Elsevier

Quaternary Geochronology

Volume 54, October 2019, 101009
Quaternary Geochronology

Corrections for initial isotopic disequilibrium in the speleothem U-Pb dating method

https://doi.org/10.1016/j.quageo.2019.101009Get rights and content

Highlights

  • Initial isotopic disequilibrium in the U-Pb system can dramatically affect the apparent speleothem age.

  • Incorporating disequilibrium information requires an expansion of the Tera-Wasserburg concordia formalism.

  • Initial 234U disequilibrium can be calculated by comparing the 238U and 235U decay chains.

  • Using 208Pb as a “non-radiogenic” reference isotope improves isochron precision and exposes heterogeneous common Pb.

  • There are now several options for creating high-quality speleothem U-Pb ages.

Abstract

The accuracy of ages produced by the U-Pb speleothem chronometer is influenced by initial disequilibrium effects resulting from weathering and transit of water through the epikarst. If some residual disequilibrium remains at the time of measurement, corrections can be made. However, once the system reaches equilibrium, the initial state has to be estimated thus propagating significant uncertainty into the age calculation. As an alternative approach, here we explore the possibility of calculating [234U/238U]i initial values by forcing concordance in the 238U to 206Pb and 235U to 207 Pb decay schemes.

Whether this method can be successfully implemented relies heavily on the quality of the linear regression enacted, particularly with respect to the 235U-207 Pb decay scheme. We therefore also explore the effect of 208Pb-based normalisation in order to optimise isochron construction. An unexpected consequence of this analysis is the observation that 208Pb/204Pb ratios may aid in the identification of heterogeneous initial Pb isotopic compositions in speleothems. Both methods have the potential to generate improved isochron fits which ultimately produce improved disequilibrium corrected speleothem ages. Initial disequilibrium values obtained in this way can be applied to other samples where disequilibrium information is unavailable.

Introduction

Speleothems (secondary cave calcites such as stalagmites, stalactites and flowstones) form robust archives of past climate and variability. Mean annual temperature, rainfall, and surface vegetation data can all be recorded in successive layers of actively growing speleothems, in certain cases with annual to sub-annual resolution. Most importantly, they are also highly suited to absolute radiometric dating techniques, in particular the familiar U-Th dating method (e.g. Richards and Dorale, 2003b). Unfortunately, this widely employed chronometer is limited to samples formed within the last ~640 ka (Cheng et al., 2013); for older materials an alternative methodology can be employed using the U-Pb decay chain (e.g. Richards et al., 1998; Woodhead et al., 2006).

While the U-Pb method shows great utility in diverse research areas from climate change, human evolution and migration, bio-diversity, tectonics to landscape evolution (e.g. Rasbury and Cole, 2009; Woodhead and Pickering, 2012), it also has limitations. The most intractable of these, at least for samples formed in the last several million years, is the difficulty in making a correction for initial isotopic disequilibrium in the 238U-206 Pb decay chain, primarily relating to the intermediate daughter isotope 234U. It is well known that speleothems can form with highly variable initial 234U/238U ratios. Uranium-234 may become enriched in groundwaters due to preferential leaching of the daughter nuclide following α-recoil, and potentially depleted in groundwaters interacting with old/previously depleted source rocks (see Porcelli and Swarzenski (2003) for detailed discussions on U-series nuclide behaviour in groundwaters). With time, and U decay, these ratios migrate towards an equilibrium state. Any initial (i.e. when the speleothem forms) excess of 234U over equilibrium values will result in calculated U-Pb ages that are too old; similarly, 234U deficits will produce erroneously young U-Pb ages (Neymark et al., 2000). Samples that retain 234U disequilibrium can, in some instances, impart age information, provided that the initial 234U/238U ratio can be constrained (e.g. Ludwig et al., 1992). Unfortunately, however, once the system attains equilibrium (i.e. 234U/238U deficit or excess can no longer be measured by mass spectrometry) knowledge of this ‘initial state’ is lost and it is no longer possible to make an accurate correction for this effect on derived U-Pb ages. Corrections are then usually based on either an estimate of the local initial state of 234U disequilibrium (for example, based upon analyses of younger speleothems with measurable 234U/238U). However, local 234U disequilibrium values often vary through time such that applied local estimates should only represent a first approximation. Samples lacking any initial disequilibrium context require a somewhat arbitrary estimate with a necessarily large uncertainty budget imposed – see later discussion.

