Newly formed oceanic crust is altered by seawater and carbonated at low temperatures over poorly defined periods of time. We applied in situ U-Pb dating to investigate 28 carbonate veins from Ocean Drilling Program Hole 801C, which is situated in the oldest Jurassic-age oceanic crust preserved in the western Pacific Ocean. Our results indicate that Pacific Ocean crust began accreting at 192 ± 6 Ma, which is ~25 m.y. earlier than previously recognized. Carbonation peaked at 171 ± 5 Ma and continued at a low rate for more than ~65 m.y. after accretion. Jurassic carbonation rates varied over ~10 m.y. timescales but encompassed a range similar to that observed today. These data suggest that carbonation rates are relatively insensitive to changes in atmospheric CO2, but confirm the longevity of seafloor alteration as a critical control in global volatile cycling.

New oceanic crust is continuously formed along the 65,000-km-long chain of volcanic spreading centers throughout the world's oceans (e.g., Bird, 2003). The oldest Jurassic-age oceanic crust that exists today is preserved in the weakly magnetic Pigafetta Basin of the western Pacific Ocean (Fig. 1). The age and alteration state of this crust is of considerable interest because it is considered to be representative of the fast-spreading oceanic crust being subducted beneath the Izu-Mariana arc today (Shipboard Scientific Party, 2000), and it anchors the Geomagnetic Polarity Time Scale in the Jurassic (Tominaga et al., 2021).

Previous studies have shown that basalts recovered from the uppermost part of Ocean Drilling Program (ODP) Hole 801C contain ~2.6 wt% CO2 in carbonate minerals, which is up to 10 times higher than in modern oceanic crust (Alt and Teagle, 1999; Gillis and Coogan, 2011). However, it is not known if the high CO2 content of the Jurassic crust is the result of long-lived alteration over many tens of millions of years or of a higher carbonation rate during the Jurassic, when atmospheric CO2 was approximately three times higher than today (Royer, 2014). Testing these alternatives is critical to our understanding of global volatile cycles, because if carbonation proceeds at a fairly uniform rate, the duration of alteration must control the amount of CO2 and other volatile elements (e.g., H2O, halogens, and noble gases) recycled into the mantle by the subduction of oceanic crust (Alt and Teagle, 1999). Alternatively, if seafloor alteration has an effective duration of less than ~20 m.y. (Coogan et al., 2016), carbonation rates must have varied over time, and the seafloor carbonation rate may provide a feedback control on atmospheric CO2 (Sleep and Zahnle, 2001).

The Jurassic age oceanic crust at ODP Site 801 comprises an ~460-m-thick sequence of sediments overlying volcanic basement (Shipboard Scientific Party, 2000). Previous 40Ar-39Ar dating suggested that the dominant mid-ocean ridge tholeiites were accreted at 167 ± 3 Ma, with minor alkali basalts emplaced in an off-axis setting at 160 ± 1 Ma (Larson and Lancelot, 1992; Koppers et al., 2003; Ludden et al., 2006). We investigated the duration of seafloor carbonation by applying recently developed laser ablation U-Pb dating to carbonate veins in the alkali basalts (sample 1r4–96) and underlying tholeiites (n = 27) recovered from ODP Hole 801C. The carbonate veins were all characterized for major and trace elements, and selected veins were characterized for Sr and clumped oxygen and carbon isotopes to enable comparison with previous studies of carbonate in oceanic crust recovered from Hole 801C and elsewhere (e.g., Alt and Teagle, 2003; Coggon et al., 2010).

We selected pure carbonate and carbonate-bearing veins from the ODP Hole 801C drill core (Table 1) and performed electron microprobe and Raman spectroscopy, demonstrating that the veins are dominated by low-Mg calcite, with dolomite present in some veins and ankerite present on some vein margins. Aragonite was detected on the margin of a vein in sample 1r4–96 from near the top of the basement (Figs. S1–S3 in the Supplemental Material1). Laser ablation trace-element and U-Pb analyses and bulk Sr isotope measurements were undertaken at the Radiogenic Isotope Facility at the University of Queensland (Brisbane, Australia). The U-Pb data were standardized using NIST SRM614 (National Institute of Standards and Technology) and AHX-1a (calcite), and the WC-1 and PTKD-2 calcite check standards gave ages within uncertainty of their reference values (Fig. S4; Roberts et al., 2017; Yang et al., 2021). The analytical uncertainties reported at the 95% confidence level in our figures and tables do not include an uncertainty of ~2.5% in the ages of the standards. However, this uncertainty is included in the text, when comparisons are made between the new U-Pb ages and old 40Ar-39Ar ages. Clumped isotopes measured at Macquarie University (Sydney, Australia) are reported in the carbon dioxide equilibrium scale (CDES) reference frame. See the Supplemental Material for further details.

