Geochimica et Cosmochimica Acta, Vol. 61, No. 16, pp.

3461-3475, 1997
Copyright © 1997 Elsevier Science Ltd
Pergamon Printed in the USA. All rights reserved
0016-7037/97 $17.00 + .00

P i I S0016-7037(97) 00169-5

Equilibrium and nonequilibrium oxygen isotope effects in synthetic carbonates

SANG-TAE KIM a n d JAMES R. O ' N E I L
Department of Geological Sciences, University of Michigan, Ann Arbor, Michigan 48109-1063, USA

(Received June 21, 1996; accepted in revised form April 21, 1997)

AbstractlA suite of divalent metal (Ca, Cd, Ba) carbonates was synthesized over the temperature
range 10-40°C by the classical method of slowly bubbling N~ through a bicarbonate solution. It was
discovered that carbonates could be precipitated reproducibly in or out of isotopic equilibrium with the
environmental solution by varying the concentrations of bicarbonate and cation. Precipitation rate had
little or no influence on the isotopic composition of the product. Relatively high initial concentrations
of up to 25 mM in both bicarbonate and cation were prepared by adding solid metal chlorides to solutions
of NaHCO3. On the basis of results of equilibrium experiments and a new determination of the acid
fractionation factor, a new expression is proposed for the oxygen isotope fractionation between calcite
and water at low temperatures:
10001na(Calcite-H20) = 18.03(103T -1) - 32.42
where ce is the fractionation factor, and T is in kelvins. Combining new data for low-temperature
precipitations and the high-temperature equilibrium fractionations published by O'Neil et al. (1969)
results in a revised expression for the oxygen isotope fractionation between octavite (CdCO3) and water
from 0 ° to 500°C:
10001nce(CdCO3-H20) = 2.76(106T -2) - 3.96
The ability to produce nonequilibrium carbonates allowed assessment to be made, for the first time,
of the temperature dependence of nonequilibrium stable isotope fractionations in mineral systems. The
temperature coefficients of c~(carbonate-water) for nonequilihrium divalent metal carbonates are greater
than those for equilibrium carbonates, a finding that may bear on the interpretation of analyses of biogenic
carbonates forming out of isotopic equilibrium in nature.
New determinations of acid fractionation factors (10001ncQ at 25°C for calcite (10.44 _+ 0.10),
aragonite (11.01 _+ 0.01), and witherite (10.57 _+ 0.16) are mildly to strongly different from those
published by Sharma and Clayton (1965) and point to a control on this fractionation by some physical
property of the mineral. Reproducible values for octavite (CdCO3) varied from 11.18 to 13.60 depending
on the conditions of preparation of the carbonate. These new values need to be considered in determina-
tions of absolute 180/160 ratios of international reference standards and in relating analyses of carbonates
to those of waters, silicates, and oxides. Copyright © 1997 Elsevier Science Ltd

1. INTRODUCTION number of complications of the method have been recog-

nized and addressed. Among the most important of these are
In his pioneering study of theoretical and experimental as- (1) validity of the original calibrations, (2) influence of
pects of the oxygen isotope geochemistry of calcium carbon- polymorphic form and chemical composition of the CaCO3
ate, McCrea (1950) used two different batches of seawater shell or cement on equilibrium fractionation factors, (3) ki-
for his experiments and obtained two different equations for netic effects of both biological (vital) and physical origin,
the fractionation of 180/160 between inorganically precipi- (4) reliability of acid fractionation factors, (5) temperature
tated CaCO3 and H20 as a function of temperature. No expla- dependence of nonequilibrium fractionations, and (6) recog-
nation was provided for why the two equations were so nition of diagenetic processes that alter the original isotopic
different. Epstein et al. (1951, 1953), working in the same compositions.
laboratory, developed an analogous carbonate-water oxygen In an attempt to address some of the above concerns
isotope temperature scale by analyzing biogenic carbonates through laboratory experimentation, O'Neil et al. (1969)
precipitated by marine organisms living either in therrno- measured oxygen isotope fractionations between alkaline-
stated t a n k s in the laboratory or under natural conditions earth carbonates and water over the temperature range 0 -
in localities where the seasonal surface temperatures were 500°C. They concluded that ( 1 ) the temperature coefficient
known. The oxygen isotope temperature scale obtained from for the CaCO3-H20 fractionation published by Epstein et al.
the biogenic carbonates was in good agreement with one of (1953) was almost the same as that determined from their
the McCrea (1950) equations. inorganic precipitation experiments and one of the equations
For over 40 years major advances in paleoceanography of McCrea (1950), (2) there are significant oxygen isotope
and paleoclimatology have been made using the oxygen iso- fractionations of up to a few permil between the various
tope paleotemperature scale of Epstein et al. (1953) or a alkaline earth carbonates at low temperatures, and (3) both
slight modification of it. Throughout the intervening years a cationic size and mass are important in controlling isotopic
3461
3462 S.-T. Kim and J. R. O'Neil

