No abstract

A new value for the C0,-H,O oxygen isotope fractionation factor of 1.041 45 f 0.000 15 ( l a ) has been determined. The data have been normalized to the VSMOWNSLAP scale and were obtained by measuring isotopic compositions with the guanidine hydrochloride and chlorine trifluoride technlques.Oxygen isotope work with small ( < 5 pL) water samples in sedimentary diagenetic, ore deposit, and paleoclimate studies has necessitated a more precise determination of the fractionation factor between carbon dioxide and water because the most common technique for such determinations involves equilibrating a small aliquot of COz of known isotopic composition with a small sample of water. The fractionation factor cy C02-H20 is defined asUntil now there have been ten different determinations of the oxygen isotope fractionation factor ranging from 1.0406 ( I ) t o 1.0424 (2). Of these only one (3) has been made by use of isotope data that are explicitly stated to be normalized to a per mil scale in which 6lSO V-SLAP = -55.5 relative t o V-SMOW(0) as discussed by Coplen e t al. ( 4 ) .Measurements made earlier a t the Southern Methodist University stable isotope laboratory gave a value for aCO,-H,O of 1.041 43 f 0.000 32 (217) (3). As this differs significantly from the values of 1.0412 (5-8) and the value of 1.041 11 f 0.00007 (9) determined by careful work in other laboratories, this paper seeks to improve upon the accuracy with which our earlier determinations have been made. EXPERIMENTAL SECTIONIn this study we have determined the starting isotopic compositions of reference waters, before and after equilibration with COz, by the guanidine hydrochloride and chlorine trifluoride techniques, the methodologies of which are described below. Oxygen isotope ratios were measured on a Finnigan MAT 251 mass spectrometer having internal precision for C 0 2 isotopes of The same mass spectrometer and mass spectrometer reference gas was used for all measurements. The high long term stability of the instrument (drift <1.4 X over 1 h) and the very low level of instrument effects (abundance sensitivity is 1.8 X cross mixing of reference and sample gases is <5 X 10-j; zero enrichment <0.004%0; backgrounds a t mass 44 and 46 are <3 X made instrument corrections unnecessary except for background. Measured ratios were corrected for ion effects after Deines (10) and Mook and Grootes (11) and included the second-order corrections. which are usually ignored. <0.004%0 (2010).Multiple determinations of three carbonate reference standards over 2 years agree extremely well with the recommended values (12) and clearly demonstrate that our results are not subject to systematic errors of measurement (Table I).The guanidine method (which is described in detail elsewhere ( 3 ) ) offers the advantages of reproducibility, batch preparation, and the ability to store the reaction tubes indefinitely after either step of reaction (2) 260 "C (NHZ)&:NH*HCl+ 2H2O <70 OC , 2NH3 + COZ + NH4C1 NH4NHzC02 + NH4C1 (2) H e 4 NHdNHPCOz + NHdCl COZ + NH4Cl + (NH4)SPOI (3)With ...

Correspondence subsequent to publication has revealed several mix ups concerning the basis for calculation of the values in Table II. An extensive review of laboratory records showed the p values to be incorrect. Also, the last two GISP results were found to be poorly documented in laboratory records.