Vein Ages

The carbonate veins have average uranium concentrations that vary between ~7 µg/g and <100 ng/g (Table 1), which is typical of carbonate in oceanic basement (Alt and Teagle, 2003; Kelley et al., 2005; Coogan et al., 2016). The measured 238U/206Pb and 207Pb/206Pb ratios have analytical uncertainties ranging from 1–2% to much higher values. For clarity, analyses with analytical uncertainty of more than 10–15% are excluded from Figure 2, unless a higher threshold was required for age calculation (see the Supplemental Material).

The veins have 238U/206Pb ratios of 0 to ~40 that are correlated with 207Pb/206Pb to varying degrees, and define mixing trends between a common Pb component with best fit 207Pb/206Pb intercepts of between 0.85 and 0.89 and the lower concordia intercept that is related to sample age (Fig. 2). Regressions with poorly defined common Pb were forced through a 207Pb/206Pb value of 0.86 that represents the average of unaltered basalt (Hauff et al., 2003; Fig. 2). The majority of the dated samples (n = 19 out of 25) have formation ages of between 192 ± 2 Ma and 163 ± 2 Ma, with two additional samples having imprecise ages of 162 Ma and 161 Ma (Fig. 2A). These data indicate that ~80% of the investigated carbonate veins formed in an ~30 m.y. period. Furthermore, they indicate that the peak in carbonate vein growth at 171 ± 5 Ma (Fig. 2A) predated accumulation of the oldest Callovian–Bathonian sediments at ca. 168–164 Ma (Shipboard Scientific Party, 2000).

The ankerite and dolomite in the oldest vein gave consistent results from 166 laser spots defining a single age of 192 ± 2 Ma (Fig. 2), demonstrating that matrix effects between carbonate minerals are smaller than the uncertainty attributable to geological scatter. Including the 2.5% uncertainty in the age of the standards gives total uncertainties in the reported ages of 5–6 m.y., meaning that six of the veins investigated have formation ages that are significantly older than the previously accepted accretion age of 167 ± 3 Ma based on 40Ar-39Ar data (Fig. 2A; Koppers et al., 2003).

Six of the veins dated (n = 6/25) have relatively imprecise formation ages of 162 Ma or younger (Figs. 1 and 2). The low precision is partly attributable to scatter resulting from high analytical uncertainty in low-U samples. Sample 5r1–104 defines a coherent isochron with a low mean square of weighted deviates (MSWD) of 1.6 that records a single generation of carbonate growth at 136 ± 12 Ma, which is 56 ± 12 m.y. younger than the oldest vein (Fig. 2; Table 1). However, the other young samples show considerable geological scatter that could be attributed to either Pb loss that shifts data points to higher 238U/206Pb ratios, or multiple generations of carbonate growth (Fig. 2). The large number of samples investigated means that detailed mapping required to confidently resolve multiple generations of vein growth was not feasible. Instead, scattered data were filtered to calculate maximum formation ages for each sample based on the data selected in Figure 2. This approach suggests that sample 35r1–88 had an initial episode of vein growth at 107 ± 17 Ma that is 85 ± 17 m.y. after accretion (Table 1). However, the scatter obtained for this sample could partly reflect later and/or multiple generations of vein growth (Fig. 2).

Multi-generational vein growth may also explain the unexpectedly high 87Sr/86Sr values of samples 38r1–57 and 7r1–141 (Fig. 3B). If the Sr in these samples was derived from seawater, it suggests episodes of vein growth at ca. 125–128 Ma in sample 38r1–57 and at ca. 25–27 Ma in sample 7r1–141 (Fig. 3B). This possibility is consistent with the scattered U-Pb data for sample 38r1–57 (Fig. 2) and the complex vein mineralogy of sample 7r1–141 (Table 1; Figs. S1 and S2). However, the high 87Sr/86Sr ratios could alternatively reflect the input of radiogenic Sr from terrigenous sediments overlying the hole.

Fluid Origins and Temperature

Most of the carbonate minerals analyzed contain ~100–200 µg/g Sr (average = 118 µg/g; Table 1), but isolated spots in samples 1r4–96 and 31r1–111 contain thousands of µg/g Sr, which indicates the presence of aragonite. Most of the samples analyzed for Sr isotopes (n = 15) have 87Sr/86Sr in the previously reported range (Alt and Teagle, 2003) that either lies close to the seawater curve at its respective U-Pb age (Fig. 2B), indicating seawater as the dominant Sr source (Fig. 3B), or lies beneath the seawater curve, requiring input of Sr from the volcanic basement (Fig. 3B).