fractionation in carbonate-water systems. Tarutani et al. calcite paleotemperature scale, but Grossman and Ku (1986)
(1969) made further studies of synthetic carbonates by found no-temperature dependence of this fi'actionalion in
slowly precipitating calcite, aragonite, and magnesian calcite their oxygen isotope measurements of aragonitic and calcitic
from various bicarbonate solutions. Their results suggested foraminifera. Furthermore, they argued that the high temper-
that, relative to calcite at 25°C, aragonite was enriched in ature data of Hoeglundina reported by Sommer and Rye
;80 by 0.6%0 and magnesian carbonates were enriched by (1978) were 2-3%0 higher than expected on the basis of
0.06%c/mo1% Mg. Their data for inorganically precipitated experimental results, because the formninifera grew during
carbonate were in accord with the CaCO3-H20 calibrations a glacial period when the isotopic composition of seawater
of Epstein et al. (1953) and O'Neil et at. (1969). was heavier.
Tarutani et al. (1969) also reported that precipitations The power and potential of oxygen isotope analysis of
from solutions with higher initial calcium concentrations biogenic carbonates remains secure, but uncertainties still
tended to yield larger fractionations and explained this obser- remain in the minds of most active workers in the field about
vation as a disequilibrium phenomenon rather than one aris- the reliability of proposed isotopic temperature scales and
ing from variations in the equilibrium constant in response the factors that influence them. The most commonly accepted
to changes in some other parameter like ionic strength or oxygen isotope paleotemperature scale in use today for bio-
pH. They measured low isotopic ratios from the outer parts genic carbonates is that proposed by Grossman and Ku
of single precipitations of the carbonates by sequential isoto- (1986) and will be discussed in more detail below. Clearly
pic analysis and did not observe any change in the fraction- there are important unresolved conflicts among the results
ation factor determined using solutions containing 36 g of published calculations, measurements of natural carbon-
NaCI/L. If this anomalous phenomenon were caused by ates, and measurements of laboratory synthesized carbon-
rapid precipitation in the first stage of precipitation, the early ates. In the present study we report results of laboratory
calcium carbonate should have had nearly the same (or experiments designed to make further tests of the importance
lighter) isotopic composition as the HCO:~ ion in solution of some chemical and physical conditions of precipitation
according to McCrea (1950). That is, the isotopic ratio to equilibrium and kinetic isotope effects in carbonate-water
should have increased from the inner parts of the crystals to systems. Carbonates were precipitated in and out of isotopic
the outer parts, because '60 species would tend to precipitate equilibrium at three different low temperatures. The impor-
faster than ~80 species as a result of the more rapid diffusion tance of rate of precipitation and initial concentrations of
bicarbonate and cation was reinvestigated. In addition, the
of the isotopically lighter species to the crystal surface
temperature dependence of nonequilibrium fractionations
(Tumer, 1982). The higher fractionation factor observed at
were determined for the first time for mineral-water systems.
higher concentrations implies that there is some other factor
Barium and cadmium carbonates, while relatively rare in
which influences the attainment of isotopic equilibrium for
nature, were investigated because of their importance to the
carbonates precipitated from aqueous solutions. In a similar
overall understanding of the oxygen isotope chemistry of
vein, O'Neil et al. (1969) also reported that the ~80/'~'O
carbonates. The CaCO3 polymorph vaterite was also synthe-
fractionation between carbonates and water decreased with
sized and its isotopic properties determined.
time during a single precipitation experiment but offered no
explanation for this phenomenon. 2. EXPERIMENTAL ME'rttODS
Golyshev et al. ( 1981 ) concluded from theoretical consid- 2.1. Preparation of Solutions
erations that the isotopic properties of carbonates are deter-
A series of Na-X-CI-HCO3solutions was made in deionized water
mined primarily by the cationic radius and that, in contrast to at equimolar concentrations of XCI2 (X = Ca 2-, Cd 2÷, or Ba 2-)
the findings of O'Neil et al. (1969), the mass effect becomes and NaHCO3 ranging from 5 to 25 mM depending on the solubility
substantial only for ions with large masses like those of Ba 2+ product of the carbonate. Purified CO2 gas (P = lbar) was normally
and Pb 2+. The Cd 2- cation takes on special significance in bubbled through stirred deionized water at room temperature for 10
min in order to adjust the pH down to a value of about 6.0 and
this respect as it has almost the same ionic radius as Ca 2' enhance solubility. The NatlCO~ was then added and, after 10 min,
in carbonate crystals but has a much larger mass. The influ- the chloride was added. Stirring was continued for an additional 10
ence of these parameters on the equilibrium isotopic proper- rain to insure that all salts were completely dissolved. The specific
ties of carbonates and other minerals needs further testing chlorides were used to make the various carbonates; for example,
barium chloride tbr witherite and cadmium chloride for octavite. An
by laboratory experiment. alternate method of preparation, used by O'Neil et al. (1969) and
Horibe and Oba (1972) reported that aragonite was de- Tarutani et al. (1969), was also employed. Solid carbonate was
pleted in '~O relative to calcite in mollusks grown at various added to 400 mL of stirred deionizcd water, and CO2 gas was bub-
temperatures and Zheng (unpubl. data), using the increment bled through the water to dissolve the carbonate to saturation. After
a minimum of 2 h of bubbling with CO2 gas, undissolved carbonatcs
method of semi-empirical calculations (Schtitze, 1980), sug- were removed by vacuum filtration, and the saturated divalent metal
gested that aragonite should be significantly depleted in t~O bicarbonate solution stored in a bottle. For CaCti, however, solu-
relative to calcite at low temperatures. Sommer and Rye tions prepared this way always produced a mixture of calcite and
(1978) and Grossman and Ku (1986), however, showed vaterite, and in small amounts. This method of preparation of bicar-
bonate solutions was, therefore, not used in preparing the majority
that the aragonitic foraminifer Hogelundia was enriched in of carbonales examined in this study.
~O relative to the calcitic foraminifer Uvigerina. The arago-
nile-calcite fractionation is now generally accepted to be 2.2. Synthesis of Carbonates
positive at low temperatures. From their analyses of natural After the bicarbonatesolutions were prepared, one aliquot of solu-
samples, Strainer and Rye (1978) proposed an aragonite- tion was taken for 6~O analysis, and 100 or 200 mL of the solution
 Oxygen isotope fractionation in synthetic carbonates 3463

was placed in a specially designed vessel that was thermostated for night. The thermally liberated CO2 gas was collected, and its volume
at least 1 h in a constant temperature bath (_0.1°C) to ensure that and isotopic composition were measured. The metal oxide remaining
the specified temperature was established before precipitation in the Ni tube was then reacted with C1F3 at 600°C for 12 h. Oxygen
started. Pre-purified nitrogen gas was then bubbled slowly through obtained from the reaction with C1F3 was converted to CO2, and its
the solutions to remove COz thus promoting supersaturation and volume and isotopic composition were measured. Generally, the
slow precipitation of the carbonates. The N2 was saturated with water yield obtained from the sum of oxygen released on thermal decompo-
vapor of the same oxygen isotope composition as the water vapor sition and fluorination of the remaining oxide was 97-100%. Com-
leaving the precipitating solution in order to maintain constant isoto- bined with the data from the phosphoric acid reaction, acid fraction-
pic composition of the solution. The bubbling rate of N2 was con- ation factors were obtained.
trolled by a regulator attached to the tank and was monitored by
counting the bubbles for periods of 30 s. Precipitations were per- 2.4. Oxygen Isotope Measurements
formed at 10, 25, and 40°C.
The time required for the first appearance of precipitate depended All isotopic measurements were made on CO2 gas using a Finni-
on the composition of the solution, temperature, and bubbling rate, gan Delta S isotope ratio mass spectrometer. 6180 values were nor-
but it took at least one day and a maximum of five days to generate malized to the recommended values for the international reference
sufficient material ( ~ 1 0 mg XCO3) for both X-ray diffraction and standards NBS-19, SMOW, and SLAP. Carbonates were analyzed
isotopic analysis. Upon completion of the experiment, a 2 mL aliquot at 25°C using the classical procedure of McCrea (1950). Acid frac-
of the final solution was taken for 6180 analysis. The solid carbonate, tionation factors used in this paper are 1.01050 for calcite, 1.01063
precipitated on the walls and at the bottom of the vessel, was re- for witherite, and 1.01146 for octavite, modified from the values
moved using a rubber policeman, filtered, rinsed several times with reported by Sharma and Clayton (1965). The previously accepted
deionized water, ethyl alcohol, and then dried on filter paper for at value of 1.01025 for calcite was used only once in this paper when
least ~ 2 h at 80°C prior to storage for isotopic analysis. The mineral- making comparisons between results obtained in this study and those
ogy of every precipitate was identified by X-ray diffraction analysis, published earlier (Fig. 1 ).
and some of the precipitates were examined by scanning electron Solutions were analyzed by the CO2-HgO equilibration method
microscopy. (Cohn and Urey, 1938) using a modification of the vacutainer
method of Socki et al. (1992). The equilibrated CO2 gas was ex-
tracted and analyzed on the mass spectrometer, and a CO2-H20
2.3. Release of Total Oxygen in Carbonates fractionation factor ( a ) of 1.0412 (O'Neil et ai., 1975) was applied
To obtain ~so/160 ratios of total oxygen of the carbonates, the to obtain the isotopic composition of the water itself. The oxygen
carbonates were decomposed thermally on a conventional silicate isotope fractionation factor a between two substances A and B is
extraction line, and the oxide residue was fluorinated to release its defined as
oxygen. Carbonate samples were placed in Ni tubes and degassed Ol = (1So/160)A/(180/160)B
at 100°C for at least 3 h. After complete degassing, the valves on
the tubes were closed and the samples were heated at 600°C over- For theoretical and practical reasons the function 1000 In a is nor-

T (°C)
40 25 I0
35
[
This study (Calcite)
O'Neil et al. (1969) (Calcite)
33 - Grossmanand Ku (1986)(Aragonite)
Patterson et al. (1993)(Aragonite) ~ , /.."

/ .-.../

27 //
/

25
Y
, , ........ i i r i i J i i J r i i i i i i i i i J i i i i i i r J i i i i i i

3.14 3.24 3.34 3.44 3.54 3.64

103/T (K)

Fig. 1. Comparison of fractionation curves for CaCO3-H20 at low temperatures.

