See below a detail of a party wall foundation. Even if the cavity is insulated there will still be a thermal bridge running down the blockwork into the foundation.

Can anyone suggest a detail to eliminate this cold bridge?

I would insulate the cavity all the way down, and use a course of 150mm aerated concrete above the beam and block floor (even though I hate the stuff).

As has been said AAC is one option (k=0.11W/mK for 2.9kN aircrete blocks) Foamglas Perinsul (k=0.058W/mK), and that Marmox Thermoblock (XPS block with concrete cylinders embedded in it - but I think that's probably not such a good bet in this case - not tall enough).

That's a lot of concrete in that strip (I assume tricky ground conditions) - were there any other options? Geotechnical survey?

Posted By: Viking House[AAC] BRE Scotland gives them a lambda of 0.25W/mK when used for foundations.

In the detail given, they'd need to be on top of the beam+block floor to do any good I think? I was thinking that this would mean above-DPC, but maybe not (looking at the detail in the drawing). Maybe one course above + one below DPC?

There are insulated beam+block systems which could be used instead I suppose (where the "blocks" are XPS, and wrap around the bottom of the beams). This might allow for insulation to wrap further down the party wall below DPC.

Posted By: Viking House

> BRE Scotland gives them a lambda of 0.25W/mK when used for foundations.

I assume that's this document:

http://www.bre.co.uk/filelibrary/pdf/rpts/BR_443_(2006_Edition).pdf

(page 24)?

It's a shame they don't state which strength AAC block that is - but I'm assuming it's 3.6N/mm²

http://www.heidelbergcement.com/uk/en/hanson/products/blocks/aircrete_blocks/thermalite_trenchblock.htm

... says:
"Mean compressive strength not less than 3.6N/mm² (7.3N/mm²: Hi-Strength)"
"Designed thermal conductivity below ground (λ) 0.23W/mK (0.31W/mK: Hi-Strength)"

... for comparison, the same people state thermal conductivity (λ) 0.15W/mK for their 3.6N/mm² block above DPC, and 0.19W/mK for the 7.3N/mm² block.


Presumably, this means that the 2.9N/mm² blocks do a little better than 0.25W/mK when (similarly) damp. I'm assuming for regularly waterlogged ground, this gets worse (Viking - I think I saw you quote 0.4W/mK somewhere?).

For the benefit of others reading this, the table I'm looking at give a (dry I presume) medium-weight concrete block a thermal conductivity of 0.5 W/mK.