the discussion here are just beautiful. Thanks, Michael.
Joshua, That is a very intriguing detail. I do have one question. How do you propose to attach the cement board that is at the base of the wall. it appears to not have a surface to nail or screw to.
Doing research at GBA for a series of MPP (Mass Plywood Panel) 1, 2 & 3-bedroom residences my company is prototyping and thought I'd share or first concept that we're engineering now upon coming across the detail that Joshua proposed. Structurally, our MPP first floor slab is only required to be 2" thick and MPP walls 2" thick. We've conceptualized a perimeter beam on Helical Piles to also avoid use of concrete. Getting this approved through our local building department will be another story...!
Intuitively I was against this approach, but the more you think about it, the less justifications for including the slab you can come up with.
The main problem with the assembly, which would make it difficult to use in some climates, is that you need to move from installing the subfloor, to having the roof on and the building weather-tight, without rain or snow.
We just did a similar floor system in Northern Minnesota. The house was designed to be completely built off the foundation, no interior walls needed to support any roof loads. The house was completely dry before the floor system was installed. We did not use a frost protected foundation, instead chose a full ICF foundation, in my climate, 5 feet to bottom of footing.
Randy, any chance you'd want to share drawings of your details? My project was the first time I had done this, and there are always things to learn. I only know of a few others who have done this type of system.
Michael,
Thinking about it again, and in light of Randy's comments, it might be as simple as using a 2"x8" sill-plate so you could build up to the roof, and still have something that tied the subfloor to the exterior walls.
We modified the rough drawing slightly, the vapor barrier is between the two sheets of EPS. We found it easier to final grade and then lay a sheet of foam down so we didn't have to walk across the graded rock fines. The entire floor assembly floats between the ICF and double bottom plate.
One other modification, the Rockwool Comfortboard 80 lines up with the ICF foam. The vented rain screen then overhangs the foundation foam by 3/4 inch.
Good looking details, Randy. I've been watching progress on your build on Instagram (@northernbuiltpro for anyone following along) but hadn't seen your building section.
Do you have any concerns about the flooring system not being tied to the foundation walls? I was concerned about it, which is why I have Advantech spanning over the mud sill, and a cast-in shelf on the interior. But both may be overkill.
I'm not concerned about the floor not being tied to the walls. We left a 1/4 inch gap between the double bottom plate and Advantech just incase we had a little expansion and made sure the soil below was well compacted. Was more concerned about having a rain event that flooded the floor assembly.
Malcolm, that's a good point. You would want to keep a giant tarp handy in my climate zone, and it might not work at all where you are in the PNW.
I'm in the middle of constructing a small accessory building for use as a B&B. Because the floor system is very close to grade, it can only be insulated from above. I've had to cover the whole subfloor with a pond-liner until I can get the roof on.
Malcolm, a couple of years ago I designed a renovation that preserved the existing walk-out basement apartment but replaced the roof with a 2 1/2 story addition. The builder covered the floor with EPDM, built the house, and left the EPDM in place as an air barrier and mass damper for sound attenuation.
Good idea about the sill plate.
Did you considered using fab-form monopour ? Should be easier to put the eps at the base of the footing.
Eric, I have not used that product. On the project pictured, most of the foundation was directly on ledge, so the ICF forms were scribed to fit and a separate footing wasn't necessary. It's not always important to fully insulate the footings, but when it is, I don't see how the Fab-Form Monopour would allow for easy insulating?
Awesome! I must try this at some point.
Mike - do you know if Glavel is making any inroads yet? They do seem like they might be a solid alternative for some of the techniques using foam below grade.
Hi Geir, I know it's available here https://foursevenfive.com/glavel-foam-glass-gravel/ but I have not spec'd it or seen it used on a project yet. It's often hard enough to get builders to order EPS instead of using readily available XPS; trying to get them to order Glavel in bags nine weeks before it's needed seems like a stretch. But it seems like a good product.
I love this sort of innovation. However, PWF or CLT foundation will severely hurt the resale value of any house.
Kevin,
I'm not sure that is true everywhere. Around here pier foundations don't seem to have an appreciable affect on real estate prices. The same seems to be true in flood prone places like New Orleans. It's probably quite dependant on where you are.
Reply to #35:
>If the main goal is to eliminate concrete, could the helical piers be used at the perimeter only to replace the ICF stem wall. Then span them with a beam the same thickness as the sub-slab (floor) foam, attach a skirt and EPS to the outside of that. Then frost protect the whole thing with wing foam and continue as Michael did above. The exterior bearing on pier and beam surrounding a frost protected floor inside. The beam would have to be treated and which point cutting right to permanent wood might make more sense. For that matter would helical piers (being metal) have less carbon footprint than concrete here?
