Two radicals threaten the fundamentals of ESR dating

In the last 25 years, ESR dating of tooth enamel has been reasonably well established in Quaternary sciences. The ESR signal of the CO2-radical in hydroxyapatite has been used for the assessment of the past radiation dose, which in turn is converted into numerical age estimates, once the dose rate parameters are assessed. ESR dating studies are conventionally carried out on powdered samples, and it has repeatedly been demonstrated that the ESR spectra recorded from fossil samples are qualitatively similar to those generated by laboratory irradiation. However, when attempting non-destructive ESR analysis, which is essential when working on fossil human remains, measurements are carried out repeatedly on tooth enamel fragments. Because of the anisotropic nature of hydroxyapatite, the ESR spectra show strong angular dependencies. In contrast to powders, the ESR spectra of fossil samples are significantly different to those generated by laboratory irradiation. These differences are caused by the occurrence of two different species of the CO2−-radical. One is embedded into the crystal structure of the hydroxyapatite, resulting in the expected anisotropic spectra, the second one is located on crystal surfaces, exhibiting a powder spectrum at all angles. While fossil samples mainly contain the former, laboratory irradiation generates a mix of both radicals. Preliminary kinetic studies imply that the surface radical is significantly less stable than the embedded one. This would lead to the prediction that all ESR dose values (including those carried out on powders) represent underestimations. Their magnitude would depend on the age of the sample as well as the relative amount of surface radicals generated in the irradiation process, which are in the range of 50%. However, life is never that easy. Our study also implies that the surface radicals can be transferred into the more stable, embedded species. The implications for the reliability of ESR dating of tooth enamel are at the moment difficult to assess. However, one should keep in mind that thorough inter-comparisons of ESR dating with other, independent methods, rarely found discrepancies larger than about 10%