3464 S.-T. Kim and J. R. O'Neil

mally employed in this paper. Oxygen isotope compositions of both fractionations were determined over a large temperature
carbonates and waters are reported in the familiar 6180 notation
relative to the SMOW standard. range and are internally consistent, and (4) the conditions
of the present experiments were tightly constrained to pro-
duce only calcite, the stable polymorph of CaCO3. These
3. RESULTS AND DISCUSSION
are only plausibility arguments as attainment of isotopic
3.1. Temperature Dependence of Equilibrium equilibrium can not be proved.
Fractionations It should be noted that an acid fractionation factor of
1.01050 (see Table 5) was used in the new determination
3.1.1. Calcium carbonates of the calcite-water calibration that is shown in Fig. 2 and
The various fractionation curves proposed for the CaCO3- represented by the following expression:
H20 system (both calcite and aragonite) are shown in Fig. 10001na(Calcite-H20) = 18.03 (103T -1) - 32.42
1. A recent determination was made by Patterson et al.
(1993) who measured the oxygen isotope composition of 3.1.2. Barium carbonate
aragonitic otoliths from freshwater fish that lived in waters
of known temperature and isotopic composition. They ob- A few checks were made of the reliability of equilibrium
tained the following relation for the temperature dependence oxygen isotope fractionations between carbonates and water
of the fractionation factor c~(CaCO3-H20): determined by the synthesis approach (Table 2). Results of
our new determinations were compared to those made many
10001nr~ = 18.56(103T -1) - 33.49 years ago in a different laboratory. O'Neil et al. (1969)
argued that the data they obtained by slow precipitations of
where T is in kelvins. They also recalculated the paleotem-
carbonates at low temperatures were equilibrium data be-
perature equation of Grossman and Ku (1986) to fit a T - 1
cause they were in good accord with data for high-tempera-
relation and obtained the following:
ture recrystallization experiments for which equilibrium
10001nc~ = 18.07 (103T -1) - 31.08 could be proved by reversing the curves. Our new low-
temperature data for BaCO3-H20 agree remarkably well with
Patterson et al. (1993) state that the slopes of these equations those published by O'Neil et al. (1969) almost 30 years
are indistinguishable, but that the intercept differs by approx- ago. The (corrected) high-temperature data of O'Neil et al.
imately +2.5 because of seasonal changes in bottom water (1969) for BaCO3 (witherite) are combined with our low
temperature, the isotopic composition of water, analytical temperature data in Fig. 3 using the new determination of
errors, and unexplained metabolic effects. In Fig. 1 we com- the acid fractionation factor for witherite (see Table 5 ). The
pare the curve determined in the present study for calcite relation obtained:
(using an acid fractionation factor of 1.01025) with the pre-
viously published curve for this system determined by 10001nc~(BaCO3-H20) = 2.63 (106T - 2 ) -- 4.04
O'Neil et al. (1969) and the two curves determined for is the same as that of O'Neil et al. (1969) within experimen-
biogenic aragonite. The temperature coefficient (slope) for tal error. T -2 is normally used to relate fractionation factors
our data on this plot is indeed very close to that of Grossman over very large temperature ranges that include high temper-
and Ku (i986) even though one is for calcite, and the other atures (e.g., O'Neil, 1986). This confirmation provides
is for aragonite. The implication of this observation is that needed credence to the approach as results of very few exper-
a(aragonite-calcite) does not change with temperature over iments of this type have been reported.
the low temperature range investigated, a conclusion that
has some basis in theory. If aragonite is indeed enriched by
3.1.3. Cadmium carbonate
0.6%o relative to calcite at 25°C as observed in the laboratory
experiments of Tarutani et al. (1969), however, the data of One of the most important carbonate systems to examine
Patterson et al. (1993) are much more in accord with our is CdCO3-H20 as the radius of the Cd 2+ ion is almost identi-
experimental data for calcite than those of Grossman and cal to that of Ca 2+ in carbonates. With our present tech-
Ku (1986) whose implied difference between aragonite and niques, the importance of cationic radius to the isotopic prop-
calcite would be much too high at approximately 1.2%o. erties of divalent metal carbonates can be assessed at low
Because of the importance of oxygen isotope analyses of temperatures where fractionation factors are larger and more
natural carbonates in several aspects of earth science, further sensitive to changes in temperature and other parameters.
experimental work on these systems is warranted. O'Neil et al. (1969) were unable to synthesize CdCO3 at
Our new data for inorganically precipitated calcite at low low temperatures by the method they used because of the
initial concentrations of 5 mM are in good accord with those low solubility of octavite. In this study, low-temperature
of O'Neil et al. (1969) above 25°C, but there is a relatively oxygen isotope fractionation factors for the CdCO3-H20 sys-
large difference in the datasets around 10°C (Fig. 1). We tem were obtained for the first time by slow synthesis from
propose that the fractionation curve for calcite-water deter- Na-Cd-C1-HCO3 solutions (Table 3). Because of the large
mined in the present study is the most reliable curve for this range of values obtained in the present study (see Table 5 ),
system on the basis of several considerations including ( 1 ) the one value of the acid fractionation factor ( a = 1.01146)
the similarity of the slope of our new curve to those deter- that agreed with the value published by Sharma and Clayton
mined for biogenic carbonates, (2) the large number of ex- (1965) was used in calculating the 6180 values of CdCO3.
periments performed provided improved statistics, (3) the Again the (corrected) high temperature equilibrium oxygen
 Oxygen isotope fractionation in synthetic carbonates 3465

T (°C)
40 25 10

33 . .1. .0. Calcite]l

. . . . . . . . . . . . . . .

~ 29

,7 // A = 18.03 (103/T) - 32.42 [

I

25
.... , / , ........... ............... '
3.14 3.24 3.34 3.44 3.54 3.64

103/'1` (K)

Fig. 2. Relation between A = 103 In a(calcite-water) and temperature at the lowest initial concentrationsof Ca 2÷
and HCO~ (5 mM).

isotope fractionations obtained by O'Neil et al. (1969) were and bicarbonate, the larger were the fractionation factors,
in excellent accord with the low temperature fractionations and reproducibly so. Solutions prepared with divalent metal
we obtained by our methods and combined to produce the chlorides and NaHCO3 have much higher ionic strengths
following relation shown in Fig. 4: than those prepared by dissolving solid carbonates in water,
and thus we were able to prepare relatively concentrated
10001na(CdCO3-H20) = 2.76( 106T -2) - 3.96 bicarbonate solutions of up to 25 mM for the more soluble
carbonates like calcite and witherite. At these concentrations
In this case, however, the above relation is significantly dif-
the solutes would have a negligible effect on the oxygen
ferent from the extrapolated high-temperature data published isotope activity of the water (e.g., O'Neil and Truesdell,
earlier, once again emphasizing the danger of extrapolating 1991 ). The low solubility of octavite precluded preparation
any type of chemical data obtained at high temperatures to of such concentrated solutions.
low temperatures as was perforce done by O'Neil et al.
The fractionation factors measured for solutions with dif-
(1969) in the absence of low-temperature data for this sys- ferent initial concentrations differed by as much as 2 - 3
tem. Despite the important change in the fractionation curve permil at a given temperature. Clearly there is only one
for CdCO3-H/O, the point made in the earlier study was equilibrium fractionation at any temperature, so most of the
confirmed: relative to cationic mass, the cationic radius plays carbonates precipitated from these solutions of varying con-
a minor role in determining the oxygen isotope properties centration were forming out of oxygen isotope equilibrium
of carbonates. with the water. Only limited criteria were available to use
in selecting which direction of change (i.e., larger or smaller
3.2. Nonequilibrium Carbonates fractionations), if any, was in the direction toward equilib-
rium. The bicarbonate solutions used in previous work, for
3.2.1. Synthesis of nonequilibrium carbonates which equilibrium was assumed, were relatively dilute be-
cause they were made by dissolving the solid carbonate in
It was noticed early on that the oxygen isotope fraction- COz-charged water. In addition, as the concentrations of our
ation factors between synthetic carbonates and water varied Na-X-C1-HCO3 solutions were lowered, the fractionation
significantly with initial concentrations of the bicarbonate factors rapidly decreased to what appeared to be a limiting
solutions. The higher the initial concentrations of metal ion value. (Experiments will be made with extremely dilute solu-
3466 S.-T. Kim and J. R. O'Neil

Table 1. Data for calcite experiments.

