Extending the amphibole sponge: The metasomatism of clinopyroxene in arcs

Volcanic arcs, like those that form the Pacific Ring of Fire, are markers for the collision and subduction of tectonic plates. These volcanic arcs are typically characterised by water-rich magma, which gives them a distinctive mineralogy and chemistry when compared to volcanoes produced from water-free magma. This water-rich character is what leads to the explosive nature of arc volcanoes, and arguably makes arc volcanoes the most hazardous on Earth. It is also responsible for their ore forming potential, most notably in the formation of copper and gold deposits.

Researchers in arc environments have recognised common chemical signatures in the erupted rocks, and have suggested that the mineral amphibole crystallises from the original magmas and is left in the lower crust as residual crystal mush. Previous geochemical studies suggest it exerts strong chemical controls on the crystallising and evolving magmas. However, despite a suggested widespread role for amphibole in these magmas, it is not particularly common in the volcanic rocks erupted at surface. If it forms so readily from the magmas to produce a mush, why do we not see it in the crystal population of the magma once it has moved away from the mush?

Experimental work on candidate "parent" magmas suggest that it is not amphibole but another mineral, clinopyroxene, that will be expected to dominate the early crystallisation and thus be the main mineral in the crystal mush. Clinopyroxene is also commonly observed in the volcanic arc rocks. There is a mismatch between observations: chemistry suggests amphibole is the important mineral, but the mineralogy of the rocks at surface suggest a greater role for clinopyroxene.

What if amphibole does not form by direct crystallisation from the melt? What if it formed by reactions between the melt and already-formed clinopyroxene? In such a scenario amphibole forms at the melt-mush interface, by altering the clinopyroxene. As melts are periodically released from this melt-mush "reaction zone", the mush (now a mixture of clinopyroxene and amphibole) is left behind. The melt moves to shallower crustal levels, and outside of the pressure stability range for amphibole. Thus, a melt has formed amphibole by reaction, left it behind in the mush, and risen to levels where more amphibole cannot form. Amphibole is not abundant in the crystal content of the melts that reach the surface. The reaction process explains the observation that amphibole is not ubiquitous in volcanic arc rocks, but can it explain the observation that amphibole is a major driver of chemistry?

A suite of samples from Savo volcano, Solomon Islands arc, will allow us to test this process. The surface rocks contain nodules of preserved crystal mush material (which are usually left behind in the crust). Detailed chemical analysis of clinopyroxene and amphibole from Savo will determine the chemical effect the reaction has on the evolving melt. Two hypotheses will be tested: 1) chemical signatures of amphibole crystallisation can instead be developed by clinopyroxene mush-melt reactions, therefore reconciling the chemistry with the minerals observed in the rocks; 2) clinopyroxene mush in the crust acts as a sponge, drawing water and copper out of the evolving melts, and thus acting as a buffer or barrier for their transfer from the mantle to the upper crust and surface.

Rather unusually, the mush nodules at Savo contain two different amphiboles - as well as the amphibole replacing clinopyroxene in the mush nodules (as per the scenario above), high water and sodium contents of melts at Savo helped to stabilise amphibole, and so it forms by direct crystallisation. This direct crystallisation amphibole will be used as a frame of reference to critically assess the two hypotheses - are the chemical effects of the two processes and produced amphiboles identical, therefore allowing the reaction process to reconcile the conflicting observations made in arc rocks?

This project will assist exploration geologists in the broad scale targeting of potential copper mineralisation targets, allowing for more focussed exploration and a better discovery rate, contributing in turn to better economic returns, lower carbon footprints and reduced environmental impacts of exploration programmes.

Grassroots exploration needs good genetic models of mineral deposit formation. This project is identifying processes that will contribute to our understanding of the precursor magmas and the first steps for the formation of a deposit (e.g. Richards 2003). Applying the clinopyroxene sponge model briefly to ore deposits: early arc magmas will be barren, due to tholeiitic chemistries, early sulphide fractionation, and water and copper loss to the juvenile clinopyroxene sponge. Later magmas will be more copper rich, due to the hydration of the sponge allowing more hydrous melts into the upper crust. Post-subduction magmas formed by the remelting of the sponge will be alkaline and copper and gold enriched (Richards 2009). Thus, exploration targets (Cu vs. Au-Cu porphyries) and host rocks (tholeitic vs. calc-alkaline vs. mildly alkaline) and the related alteration signatures are changed as a result of a genetic model. This project will therefore contribute to the wider industrial effort to understand and discover copper and gold mineralisation in arc terrains.

Discoveries of gold and copper deposits in particular have not kept pace with demand, despite high prices (even during the economic downturn). Hall (2010) calculated that the major gold discoveries of the period 1997-2007 contained 377 Moz total anticipated recoverable reserves - less than half that produced from active gold mines in the same period. Grassroots exploration (that is, exploration for mineralisation in frontier countries or areas without previously recognised mineralisation) in particular has declined (Sillitoe 2010). Improving discovery rate, cost per discovery, reducing the environmental impact and carbon footprint of exploration, and encouraging and supporting higher-risk greenfield exploration depends on a number of factors, but good underpinning science plays a crucial role. In summarising the exploration techniques that contributed to the major ore deposit discoveries of the last forty years, Sillitoe and Thompson (2006) stated that "perhaps the greatest geologic influence on exploration practice has resulted from radical reinterpretation or modification of the empirical and/or genetic models for particular ore deposits types".

Hall DJ (2010) Exploration and discovery: Paradigm shift required. SEG Newsletter (82):15-17
Richards JP (2003) Tectono-magmatic precursors for porphyry Cu-(Mo-Au) deposit formation. Econ Geol 98(8):1515-1533
Richards JP (2009) Postsubduction porphyry Cu-Au and epithermal Au deposits: Products of remelting of subduction-modified lithosphere. Geology 37(3):247-250
Sillitoe RH (2010) Grassroots exploration: Between a major rock and a junior hard place. SEG Newsletter (83):11-13
Sillitoe RH and Thompson JFH (2006) Changes in mineral exploration practice: consequences for discovery. SEG Special Publication 12: 193-219