Dear Julio,

we operate an auto-prep device from Finnigan which we have modified to
achieve a standard reproducibility of +/- 0.02 per mil. or better. The
design follows basically the one of Wolfgang Roether (1970, Int. J. of
Applied Rad. and Isotopes,  pp.379-387). We use stainless steel
capillaries to suck the air out of the vials and the water samples such
that time consuming freezing of the samples is no longer necessary.
Automated valves sit on top of the vials and  CO2 can be taken from the
vials while they are still in the water bath.

The most sensitive parameter during equilibration is temperature. To
keep temperature as constant as possible I recommend to
1. set equilibration temperature to mean air temperature in the lab
(e.g. 284K)
2. isolate the equilibration bath, protect vials from any draught in the
lab.
3. if you follow #1 and #2, no condensation should occur in the vials
after equilibration and no temperature gradient exists in the water
bath.

other important issues:
1. pressure readout for CO2 should be of an accuracy of +/- 1 mbar
2. shake samples to reduce equilibration time
3. all capillaries should have exactly the same length (and diameter)
and the residual volume between capillary and valve should be minimal.
4. capillaries require cleaning maybe twice a year when used with salty
water.

This is (very brief) what I believe is most important for an automated
CO2-water-equilibration unit. For further questions, please don't
hesitate to contact me directly via email.

Thorsten Agemar.