<p>Sulphur and molybdenum trace impurities in speleothems (stalagmites
and stalactites) can provide long and continuous records of volcanic activity,
which are important for past climatic and environmental reconstructions. However,
the chemistry governing the incorporation of the trace-element bearing species
into the calcium carbonate phases forming speleothems is not well understood.
Our previous work has shown that substitution as tetrahedral oxyanions [<i>X</i>O<sub>4</sub>]<sup>2-</sup> (<i>X</i>=S,
Mo) replacing [CO<sub>3</sub>]<sup>2-</sup> in CaCO<sub>3</sub> bulk
phases (except perhaps for vaterite)
is thermodynamically unfavourable with respect to the formation of competing
phases, due to the larger size and different shape of the [<i>X</i>O<sub>4</sub>]<sup>2-
</sup>tetrahedral anions in comparison with the flat [CO<sub>3</sub>]<sup>2-</sup>
anions, which implied that most of the incorporation would happen at the
surface rather than the bulk of the mineral. Here we present an ab initio
molecular dynamics study exploring the incorporation of these impurities at the
mineral-water interface. We show that the oxyanions substitution at the aqueous
calcite (10.4) surface is clearly favoured over bulk incorporation, due to the
lower structural strain on the calcium carbonate solid. Incorporation at
surface step sites is even more favourable for both oxyanions, thanks to the additional
interface space afforded by the surface line defect to accommodate the
tetrahedral anion. Differences between sulphate and molybdate substitution can
be mostly explained by the size of the anions. The molybdate oxyanion is more
difficult to incorporate in the calcite bulk than the smaller sulphate oxyanion.
However, when molybdate is substituted at the surface, the elastic cost is
avoided because the oxyanion protrudes out of the surface and gains stability
via the interaction with water at the interface, which in balance results in
more favourable surface substitution for molybdate than for sulphate. The detailed
molecular-level insights provided by our calculations will be useful to understand
the chemical basis of S- and Mo-based speleothem records.</p>