High-precision Triple Oxygen Isotope Research Articles

High precision triple oxygen isotope composition of small size urban micrometeorites indicating constant influx composition in the early geologic past

AbstractIn this study, we present a method for high precision Δ′17O (Δ′17ORL = ln(δ17O + 1) – λRL ln(δ18O + 1)) analysis of small mass silicate and oxide materials. The analyses were conducted by laser fluorination in combination with gas chromatography and continuous flow isotope ratio monitoring gas spectrometry. We could analyze the oxygen isotope composition of samples down to 1 μg, which corresponded to about 13 nmol O2. The analytical error (we report the 1σ external reproducibility of a single analysis) in δ18O increases with decreasing sample sizes from ~0.2‰ for ~20 μg samples to ~0.9‰ for 1 μg samples. For Δ′17O, we achieved an external reproducibility of 0.04‰ for a sample mass range between 1 and 27 μg. The uncertainty in Δ′17O is smaller than the uncertainty in δ18O due to the correlated errors in δ17O and δ18O. We applied the method to urban micrometeorites, that is, small meteorites (<2 mm) that were sampled from a rooftop in Berlin, Germany. A total of 10 melted micrometeorites (S‐type cosmic spherules, masses between 11 and 22 μg) were analyzed. The oxygen isotope compositions are comparable to that of modern Antarctic collections, indicating that the urban micrometeorites sample the same population. No indication for terrestrial weathering had been identified in the studied set of urban micrometeorites making them suitable materials for the study of micrometeorite origins.

A novel method for high-precision triple oxygen isotope analysis of diverse Earth materials using high temperature conversion–methanation–fluorination and isotope ratio mass spectrometry

Triple oxygen isotope geochemistry is a growing field that investigates mass dependent and mass independent fractionation in 17O/16O relative to 18O/16O imparted by geochemical, physical, and biological processes. Many applications require 10 ppm-level precision in measurement of the Δ′17O parameter. Such analytical precisions are rarely realized for some materials (e.g. sulfates), and no method presently exists for many organic compounds. We describe a new analytical method that combines the versatility of high-temperature conversion (HTC, also known as Temperature Conversion Elemental Analysis, TC/EA) with the high-precision of the methanation-fluorination method described by Passey et al. (2014). In this method, materials including organics, phosphates, sulfates, nitrates, carbonates, and silicates undergo high-temperature conversion of their oxygen to CO in a glassy carbon reactor at temperatures >1400 °C. This CO is then reacted with hydrogen at 560 °C over an iron catalyst to transfer CO oxygen to oxygen in H2O (a methanation reaction). This H2O is subsequently fluorinated by CoF3 to yield O2, which is then analyzed by mass spectrometry to determine δ17O and δ18O. Sample oxygen conversion is quantitative (or nearly so) for most compatible materials, resulting in excellent reproducibility in Δ′17O (<10 ppm 1σ) and no clear evidence of fractionation affecting Δ′17O. We report the first (to our knowledge) high-precision Δ′17O measurements for organics, including benzoic acid (IAEA-601, IAEA-602), cellulose (IAEA-CH-3), keratin (USGS KHS), and honey (USGS 83). We present results for silver phosphates (NBS 120c, USGS 80, USGS 81, B2207), carbonates (IAEA-603, IAEA-C1, NBS 18), sulfates (IAEA-SO-5, IAEA-SO-6), a silicate (NBS 28), and waters (VSMOW2, SLAP2). We also present results from several in-house reference materials, including a silver nitrate and a phosphate that was equilibrated with isotopically exotic waters. Analysis of carbonates, sulfates, and silicates requires accelerants (e.g., graphite and AgCl or KF) to achieve high yields and ppm-level precision in Δ′17O. Waters can be analyzed using the same method during the same analytical sessions as solid materials, making determination of water-mineral fractionation factors straightforward and enabling values to be placed directly on the VSMOW-SLAP scale by concurrent analyses with those reference materials. We find that the method is especially well-suited for organics, sulfates, phosphates, and nitrates, however traditional fluorination methods may be better suited for many silicate minerals.

High-Precision Triple Oxygen Isotope Analysis of Carbon Dioxide by Tunable Infrared Laser Absorption Spectroscopy.

