Obsidian Diffusion Dating by Secondary Ion Mass Spectrometry: A Test using Results from Mound 65, Chalco, Mexico

Secondary ion mass spectrometry (SIMS) was used to measure hydrogen and other elemental concentrations as a function of depth in ten obsidian artifacts (Pachuca Source), each with a well-constrained 14C date, from Mound 65, Chalco, Mexico. Hydrogen depth profiles for the different artifacts all display a characteristic S-shape, and increasing maximum hydrogen content in each profile and profile depths are well correlated with time. These data are used to investigate the potential use of Obsidian Diffusion Dating by SIMS (ODDSIMS) for both extrinsic and intrinsic dating of obsidian artifacts. Using "characteristic points" on the hydrogen profile (half-fall depth, inflection point depth), simple hydration rate equations were evaluated against time constraints provided by associated 14C dates. We demonstrate that neither the traditional OHD equation for depth (x) as a function of the square root of time (t1/2) nor a linear function (t1) fit the data. Solving the more generalized tn function provides an excellent fit between characteristic point depths and 14C dates (for n≈0·75), and meets the constraint that at time equal to zero, the depth of the hydration profile must also be zero. However, this may be an average coefficient over the range of ages available, and may not accurately reflect rates at shorter or longer times. Using only two obsidian samples and their associated 14C dates, a calibration curve can be derived that provides ODDSIMS dates for the other pieces that are in excellent agreement with associated14 C dates, indicating that empirical application of the technique is potentially feasible, at least at individual sites. The underlying processes governing hydrogen transport into the obsidian were also investigated by using finite difference modelling to reproduce the shape of the hydrogen depth profile. Excellent fits were obtained by assuming concentration-dependent diffusion, and dates that agree well with associated 14C dates can also be extracted from the finite difference profiles. Although considerable additional work needs to be done, the success of the finite difference modelling suggests that development of an independent, intrinsic ODDSIMS model may be possible.