Sample Temp. (°C) 6~80 (Water) 5180 (CaCO3) acac%_iJ2o 1 0 3 1 h a Init. conc. (raM)

ST-18 10 -7.74 25.56 1.03356 33.01 15

ST-20 10 -7.74 25.69 1.03369 33.14 25
ST-24 10 -7.74 26.50 1.03451 33.93 25
ST-25 10 -7.74 26.64 1.03465 34.06 25
ST-36 10 -7.74 25.32 1.03332 32.77 25
ST-37 10 -7.74 25.46 1.03346 32.91 25
ST-38-S 10 -8.30 25.86 1.03445 33.87 25
ST-38-N 10 -8.30 25.12 1.03370 33.14 25
ST-39 10 -8.30 25.72 1.03430 33.73 25
ST-52-10 10 -8.12 23.47 1.03185 31.35 5
ST-54-10 10 -8.23 23.21 1.03170 31.21 5
ST-18 25 -7.74 21.52 1.02949 29.06 15
ST-19 25 -7.74 21.19 1.02916 28.74 15
ST-20 25 -7.74 22.06 1.03003 29.59 25
ST-22 25 -7.74 21.45 1.02942 28.99 15
ST-24 25 -7.74 21.58 1.02955 29.12 25
ST-25 25 -7.74 21.97 1.02994 29.50 25
ST-36 25 -7.74 21.51 1.02948 29.05 25
ST-37 25 -7.74 21.65 1.02962 29.19 25
ST-38-S 25 -8.30 21.46 1.03001 29.57 25
ST-38-N 25 -8.30 21.15 1.02970 29.26 25
ST-39 25 -8.30 21.32 1.02987 29.43 25
ST-45-25-N 25 -8.30 19.73 1.02826 27.87 5
ST-50-25 25 -8.25 20.23 1.02872 28.31 5
ST-52-25 25 -8.12 20.00 1.02835 27.96 5
ST-54-25 25 -8.23 20.03 1.02849 28.10 5
ST-18 40 -7.74 18.06 1.02600 25.67 15
ST-20 40 -7.74 18.40 1.02634 26.00 25
ST-22 40 -7.74 17.94 1.02588 25.55 15
ST-24 40 -7.74 18.41 1.02635 26.01 25
ST-25 40 -7.74 18.42 1.02636 26.02 25
ST-36 40 -7.74 18.82 1.02677 26.42 25
ST-37 40 -7.74 18.60 1.02655 26.20 25
ST-38-S 40 -8.30 18.11 1.02663 26.28 25
ST-38-N 40 -8.30 17.98 1.02650 26.15 25
ST-39 40 -8.30 18.01 1.02653 26.18 25
ST-50-40 40 -8.20 17.06 1.02547 25.15 5
ST-52-40 40 -8.12 17.24 1.02557 25.25 5
ST-54-40 40 -8.23 17.01 1.02545 25.13 5

Temp. = Temperature.
Init. conc. = Initial concentration (Ca2+) = (HCO3).

Table 2. Data for witherite experiments.

Sample Temp. (°C) 6~80 (Water) 6180 (BaCO3) aB~CO3_H2

o 1031ha Init. conc. (raM)

C30-10 10 -7.80 23.12 1.03116 30.69 25

C32-10 10 -7.80 22.61 1.03065 30.19 15
C33-10 10 -7.80 22.68 1.03072 30.26 15
C34-10 10 -7.80 22.51 1.03055 30.09 15
C48-10-N 10 -8.30 20.58 1.02912 28.71 5
C49-10-N 10 -8.30 20.80 1.02934 28.92 5
C30-25 25 -7.80 19.19 1.02720 26.84 25
C32-25 25 -7.80 18.83 1.02684 26.49 15
C33-25 25 -7.80 19.05 1.02706 26.70 15
C34-25 25 -7.80 18.84 1.02685 26.50 15
C48-25-N 25 -8.30 17.52 1.02604 25.70 5
C49-25-N 25 -8.30 17.36 1.02587 25.55 5
C32-40 40 -7.80 15.76 1.02375 23.47 15
C33-40 40 -7.80 15.82 1.02381 23.53 15
C34-40 40 -7.80 15.75 1.02374 23.46 15
C49-40-N 40 -8.30 14.43 1.02292 22.66 5
 Oxygen isotope fractionation in synthetic carbonates 3467

35
I I
(3 This Study Z
30 " " • O'Neil et al. (1969) /-
25 /
20
/
"" 15
/
10 /
5
/ r 2 = 0.999

0 , , Ir . . . . . . . . . , , , [ , , , -,
0 2 4 6 8 10 12 14

106/T 2

Fig. 3. Oxygen isotope fractionation between barium carbonate and water. The initial concentrations of Ba 2+ and
HCO3 were 5 mM. High temperature data from O'Neil et al. 1969).

tions in the future to test that we h a v e indeed reached a j u d g m e n t m a y seem counterintuitive to those w h o would
limiting value.) On the basis of analyses of natural materials, expect kinetically-controlled, nonequilibrium fractionations
the larger fractionation factors measured were greater than to b e too small (i.e., the carbonate would b e isotopically too
d e e m e d possible. Finally we place emphasis on those condi- light) as is seen in some biogenic carbonates like coralline
tions of precipitation that produce only calcite, the stable carbonate. It is well to point out that the good reproducibility
phase. Taking all these points into consideration, we j u d g e obtained for the fractionation factors of what we interpret to
the smallest fractionation factors obtained to b e the best be nonequilibrium carbonates m a y indeed b e reflecting some
representations of equilibrium fractionation factors. This equilibrium system we have not yet identified.

Table 3. Data for octavite experiments.

Sample Temp. (°C) 6180 (Water) 6180 (CdCO3) acdcorn2o 1031na Init. conc. (raM)

40-10 10 -8.30 22.43 1.03099 30.52 5

42-10 10 -8.30 22.45 1.03101 30.54 10
44-10 10 -8.30 22.35 1.03091 30.44 5
46-10 I0 -8.30 22.51 1.03107 30.60 5/10
51-10 10 -8.30 22.53 1.03109 30.61 5
40-25 25 -8.30 18.73 1.02726 26.89 5
42-25 25 -8.30 19.39 1.02792 27.54 10
44-25 25 -8.30 18.89 1.02742 27.05 5
46-25 25 -8.17 19.18 1.02758 27.20 5/10
51-25 25 -8.30 19.08 1.02761 27.23 5
53-25 25 -8.30 19.15 1.02768 27.30 5
40-40 40 -8.30 15.85 1.02435 24.06 5
42-40 40 -8.30 16.49 1.02500 24.69 10
44-40 40 -8.13 15.90 1.02423 23.94 5
46-40 40 -8.30 16.13 1.02463 24.34 5/10
51-40 40 -8.30 16.13 1.02463 24.34 5
53-40 40 -8.30 16.32 1.02483 24.52 5

Temp. = Temperature.
Init. conc. = Initial concentration (X 2+) = (HCO3).
3468 S.-T. Kim and J. R. O'Neil

35

30 -
I
© This Study
I
/
/
• O'Neil et al. (1969)

25

20 /
15 /
10 /
0 i l l
/
/ i i i ii i i i i
A = 2.76 (106/T2) - 3.96
r2 = 0.999

i i i ~ I r i i i

0 2 4 6 8 10 12 14
106/T2

Fig. 4. Oxygen isotope fractionation between cadmium carbonate and water. The initial concentrations of Cd 2+
and HCO~ were 5 mM. High temperature data from O'Neil et al. (1969).

