Dear all,

this question of the appropriate fractionation factor to use is of great
interest. The recommendation is that the carbonate-CO2 fractionation factor
at 25 degrees C for calcite is 1.01025. As pointed out by Bryan, Kim and
O'Neil (1997) recently redetermined a fractionation factor 0f 1.0105. In a
fairly recent paper Swart and co-workers (Chemical Geology 1992? I think)
showed that different fractionation factors were found to apply for
different carbonate-acid reaction methods. The differences between the
fractionation factors were significant and as much as 0.28 per mille
depending on reaction temperature. They attributed the effects to (i)
isotopic exchange between the CO2, phopshate polymers and water produced in
the reaction; (ii) dissolution of CO2 in the acid, and; (iii) incomplete
removal of CO2 from the acid. Wachter and Hayes (1985 Chem. Geol.) also
demonstarted the importance of exchange between the acid/water and CO2.

I suspect that there are as many different fractionation factors as there
are labs with different methods of preparing carbonates. This was brought
home to me when considering the results of an inter-lab comparison of the
carbon and oxygen isotope composition of a series of CO2 gas samples. Labs
were asked to calibrate using NBS-19 and an acid reaction fractionation
factor of 1.01025. The range in measured delta 18-O compositions of any one
gas was ca. 0.3 per mille. Gases in the test ranged over a 25 per mille
delta 18-O compositional range. Over this range there were consistent
differences between labs. i.e. irrespective of the gas composition one lab
would consistently report data 0.1 per mille heavier than another etc. This
I think can only be due to varying acid fractionation factors that apply to
each lab.

Given this it is dangerous to base a calibration scheme for CO2 on a
carbonate reference material. Internally consistent results are achievable
when measuring just carbonates. In my lab we obtain excellent results for
carbonates using alpha = 1.01025. For example a recent measurement of NBS18
gives results of -22.947, -22.998 and -23.000 (IAEA recommended value
is -23.000). However I am certain that the true fractionation factor
operating in our reaction system is not 1.01025!

The problem comes when you want to measure the composition of say a CO2
sample derived from another sample material, say methane or water. To get
accurate results it is necessary to calibrate to a known CO2 composition.
This is not possible using carbonate standards and there is a pressing need
for CO2 reference materials.

Finally it also means that conversion between VPDB and VSMOW is not simply a
question of applying a standard formulae. There are as many formulae as
there are labs!

I would be interested to hear any other comments about this matter. It is a
concern of ours where we are looking for accurate measurments of CO2 sourced
from anything other than carbonates.

Best wishes to all,

Paul

Paul Dennis
Stable Isotope Laboratory,
Environmental Sciences,
UEA, NORWICH, UK