The carbonates have rare earth element (REE) abundance patterns between those of seawater and the basaltic host rocks that do not vary as a function of carbonate mineralogy (Fig. 3C). They preserve heavy REE enrichments and positive Y anomalies similar to seawater, but they lack the negative Ce anomaly of seawater and are depleted in light REEs and have very small positive Eu anomalies, similar to basalt (Fig. 3C). The carbonate veins have δ13C of 2.9‰ to −3.5‰, values typical of marine carbonate and similar to previous 801C carbonate analyses (Fig S6; Alt and Teagle, 2003). The carbonate veins have δ18O that ranges from 25.4‰ to 32.1‰ relative to Vienna Standard Mean Ocean Water (VSMOW), which overlaps the previously reported range of 22.3–31.4‰ (Fig S6; Alt and Teagle, 2003; Coggon et al., 2010). The carbonate veins have Δ47 that varies from 0.694 ± 0.028‰ to 0.593 ± 0.008‰ CDES (Table 1). Based on the calcite calibration of Bernasconi et al. (2018), these values correspond to formation temperatures of 19 ± 8–51 ± 3 °C (Table 1), which are indistinguishable from previous 18O isotope temperatures (Alt and Teagle, 2003; Coggon et al., 2010). The δ18Ofluid calculated from the Δ47 temperature is within uncertainty of the 0‰ seawater value for most samples. However, three samples give elevated δ18Ofluid values of up to 3.2 ± 0.9‰ that could reflect the input of hydrothermal fluids in some of the oldest samples (Fig S7). Neither the formation temperature nor the carbonate mineralogy varies systematically as a function of age (Fig S7). However, the Fe-rich carbonates are always found on the vein margins, indicating that they were the first carbonate phase formed within each vein (Figs. S1 and S2), and the Fe-Mg carbonates all come from the top of the hole (Fig S8).

An unexpected result of the current study is the revelation that the Pacific Plate is ~25 m.y. older than previously recognized. The oldest U-Pb vein age of 192 ± 6 Ma (including 2.5% uncertainty in the standard) provides a new minimum accretion age for Hole 801C crust, because the vein occurs only 160 m below the basement and its formation temperature of 36 ± 6 °C (Table 1) would be expected a couple of hundred thousand years after accretion (Stein and Stein, 1992). The new accretion age is substantially older than the previously preferred 40Ar-39Ar age of 167 ± 3 Ma obtained from two tholeiitic basalts (Koppers et al., 2003). However, all previous 40Ar-39Ar ages were based on discordant and difficult-to-interpret, saddle-shaped 40Ar-39Ar age spectra with total fusion ages of up to 180 ± 2 Ma (see Pringle, 1992; Koppers et al., 2003). The combined data can therefore be readily explained by crustal accretion at 192 ± 6 Ma with subsequent 40Ar loss occurring during alteration, which is suggested to have peaked at ca. 170 Ma by both the U-Pb data (Fig. 3A) and Sr isotope modeling (Coggon and Teagle, 2011). This evidence for disturbance of the Ar-system also implies that the alkali basalts may have been emplaced much earlier than previously estimated.

The second revelation is that while the vast majority of seafloor alteration at Site 801 occurred within a ~30 m.y. period following accretion at 192 ± 6 Ma (~80% of dated carbonate veins), seafloor alteration continued for at least 85 ± 17 m.y. until 107 ± 17 Ma (Fig. 3A) and possibly much longer. Furthermore, alteration did not follow an exponentially decreasing rate similar to that suggested for other settings (Coogan et al., 2016; Laureijs et al., 2021). Combining the cumulative distribution of U-Pb vein ages (Fig. 3A) with the previously estimated crustal concentration of 2.6 ± 0.8 wt% CO2 in Hole 801C basalts (Gillis and Coogan, 2011) suggests an initial carbonation rate of the uppermost oceanic crust of 300 ± 100 µg/g of CO2 per m.y. that increased some 13 m.y. after accretion to a maximum rate of 1000 ± 400 µg/g CO2 per m.y. between ca. 179 Ma and 161 Ma (Fig. 3A; Table S1). The reason for the delay in peak carbonation is unclear; however, the maxima might have followed emplacement of the alkali basalts in an off-axis setting. The carbonation rate of the uppermost oceanic crust decreased even further to a mean rate of 100 ± 40 µg/g of CO2 per m.y. between ca. 161 Ma and 107 Ma (Fig. 3A), which is presumably related to the accumulation of sediments after ca. 168–164 Ma (Shipboard Scientific Party, 2000), which restricted fluid infiltration.