Andy, that's a very interesting idea. It wouldn't have worked on my project because piers and ledge aren't a good mix, but in some cases that might be a better approach than either a conventional or frost protected permanent wood foundation. I've seen stats that steel and Portland cement both release somewhere between 1 and 2 pounds of CO2 for every pound of material. Considering the weight of a concrete foundation vs. a pier foundation, I'd go with steel. I'm not sure what the carbon footprint is for modern pressure treated lumber--something I might dig into. (and thanks for the kind words!)
Michael,
Thanks for this article--I've wondered for a long time about a viable low-energy alternative to the ubiquitous slab-on-grade approach. A question: you specify a "heavy duty vapor retarder" rather than a vapor barrier, so that's a material between .1 and 1 perm. I can't think why heavy duty poly (a "barrier" with a perm rating maybe .01) wouldn't work here. Is some small amount of vapor transmission desirable? Am I missing something?
Norman, like many building nerds I try to be consistent and never use the term "vapor barrier." There are class 1, class 2 and class 3 vapor retarders, and vapor-open assemblies. Vapor retarders used under slabs are class 1, less than 0.1 perms. I typically spec either Stego or Viper 10- to 15-mil polyethylene vapor retarders, depending on the situation.
Thanks for the clarification!
It sounds like that vapor barrier is also the primary air-barrier for the floor, is that correct?
In terms of moisture considerations, do you feel like above the foam is the better place for the poly, or did it have more to do with the way it integrated with the rest of the build?
I recall other articles saying 'it doesn't really matter what goes down first (poly or foam)' when insulating above a slab. I've felt like poly above is better if doubling as an air barrier to keep interior air from condensing underneath the rigid (taping those seems though would perhaps help/suffice), and perhaps condensation there isn't a huge concern.
Tyler, yes, the poly is also the airtight layer at the floor. The main reason I thought it would be best above the foam is that it could drape cleanly over the perimeter wall, instead of needing to be tucked, folded and sealed. Plus it's habit; that's where it should go when there is a concrete slab as well. But due to the thick layer of crushed stone and good perimeter drainage it probably would have been fine under the foam as well. The Advantech subfloor is also pretty airtight, so I don't think there will be much indoor air getting into the floor system.
I wonder how this may work in a more cooling centered zone 3/4 climate such as Raleigh, NC where insulation may not be desired sub slab. The frost line is 12''. Would it work to have just well packed gravel topped by a vapor barrier and advantech or is the EPS needed for stability?
Eli, I don't design for warm climates so I'm not sure of the energy balance but the IRC requires R-10 sub-slab insulation in Raleigh, NC (climate zone 4A). In any case, the foam is not required for stability but I would not place wood in a location where it is likely to fall below the dewpoint temperature, so I would not recommend placing the subfloor directly on gravel (with or without a vapor retarder).
Michael, thanks so much for sharing this. This is quite along the lines of what I'd like to do for my new build. My question is regarding the building officials and code compliance. The FHB version of your article mentions this method is code compliant. Could you share a bit more on the sections of the IRC you used to make your case to the inspector, and how that conversation went? My county official isn't exactly the most open-minded guy :) but I'd really like to eliminate the concrete slab if I can. Like your project, I have stem/frost walls and a floating slab in-between & above grade. Nothing fancy.. Thanks!
Brett, there is nothing in the building code that requires a concrete slab, so there's nothing I can point you to there! If you are in a seismically active zone or if you have expansive soil, you may need the slab to resist lateral loads, but I know that others such as Josh Salinger in Oregon have done it without a slab.
I used (and often use) a frost-protected shallow foundation, as allowed here: https://codes.iccsafe.org/content/IRC2021P1/chapter-4-foundations#IRC2021P1_Pt03_Ch04_SecR403.3. If your code official balks, you could work with a licensed structural engineer who would use document ASCE 32-01 to design a foundation that works.
It's common to ask a code official which part of the code they are using to reject a design. Have you tried doing that? They may get defensive, and to be fair, they have a difficult job and it's easy to go with what is commonly done. But if they can't point to a place in the code that prohibits a design, it should be allowed to be built.
Michael, thanks for the quick reply! This is very helpful.
I'm definitely approaching the official with a very friendly demeanor and not looking to rile anyone :) I'm in the Midwest outside of any seismic zone. So as we move into spring here, soon, I'll be getting started on my project, hopefully in April, but that will be determined by how fast things dry out.
Thanks again for your valuable advice!
You're welcome! Everyone is in a seismic zone, though: https://codes.iccsafe.org/content/IRC2021P1/chapter-3-building-planning#IRC2021P1_Pt03_Ch03_SecR301.2. Figure R301.2.2(5) shows that you're probably in seismic zone A, which basically has no risk of activity. ;-)
If you want to really make your code official happy, it would be best to find a licensed engineer to approve your design, so you aren't asking the code official to take on risk with something they probably don't fully understand. Not every engineer is free-thinking enough to accept a concrete-free slab, so you might have to look around a bit.
Log in or become a member to post a comment.
Sign up Log in