|dc.description.abstract||Speleothems are important archives of terrestrial paleoclimate due to their wide geographical coverage and the ability to date their growth layers using U-Th and U-Pb radiometric dating. Extensive research into various speleothem-based geochemical proxies has been undertaken in the last half century, and a significant number of proxies are now used routinely for reconstructions of paleoclimate including stable oxygen and carbon isotopes (known to respond to variations in rainfall and temperature), fluorescence (indicative of the quality and quantity of organic matter present in cave dripwaters) and speleothem minerology (e.g. fabric defects and nano-crystal aggregation). Trace element proxies in speleothems have also garnered attention, but the suite of trace elements used to date has primarily been limited to the alkali earth metals as they exhibit simple partitioning and are characterised by partition coefficients (Kd values) less than 1. Their role in karst systems is thus comparatively well constrained.
In contrast, the first-row transition metals have been the focus of only a few speleothem-based studies, despite recent evidence suggesting some of these metals (namely Co, Ni and Cu) may be useful additions to the current suite of trace element proxies already utilised. Thus, although theoretical distribution coefficient data are available for these metals, their partitioning behaviour into speleothem calcite is not well understood. Theoretical distribution coefficients are useful in understanding partitioning behaviour under certain conditions, but the non-thermodynamic nature of partition coefficients during calcite precipitation in most natural contexts requires partitioning behaviour under speleothem-specific conditions to be established. For instance, despite theoretical distribution coefficients being greater than 1 for Co, Ni and Cu, apparent partition coefficients (Kd app) calculated for these metals in actual speleothem samples are frequently less than 1 due to the dependence of Kd app on a number of factors including calcite precipitation rate, crystal morphology, and complexation reactions between natural organic matter (NOM) and trace ions. Indeed, complexation of first-row transition metals in speleothem-forming dripwaters has been shown to significantly alter the availability of these metals for incorporation into speleothem calcite. This study therefore aimed to investigate the partitioning behaviour of Co, Ni and Cu into calcite precipitated under karst-analogue conditions and in the presence of organic ligands to help establish these metals as viable speleothem-based paleoclimate proxies.
In order to do this, a suitable method for growing calcite under karst-analogue conditions was first established. All experiments were undertaken inside a purpose-built chamber that ensured precise control of temperature, humidity and pCO2, and calcite was grown in a manner analogous to that observed for natural speleothems, with calcite precipitation occurring as a result of CO2 degassing from a thin solution film. Observed growth rates and structural characterisation of the precipitated CaCO3 indicated the method developed was suitable for the desired purpose, and the partitioning data obtained subsequently was considered to be applicable to ‘real-world’ speleothem and cave dripwater samples.
The partitioning behaviour of Co, Ni and Cu into calcite grown under speleothem-like conditions was assessed, and speleothem-specific Kd values of ~ 4, 1 and 44, respectively, were determined. The relatively high inorganic Kd values determined for Co and Cu illustrate the reason why these divalent metals have frequently been overlooked as paleoclimate proxies in speleothems: metals with high partition coefficients are thought to exhibit complex partitioning behaviour and are significantly affected by PCP. However, further experiments carried out with organic ligands present in solution illustrated that the complexation of Co, Ni and Cu by nitrilotriacetic acid (NTA) and Suwanee River fulvic acid (SRFA) significantly altered their partitioning behaviour, with apparent partition coefficients reduced below 1 for all three metals. It was shown that partitioning of Co, Ni and Cu into calcite was controlled not only by their affinity for the calcite lattice (i.e. inorganic Kd values) but also by the dissociation of the metal-ligand complexes, and that the amount of ‘free’ metal available for incorporation into calcite was dependent on residence time and the dissociation constants of the complexes.
The results found in this study highlight the fact that the incorporation of the divalent first-row transition metals into speleothem calcite cannot be considered in terms of simple inorganic partitioning, as the widespread presence of NOM in karst systems will alter their partitioning behaviour significantly. Thus, although the use of these metals as speleothem-based proxies has previously been overlooked in favour of elements such as the alkali earth metals which exhibit simple partitioning (i.e. Kd << 1), in reality the NOM present in karst systems will reduce the effects of PCP on these divalent first-row metals, and they should in fact be considered as useful additions to the current range of paleo-proxies used.||