Elsevier

Radiation Measurements

Volume 79, August 2015, Pages 24-32
Radiation Measurements

Evaluating the accuracy of ESR dose determination of pseudo-Early Pleistocene fossil tooth enamel samples using dose recovery tests

https://doi.org/10.1016/j.radmeas.2015.06.004Get rights and content

Highlights

  • Dose recovery tests performed on fossil tooth enamel with DE values >1,000 Gy.

  • Several fitting functions were used.

  • The SSE function does not correctly describe the behavior of the ESR signal.

  • The most accurate DE results are obtained by using a DSE with data weighted by 1/I2.

  • The SSE might nevertheless produce fairly consistent results under certain conditions.

Abstract

In ESR dating of Early Pleistocene fossil tooth enamel samples, the fitting function used for the evaluation of the DE value is undoubtedly among the major sources of uncertainty. Dose recovery tests performed on fossil tooth enamel showing DE values >1,000 Gy demonstrate: (i) that high precision ESR measurements (<0.5%) and high DE reproducibility (<5%) may be achieved; (ii) the appropriateness of the Double Saturating Exponential (DSE) fitting function for ESR dose reconstruction. In contrast, the SSE function, which has been almost exclusively used so far, does simply not correctly describe the behavior of the radiation induced ESR signal of tooth enamel with the dose.

Several fitting functions and data weighting options were tested and the combination of a DSE with data weighted by the inverse of the squared intensities is the procedure providing the most accurate DE results. However, the SSE may nevertheless sometimes produce consistent results if Dmax does not exceed 6*DE. Further work is required in that direction in order to determine more precisely in which conditions the SSE could be used as a fair approximation of the DSE function for these samples.

Introduction

Over the last decade, an extensive work has been dedicated to the evaluation of the exact potential and the current limitations of the ESR dating method specifically applied to Early Pleistocene (>0.78 Ma) fossil tooth enamel samples, with the idea to reach a better understanding of the method and identify the main sources of uncertainty involved in the dating process (e.g. Duval et al., 2009, Duval et al., 2011a Duval et al., 2011b, Duval et al., 2012a Duval et al., 2012b, Duval et al., 2013, Duval et al., 2015). The long term objective of this ongoing investigation is to improve the reliability and accuracy of this dating method, which is a major challenge in archaeology given that ESR is one of the very few numerical methods that can be used to date the earliest hominin occupations in non-volcanic context in Europe (e.g. Duval et al., 2012), Africa (e.g. Curnoe et al., 2001, Schwarcz et al., 1994) and Asia (e.g. Shao et al., 2014, Han et al., 2012, Tiemei et al., 2001).

A synthesis based on 20 Early Pleistocene teeth from various Spanish localities recently published by Duval et al. (2012) showed that these “old” samples are usually mainly characterized by: high 230Th/234U activity ratios that may sometimes exceed secular equilibrium, high Uranium concentrations (several tens of ppm or more in dentine) and high equivalent dose (DE) values (in most cases >1000 Gy). In particular, the magnitude of these DE values raised many new questions about the reliability of the ESR dose reconstruction procedure that has been routinely used so far, in particular regarding the selection of an appropriate fitting function (e.g. Duval et al., 2009).

Since the first ESR dating applications to fossil tooth enamel in the mid 1980s' (e.g. Grün et al., 1985), the use of a linear fitting function has been rapidly abandoned (Grün and MacDonald, 1989, Grün, 1996) and replaced by a single saturating exponential function (SSE) in order to take into consideration the saturation of the ESR signal that occur at high irradiation doses. From a physical perspective, this saturation may be explained by the presence of a finite number of precursors available in a given sample and thus a decreasing probability to create radicals with the increasing irradiation dose (Apers et al., 1981). The use of this SSE was also a way to indirectly assume that the radiation induced signal was dominated by a single component, which was found later to be incorrect (e.g. Vanhaelewyn et al., 2000). So far, the SSE function has been exclusively used for ESR dose reconstruction of fossil tooth enamel during several decades, but some limitations have been progressively observed, in particular in the case of samples with high equivalent dose (DE) values (Chen, 1997). In that regard, Duval et al. (2009) highlighted some clear issues with the use of the SSE in the specific case of Early Pleistocene teeth. For example, these authors showed the strong impact of the maximum irradiation dose (Dmax) on the DE value, which was constantly increasing with the successive addition of ESR points at high doses. This systematic trend demonstrated the impossibility for this function to describe the behavior of the ESR signal at high irradiation doses. To address this issue, they proposed the use of a Double Saturating Exponential function (DSE) instead, which was found not only to be more appropriate for fitting the experimental data, but also more in agreement with the recent advances on the understanding of the ESR signal of fossil tooth enamel and the identification of several types of radiation-induced CO2 radicals contributing to the main signal (Joannes-Boyau and Grün, 2011 and references therein). More recently, Duval et al. (2013) completed the study by: (i) proposing the combination of two types of DSE functions in order to fully encompass the uncertainty associated to the variability of the relative proportions of the various types of CO2 among the samples, and (ii) defining some empirical criteria to ensure a reliable fitting with the DSE.

