I am building my own home, just west of Minneapolis, MN in the city of Minnetrista. There is no basement and a 4' frost-wall in clay. There will be in-floor radiant, possible geothermal and some passive solar.

Should I put in 2" or 4" of high density sub-slab XPS insulation, the cost is about $1200 for 2"?

2" is the standard here. You should have many more questions. Who is designing the HVAC system complete? How big is the house?

I would use EPS foam instead of XPS. Cheaper, greener, nearly the same R-value over time.

2" is standard but I am seeing a lot more 4" and thicker these days, and in my own "basement" experience, I'm not really convinced anymore that the models we have been using are really all that accurate.

you could probably get 4" of EPS for the cost of 2" XPS.

Not a big ran of eps below grade in midwest. I use 2" xps as a min. 4" 1ft down on slab edge 2" down to footing and 2" horizontal and keep 1st pex tube in around 1ft from edge. Only use 25# xps in garage or shop applications...Eric

what's your perceived advantage to XPS there?

its my understanding eps has a much higher % of open cells and absorbs water reducing r value

EPS is open cell, but it retains its R-value when wet and doesn't lose R-value to offgassing like XPS does.

if air is replaced by water then it would lose r value and i believe xps stabilizes after initial of gassing, although here is a study that says you r correct Rob.http://www.epsmolders.org/PDF_FILES/EPS%20Below%20Grade103.pdf

XPS does stabilize at levels near EPS (as do most closed cell foams.. they are better for R, but not as much better as their initial R-values indicate), but open cell foam does hold R value when wet.

well at this point I will have to look into this further as I understood xps to hold app 90% r value and I be less vulnerable to insect and rodent damage...Eric

another interesting article.... http://www.earthcoresips.com/wp-content/uploads/2010/03/18_8-Checklist_-Avoid-XPS-and-EPS-that-is-treated-with-HBCD1.pdfng articale here

that was interesting, thanks for sharing. untreated foam sounds like a big improvement over fire treated foam. I wonder about the other alternatives as well.

ICFs are EPS and they are used below grade. EPS is generally a closed cell foam.

Thats why I like coming here to gbt as most everyone is willing to listen and debate as well as give opinion and state facts...Eric
not all icfs are eps. I only see a few every year and most have been xps in N IL

XPS is higher density, 2.5 psi, most EPS is 1.0 psi, most ICFs are 1.5 psi. However ICFs are vertical and don't have weight on them.

I would like to know what the insulation does under the weight of concrete and live load. I figure 4" of concrete weighs about 4 psi, and that is without live load.

you definitely need a psi rating appropriate for what you are using the foam for. Most EPS I've seen for slab installs are rated to at least 15 psi.

Our Crete-Heat has a standard 25 psi, can be made heavier up to 40 psi for heavy loads.
The Barrier has 15 psi.

Brad, try this http://www.epsmolders.org/4-ftc.htmlm as I'm no expert...xps standard is 15psi then 25 and 40 and I believe 60 is available but I have worked on jobs with 40 under 8 to 12" of heavily reinforced concrete...Eric

I think I got that wrong;
"XPS is higher density, 2.5 psi, most EPS is 1.0 psi, most ICFs are 1.5 psi. However ICFs are vertical and don't have weight on them."
The numbers are a measure of material weight and is per cubic foot, so EPS weighs 1.0 to 1.5 lbs per cubic foot. For compression strength the XPS I use is good for 25-40 pounds per square inch. So what is the average EPS sheets good for?

Insect & rodent damage aren't nearly as likely under a slab as they would be on the exterior side of stemwalls, eh?

Borate loaded EPS is available for use on stem walls & footings which is fairly ant & termite resistant.

There's a common misconception that EPS is open-cell, but it is in fact closed cell at the micro-level, but with interstitial spaces between the beads where water can penetrate (single-digit percentages by weight at saturation). But the low cross sectional area of those water-passages is not enough to change the insulating characteristics much. In total aged-R/$ it's a value-leader in applications where the R/inch is less important.

The % open-cells argument also fails- the manufacturing processes of EPS put less stress on the cell structure than extrusion processes do. It's the lack of interstitial space that gives XPS it's lower vapor permeance, higher R value/inch and slower uptake/release of moisture, but in long term saturation environments those advantages dissipate.

Also, under residential slabs the EPS need not be as high-density as it would be under footings. 2lb goods are the standard but may be overkill. 1.25-1.5lb can also work under slabs, since slabs distribute the dynamic loads well. (Even 1lb EPS is sufficient for side loads below grade on stem walls.) But even 2lb EPS is still cheaper than XPS by a good margin.

Assuming it's the same money for 4" of EPS vs. 2" of XPS (per Rob's estimate) you're getting a 40% reduction in initial heat loss by going with the 4" of EPS, with an even greater delta over time as they saturate (per the epsmolders.org document).