3.2.2. Physical properties of nonequilibrium carbonates pending on the mineral, in order to obtain enough material
for analysis:
In an attempt to understand the nature of the carbonates
precipitated out of isotopic equilibrium with water, the prod- 3.2.3. Temperature dependence of a
ucts were examined under the petrographic microscope, with
the scanning electron microscope (SEM) and by X-ray dif- The ability to prepare carbonates out of isotopic equilib-
fraction (XRD) analysis. The crystaUinities of the synthetic rium in a reproducible fashion provided us with the opportu-
carbonates, determined from ratios of specific peak heights nity to test the assumption often made that the temperature
on the diffractograms, were identical to those of natural min- coefficient of the CaCO3-H20 fractionation is the same
erals so that any difference in fractionation factors (between whether the carbonate is precipitated in or out of equilibrium.
carbonates and water or between acid-liberated CO2 and car- Using this assumption, some workers have used changes in
bonate) could not be explained by poor crystallinity of syn- the oxygen isotope compositions of coralline carbonate, for
thetic materials. The expected external forms of the minerals example, to reflect changes in temperature of the environ-
(e.g., well-formed rhombs of calcite) were always observed mental waters (e.g., Druffel, 1985; Shen et al., 1992; Dunbar
under the scanning electron microscope. et al., 1994). In fact, on theoretical grounds, it is even possi-
The only identifiable change in physical property of the ble that the fractionation factors of materials precipitated out
products obtained was a variation in crystal size. For calcite of equilibrium would be temperature independent or only
and witherite, the lower the initial concentration and the weakly so. The fractionation curves we determined for these
higher the temperature of precipitation, the larger the crystal carbonates are the first ever reported for nonequilibriummin-
size. For octavite, there was a slight increase in crystal size eral systems.
with increasing temperature but, in contrast to the other min- The oxygen isotope fractionations measured between non-
erals examined, there was a marked increase in crystal size equilibrium calcite and water from 10° to 40°C are given in
with increasing initial concentration. In addition, the size of Table 1 and shown in Figs. 5 and 6. The most striking
the octavite crystals was markedly smaller than those of the features of the data are the unusually large fractionations
other minerals regardless of the conditions of precipitation. that increase with increasing initial concentrations and their
The length of time required for the first appearance of precip- regular decrease with increasing temperature. The slopes of
itate was shorter (minimum 5 h; maximum 2 days) for octav- the nonequilibrium fractionation curves are much steeper
ite than for calcite and witherite (minimum 10 h; maximum than those of the assumed equilibrium curve (Fig. 6). In
5 days). The precipitation rates were always slow and the addition, the slopes of the two nonequilibrium curves are
precipitations were allowed to proceed for several days, de- almost identical and readily distinguishable from the equilib-
 Oxygen isotope fractionation in synthetic carbonates 3469

T (°C)

35
40

[ <> Calcite 1
25 10

/,
33

31

29

27
/
_ _ / r2A=1 21'516
0.988
2 ' (103/T)
8 0 -=

2 5 q i i i i i , , , p i i , , , ~ i ~ , , , , i , i i i t r t t t r i i

3.14 3.24 3.34 3.44 3.54 3.64

103/T (K)

Fig. 5. Relation between A = 1031na(calcite-water) and temperature at the highest initial concentrations of Ca 2+
and HCOf (25 raM).

rium slope. If the results of these experiments can be applied the observation that equilibrium precipitation is favored from
to natural biogenic carbonate systems, and that application solutions that are relatively dilute, or initially undersaturated
is questionable, the practical effect would be that a given with respect to the solid carbonate, was confirmed as a gen-
change in ~5i80 of nonequilibrium carbonate like coralline eral phenomenon for all carbonates and perhaps for other
carbonate would correspond to a smaller change in tempera- minerals as well.
ture than indicated from the conventional oxygen isotope
paleotemperature curves in use. All else being equal, a
change of 1 permil in ~5180 of nonequilibrium carbonate 3.3. Effect of Precipitation Rate on CaCO3-H20
would correspond to a change of about 3°C whereas the Fractionation
same change in 6180 for an equilibrium carbonate would
3.3.1. Biogenic carbonate
correspond to a change of about 4°C.
Attempts were also made to synthesize nonequilibrium The effect of the rate of precipitation on carbon isotope
witherite and octavite. The data are given in Tables 2 and 3 fractionation has been studied many times and, for most
and the fractionation curves are shown in Figs. 7 and 8. geologic processes, carbon isotope fractionation is believed
Witherite is soluble enough that the relatively highly concen- to be independent of rate (Romanek et al., 1992), except
trated solutions necessary for nonequilibrium precipitation for a few unusual cases (e.g., O'Neil and Barnes, 1971).
could be prepared. Again the fractionation factors for the On the other hand, only a few studies have been made of the
nonequilibrium cases were higher, but the difference be- effect of precipitation rate on oxygen isotope fractionation
tween the data for initial concentrations of 15 and 25 mM between carbonates and water. Most of these investigations
were not as pronounced as they were for calcite and signifi- dealt with natural biogenic samples such as corals, and only
cant only at the lowest temperature where fractionations are one experimental study, using inorganically synthesized ma-
highest in general. Octavite is the least soluble carbonate terial, was made by Tarutani et al. (1969). The importance
examined in this study, and only relatively dilute solutions of precipitation rate to the fractionation of both stable isotope
could he prepared. In this case, the difference between proba- and elemental ratios between natural carbonates, and their
ble nonequilibrium and equilibrium precipitations could be environmental solutions remains controversial.
seen only at the highest temperatures, a phenomenon com- McConnaughey (1989a) sampled skeletal materials from
pletely unexpected and difficult to understand. In any event, corals that grew at two different sites and reported a correla-
3470 S.-T. Kim and J. R. O'Neil

T(°C)
40 25 10
34
/
Init. Conc. 5
/ ."•
I.nit. Conc. 15
/
32
- - ' Init.Conc. 25 j .."

/ ,
/ .•'"
¢ ,.•'
30
~3

/ ..•'"
28

26
<
ll,bl,,b ,i,llq,I,

3.14 3.24 3.34 3.44 3.54 3.64

103/T (K)

Fig. 6. Relation between A = 1031na(calcite-water)and temperature at different initial concentrations of Ca 2+

and HCO 7.

tion between the 6180 value of skeletal carbonate and growth passing the N2 through a tube with no constriction. Type I I
rate of the coral. The 6180 values are lower than equilibrium bubbling is characterized by the production of many smaller
values and remain constant between 15 mm/yr and 5 mm/ bubbles when the N2 passes through a fine frit located at the
yr and then increase dramatically as the growth rate de- bottom of the vessel. Because the apparatus is connected up
creases. The 6180 values approach equilibrium values at in series, the same amount of nitrogen gas flows through
growth rates below 2 ram/yr. In contrast, De Villers et al• both sections of the apparatus, but vastly different sizes and
(1995) found a notable effect of growth rate on 6180 values numbers of bubbles are generated in the two sections. The
and Sr/Ca ratios in corals that grew at 6 mrn/yr and 12 ram/ final pH of the solutions where Type II bubbles were gener-
yr and argued against the McConnanghey (1989a) sugges- ated was about 0.5 units higher than that of the other solution,
tion that the 8180 value is independent of growth rate over which means that Type II bubbles remove CO2 more effi-
the range of 5 mm/yr to 15 ram/yr. From the trace element ciently than Type I bubbles as anticipated• We used two
literature, the consensus view at the present time is that the different N2 flow rates to precipitate pure calcite from the
slower the precipitation rate, the closer is the approach to bicarbonate solution. Slow is defined as approximately 10
chemical equilibrium (e.g., Dromgoole and Walter, 1990). bubbles per 30 s, and fast as 24 bubbles/30 s. The flow rate
was held constant throughout the duration of the experiment.
After several tests, we chose between 9 and 24 bubbles per
3.3.2. Inorganic carbonate
30 s flow rates to obtain calcite-only precipitation• Precipita-
Tarutani et al. (1969) synthesized inorganic carbonates at tions, made with higher than 24 or lower than 10 bubbles/
two different growth rates by changing the N2 flow rate. In 30 s flow rates, produced mixtures of calcite and vaterite.
contrast to the results obtained for biogenic carbonates, these Our bubbling rates are only qualitatively related to actual
authors did not detect any influence of precipitation rate on rates of precipitation.
the oxygen isotope fractionation factor between CaCO3 and The relation between precipitation (bubbling) rate and
H20. In our experiments, we reinvestigated a possible rela- oxygen isotope fraetionation is shown in Fig. 10. The pH
tion between oxygen isotope fractionation and growth rate values of the final solutions are plotted as well. The pH of
at 25°C using an apparatus specially configured to test two the starting solution was 5•6, and the final pH values were
different types of bubbling (Fig. 9). In Type I bubbling, between 7.5 and 8•5, depending on the flow rate. As CO2 is
relatively large bubbles are generated, one at a time, by carried away in the stream of nitrogen, the pH of the solution
 Oxygen isotope fractionation in synthetic carbonates 3471

