Regarding calibrated CO2 tanks.  What we've done here is 
to evacuate a tank, fill it with a small amount of CO2 and 
then backfill it with dry tank air.  Then, after 
equilibration for a day or two, we expand and aliquot of 
gas into a calibrated volume, and measure the pressure 
with a very accurate Baratron gauge.  Next, we slowly 
freeze out the CO2 on an LN2 trap and measure the pressure 
of the remaining CO2 (in a much smaller volume in order to 
have the same pressure range.  In this way, we have been 
able to make a full range of tanks an accuracy (CO2 ppm) 
of ~1 ppm.  It takes time, but once it is set up, it works 
well.

Good luck,

Zach