Dear Colleagues, I found that the article of Halas et al. (GCA, 61, 2691-2695, 1997) was recommended as a reference for isotopic fractionation between CO2(s) and carbonate, etc. I have gone through the article of Halas et al. (GCA, 61, 2691-2695, 1997). They have problems in the chemistry of the system; they used to measure the isotopic fractionation coefficient. I wrote a comment for it (GCA, 64, 573-574, 2000). I got surprised to read their reply (GCA, 64, 575-576). During writing the comment I thought that the error was a mistake by the authors and was escaped by the reviewers. It can be done by anyone. After reading the reply my opinion has changed. It is really incredible that no one checks the basic chemistry and mathematics even after pointing the error. I still do not believe that it could be possible in such a reputed Journal (GCA). They have problem in chemistry of the system, so their data may not be reliable. Let me write first, how I solved the chemistry of the system considered by Halas et al. The procedure is well established in the basic books on chemistry like Aquatic Chemistry by Stumm and Morgan, 1981. There were seven equations in total to solve a system with seven variables (H, OH, H2CO3, HCO3, CO3, alk(Na), pCO2 (g)). The five equations, which were written by Halas et al. (2000) and one for total dissolved carbonic concentration (CT=[H2CO3]+[HCO3]+ [CO3]+ in vapor phase (negligible for alkaline solution)) and other alk=Na(given). Halas et al. indirectly used the equations. These equations can be solved with sufficient accuracy for this system with considering molar concentrations instead of activity. As I did. The curves for the variation of carbonic species with pH are well standardized and can be found in all the books on the carbonic chemistry. If the ratio (or values) of [HCO3] and [CO3] is known, one can read the pH through the figure (graphic method presented by Stumm and Morgan). A question comes in mind why there is inconsistency in pH values with Halas et al. calculations. They are using activity instead of molar concentration. Their results should be more accurate, if they worked out correctly. The concentration of HCO3 and CO3 are sufficiently close together even using slightly different values of the constants. So both the approaches (molar concentration and activity) should give similar results. Why is there difference in the pH values? Let me point out their basic problems. 1. In the equation 8 of their reply, they should include terms corresponding to HCO3 and OH, as they have calculated the proportion of HCO3 and CO3 as 0.18 and 0.82, respectively. At high pH the concentration of H is insignificant, but it is not the case for OH. 2. Secondly I did not check their whole mathematics. They got the activity of H2O as 0.98, which is impossible for this system. It is the activity of H2O in seawaters. 3. I wrote the values in two decimal places in my calculation. It was just for the sake of comparison. The accuracy depends on the values of the constants and analytical errors. The values of these constants are pK1=6.35, pK2=10.33, pKW=13.998 in the Handbook of Chemistry and Physics (David R. Lide, CRC press). The values of Henry constant is not given in this form, but the accuracy may not be better than the three significant digits. If you fit a regression equation, it does not mean that the accuracy improve in the values of constants as reported by Halas et al. Anyway the calculation will not affect much. 4. It is good that the pH meters do not have 10 digits display. Otherwise it could be possible do isotope geochemistry with chemical analysis. I wish that the authors do some exercise to calculate the error in the fractionation coefficient values even with an error of 0.01 in pH, when the concentration of HCO3 and CO3 are in the proportion 0.1827:0.8172. There are many problem associated the pH measurement with an electrode. For example, at pH=10.2, the concentration of H will be 10^-10.2, so you are measuring very low concentration with an electrode. The pH-meter works better for buffer solutions. They have electronics problems. Giggenbach (IAEA-TECDOC-641, Vienna, p.439-456, 1989) did an interlab comparative study of geothermal water analysis. The values of HCO3 and SiO2 were having maximum discrepancy. The values of pH were reported only up to first decimal place. For example, the values were in the range 7.6 to 8.4 for a samples and had error (Standard Deviation) much higher than 0.1. With the above arguments I say with confidence that their values of the fractionation coefficient are not reliable. So, I recommend giving a second thought to use the data from Halas et al. (GCA, 61, 2691-2695, 1997). Cheers, Mahendra P. Verma "=?iso-8859-1?Q?=22Michael_E._B=F6ttcher=22?=" escribis: > Dear Sonja, > > the C isotope fractionation between carbonate ion (CO3--) and the rest of > DIC is still in debate, even for simple aqueous solutions. There are > several experimental studies dealing with the carbon isotope fractionation > between CO3-- and CO2-gas (e.g., Thode et al., 1965; Halas et al., 1997; > Lesniak & Sakai, 1989) and additionally Thode made some theoretcial > calculations. > The situation for fractionation between CO2(aq)-HCO3- and CO2 gas seems to > much more consistent and most studies report quite similar values. This is > actually expected , considering the typically small concentrations of CO32-- > > C isotope fractionation in the carbonate system of seawater has been > investigated by Zhang et al. (1995) in seawater. They have found some > differences between seawater matrix and simple electrolytic solutions, > probably dut to more important complex formation of carbonate species in > seawater (and associated different individual isotope fractionation > factors). Complex formation was also discussed in the fundamental paper by > Thode et al. > > Good luck, Michael > > ********************************* > Dr. Michael E. Bvttcher > Department of Biogeochemistry > Max-Planck-Institute for Marine Microbiology > Celsiusstr.1 > D-28359 Bremen > Voice: (+49)421-2028-632 > Fax: (+49)421-2028-580 > *********************************