Tourmaline geochemistry and cassiterite geochronology of highly evolved tin granites and their hydrothermal systems in eastern Australia
Abstract
Three models have been proposed for cassiterite (SnO2)
mineralisation in magmatic–hydrothermal environments: (1)
magmatic crystallisation from a granitic melt, (2) late-stage
magmatic partition of Sn into a fluid or vapour phase and
subsequent cassiterite deposition, and (3) hydrothermal leaching
of Sn from granite and/or country rocks and subsequent
deposition. The complex chemistry of the ‘tin’ granites, and
the large and pervasive hydrothermal systems which can overprint
and destroy primary features make understanding the processes
responsible for Sn enrichment difficult.
Two new analytical methods were developed. Firstly, a method for
the determination of Rb–Sr and Sm–Nd isotopic compositions of
magmatic and hydrothermal tourmalines, which can record the
compositional evolution of magmas and their hydrothermal fluids.
Secondly, cassiterite U–Pb geochronology to constrain the
absolute age and duration of magmatic–hydrothermal Sn systems.
These data, together with major and trace element compositions of
tourmaline, whole-rock geochemistry, quartz δ18O values and
zircon U–Pb geochronology are applied to two Sn deposits
associated with the Ardlethan and Mole granites of eastern
Australia.
The geochemical and isotope data of tourmaline show large
compositional changes across the magmatic–hydrothermal
transition. In the Ardlethan Granite, tourmaline 87Rb–86Sr
isotope compositions, which provide robust estimates of
87Sr/86Sr(i) because of their low 87Rb/86Sr, are used to model
the assimilation and fractional crystallisation processes that
lead to a 30-times enrichment of Sn in residual melts relative to
the source rocks. However, caution must be taken with
interpreting 87Sr/86Sr(i) tourmaline data as high 87Rb/86Sr of
parental melts and fluids can lead to significant in-situ decay
of 87Rb prior to tourmaline precipitation. This phenomon is
hypothesised for the parental melts of the Mole Granite which due
toextreme fractional crystallisation have extreme 87Rb/86Sr of
~900. Subsequently 87Sr/86Sr(i) tourmaline compositions are more
evolved the the whole rock composition.
The Sn concentration of tourmaline increases from magmatic to
hydrothermal settings within the Ardlethan and Mole granites,
recording the exsolution of a fluid from a silicate melt. The
enrichment of Sn during fluid fractionation, recorded by
tourmaline, agrees with experimentally determined melt–fluid
partitioning coefficients. Fluid fractionation is the dominant
enrichment process for greisen deposits of the Ardlethan Granite,
and all deposits of the Mole Granite.
Fluid leaching of host rocks is evidenced by convergence of
Fe/(Fe+Mg), Sr, 87Sr/86Sr(i) and εNd(i) in hydrothermal
tourmaline from the original source rock composition to the host
rock composition. At Ardlethan, the host rock of mineralised
breccia pipes is enriched in Sn (~50 ppm) and fluid leaching
results in an increase of Sn in the mineralising fluids. Although
fluid leaching occurs around the Mole Granite, the low Sn
concentrations in the host rocks limits Sn enrichment.
Melt/fluid-mineral partitioning is a major uncertainty in the
interpretation of tourmaline trace element geochemistry. Natural
studies performed here provide some constraints, however, more
targeted experimental work is required.
A new method for U–Pb characterisation of cassiterite by
ID-TIMS has provided a matrix-matched reference material for
in-situ techniques. However, common-Pb corrections of in-situ
techniques remain a large uncertainty in cassiterite
geochronology. At Ardlethan, the common-Pb compositions are
appropriately estimated by terrestrial Earth models and are more
precise than isochron ages. Conversely, the common-Pb associated
with the Mole Granite appears variable between a terrestrial
Earth composition and a highly evolved composition.
Cassiterite U–Pb geochronology of both the Ardlethan and Mole
granite mineralisation systems indicate precipitation synchronous
with emplacement. The method does not have sufficient precision
(~4 % absolute) to distinguish the age of cassiterite
precipitation from that of zircon, however, the
magmatic–hydrothermal systems of the Ardlethan and Mole
granites persisted for a maximum of 4.2 Ma.