The Petrogenesis of the S-Type Cape Granite Suite, South Africa: Enlightenment From Xenocrysts and Initial Melt Compositions
Abstract
Panafrican S-type granites from the Cape Granite suite (CGS) intrude the low grade metasediments of the Malmesbury supergroup. They are strongly peraluminous (1 < A/CNK < 2), K-rich granites with total FeO + MgO values ranging between 0.8 and 9 wt%. Strongly peraluminous granites are typically interpreted to be the result of incongruent, granulite facies partial melting of aluminous metasediments, and the Peninsular and Darling plutons contain generations of cordierite and garnet that appear to confirm this. Experimental melts compositions from a variety of sources and PT conditions (750 to 1000 oC and 0.3 to 1.0 GPa) have been compared with 212 CGS analyses. While all the experimental melts are strinkingly similar and correspond to peraluminous leucogranites, only one third of the CGS samples plot in the experimental field. In addition, about half the experimental melts plot in a field that coincides with no natural granite compositions. Typically, experimental glasses are more leucocratic and aluminous than the natural granites. The major element compositions of most of the natural granites can be successfully modelled as mixtures of the experimental glass compositions and up to 15 wt% cordierite and/or garnet, common phases produced in equilibrium with the melts in the incongruent melting experiments. This appears to be a more effective mechanism for accounting for compositional variation within the CGS than any combination of AFC processes, which all require a balancing population of contaminant material of composition unknown either in the CGS magmas or in the surrounding country rocks. Garnet from the CGS is generally Fe-rich (10 < Mg/# < 25) with narrow (20-30 m) rims that are Mn and Fe-enriched. Their major element and REE core to rim zonation patterns are flat, The garnet crystals are very strongly enriched in HREE and display strong negative Eu anomalies. HREE concentrations are several times higher than would be predicted from REE concentrations of the bulk rock and accepted garnet-melt partition coefficients. Collectively, these features suggest that the garnet crystals did not crystalise from the magmas that formed these plutons, but rather, equilibrated under granulite facies metamorphic conditions. Similarly, these plutons both contain two generations of cordierite, one of which forms large and euhedral crystals that are interpreted to have crystallized from the magmas; the other one consists in Mg-richer cordierites that occur as rounded aggregates containing inclusions of hercynitic spinel and relatively Mg-rich biotite. The hercynite-cordierite association is well know from medium- to low-pressure anatectic terrains, and it is proposed that these mineral aggregates are also inherited from the source zone of the magmas. Both garnet and cordierite are rimmed by biotite, and appear to not have been in equilibrium with the magma at the site of intrusion. REE modelling illustrates that the xenocrystic garnets documented in this study are not the source of Fe and Mg enrichment in the magmas, as this would have produced substantially higher HREE bulk compositions than exist. A combination of these garnet crystals and cordierite, or peritectic garnet produced in the melting reactions and thus, of different trace element composition allow the CGS major element and REE compositions to be modelled as mixtures of the experimental melts and this very specific version of restite.
- Publication:
-
AGU Spring Meeting Abstracts
- Pub Date:
- May 2004
- Bibcode:
- 2004AGUSM.V12A..05S
- Keywords:
-
- 3600 MINERALOGY AND PETROLOGY (replaces;
- 3640 Igneous petrology;
- 3655 Major element composition