Precision measurements of the stable isotope ratios of oxygen (18O/16O and 17O/16O) in CO2 are critical to atmospheric monitoring and terrestrial climate research. High-precision 17O measurements by isotope ratio mass spectrometry (IRMS) are challenging because they require complicated sample preparation procedures, long measurement times, and relatively large samples sizes. Recently, tunable infrared laser direct absorption spectroscopy (TILDAS) has shown significant potential as an alternative technique for triple oxygen isotope analysis of CO2, although the ultimate level of reproducibility is unknown, partly because it is unclear how to relate TILDAS measurements to an internationally accepted isotope abundance scale (e.g., VSMOW2-SLAP2). Here, we present a method for high-precision triple oxygen isotope analysis of CO2 by TILDAS, requiring ∼8-9 μmol of CO2 (or 0.9 mg carbonate) in 50 min, plus ∼1.5 h for sample preparation and dilution of CO2 in N2 to a nominal 400 μmol mol-1. Overall reproducibility of Δ'17O (CO2) was 0.004‰ (4 per meg) for IAEA603 (SE, n = 6) and 10 per meg for NBS18 (SE, n = 4). Values corrected to the VSMOW2-SLAP2 scale are in good agreement with established techniques of high-precision IRMS, with the exception of Δ'17O measured by platinum-catalyzed exchange of CO2 with O2. Compared to high-precision IRMS, TILDAS offers the advantage of times less sample, and greater throughput, without loss of reproducibility. The flexibility of the technique should allow for many important applications to global biogeochemical monitoring and investigation of 17O anomalies in a range of geological materials.

Triple Oxygen Isotopic Compositions of Ocean Water from the Mariana Trench

High-precision triple oxygen isotope data on 40 water samples (5–10 923 m in depth) collected at the Challenger Deep of the Mariana Trench are reported in this study. The analyses at Laboratoire des Sciences du Climat et l’Environnement yield mean isotope values δ18O = −0.084 ± 0.224‰, δ17O = −0.061 ± 0.117‰, and 17O-excess = −17 ± 5 per meg, with standard deviations reported at 1σ. The range of δ18O (from −0.480 to 0.544‰) falls within records of the Global Seawater Oxygen-18 Database at the Mariana Trench. The average 17O-excess value at the Mariana Trench is more negative than the average 17O-excess value of −5 ± 4 (1σ) per meg in the only prior data set including deep ocean samples. The slope (λ) of the three-isotope plot of Mariana Trench water is 0.521 ± 0.003 (1σ), lower than λ of 0.528 ± 0.001 (1σ) for ocean water in the prior study. The new data set matches the prediction of the ocean isotope mass balance model, suggesting that it may represent a more appropriate ocean endmember for triple oxygen isotope thermometry. The 17O-excess of ocean water with respect to Vienna Standard Mean Ocean Water (VSMOW) is recognized to be a necessary correction in quantifying gross oxygen productivity of euphotic regions and relative humidity of moisture source regions.

Comparison of Three Measurement Principles on Water Triple Oxygen Isotopologues

The widespread method for measuring Δ17O (17O-excess) is an offline CoF3 (Cobalt tri-fluoride) conversion of water to molecular oxygen with subsequent isotope determination by dual inlet mass spectrometry. High precisions for Δ17O measurements, using CoF3 water conversion, are so far only possible with off-line methods. Here we report on an improved and modified online continuous flow method intended for high precision triple oxygen isotope analysis. This method is improved by optimizing the reactor (site for conversion of H2O into oxygen through the chemical reaction) compositions, size of the fused silica capillary, flow regulator, and data treatment. Our modified online continuous method was further compared with the recently developed cavity ring down measurement principle. The precision is significantly better for the commercially available laser-based system than our current version of improved online CoF3 conversion method using mass spectrometry. Factors identified for limiting precision in our continuous flow system are: (i) compaction of the reactor with time that leads to the restriction of flow rate of carrier gas, (ii) the CoF3 treatment, (iii) the amount of CoF3 inside the reactor, (iv) the pore size of the steel frit, and (v) the metallic tube. Changes in all of these items as well as the dimension of the fused silica capillary, the positioning of the fused silica capillary in the open split, and the memory effect can also lead to a declining precision. These limiting factors for precision still provide us enough space for further improvement of our improved online method which will be worthwhile for the measurement of smaller aliquot samples as fluid inclusions for palaeoclimatic applications. With present improvement, multiple injections (n = 15 or even more) should be applied to obtain a precision better than 10 per meg for Δ17O. Furthermore, a comparison of the laser-based system with an improved conventional equilibration method has been made on precipitation samples originating from Jungfraujoch.