T (°C)
40 25 10
32

• Init.Cone. 5 A
30 • Init. Cone. 15
• Init. Cone. 25 ~ /

C,
C,

22

2 0 , , , , , , , , , , , , , , , , , , , ,i , ,t ,i , ,I , ,, , , , . . . . . . . . . .

3.14 3.24 3.34 3.44 3.54 3.64

103fr (IO

Fig. 7. Relation between A = 1031na(witherite-water) and temperature at different initial concentrations of Ba 2+

and HCO 3"

gradually increases until the requisite supersaturation occurs, in nature, it is important to our understanding of the theory
and the CaCO3 precipitates. As can be seen in Fig. 10, the of isotopic fractionation to determine, if possible, the equilib-
rate of precipitation has little bearing on the oxygen isotope rium isotopic properties of this mineral, if indeed equilibrium
fractionation within the flow rate range of calcite-only pre- properties of an unstable mineral can be determined. We
cipitation. It is noteworthy that inorganic calcite was not first synthesized vaterite using the method of Easton and
isotopically light compared to the calcite precipitated by Claugher (1986) because relatively large, well-formed crys-
McConnaughey (1989a,b). According to McConnaughey tals are produced very slowly (several days to a week).
(1989b), isotopic disequilibrium between synthetic carbon- The beautiful crystals of vaterite produced, however, were
ates and water occurs when CO2 undergoes reactions to form isotopically very light and far removed from isotopic equilib-
HCO3. It seems clear from the results of the present experi- rium with the water. The isotopic composition of such vater-
ments and those performed earlier by Tarutani et al. (1969) ite is clearly controlled by the differing rates of diffusion of
that precipitation rate has little or no bearing on the oxygen the different isotopic species of CO2 and useless for our
isotope fractionation between calcite and water. Caution purposes. This behavior is reported here only as an interest-
should be exercised, as always, comparing the results of ing curiosity.
laboratory systems to those of natural systems, particularly Vaterite was always produced with calcite when a satu-
when biological reactions are involved. Also of importance rated calcium bicarbonate solution was prepared by dissolv-
is the fact that our experiments were specifically controlled ing CaCO3. In contrast, only calcite was precipitated from
to produce the thermodynamically stable polymorph of solutions containing relatively high concentrations of CaC12
CaCO3 whereas the coralline data are for aragonite. It is also and NaHCO3. For solutions less: concentrated in chloride,
possible that the rate of precipitation was not changed either mixtures of calcite and vaterite, or pure calcite was
enough in the experiments to produce a measurable effect, synthesized. While it is not cleari which factors control the
but as described below, when the rate of precipitation is very proportion of the two polymorph!c forms of CaCO3, it was
rapid, vaterite-calcite mixtures were produced. a general observation that the proportion of vaterite, precipi-
tated from dilute Na-Ca-C1-HCO31 solutions, increased when
3.4. Precipitation of Vaterite the precipitation rate was very sloW. Rather special physical
and chemical conditions seem to b e necessary for the vaterite
Vaterite, one of the three polymorphs of calcium carbon- polymorph to precipitate, presumably due to its high instabil-
ate, was first identified in the repair tissue of young gastro- ity in aqueous solutions.
pods by Mayer and Weineck (1932). Regardless of its rarity Tarutani et al. (1969) reported that aragonite was enriched
3472 S.-T. Kim and J. R. O'Neil

T (°C)
40 25 10
, l

i
O Init.Cone. 5 J
/-

/
31
• Init.Cone. 10
• Init.Cone. 5/10

29

27 J/
25

/
23 ' ' ' ' ' ' ' ' ' "''''~' '''''''' ........ ' ..........
3.14 3.24 3.34 3.44 3.54 3.64

103/T (K')

Fig. 8. Relation between A = 1031na(octavite-water)and temperature at different initial concentrations of Cd 2+

and HCO~-.

by 0.6%o relative to calcite at 25°C; furthermore, they men- be conducted over a larger range of temperatures at a later
tioned that vaterite was enriched in 180 by 0.5%o relative to date. It is well to keep in mind at this juncture that vaterite
the associated calcite that was separated from mixtures of is a thermodynamically unstable mineral under the condi-
the two polymorphs produced in two separate experiments tions of the experiment and that it is not possible to prove
at 25°C. In the present study, oxygen isotope compositions that isotopic equilibrium was attained between vaterite and
were measured of preparations of pure vaterite and of vater- water in these experiments.
ite-calcite mixtures of different proportions at 25 and 40°C
(Table 4). At both temperatures, vaterite was found to be 3.5. Acid Fractionation Factors
0.6%o enriched relative to calcite, confirming the earlier re-
The oxygen isotope composition of carbonates can be
suits of Tarutani et al. (1969), but with the additional intrigu-
analyzed in a variety of ways, but some modification of
ing suggestion that the vaterite-calcite fractionation is tem-
the original phosphoric acid reaction developed by McCrea
perature independent, at least over this range of 15°. Because
(1950) is used in almost every stable isotope laboratory in
of the theoretical significance of these preliminary data, addi-
the world. C02 liberated by the reaction of the acid with the
tional experiments on this question are warranted and will
carbonate constitutes only two-thirds of the oxygen in the
carbonate so an acid fractionation factor must be applied to
the analysis of the C02 to obtain the isotopic composition
N2 + CO2 ÷ H2Ov Na + CO2 + H20* Na ÷ CO14. Hat*
N2 of the carbonate. At 25°C the liberated CO2 is around 10%o
heavier than the carbonate, and this number will vary by a
couple of permil depending on the carbonate and the temper-
ature of the reaction. Accurate knowledge of the acid frac-
tionation factors is essential to interpreting natural data not
only for the CaCO3 polymorphs but for other common car-
bonates like dolomite, rhodocrosite, ankerite, and siderite.
Sharma and Clayton (1965) measured the 180/160 ratios
Frit T y p e II Type I
of total oxygen of alkaline earth and transition metal carbon-
Fig. 9. Apparatus used in the synthesis of divalent carbonates ates and concluded that acid fractionation factors depend on
illustrating two different types of bubbling employed. See text. the chemical composition of the carbonates. O'Neil (1986)
 Oxygen isotope fractionati0n in synthetic carbonates 3473

