I am in Vancouver Canada, and am planning on doing an upgrade to the first floor of a slab-on-grade house.  Part of the upgrade will involve removing existing carpet and foam underlay (on bare concrete), and laying down insulating subfloor plus a floating floor of some kind (probably engineered hardwood).  In the upgrade, I am probably going room by room as time permits.

A question I have is what are the building science implications of using foam contact with the slab (eg, OVRX Barricade or full slabs of 2'x8' XPS) as the base of my plywood subfloor vs. one of the air-gap products (eg, Dricore or the equivalent built up using basement moisture barrier).  Putting aside the cost vs. convenience differences of the subfloor panels vs. using sheet goods, both OVRX and Dricore, for example, make diametrically opposing claims on their building science.  OVRX claims that one wants to have full contact with the slab in order to avoid having a space where moisture and mould can accumulate.  Dricore claims that one wants to have an air gap so that any moisture from the concrete essentially can dissipate into the room.  BSC doesn't seem to have specific guidance for slab-on-grades, just basements.

It seems to me that the best thing to do would be to use sheets of XPS (eg, 1" Foamular C-300) bonded and taped together, and foamed to the walls.  This would force moisture out the side of the slab, with the bonded sheets of foam acting as a moisture barrier.  The air gap idea seems to depend on vapor pressure from underneath the subfloor pushing out through to the edges of the subfloor.

I haven't yet torn up the carpet, but as far as I can tell, there are no tell-tale signs of moisture problems through the slab.

Does anyone have any opinions or experience on this?

In my opinion, OVRX has the better argument.

My only issue with OVRX is that it's only R3.2 under the floor, which isn't much, and R12 of XPS is a hair less vapor-permeable than some folks recommend (but close enough- most say R10 max, if XPS.)  But their moisture control reasoning is sound.  The air gap idea is brain-dead unless it's fully vented to the outdoors somehow.

On the slab R5-R10 of (1-2") of XPS sheathing would be more appropriate for Vancouver's subsoil temps & heating fuel costs.  Foamular 150 is good enough- you don't need the high compressive strength of C-300 in a residential flooring app. with a full subfloor above it.  This is a house, not a road-bed or airport runway.)  Cheaper still would be to use Type-II EPS (bead-board) under the sub-flooring, but it takes up comparatively more headroom, and it takes some long tapcons to go higher than R16 (~4").  But if you have the headroom, there's an economic argument for R16 in EPS where there isn't if XPS.

Similarly on the walls, R12 is a bit on the minimalist side. R16-20 is reasonable if you do it cheaply.  You can get there with 1-1.5" of XPS (R7.5) glued to the foundation wall, with a batt-insulated 2x4 studwall w/batts snugged right up against it.  In Vancouver's climate as long as the foam/fiber R ratio is 1/3 or higher  the cold edge of the studs will stay well above the dew point of the interior air, and NO interior vapor barrier should be used- not even kraft facers and certainly no poly.  (This is in contravention to Canadian code, but valid building science.)  The XPS  is permeable enough to allow ground moisture to continue to dry toward the interior rather than rise higher in the foundation wall putting the foundation sill & band joist at risk, and with the mild winter temperature averages of the coastal climate there is little to no condensation/mold risk as long as the studwall can dry toward the interior.  The argument is best made to the inspector that the XPS is the vapor barrier (which it is, at a perm rating about an order of magnitude higher than poly.) A 16" on center studwall with R13 batts comes in at about R10-R11 when the thermal bridging of the studs and plates are factored in, so with an inch of XPS you're looking at ~ R16 total.  If you went with 2" of unfaced low density Type-I EPS rather than an inch of XPS you'd be looking at R19-R20 for similar money, with better interior drying capacity, along with even fewer condensing hours at the foam/fiber interface. (Avoid vinyl, poly, or foil faced goods for the wall insulation to keep the foundation's moisture content low enough to protect the sill.)