Triple oxygen isotopes of cherts through time

It has long been recognized that the stable oxygen isotope 18O/16O ratios of chemical sediments show a systematic decrease with increasing age. This continuous, 3.8 Ga long secular trend has been attributed to 80 °C hot early Earth oceans, a ~15‰ lower 18O/16O ratio of Archean seawater, or diagenetic obliteration of the original isotope signal. We demonstrate that high-precision triple oxygen isotope measurements including the rare third stable oxygen isotope 17O on silica rocks (cherts) can be used to shed light on this major problem in geochemistry. Our triple oxygen isotope data do not support 80 °C hot oceans in the Archean. The data also exclude that the 18O/16O of seawater was much lower in the Archean. Also, a simple freshwater-origin of cherts is not supported by the new data. We discuss the effects of diagenetic alteration on the triple oxygen isotope compositions of cherts and show that the Archean and Proterozoic cherts could have equilibrated with fluids with higher 18O/16O than today's oceans. For the Archean cherts, the high 18O/16O fluids required to explain the observations could reflect the composition of ambient seawater or of modified hydrothermal fluids. Oceans with higher 18O/16O ratio would be expected for an Archean “water world” without large continents. High 18O/16O fluids, however, could also have formed locally by reactions between seawater and the oceanic crust. Our study shows that high-precision triple oxygen isotope data provide new understandings of processes deep in time.

Calibration of carbonate-water triple oxygen isotope fractionation: Seeing through diagenesis in ancient carbonates

High precision triple oxygen isotope measurements of carbonates can better constrain temperatures and oxygen isotope compositions of seawater through geologic time than 18O/16O measurements alone, but lack of a definitive calibration has hindered progress. In this study, we fluorinated both carbonate and water samples to measure quantitatively the triple oxygen isotope composition of each phase. We compared the oxygen isotope fractionation between carbonate and water for different carbonate materials: calcite synthesized with and without carbonic anhydrase, abiogenic calcite from Devils Hole, and extant biogenic calcite and aragonite of marine origin. We found similar 1000lnα18Occ-wt values for all materials and combined the results with the high temperature experimental data of O'Neil et al. (1969), resulting in the following fractionation equation (T in Kelvins) 1000lnα18Occ-wt=2.84(±0.02)×106T2-2.96(±0.19). The calcite triple oxygen isotope values yielded a θ-T relationship of θcc-wt = –1.39(±0.01)/T + 0.5305 whereas the aragonite triple oxygen isotope values yielded a θ-T relationship of θara-wt = –1.53(±0.02)/T + 0.5305. The calcite-water triple oxygen isotope equilibrium fractionation equation for natural samples is Δ17′Occ-Δ17′Owt=2.84(±0.02)×106T2-2.96(±0.19)-1.39(±0.01)T+0.5305-λ. The combined 1000lnα18O and 1000lnα17O relationships can be used to assess equilibrium in ancient samples and to evaluate potential secular changes in the δ18O value of seawater. Most of the Phanerozoic samples analyzed in this study, which were determined to be pristine in previous studies, have undergone some level of diagenesis. Two samples appear to preserve their original oxygen isotope compositions and suggest a cool ocean with a δ18O value similar to the modern ocean. Using a fluid-rock interaction model, we can “see through” the diagenetic process and estimate the triple oxygen isotope composition of the carbonate prior to alteration. In doing so, we show that for the time intervals and sample locations measured in this study, Phanerozoic oceans had a comparable range of oxygen isotope compositions and temperatures as modern seawater.

Deuterium excess and 17O-excess variability in meteoric water across the Pacific Northwest, USA