30 fractionation factor between divalent metal carbonates and

water.
In an attempt to resolve such apparent conflicts, we re-
29
O Type I
peated the experiments of Sharma and Clayton (1965) for
Q Type II a few selected carbonates using the method described in
section 2.3. For repeat determinations of the same prepara-
tion, reproducibility was generally excellent (e.g., 11.16,
28 A O~ O 11.23; 11.21, 11.15; 11.96, 11.82) so this method appears
~5 to be reliable. In agreement with the findings of Sharma and
Clayton (1965), there are differences in acid fractionation
27 factors (25°C) among the carbonates, but the values obtained
in the present study are different from those obtained earlier,
and sometimes they are significantly different (Table 5).
O Slow(9 bubbles/30see.) The most disturbing conclusion reached from the new
26 determinations is that the fractionation factors obtained de-
O Fast(24 bubbles/30see.)
pended on the material used. Most of the materials examined
were the carbonates synthesized in this study. Note that there
25
. . . . . . . . . . . . , , , , , , , , , , , , , . . . . . . . . . . . . .
is a tendency for 10001na to increase with decreasing tem-
7.5 7.7 7.9 8.1 8.3 8.5 perature of precipitation of the carbonate. A glaring example
of this variation is provided by the determinations of 1000
pH (final solution)
In a for octavite, a well-crystallized mineral that we con-
Fig. 10. Effect of bubbling rate and type on 1031na(calcite-water) cluded was precipitating in near equilibrium in all cases. The
at 25°C. values range widely from an average value of 11.18 for
two determinations using a commercial preparation (sample
CdCO3 of Table 5) to a single determination of 13.60 for
argued from some unpublished data that poorly crystallized sample 42-10. The (corrected) value reported by Sharma
naturally-occurring materials like certain protodolomites and and Clayton (1965) for this mineral was 11.39, a value close
some laboratory synthesized carbonates seemed to have acid to the value we obtained for the commercial preparation.
fractionation factors that were similar to that of calcite as if Our average value for witherite is 10.57 - 0.16 and is 0.34
the rate of reaction was as important or more important than per mil smaller than the value reported by Sharma and Clay-
the chemical composition of the carbonate. O'Neil (1986) ton (1965). Variation of over 2 permil in the fractionation
suggested, furthermore, that if the same acid fractionation factor for a mineral like octavite means that such a variation
factor were applied to all synthetic carbonates, there would could exist for other minerals as well, including the poly-
be no significant oxygen isotope fractionation between morphs of CaCO3.
CaCO3 and CdCO3, and thus the cationic radius would be The average of our eight determinations of the acid frac-
the controlling factor in determining the magnitude of the tionation factor for calcite is 10.44 with an average deviation

Table 4. Data for other experiments.

Sample T e m p .(°C) 6180 (Water) d~180(CaCO3) OZCaCO3.H20 1031na Mineralogy

Calcite-Vaterite Mixtures
60-R1-S 25 -8.30 20.15 1.02869 28.28 Mixture
60-R2-S 25 -8.30 20.59 1.02913 28.72 Mixture
60-R3-S 25 -8.30 20.66 1.02920 28.78 Mixture
60-R1-N 25 -8.30 20.44 1.02898 28.57 Mixture
60-R2-N 25 -8.30 20.78 1.02932 28.90 Vaterite
60-R3-N 25 -8.30 20.29 1.02883 28.42 Mixture
45-40-N 40 -8.30 17.85 1.02637 26.03 Vaterite
58-25-N 25 -8.30 20.76 1.02930 28.88 Mixture
58-40-N 40 -8.30 17.81 1.02633 25.99 Mixture
58-25-S 25 -8.30 20.92 1.02946 29.04 Mixture
58-40-S 40 -8.30 17.80 1.02632 25.98 Mixture

Rate Effect
C61-R1-N 25 -8.30 19.97 1.02851 28.11 Calcite
C61-R2-N 25 -8.30 19.87 1.02841 28.01 Calcite
C61-R1-S 25 -8.30 19.91 1.02845 28.05 Calcite
C61-R2-S 25 -8.30 19.89 1.02843 28.03 Calcite

Temp. = Temperature.
Mixture = Calcite and Vaterite Mixture.
3474 S.-T. Kim and J. R. O'Neil

Table 5. Acid fractionation factors.

Sample 6180 (C02) 6180 (XCO3) a(CO2-XCO3) 1031na Mineralogy

LV3 36.40 25.57 1.01056 10.50 Calcite

24-10 37.28 26.48 1.01052 10.47 Calcite
24-10 37.28 26.50 1.01050 10.44 Calcite
47-10 36.42 25.36 1.01078 10.72 Calcite
38-25-N 31.87 21.22 1.01043 10.38 Calcite
47-25 31.82 21.07 1.01052 10.47 Calcite
38-40-N 28.67 18.24 1.01024 10.19 Calcite
38-40-N 28.67 18.05 1.01043 10.38 Calcite
syn-arag-1 21.77 10.60 1.01105 10.99 Aragonite
syn-arag-2 21.88 10.69 1.01108 11.02 Aragonite
CdCO3 28.03 16.62 i .01122 11.16 Octavite
CdCO3 28.03 16.54 1.01130 11.23 Octavite
44-10 33.97 20.35 1.01335 13.26 Octavite
42-10 34.08 20.12 1.01369 13.60 Octavite
46-10 34.23 21.02 1.01294 12.85 Octavite
44-25 30.47 17.59 1.01267 12.59 Octavite
42-25 30.98 18.00 1.01275 12.67 Octavite
46-25 30.86 18.61 1.01203 11.96 Octavite
46-25 30.86 18.75 1.01189 11.82 Octavite
44-40 27.45 15.68 1.01159 11.52 Octavite
42-40 28.05 15.71 1.01214 12.07 Octavite
46-40 27.77 16.31 1.01127 11.21 Octavite
46-40 27.77 16.37 1.01122 11.15 Octavite
32-10 33.48 22.52 1.01072 10.66 Witherite
32-10 33.48 22.40 1.01084 10.78 Witherite
32-25 29.66 18.67 1.01078 10.73 Witherite
32-25 29.66 18.95 1.01051 10.45 Witherite
49-25-N 28.18 17.33 1.01066 10.60 Witherite
32-40 26.56 16.12 1.01027 10.22 Witherite

X = cation (Ca2+, Cd2+, Ba2+).

of _+0.10. This value is 0.24 permil larger than that reported An expression that combines present low-temperature frac-
by Sharma and Clayton (1965). Our value of 11.01 _ 0.01 tionations with the equilibrium high-temperature fraction-
for aragonite is significantly larger than the value of 10.29 ations of O'Neil et al. (1969) is also proposed for the octav-
reported by Sharma and Clayton (1965). Because the same ite-water fractionation over the temperature range of 0 -
fractionation factor, correct or not, is used in the analysis 500°C:
of all natural carbonates, the significant differences we are
10001na(CdCO3-H20) = 2.76(106T -2) - 3.96
reporting for determinations of acid fractionation factors are
usually of minor importance in the resolution of specific New determinations of acid fractionation factors ( 1000 In
geochemical problems. Where these differences become im- a ) at 25°C for calcite (10.44 + 0.10), aragonite (11.01
portant is in the matter of standards, absolute isotopic ratios, _+ 0.01 ), and witherite ( 10.57 _+ 0.16) are mildly to strongly
and relating oxygen isotope analyses of carbonates to those different from those published by Sharma and Clayton
of water, silicates, and oxides. (1965) and point to a control on this fractionation by some
physical property of the mineral. Reproducible values for
4. CONCLUSIONS octavite (CdCO3) varied from 11.18 to 13.60 depending on
the conditions of preparation of the carbonate. While these
Carbonates slowly precipitated at a given temperature
differences will not affect the interpretation of oxygen iso-
from solutions of different initial concentrations of cation
tope variations among natural carbonates in any major way,
and bicarbonate have markedly different oxygen isotope
these new values will need to be considered in establishing
compositions. Carbonates precipitated from the more con-
absolute 180/160 ratios of international reference standards
centrated solutions employed in this study are interpreted
and in relating analyses of carbonates to those of waters,
to have formed under nonequilibrium conditions. Previous
silicates, and oxides.
workers using the N2 bubbling technique of carbonate precip-
Stable isotope geochemists use a number of published
itation noted that vaterite-calcite mixtures of various uncon-
fractionation factors that may be seriously in error, not only
trollable proportions were the rule. By choosing appropriate
for carbonates as shown in the present study, but possibly
conditions of precipitation, it is possible to produce pure
for other systems as well. It is accepted that new determina-
calcite that is in apparent isotopic equilibriumWith the water.
tions of thermodynamic or physical constants are often dif-
A new expression is proposed for the calcite-water fraction-
ferent fi'om those published previously, but ideally these
ation at low temperatures:
differences will be small. Some of the differences observed
10001na(Calcite-H20) = 18.03 (103T -1) - 32.42 in the present study are important and concern the CaCO3
 Oxygen isotope fractionation in synthetic carbonates 3475