Future studies will undoubtedly refine the estimated crustal carbonation rates, particularly for the initial and final periods, which are constrained by relatively few samples (Fig. 3A). Nonetheless, it is apparent that the initial carbonation rate of Hole 801C crust of 300 ± 100 µg/g of CO2 per m.y. is similar to the modern rate of ~350 µg/g of CO2 per m.y. in 6 Ma basalts recovered from ODP Hole 504B with 0.21 wt% CO2 (Alt and Teagle, 1999). Furthermore, the maximum carbonation rate recorded by Hole 801C crust of 1000 ± 400 µg/g of CO2 per m.y. is only slightly higher than the rate of ~830 µg/g of CO2 per m.y. in ca. 6 Ma basalts recovered from ODP Hole 896A with 0.51 wt% CO2 (Alt and Teagle, 1999). The similarity of the Jurassic carbonation rates recorded in Hole 801C and modern carbonation rates recorded in ODP Holes 504B and 896A suggests that seafloor carbonation has been fairly insensitive to the approximate threefold decrease in atmospheric CO2 since the Jurassic (Royer, 2014). Instead, the carbonation rate of the uppermost oceanic crust is probably limited by fluid infiltration related to sedimentation, the rate of basalt dissolution required to provide cations for carbonation (Caldeira, 1995), and fluid advection related to initial cooling and possibly either sustained or rejuvenated by intraplate volcanism that peaked between ca. 125 Ma and 75 Ma in the western Pacific Ocean (Abrams et al., 1993; Stadler and Tominaga, 2015). Therefore, the current data do not favor seafloor carbonation as an important feedback control on atmospheric CO2 (cf. Sleep and Zahnle, 2001).

In combination with previous studies, our new data suggest that the duration of seafloor carbonation provides an important control on the volatile content of oceanic crust subducted into the mantle (Alt and Teagle, 1999). The main alteration stage of oceanic crust recovered from Hole 801C in the Western Pacific lasted ~30 m.y., which is two to three times longer than the 10–15 m.y. estimated from early calcite-Sr-isotope stratigraphic ages and Rb-Sr dating of celadonite of ocean crust recovered from the western Atlantic (Richardson et al., 1980; Staudigel et al., 1981). In addition, it is 10 m.y. longer than the duration of alteration estimated from previous calcite U-Pb dating and in situ Rb-Sr dating of celadonite (Coogan et al., 2016; Laureijs et al., 2021). The total duration of carbonation indicated as 85 ± 17 m.y. is consistent with oceanic crust younger than ca. 65 Ma being colder than predicted by conductive cooling models (Stein and Stein, 1992), and the observed slight change in crustal microporosity as a function of age (Jarrard et al., 2003), which both favor advective cooling of the crust over periods of at least 65 m.y. Further studies are required to better characterize the duration and rate of carbonation in the modern and ancient oceanic crust and its relationship to sedimentation and/or intraplate volcanism (Abrams et al., 1993; Stadler and Tominaga, 2015). However, the current data imply that modern oceanic crust similar to that recovered from ODP Holes 504B and 896A in the eastern Pacific may ultimately achieve a CO2 content similar to that of Jurassic crust recovered from ODP Hole 801C.

The International Ocean Discovery Program (IODP) provided samples, and the Australia–New Zealand IODP Consortium (ANZIC) provided post-expedition funding. ANZIC is supported by the Australian government through the Australian Research Council's Linkage Infrastructure, Equipment and Facilities funding scheme (LE160100067) and the Australian and New Zealand consortium of universities and government agencies. We thank Sean Murrey (Macquarie University) for clumped isotope analysis, and Ying Yu and Faye Liu for technical assistance at the University of Queensland's Centre for Microscopy and Microanalysis and Radiogenic Isotope Facility (Brisbane). We thank Randall Parrish and two anonymous reviewers for constructive comments that improved the manuscript.

1Supplemental Material. Description of the methods, Figures S1–S8, and Dataset S1 (an Excel data sheet with all U-Pb, trace element, and clumped isotope data). Please visit https://doi.org/10.1130/GEOL.S.20466036 to access the supplemental material, and contact editing@geosociety.org with any questions.
Gold Open Access: This paper is published under the terms of the CC-BY license.