High-precision triple oxygen isotope analysis of water has given rise to a novel second-order parameter, 17O-excess (often denoted as Δ17O), which describes the deviation from a reference relationship between δ18O and δ17O. This tracer, like deuterium excess (d-excess), is affected by kinetic fractionation (diffusion) during phase changes within the hydrologic cycle. However, unlike d-excess, 17O-excess is present in paleowater proxy minerals and is not thought to vary significantly with temperature. This makes it a promising tool in paleoclimate research, particularly in relatively arid continental regions where traditional approaches have produced equivocal results. We present new δ18O, δ17O, and δ2H data from stream waters along two east–west transects in the Pacific Northwest to explore the sensitivity of 17O-excess to topography, climate, and moisture source. We find that discrepancies in d-excess and 17O-excess between the Olympic Mountains and Coast Range are consistent with distinct moisture source meteorology, inferred from air-mass back trajectory analysis. We suggest that vapor d-excess is affected by relative humidity and temperature at its oceanic source, whereas 17O-excess vapor is controlled by relative humidity at its oceanic source. Like d-excess, 17O-excess is significantly affected by evaporation in the rain shadow of the Cascade Mountains, supporting its utility as an aridity indicator in paleoclimate studies where δ2H data are unavailable. We use a raindrop evaporation model and local meteorology to investigate the effects of subcloud evaporation on d-excess and 17O-excess along altitudinal transects. We find that subcloud evaporation explains much, but not all of observed increases in d-excess with elevation and a minor amount of 17O-excess variation in the Olympic Mountains and Coast Range of Oregon. Key Points 17O-excess correlates spatially with relative humidity across the Pacific Northwest, supporting its use as an aridity indicator in paleoclimate studies. Discrepancies in d-excess and 17O-excess between the Olympic Mountains and Oregon Coast Range suggest that their moisture source is different. Subcloud evaporation explains most of observed increases in d-excess with elevation, and a minor amount of 17O-excess variation in the Olympic Mountains and Oregon Coast Range.

An internally consistent triple oxygen isotope calibration of standards for silicates, carbonates and air relative to VSMOW2 and SLAP2

High precision triple oxygen isotope measurements are becoming a more common analysis in laboratories. There is a lack of calibrated standards to use for triple oxygen isotope measurements and this has led to data being presented on different scales rather than to the traditional VSMOW2-SLAP2 scale. Here we present triple oxygen isotope values of standard carbonates, CO2 liberated from carbonates, silicates and air calibrated to the VSMOW2-SLAP2 scale. We analyzed VSMOW2 and SLAP2 to calibrate our reference gas. Our measured δ18O value of SLAP2 is −55.55‰, indistinguishable from the accepted value of −55.5‰. Our Δ′17O value of SLAP2 (λ = 0.528) is not zero, but rather −0.015‰, corresponding to a δ17O value of −29.741‰. The Δ′17O values of carbonate standards NBS19, IAEA603 and NBS18 are −0.102, −0.100 and − 0.048‰, respectively (±0.010). For CO2 of calcite liberated by phosphoric acid digestion at 25 °C, the θACID value at 25 °C is 0.5230 ± 0.0003. These results can be used to correct triple oxygen isotope measurements of CO2 released by phosphoric acid digestion in other laboratories. We present triple oxygen isotope values for UW Garnet-2, NBS-28, San Carlos Olivine (NM-SCO), and our in-house quartz standard (NM-Q). Aliquots of NM-Q and NM-SCO are available from the Center for Stable Isotopes (CSI), New Mexico for interlaboratory comparison. Our δ17O and δ18O values for air are 12.178‰ (±0.066) and 24.046‰ (±0.117), respectively, with a corresponding Δ′17O value of −0.441 ± 0.012‰. With the availability of common standards, all laboratories making δ17O-δ18O measurements can calibrate their reference gas relative to the VSMOW2-SLAP2 scale. Laboratories making triple oxygen isotope measurements on CO2 released from carbonates using phosphoric acid digestion can correct to the bulk carbonate value.

Triple oxygen isotope systematics as a tracer of fluids in the crust: A study from modern geothermal systems of Iceland