polymorphs, arguably the most studied mineral in the field McConnanghey T. (1989b) 13C and 180 isotopic disequilibrium in
biological carbonates: II. In vitro simulation of kinetic isotopic
of stable isotope geochemistry. Further determinations of the
effects. Geochim. Cosmochim. Acta 53, 163-171.
type presented here are necessary to confirm or disprove McCrea J. M. (1950) On the isotopic chemistry of carbonates and
frequently used fractionation factors. a paleotemperature scale. J. Chem. Phys. 18, 849-853.
O'Neil J. R. (1986) Theoretical and experimental aspects of isotopic
Acknowledgments The authors thank L. Walter, H. Pollack, J. Hor- fractionation. In Stable Isotopes in High Temperature Geological
ita, S. Paulsen, and an anonymous reviewer for their valuable and Processes (ed. J. W. Valley et al.); Rev. Mineral. 16, 1-40.
insightful comments on earlier versions of the manuscript and M. O'Neil J.R. and Barnes I. (1971) C 13 and 018 compositions in
Brandriss, T. Vennemann, and K. Lohmann for their help and advice some fresh-water carbonates associated with ultramafic rocks and
in the laboratory. serpentinites, western United States. Geochim. Cosmochim. Aeta
35, 687-697.
O'Neil J. R. and Truesdell A. H. (1991) Oxygen isotope fraction-
Editorial handling: H. Ohmoto ation studies of solute-water interactions. In Stable Isotope Get-
chemistry: A Tribute to Samuel Epstein (ed. H. P. Taylor, Jr. et
REFERENCES al.); Geochem. Soe. Spec. Publ. 3, 17-25.
O'Neil J. R., Clayton R. N., and Mayeda T. K. (1969) Oxygen iso-
Cohn M. and Urey H.C. (1938) Oxygen exchange reactions of tope fractionation in divalent metal carbonates. J. Chem. Phys.
organic compounds and water. J. Amer. Chem. Soc. 60, 679-687. 51, 5547-5558.
De Villiers S., Nelson B. K., and Chivas A. R. (1995) Biological O'Neil J. R., Adami L., and Epstein S. (1975) Revised value for
controls on coral Sr/Ca and 6180 reconstructions of sea surface the O 18 fractionation between CO2 and H20 at 25°C. J. Res. US
temperatures. Science 269, 1247-1249. Geol. Surv. 3, 623-624.
Dromgoole E. L. and Walter L. M. (1990) Iron and manganese in- Patterson W. P., Smith G. R., and Lohmann K. C. ( 1993 ) Continen-
corporation into calcite: Effects of growth kinetics, temperature, tal Paleothermometry and seasonality using the isotopic composi-
and solution chemistry. Chem. Geol. 81, 311-336. tion of aragonitic otoliths of freshwater fishes. In Climate Change
Drnffel E. R. M. (1985) Detection of E1 Nifio and decade time scale in Continental Isotopic Records (ed. P. K. Swart et al.); Geophys.
variations of sea surface temperature from banded coral records: Monogr. Ser. 78, 191-202.
Implications for the carbon dioxide cycle. In The Carbon Cycle Romanek C. S., Grossman E. L., and Morse J. W. (1992) Carbon
and Atmospheric C02: Natural Variations Archaen to Present isotopic fractionation in synthetic aragonite and calcite: Effects
(ed. E.T. Sunquist and W. S. Broecker), pp. 111-122. Ameri. of temperature and precipitation rate. Geochim. Cosmochim. Acta
Geophys. Union. 56, 419-430.
Dunbar R. B., Wellington G. M., Colgan M. W., and Glynn P. W. Schtitze H. (1980) Der Isotopenindex - eine Inkrementenmethode
(1994) Eastern Pacific sea surface temperatures since 1600 AD: zur nghernngsweisen Berechnung von Isotopenaustauschgleich-
The 6180 record of climate variability in Gal~ipagos corals. gewichten zwischen kristallinen Substanzen. Chem. Erde 39, 321-
Palaeoceanography 9, 291-315. 334.
Easton A. L and Claugher D. (1986) Variations in a growth form Sharma T. and Clayton R. N. (1965) Measurement of O 18/O 16ratios
of synthetic vaterite. Mineral. Mag. 50, 332-336. of total oxygen of carbonates. Geochim. Cosmochim. Acta 29,
Epstein S., Buchsbaum R., Lowenstam H., and Urey H. C. (1951) 1347-1353.
Carbonate-water isotopic temperature scale. Bull. Geol. Soc. Amer. Shen G. T., Cole J. E., Lea D. W., Linn L. J.. McConnaughey T. A.,
62, 417-426. and Fairbanks R. G. (1992) Surface ocean variability at Galapagos
Epstein S., Buchsbaum R., Lowenstam H., and Urey H. C. (1953) from 1936-1982: Calibration of geochemical tracers in corals.
Revised carbonate-water isotopic temperature scale. Bull GeoL Palaeoceanography 7, 563-588.
Soc. Amer. 64, 1315-1326. Socki R. A., Karlsson H. R., and Gibson E. K.. Jr. (1992) Extraction
Golyshev S. I., Padalko N. L., and Pechenkin S. A. (1981) Fraction- technique for the determination of oxygen-18 in water using pre-
ation of stable oxygen and carbon isotopes in carbonate systems. evacuated glass vials. Anal Chem. 64, 829-831.
Geochem. Intl. 18, 85-99. Strainer M.A. and Rye M. (1978)Oxygen and carbon isotope
Grossman E. L. and Ku T.-L. (1986) Oxygen and carbon isotope internal thermometry using benthic calcite and aragonite foramini-
fractionation in biogenic aragonite: Temperature effects. Chem. fera pairs. In Short Papers 4th International Conference, Geochro-
Geol. 59, 59-74. nology, Cosmochemistry, Isotope Geology; USGS Open-File Rept.
Horibe Y. and Oba T. (1972) Temperature scales of aragonite-water 78-701, 408-410.
and calcite-water systems. Fossils 23/24, 69-79. Tarutani T., Clayton R. N., and Mayeda T. K. (1969) The effect
Mayer F. K. and Weineck E. (1932) Die Verbreitung des Kalziumk- of polymorphism and magnesium substitution on oxygen isotope
arbonates im Tierreich unter besonderer Beriacksichtigung der fractionation between calcium carbonate and water. Geochim.
Wirbellosen. Jena. Z Naturw. 66, 199-222. Cosmochim. Acta 33, 987-996.
McConnaughey T. (1989a) 13C and 180 isotopic disequilibrium in Turner J. V. (1982) Kinetic fractionation of carbon-13 during cal-
biological carbonates: I. Patterns. Geochim. Cosmoehim. Acta 53, cium carbonate precipitation. Geochim. Cosmochim. Acta 46,
151-162. 1183-1191.