Since the studwall isn't structural, it's fine to put the bottom plate directly on the floor's insulation layer or on the subfloor (which would only extend up to the wall's foam layer, not the concrete), which gives it a good capillary break from the slab.  You can also increase the stud spacing from 16" to 24" to reduce thermal bridging, at the cost of some wall-flatness.  Similiarly, the top & bottom plates need not be doubled up as they might be in a structural wall reducing the thermal bridging.

The moisture dynamics in this analysis of high-R basements was for a much colder climate than Vancouver's.  What I've recommend above is similar to Case 4 and Case 8.  When reviewing the other possibilities, pay close attention to the location of the capillary breaks installed at both the footings, slab edges and foundation sill.  Unless your house was built that way (probably not), you're better off keeping the vapor permeability of the wall high- no more than 2" of XPS (or 4" of EPS), and dealing with the winter moisture accumulation issues in the studwall with the balance of foam/fiber R values.  (In that analysis R10 foam/R13 fiber works for Minneapolis, but R5/R13 wasn't so great.  But since Vancouver's winter outdoor temperature averages are on the order 10C higher than Minneapolis,  the R5/R13 ratio works just fine in Vancouver.)

Once again, thanks for the very thorough reply, Dana.

I totally agree with the fishiness of the Dricore air gap idea -- I couldn't figure out where the moisture was supposed to go to, unless I had a dehumidifier running 24/7 in the room. (in which case, where would I drain the water?) As regards the OVRX product, the main utility I could see to it would be in the case where some catastrophic local water event happened, it'd probably be easier to take up and replace, but then again, using slabs and big sheets would be cheaper enough that I might as well redo the whole room if that happened.

I'll likely be leaving the walls alone for now -- from the shape of the vinyl siding, it looks like I'm going to have to redo the siding sometime in the next 5-10 years, anyways, and it'd be more economic to simply put on extra foam at that point all along the outside, rather than in just the few rooms I'm redoing. That would, according to BSC, also have the beneficial effect of keeping the dew line well away from the main structure. There doesn't seem much point in increasing the insulation in the walls in just a few rooms.

The question I had was for the case where I'm simply laying down a new floor and trying to warm up the floor's feel (eg, to bare feet) without redoing either the exterior- or interior-facing walls. I'm already planning to lay down beads of acoustic sealant for air infiltration reasons along the bottoms and tops of the walls as I go through this project. My original question about sealing was whether it made sense to seal the foam on the floor (Foamular) to the existing walls (sill plates, I guess), or if I should leave a gap for moisture (?) to go around, or maybe the gaps involved are so small it doesn't matter?

Air sealing the floor foam to the slab or walls would be of negligible consequence, even in bulk-water intrusion/flood event. Liquid water won't wick up the seams of XPS or EPS, and most of the moisture transfer from the slab would be via vapor diffusion through the foam, which will be slower than the drying rate through sub-flooring to the room. And liquid water would still wick up the concrete wall bypassing the edge sealant, and dry toward the room.

The economic value of the insulation in fuel savings terms is on a per-unit-area basis, there IS a point to it even on a piecemeal basis. Unless you're planning on digging right down to the footing with the exterior foam, the heat loss out of the lower half of the wall is still significant enough to warrant R7.5 of XPS or R10 of EPS at Vancouver's subsoil temps in a longer term present-value calculation. If you later end up with foam on both the interior and exterior for part of the wall, that's fine (that's what insulated concrete forms do for you), and the location of the interior dew point line becomes less relevant, since the foam is an interior vapor-retarder, and the primary moisture drive into the concrete would always be ground moisture. (At Vancouver's mild winters the average January dewpoint line of the interior air is in the outer 25% of the R value anyway.)

Even if it's not the full foundation, on the rooms that you're finishing DO put up an inch or two of XPS on the walls, even if you're only holding it in place with furring through-screwed to the concrete on which to hang the wallboard for minimum loss of floor area. Beyond the fuel savings, at Vancouver climate & subsoil temps if you insulate the walls it'll be noticeably more comfortable in there year-round.