Triple oxygen isotope systematics of ancient hydrothermally altered rocks has been previously used to constrain environmental conditions of the Precambrian. To validate those studies, we report high-precision triple oxygen isotope measurements (expressed as Δ′17O with reference slope 0.528) of quartz, epidote and well fluids from modern geothermal areas of Iceland, Krafla and Reykjanes, as well as measurements from the extinct 6 Ma Geitafell hydrothermal system. At these systems, basalts reacted with distinct fluid sources at temperatures ranging from 250 to 400°C. Resolvable difference between isotope compositions of surface waters and rocks enables novel insights into boiling, isotope exchange at variable water-rock ratios, and remelting of altered rock. Our measurements of δD, Δ′17O, and δ′18O in well fluids show that the reactions proceeded at water-rock ratios of 0.1 to 2, and reveal the addition of meteoric water in the Reykjanes system, and near-surface boiling and steam-liquid separation at Krafla. The δ′18O and Δ′17O values of fluids shift due to exchange with rocks at high temperature following the slope 0.51 in the triple isotope space. Near-surface boiling and steam-liquid separation cause shifts in δD and δ′18O values of the fluids. Their Δ′17O values do not change significantly owning to the shallow slope of liquid-vapor equilibrium fractionation relative to the reference line slope. Epidote δ′18O and Δ′17O values in all three localities closely resemble the isotope composition of local fluid sources. The measured slope of triple oxygen isotope fractionation between quartz and epidote at 250–400°C is 0.526±0.001. We suggest that triple oxygen isotope composition of epidote can be used as a direct first-order proxy for the equilibrium fluid δ′18O and Δ′17O values. The calibrated quartz-water equilibrium fractionation for triple oxygen isotopes yields general agreement with the local fluid sources, within ±1.5‰ of their δ18O values, while the Δ′17O agree within ±0.02‰. We present in situ δ18O measurements in a quartz crystal from Krafla that show several ‰ heterogeneities which may affect the reconstructed equilibrium fluid values. We tested the effect of shallow crustal contamination on the Δ′17O values of rhyolitic glasses from Krafla, including those quenched and extracted by drilling, that likely formed by assimilation of low-δ18O hydrothermally altered crust. Our Δ′17O measurements constrain the degree of crustal assimilation to 10–20 %. Our study shows that the Δ′17O values measured in geothermal fluids, secondary minerals and low δ18O contaminated magmas can provide key information on the conditions of water-rock reaction and magma genesis, and contain additional details that were not accessible through conventional analyses of δD and δ18O.

Insights into the origin of carbonaceous chondrite organics from their triple oxygen isotope composition

Dust grains of organic matter were the main reservoir of C and N in the forming Solar System and are thus considered to be an essential ingredient for the emergence of life. However, the physical environment and the chemical mechanisms at the origin of these organic grains are still highly debated. In this study, we report high-precision triple oxygen isotope composition for insoluble organic matter isolated from three emblematic carbonaceous chondrites, Orgueil, Murchison, and Cold Bokkeveld. These results suggest that the O isotope composition of carbonaceous chondrite insoluble organic matter falls on a slope 1 correlation line in the triple oxygen isotope diagram. The lack of detectable mass-dependent O isotopic fractionation, indicated by the slope 1 line, suggests that the bulk of carbonaceous chondrite organics did not form on asteroidal parent bodies during low-temperature hydrothermal events. On the other hand, these O isotope data, together with the H and N isotope characteristics of insoluble organic matter, may indicate that parent bodies of different carbonaceous chondrite types largely accreted organics formed locally in the protosolar nebula, possibly by photochemical dissociation of C-rich precursors.

Atmospheric measurements of Δ17O in CO2 in Göttingen, Germany reveal a seasonal cycle driven by biospheric uptake

The triple oxygen isotope composition of tropospheric CO2 might be a promising new tracer for terrestrial gross carbon fluxes. This notion is based on global box modeling of its abundance, and on highly challenging and therefore very sparse measurements of 16O, 17O and 18O in CO2 in the lower atmosphere. Here, we present additional high-precision triple oxygen isotope measurements of ambient air CO2 sampled in Göttingen (NW Germany) over the course of 2years and of two air samples taken on top of the Brocken Mountain (1140m, NW Germany). Göttingen differs from other locations where Δ17O was measured by its proximity to both urban sources of CO2, and to extensive uptake of CO2 by vegetation. In our analysis, we specifically try to discern this latter influence on our measurements, and to distinguish it from other known sources of variation in Δ17O.Our triple oxygen isotope data are reported as Δ17O values relative to a CO2-water equilibration line with Δ17O=ln (δ17O+1)−0.5229×ln (δ18O+1). We report an average of -0.02±0.05‰ (SD) in the first year and -0.12±0.04‰ (SD) in the second year of our measurements. This year-to-year difference is higher than expected based on other available Δ17O records, but careful scrutiny of our measurement approach did not reveal obvious analytical biases, leaving this aspect of our record unexplained. After removing the year-to-year trend, our time series shows a statistically robust seasonal cycle with maximum values in June/July and an amplitude (peak-to-trough) of 0.13±0.02‰. We compare our observational data to a revised triple oxygen isotope mass balance “box” model of tropospheric CO2 where we reconcile both 18O/16O and 17O/16O fractionation processes. We also compare them to Göttingen-specific output from a three-dimensional transport model simulation of Δ17O in CO2 performed with the Tracer Model 5 (TM5). Both the modeled isofluxes at the surface, and the modeled stratospheric, fossil, and biospheric Δ17O components in the atmosphere at Göttingen confirm that the observed seasonal cycle in Δ17O is driven primarily by the seasonal cycle of gross primary productivity (GPP), and that the seasonal variations in both stratospheric transport and fossil fuel emissions play a minor role at our location. Our results therefore strengthen earlier suggestions that GPP is reflected in Δ17O, and call for more seasonally resolved measurements at continental locations like Göttingen.