In ESR dose reconstruction of fossil tooth enamel, the equivalent dose (DE) is systematically obtained via the additive dose method, i.e. through a back extrapolation of the fitting function to the X-axis. Consequently, and in contrast with the regenerative dose method, the DE value is thus very dependent on the choice of the fitting function. For example, Duval et al. (2009) showed that the relative deviation between the DE values obtained from different functions is significantly increasing along with the magnitude of the DE value. In the case of samples showing DE values >1000 Gy, the systematic deviation between the DE results derived either from a SSE or a DSE function was estimated to be around 23% in average, while it was <10% for DE values <300 Gy. In other words, the further goes the back extrapolation (= the larger is the DE), the stronger is the influence of the fitting function on the final DE value. Consequently, if the fitting function must be considered as a significant source of uncertainty in ESR dating of fossil tooth enamel in general, its importance is even greater in the specific case of Early Pleistocene samples.

However, despite a series of evidence showing the appropriateness of the DSE and the major limitations of the SSE (e.g. Duval et al., 2009, Duval et al., 2013), there is still some uncertainty about the ability of the DSE function to yield correct DE values. Actually, the only way to check whether the DSE does not systematically provide overestimated or underestimated results would be to work with known-dose samples. This may be done via “dose recovery” tests that are usually designed to assess the appropriateness of a standard analytical procedure by evaluating whether a laboratory given dose may be experimentally recovered with accuracy. This test is easily performed when the signal may be zeroed without modifying the structure and chemical composition of the sample, like in ESR and OSL dating of optically bleached quartz and feldspar grains (e.g. Asagoe et al., 2011, Murray and Wintle, 2003). However, this is not the case for tooth enamel, since the ESR signal cannot be fully reset by sunlight or by heat. UV irradiations actually create a radiation-induced signal (e.g. Nilsson et al., 2001), while thermal annealing at increasing temperature induces a decrease of the ESR intensity but is accompanied with some irreversible major changes in the crystalline structure of tooth enamel (e.g. Brik et al., 1996, and 1997), making thus impossible the use of these procedures for dose recovery tests. Consequently, the best strategy for working with known-dose samples consists in selecting tooth samples that have accumulated a very small dose, so that it could be considered as negligible in comparison with a geological dose accumulated over a hundreds of ka. In that regard, recent fossil teeth should be preferred in first instance to modern enamel samples given the existing differences in structure and composition (e.g. Kohn et al., 1999) that may induce specific behaviors with the irradiation dose, even though it is worth mentioning that the very few comparison studies published on this aspect did not lead to identify significant differences in terms of radiation sensitivities (Rink and Schwarcz, 1994, Rink and Schwarcz, 1995).

Consequently, three recent fossil teeth were selected for the present study and gamma irradiated to ∼1500 Gy, in order to simulate a DE value that may be commonly found in Early Pleistocene tooth samples (e.g. Duval et al., 2013). Samples were then analyzed following a Multiple Aliquot Additive (MAA) dose approach, the standard procedure in ESR dating of fossil teeth (e.g. Duval et al., 2011b, Duval et al., 2012b) and several fitting functions were finally used to derive DE values in order to see whether the laboratory given dose could be accurately recovered. The deviations observed between the calculated DE value and the expected DE are presented and discussed in this work.

Section snippets

Material and methods

Three teeth (MOD1301, MOD1302 and MOD1303) were collected from a protohistoric site in France and prepared following a standard ESR dating procedure. The enamel layer of the vestibular side was extracted and then cleaned using a dentist drill. Clean enamel fragments were then ground and sieved to obtain a powder of 100–200 μm.

The first step of the analytical procedure was to make sure that the natural dose registered over time by the samples would not significantly interfere with the laboratory

ESR measurement repeatability

Each sample was measured three times on different days in order to evaluate the repeatability of the ESR measurements. From one measurement to the other, the ESR intensities of each aliquot showed a very small variation of 0.5%, 0.4% and 0.3% on average (1 standard deviation) for samples MOD1301, MOD1302 and MOD1303, respectively. For a given aliquot, variations were systematically <0.9%, except for one aliquot (1.7%), demonstrating that ESR measurements were highly repeatable.

The DE value was

Conclusion

The results obtained for the present study confirm those previously shown in Duval et al., 2009, Duval et al., 2013 and definitely demonstrate the appropriateness of the DSE fitting function for ESR dating/dosimetry of tooth samples with DE values >1,000 Gy. Consequently, they also suggest that the behavior of the radiation induced ESR signal with the dose is best described by the sum of two exponential components showing distinct saturation levels. If this general behavior that has been so far

Acknowledgments

I would like to thank M. Lebon (MNHN, France) for providing the tooth samples and V. Guilarte (CENIEH, Spain) for her technical support in the early stage of the analytical procedure. The research leading to these results has received funding from the People Programme (Marie Curie Actions) of the European Union's Seventh Framework Programme (FP7/2007-2013) under REA Grant Agreement n° PIOF-GA-2013-626474. I thank the two anonymous reviewers for constructive comments.

References (35)

Cited by (12)

  • U-series and ESR/U-series dating of the Stegodon–Ailuropoda fauna at Black Cave, Guangxi, southern China with implications for the timing of the extinction of Gigantopithecus blacki

    2017, Quaternary International
    Citation Excerpt :

    The SSE function yielded De results in agreement with those by E + L function for BC-1, BC-2 and BC-3 in error range, but significantly higher for BC-4 and 5 (>2σ). The DSE also yielded De-values completely consistent with the E + L ones but with larger error (see Table S1), which is probably due to the small number of data points and the limited maximum given dose (Duval et al., 2009; Duval, 2015). The fossil teeth from Black Cave have averagely 2.7 ppm U in enamel and 129 ppm in dentine (Table 2).

View all citing articles on Scopus
View full text