Trace element incorporation in silicate melts and glasses at high pressure
de Grouchy, Charlotte J. L.
Trace elements are highly fractionated during large-scale melting associated with planetary differentiation events. The resulting partition coefficients are used to constrain a range of geological processes and are known to be influenced by pressure, temperature, and compositional changes in crystalline structures. Although recent studies have shown that melt compositional changes affect the partitioning of trace elements, the degree to which these ratios are influenced by alterations in the melt structure, especially with increasing pressure, is poorly constrained due to the difficulty of collecting structural information on bonding environments in situ. A basic understanding of how these elements are incorporated in silicate melts is critical to interpreting early planetary differentiation and crust forming events. This thesis presents results from both x-ray diffraction and absorption techniques on trace element (Y, Zr, Lu and Nd) incorporation in silicate melt structures. The structure of two rare Earth element doped model end member silicate liquids, a highly polymerised haplogranite (Si- Al-Na-K-O) and a less polymerised anorthite-diopside (Si-Al-Mg-Ca-O), have been studied. The results are the first to identify trace rare Earth element (REE) incorporation in silicate melts at high pressure using x-ray diffraction techniques. The local melt structure around Y and Zr in a highly polymerised haplogranite has been studied using x-ray absorption spectroscopy up to 8GPa and 1650 K. Both elements appear to adopt 8-fold coordination within the melt structure with no variation over the pressure range studied. This was also found for the Lu bonding environment in the same composition where the coordination number of Lu-O was found to be 8, with a bond distance rLu-O = 2:36A in the haplogranite melt. At low pressures, < 5GPa, the bonding environment of Lu-O was found to be dependent on composition with coordination decreasing to CNLu-O = 6 and rLu-O = 2:29A in the anorthite-diopside melt. This compositional variance in coordination number at low pressure is consistent with observations made for Y-O in glasses at ambient conditions and is coincident with a dramatic increase in the partition coefficients previously observed for rare Earth elements (REE) with increasing melt polymerisation. However, an abrupt change in both Lu-O coordination and bond distance is observed at 5GPa in the anorthite-diopside melt, with CNLu-O increasing from 6 to 8-fold and rLu-O from 2.29 to 2.39A. This occurs over a similar pressure range where a reduction in the reported heavy REE partition coefficients is observed. X-ray diffraction experiments up to 60GPa and 2000K have also been performed on the incorporation of the larger light REE, Nd, in basaltic-like melts. The results presented show that incorporation within the anorthite-diopside composition is dependent on the size of the REE. Nd-O initially shows the same 6-fold coordination as Lu-O at ambient conditions, although the change to 8-fold coordination appears to occur at considerably lower pressure between 1-2GPa. Coordination change in both cases can be attributed to collapse of the silicate network and an increase in the average number of available 'crystal like' sites in the liquid, with ionic radius of the REE controlling at which pressure the preference for these sites in the melt occurs. Published mineral-melt partition coefficients for Nd, with major mineral phases such as garnet, show very little variation with pressure, in contrast to Lu. The difference in structural incorporation of Lu and Nd in the melts presented in this thesis could explain this partitioning behaviour. Overall this thesis highlights that important structural changes of the trace element bonding environment in silicate melts occur with both compositional variation and pressure. Melt structural changes with pressure cannot be neglected in predictive models of trace element behaviour, and using a single melt term to normalise the effects of melt on trace element partitioning will not accurately predict partitioning behaviour at depth during magma formation or differentiation.