High-precision measurement and standard calibration of triple oxygen isotopic compositions (δ18O, Δ′17O) of sulfate by F2 laser fluorination

A new approach for measuring the triple oxygen isotope composition of sulfate minerals was developed using fluorine gas and an infrared laser to generate O2. A correction for the mass-dependent isotope effect observed during barite fluorination was rigorously calibrated. Analyte gas purification was performed via numerous cryofocus steps and the introduction of an in-line gas chromatograph. As with previous methods, our fluorination technique still requires pairing with independent δ18O measurements made by TC/EA conversion to carbon monoxide, but leads to higher yields than BrF5-based methods. We first calibrated our method against known silicate standards (UWG-2, SCO, NBS-28). Following from this, we report high-precision triple oxygen isotope ratios for international sulfate standards (IAEA-SO-5, IAEA-S O-6, NBS-127), and an internal laboratory standard, JMG. Replicate analyses of JMG yielded a Δ′17O value of −0.057 ± 0.004‰ (standard error), demonstrating per-meg level precision that approaches instrumental limits. We provide δ17O and δ18O values that can be translated into any preferred reference frame for comparison with gases, water, or rocks/minerals. Quantification of differences in triple oxygen isotope composition of the standard reference materials further enables application of this measure and approach on environmental and geological materials.

Continental scale variation in 17O-excess of meteoric waters in the United States

High-precision triple oxygen isotope analysis of waters is an emerging tool in hydrological and paleoclimate research. The existing research on 17O-excess in waters includes surveys of meteoric waters and region-specific studies of high-latitude snow and tropical storms. However, a better understanding of the variation in 17O-excess of waters across large geographic regions is needed to expand the utility of triple oxygen isotope measurements. Here we present 17O-excess data from tap waters across the continental U.S., which we used as a proxy for precipitation. The 17O-excess values of tap waters ranged from −6 to +43 per meg and averaged 17±11 per meg which is lower than the average 17O-excess reported for global meteoric waters, but overlaps with reported 17O-excess values of rainfall from the tropics. We observed relatively high 17O-excess values (>25 per meg) of tap waters in the northwestern U.S. and some of the lowest 17O-excess values (<5 per meg) in the states bordering the Gulf of Mexico. The latitudinal variation of 17O-excess among tap waters likely reflects the different controls on 17O-excess in precipitation. For example, re-evaporation of precipitation and convective processes influence the isotopic composition of tap waters from the southern portions of the U.S., resulting in relatively low 17O-excess values. In contrast, these effects are reduced in tap waters from the northern portions of the U.S. where snow and cold-season rainfall are primarily responsible for the majority of annual precipitation. Exceptions to the latitudinal trend are prevalent in the central portions of the U.S., where mixing and convection are likely responsible for 17O-excess values that are lower than would be expected at their latitudes. The results of this study provide both a first look at the variation of 17O-excess in meteoric waters on a continental scale and a predictive map for 17O-excess of meteoric waters in the U.S.

Triple oxygen isotope equilibrium fractionation between carbon dioxide and water

The growing number of recent high precision triple oxygen isotope data of water and atmospheric gases requires the reinvestigation of the major isotope fractionation processes between the hydro-, bio- and atmosphere. Here we present the experimental triple oxygen isotope fractionation factor between CO2 gas and water in the temperature range between 2 and 37°C. We have determined the triple oxygen isotope fractionation factor, θ=ln17α/ln18α, for CO2–water equilibrium and found that θCO2–water=0.522±0.002 (t0.95×SEM). No temperature dependence was resolved. Knowledge on the triple oxygen isotope equilibrium fractionation between CO2 and water is crucial for estimating the triple oxygen isotope composition of tropospheric CO2. Model predictions combined with high-precision Δ17O analysis of tropospheric CO2 might reveal new information on biosphere–atmosphere interactions or the influx of anomalous stratospheric CO2.