Dating Rocks With Uranium-235: A Sharper Clock for Ancient Cave Carbonates

Carbonates — limestones, calcite cements and cave deposits such as speleothems — hold some of the most valuable archives of Earth’s past climate, landscapes and even early human evolution. But pinning down when these rocks formed has long been a stubborn analytical problem, especially for samples younger than a few million years that grew from freshwater rather than seawater. A new study published in the journal Geochronology argues that the fix lies in switching the isotope clock that geochronologists rely on, and it lays out a detailed playbook for doing so.

The research, led by Dawid Szymanowski of ETH Zurich with colleagues Marcel Guillong, Pieter Vermeesch and Philip J. Hopley, examines analytical strategies for carbonate geochronology based on the ratio of lead-207 to uranium-235. This pairing is an alternative to the far more commonly used lead-206 to uranium-238 system, and the authors show that with modest adjustments to existing laboratory workflows it can deliver accurate dates across a much wider range of samples than previously appreciated.

The core of the problem is uranium-234, an intermediate daughter in the uranium-238 decay chain. In low-temperature environments, uranium-234 is frequently fractionated from its parent, so the amount present when a carbonate forms rarely matches the secular equilibrium value. That initial imbalance eventually translates into an excess or deficit of lead-206, the final daughter of the chain, which biases raw uranium-238-based dates. In modern seawater the effect is modest: the initial uranium-234 to uranium-238 activity ratio averages about 1.14. In freshwater, however, ratios can climb as high as 12, which according to earlier theoretical work by Vermeesch and colleagues can skew raw dates by as much as 4 million years.

Correcting for this disequilibrium is possible, but only up to a point. Corrections are considered robust for samples no older than roughly 1.5 million years, and even then they demand extremely precise measurements of the residual uranium-234 to uranium-238 ratio — a tricky determination because the dynamic range involved is enormous, with uranium-238 outnumbering uranium-234 by a factor of about 18,200 at equilibrium. Beyond that age threshold, the uncertainties introduced by the correction balloon to the point of making the dates effectively unusable.

The authors illustrate the stakes with a South African case study. The Hoogland basal flowstone sample HL1, dated by solution mass spectrometry in 2019, had a raw Tera-Wasserburg intercept age of 7.407 ± 0.080 million years. Applying a disequilibrium correction with a measured uranium-234 to uranium-238 ratio of 1.0016 slashed the date to 3.145 ± 0.243 million years — a 60 percent reduction — while implying an extreme initial activity ratio of about 12.9. A newer Bayesian uncertainty analysis applied to the same data yields the same corrected date but with a 95 percent credible interval stretching from 2.94 to 7.69 million years, a range spanning 150 percent of the corrected age. Such uncertainty, the authors note, severely limits what the date can actually tell us.

The appeal of the lead-207 to uranium-235 system is that it sidesteps this headache entirely. The only long-lived intermediate daughter in the uranium-235 chain is protactinium-231, whose extremely low solubility in water means the initial disequilibrium can safely be assumed to be zero. Even if that assumption were wrong, the maximum bias from an inaccurate protactinium correction is likely less than 50,000 years. The historical obstacle has been analytical: uranium-235 is about 138 times less abundant than uranium-238 in nature, and lead-207 is correspondingly scarcer than lead-206, so the raw measurements carry larger uncertainties.

To close that gap, the team evaluated the two techniques they consider most useful and complementary. Isotope dilution thermal ionisation mass spectrometry, or ID-TIMS, involves dissolving milligram-sized sample fragments, adding an enriched isotope tracer and purifying uranium and lead before analysis. It offers the highest precision and accuracy and, crucially, allows accurate measurement of lead-204, a stable non-radiogenic isotope that serves as a tracer of initial, non-radiogenic lead. Laser ablation inductively coupled plasma mass spectrometry, by contrast, samples much smaller volumes with a focused beam, offering far better spatial control and a wider spread of data along the isochron, but it generally cannot measure lead-204 reliably because of interference from mercury-204 in the argon gas, forcing reliance on lead-208 instead.

The researchers put both approaches to the test on reference materials and on a new sample, X4, from the same Hoogland flowstone as the problematic HL1. Their new ID-TIMS analyses returned lead-207 to uranium-235 isochron dates of 5.29 ± 0.15 and 5.35 ± 0.13 million years, consistent with an earlier date on the same sample and with their own uncorrected uranium-238-based isochron of 5.349 ± 0.085 million years — indicating that initial uranium-234 disequilibrium was negligible in this particular flowstone. Reprocessed data for the roughly 322-million-year-old marine calcite RA138 and the roughly 3-million-year-old speleothem ASH15 likewise produced lead-207 to uranium-235 dates consistent with published values. Across all three cases, the lead-207 to uranium-235 uncertainties were 1.5 to 6 times larger than the uranium-238-based counterparts, but the data dispersion was consistent with a single age population, whereas the uranium-238 data were overdispersed.

A particularly elegant contribution is the use of measured lead-208 to lead-204 ratios as a contamination detector. An isochron, by definition, is a mixing line between domains of the same age sharing a single initial lead composition, so any aliquot contaminated with extraneous lead can silently bias the date in ways that standard fit statistics may not reveal. Because lead-208 and lead-204 should both remain invariant in thorium-poor carbonates after the system closes, an anomalous lead-208 to lead-204 ratio flags suspect analyses. For the Hoogland sample, three analyses trended toward a low-ratio endmember consistent with contamination from the ancient, uranium-poor dolomite host rock, and excluding them improved the isochron fit. The authors argue this geology-based filtering is a more principled tool than simply discarding statistically poor points.

On the laser ablation side, the team adopted an unusually large 380-micrometre spot size to maximise ion yields, since the standard 100-micrometre spots produced unacceptably large uncertainties in the youngest samples. Laser results for the Hoogland sample gave 5.09 ± 0.26 million years, in excellent agreement with the ID-TIMS date, and ASH15 yielded 3.04 ± 0.25 million years, also consistent. The laser-derived uranium-235 to lead-207 ratios for the Hoogland sample all exceeded those from solution analyses, underscoring laser sampling’s ability to avoid visibly contaminated material. The authors conclude that lead-207 to uranium-235 dating is feasible now and recommend it for all carbonates formed from meteoric water, while noting that marine carbonates, with better-constrained seawater uranium-234 values, may still tolerate the uranium-238 methods if appropriate uncertainty is assigned. Future work, they add, should collect additional compositional information to robustly identify lead-contaminated domains.

Subject of Research: Analytical strategies for uranium-235–lead-207 dating of carbonate rocks to avoid biases from initial uranium-234 disequilibrium

Article Title: Analytical strategies for 207Pb ∕ 235U carbonate geochronology

Article References: Szymanowski, D., Guillong, M., Vermeesch, P., & Hopley, P. J. (2026). Analytical strategies for 207 Pb ∕ 235 U carbonate geochronology. Geochronology, 8(4), 679-692. https://doi.org/10.5194/gchron-8-679-2026

Image Credits: AI Generated

DOI: 10.5194/gchron-8-679-2026

Keywords: carbonate geochronology, uranium-lead dating, speleothems, lead-207, uranium-235, uranium-234 disequilibrium, ID-TIMS, laser ablation ICP-MS, isochron, lead contamination, freshwater carbonates, Geochronology