PII S0016-7037(01)00833-XTemperature dependence of oxygen isotope fractionation of CO2from magnesite-phosphoric acid reactionS. DAS SHARMA,* D. J. PATIL, and K. GOPALANNational Geophysical Research Institute, Uppal Road, Hyderabad 500 007, India(Received March 15, 2001; accepted in revised form September 18, 2001)Abstract—New experimental results are reported on oxygen isotope fractionation factors,␣T, between the␦18O compositions of carbon dioxide liberated by phosphoric acid in the temperature interval of 323 to 373Kand that of total oxygen from a natural magnesite (MgCO3). These results are distinctly different from somepreviously published mutually inconsistent data, and can be expressed as a linear relationship: 103ln␣T⫽[{(6.845 ⫾ 0.475)*105}/T2] ⫹ (4.22 ⫾ 0.08), where 103ln␣Trefers to fractionation at different temperaturesT in Kelvin.Fractionation factors have also been determined at 323 and 368K on a natural calcite. The results on calciteare in excellent agreement with previously published data and can be written as:103ln␣T⫽ [{(5.608 ⫾0.151)*105}/T2] ⫹ (3.89 ⫾ 0.08).The combined results on magnesite and calcite yield a computed value of␣⫽ 1.01117 for dolomite at298K, assuming equal proportion of 0.5 mole of magnesium and calcium in dolomite, the previously reportedexperimental values being 1.01109 and 1.01110. Copyright © 2002 Elsevier Science Ltd1. INTRODUCTIONThe standard procedure for determining carbon and oxygenisotope composition of carbonates involves digestion of sam-ples in a closed reaction vessel with 100% phosphoric acid at aconstant temperature of 25°C (McCrea, 1950). Since this pro-cess liberates two-thirds of the total oxygen as carbon dioxide,fractionation of oxygen isotopes takes place between theevolved carbon dioxide and remaining oxide. Sharma and Clay-ton (1965) performed experiments on several alkaline earth andtransition metal carbonates and found that the fractionationfactors between acid liberated and total oxygen [␣⫽{(18O/16O)acid liberated at 25°C}/{(18O/16O)total}] vary considerably,indicating that␣is dependent on the composition of the car-bonate mineral. Several workers have also observed that phos-phoric acid reacts sluggishly at 25°C with carbonates likesiderite, ankerite, dolomite, and magnesite (Becker, 1971;Perry and Tan, 1972; Rosenbaum and Sheppard, 1986). How-ever, the rate of reaction can be enhanced if higher temperatureis employed (Rosenbaum and Sheppard, 1986; Carothers et al.,1988). Bo¨ttcher (1996) has recently estimated the oxygen iso-tope fractionation of calcite, rhodochrosite, kutnahorite, with-erite, and strontianite during phosphoric acid reaction as afunction of temperature in the range of 20 to 90°C. He foundthat the temperature dependence of oxygen isotope fraction-ation of the studied carbonates is characterized by specificslopes. It is therefore essential to estimate the fractionationfactor␣Tfor particular carbonate mineral of interest at variousreaction temperatures T (in Kelvin). While the temperature-dependent oxygen isotope fractionations of siderite, ankerite,and dolomite during acid reaction have been well established(Rosenbaum and Sheppard, 1986; Carothers et al., 1988), sim-ilar systematic determinations have not been performed formagnesite. A careful examination of available␣Tdata onmagnesite reveals that there exists a large discrepancy amongthe reported values (Table 1).In the present study carbon dioxide has been extracted using100% phosphoric acid from a natural magnesite sample at fourdifferent reaction temperatures in the interval of 50 to 100°C.Values of␣Thave been determined from acid liberated andtotal oxygen isotope compositions, and these are expressed interms of an equation between 103ln␣Tand 1/T2. To authenti-cate the data on magnesite, total oxygen has also been liberatedfrom a natural calcite sample along with phosphoric acid reac-tion at 25 and 95°C to evaluate the values of␣T. These valuesare compared with published data reported by Friedman andO’Neil (1977), Swart et al. (1991), and Bo¨ttcher (1996).2. SAMPLING AND EXPERIMENTAL PROCEDUREMagnesite occurs in the form of veins, 2 to 100 cm inthickness within dunite and serpentinite at Karya deposit nearMysore, Karnataka State in southern India. The sample col-lected for the present study is compact, massive, snow white incolor, and free from other materials. Petrographic examinationshows that the sample is cryptocrystalline and homogeneous innature. X-ray diffraction results confirm the presence of onlymagnesium carbonate. The chemical compositions of the sam-ple, along with a limestone standard GSR-6, were determinedon a Phillips MagiX PRO 2440 X-ray fluorescence spectrom-eter. CO2content was determined by loss-on-ignition method.The chemical analyses are presented in Table 2.Minor modification was made in the h-shaped glass reactionvessel of McCrea (1950) so that the cone-socket grease joinwas replaced by a Swagelok join with teflon ferrules to providevacuum sealing. About 7- to 10-mg finely crushed (⬃60m)sample powder and 100% phosphoric acid were reacted andkept for reaction for 4 to 240 h, depending on temperature (seeTable 3). Constant temperature bath (Lauda) with ethyleneglycol and water in the ratio of1:1wasused as bath fluid* Author to whom correspondence should be addressed ([email protected]).PergamonGeochimica et Cosmochimica Acta, Vol. 66, No. 4, pp. 589–593, 2002Copyright © 2002 Elsevier Science LtdPrinted in the USA. All rights reserved0016-7037/⫺1898 $22.00 ⫹ .00589during the reaction. The temperature was kept constant towithin ⫾ 0.1°C of the set temperature for all the runs. Tworepresentative runs with sample amounts of 25-mg each at 50and 75°C were performed to establish the relationship amongpercentage yield and␦18O as a function of reaction time (Fig.1a,b). The␦18O values of CO2liberated at different timeintervals at a particular temperature were close to each other(within ⫾ 0.08 and ⫾ 0.04 at 50 and 75°C, respectively). Thisindicates that as far as the present sample was concerned, theoxygen isotopic composition is independent of yield. Aharon(1988) and Abu-Jaber (1991) also noted similar observationswith their magnesite samples.Total oxygen was liberated as CO2in two steps: (1) thermaldecomposition of magnesite up to a temperature of ⬃650°Cand (2) fluorination of magnesium oxide using bromine pen-tafluoride and conversion of
View Full Document