In contrast to these complexities, the 235U-207 Pb decay chain is not limited by such effects (with the possible exception of minor 231Pa disequilibrium, section 3.5) and so, theoretically, ages derived from this decay scheme could be considered more accurate than those from 238U-206 Pb decay. Unfortunately, however, concentrations of these isotopes in typical samples are far lower than their 238U-206Pb counterparts and produce lower count rates that limit analytical precision. In this study we explore this trade-off in an attempt to provide optimal strategies for U-Pb dating of speleothems. In this context we also explore the use of 208Pb-as opposed to 204Pb-normalization as a means of reducing uncertainty on the isochron line fits: constraining linear regression is an essential component of employing the 235U-207Pb scheme in this way.

Three speleothem samples with different ages and U-Pb concentrations (based on reconnaissance studies) were selected for detailed analysis. Samples 73052 from the Nahanni caves in the Mackenzie Mountains, Northwest Territories of Canada, and CC12-f-2 from Monte Corchia cave system in Italy are both of early Quaternary age with high U content (10–15 ppm) but the former has a substantially larger Pb content (~0.01–0.2 ppm) compared to the latter (~0.004–0.01 ppm). Sample 483-9 from the Nullarbor Plain of southern Australia is of Pliocene age but with relatively low U content (~0.1 ppm) and very minor common Pb component (~0.002 ppm), limiting the range of U/Pb ratios. Together these samples encompass a variety of speleothem compositions likely to be encountered in routine U-Pb dating programs.

Section snippets

Sampling

Speleothems were cut perpendicular to the growth axis and sampled using a dental drill along the calcite growth bands considered to be least contaminated by detrital material (i.e. the most translucent layers). Five or more sub-samples, each of ~100 mg size, were extracted from each growth layer of interest. After sub-sample extraction, all subsequent work was conducted in a multiple HEPA filtered environment using 2- or 3-times distilled acids and ultrapure water.

Chemistry and mass spectrometry

Uranium and Pb were extracted

U-Pb lower intercept ages with assumed initial 234U

As a baseline with which to compare alternative approaches, Tera-Wasserburg ages are first calculated assuming no disequilibrium effects in either U decay chain ([234U/238U]i = [231Pa/235U]i = [230Th/238U]i = 1), followed by correction for the estimated disequilibrium values (Fig. 2). Calculated ages are presented with 2σ absolute errors. Thirteen data points for sample 73052 yield a well-defined regression with equilibrium lower intercept age of 2.339 ± 0.041 Ma. Applying the assumed

Conclusions

U-Pb dating of late Neogene - Quaternary speleothems is both based on, and limited by, disruption of the U-series decay chains prior to and during cave calcite deposition. If these fractionation effects are not recognised and corrected, the apparent Tera-Wasserburg concordia U-Pb age will under/overestimate systems with intermediate daughter deficiency/excess, most importantly 234U in the 238U decay series. We have considered the effects that 231Pa and 230Th (assumed zero via clay adsorption

Declarations of interest

None.

Acknowledgements

We thank members of the Gruppo Speleologico Lucchese and Leonardo Piccini (Earth Sciences, University of Florence) for assistance in the recovery of sample CC17-f-2 from Antro del Corchia. Cameron Patrick (University of Melbourne Statistical Consulting Centre) provided advice and assistance during development of the Monte Carlo simulation. We also thank two anonymous reviewers for their help and insight. This research was facilitated by Australian Research Council grant FL160100028 to JDW.

References (40)

Cited by (0)

View full text