Jump to Search Jump to Navigation Jump to Content

Detrital Zircon Chemistry

Sep 02, 2016

Robert R. Loucks1, Marco L. Fiorentini1, and Yong-Jun Lu2

1Centre for Exploration Targeting, The University of Western Australia
2 Geological Survey of Western Australia, 100 Plain Street, East Perth, WA 6004, Australia


We introduce new indicators of magmatic fertility for copper metallogeny at convergent plate margins using trace elements in zircon.  Based on theoretical considerations, the ratios (EuN/Eu*)/YbN and (CeN/Ce*)/Y, or equivalently (CeN/NdN)/Y, (wherein the subscript N denotes normalisation to a reference material) are formulated as proxies for temperature and hydration state of the silicate melt as zircon crystallised. Unusually high dissolved H2O in residual felsic melts develops during unusually deep magma chamber evolution; high dissolved H2O lowers the crystallisation temperatures of silicate melts and endows granitoid magmas with copper metallogenic fertility by order-of-magnitude increase in the efficiency of Cu scavenging from the silicate melt by exsolving hydrothermal fluid. New analyses and a compilation of published analyses of the trace-element compositions of zircons from 23 copper-ore-bearing igneous complexes are compared with new analyses and a compilation of published zircon trace-element analyses from many unmineralised igneous complexes in convergent-plate-margin, divergent-plate-margin, and intra-plate tectonic settings.  The comparisons show that the parameters (EuN/Eu*)/YbN and (CeN/Ce*)/Y and (CeN/NdN)/Y are effective discriminants of copper-fertile igneous complexes.

Besides high content of dissolved H2O, the other principal determinant of copper metallogenic fertility of arc magmas is high content of dissolved sulfur in the melt.  The sulfur-carrying capacity of oxidised granitoid magmas is limited by igneous anhydrite saturation and is about 10 times greater than the sulfur-carrying capacity of reduced granitoid magmas, which is limited by igneous pyrrhotite saturation.  We introduce a new indicator of magmatic sulfur-carrying capacity using trace elements in zircon.  It is applicable to detrital zircons to further characterise the copper metallogenic fertility of the zircon’s source igneous complex.


In exploration for magmatic-hydrothermal ore deposits of copper, gold, silver, molybdenum, tin, tungsten and others, the search for reliable means of geochemically distinguishing metallogenically fertile igneous complexes from infertile ones has a long and chequered history.  Loucks (2012a, 2012b, and 2014) introduced and discussed some of the more successful whole-rock igneous geochemical guides for copper and gold exploration.  Minerals that are capable of analytically significant chemical variation inherit variations in composition of their parent magmas via crystal/melt partition coefficients, which can vary complexly with magma temperature, oxidation state, and major-element composition. Due to pitfalls in deciphering the complexity of trace-element compositions of accessory igneous minerals, the use of chemical variations in detrital apatite, tourmaline, rutile, and zircon as guides to ore has a comparably chequered history.

For the past two decades, microbeam sampling of mineral grains or glass or fluid inclusions by far-ultraviolet excimer-laser ablation of spots ~10–100 microns diameter and analysis of the ablated material by fast-scanning quadrupole mass spectrometry has afforded the capability of doing, at relatively low cost, 300-500 quantitative multi-element chemical analyses per day of up to ~25 elements per ablation spot at concentration levels spanning a range from a few parts per billion to percent levels (Loucks et al, 1996).  However, this instrumental capability has not found widespread application in mineral exploration.  We show herein that zircons from copper (±gold±molybdenum)-ore-productive granitoid intrusions at convergent plate margins have proportions of lithophile trace elements that are usefully distinct from zircons in ordinary, Cu-infertile igneous suites at plate-margin and intra-plate tectonic settings.  Because detrital zircons survive transport distances longer than water and soil geochemical dispersion anomalies, and zircons can be separated from sediment samples quickly and inexpensively, the distinctive chemical features of zircons from Cu-fertile granitoid intrusions and coeval volcanics can be exploited cost effectively to identify watersheds containing Cu-prospective igneous complexes that were eroding at the time of sedimentation, but may be covered now.  Building upon our success in identifying Cu-fertile geochemical signatures in zircon, described herein, we hope to extend the research program to identification of Au, Mo-, Sn- and/or W-fertility geochemical signatures in zircon and other minerals by sampling ore-productive intrusions in many mining districts.  

Distinctive REE patterns of Cu-infertile and ore-productive arc magmas and their zircons

Loucks (2012a, 2014) summarised some results of his research over the prior 15 years on characteristics and petrogenesis of copper-ore-forming arc magmas which demonstrates that igneous intrusions that generate magmatic-hydrothermal copper-rich ore deposits at convergent plate margins have distinctive proportions of lithophile trace elements that usefully distinguish them from ordinary tholeiitic and calc-alkalic arc igneous suites in well explored continent-margin and island arc segments that are apparently devoid of economically significant copper mineralisation.  As illustrated in the seminal paper by Lang and Titley (1998) on Laramide igneous complexes in Arizona and in Figure 4 of the March 2012 issue of the CET Quarterly News on a global dataset (Loucks, 2012a), samples of ore-stage intrusions that retain fresh or little-altered igneous plagioclase have chondrite- or NMORB-normalised rare-earth-element (REE) patterns that are hook- or “Nike swoosh”-shaped, with no substantial Eu anomaly (Eu/Eu* ≈ 1 ± 0.2) and with a minimum around Ho, Er, or Tm. Figure 1 illustrates the contrast in REE patterns of a typical Cu-infertile arc magma and little-altered samples of Cu-ore-productive intrusions from two major high-sulfidation and porphyry ore deposits, and compares the shapes of REE patterns in their contained zircons. 

Fig. 1. Distinctive features of typical whole-rock and zircon REE patterns of copper-infertile (A, B) and copper-ore-productive (C, D, E, F) felsic igneous suites at convergent plate margins.  (A) Whole-rock REE abundances, normalised to abundances in average N-type mid-ocean-ridge basalt (Sun & McDonough, 1989), of a caldera-forming dacitic ignimbrite erupted at 34.3 ± 0.1 Ma and overlying rhyolitic ignimbrite (>32 Ma) in the Tilzapotla caldera in the Sierra Madre del Sur, Guerrero, Mexico (Mori et al., 2012) illustrate features characteristic of Cu-infertile felsic arc magmas. The Eu anomaly in whole-rocks and in zircons is the ratio of analytical Eu to Eu*, wherein Eu* = √(SmN X GdN) is the expected value of Eu if it were entirely trivalent like adjacent REEs.  These features produce a low value of the REE pattern shape factor (EuN/Eu*)/YbN. (B) REE abundances in representative zircon grains from the caldera-forming Tilzapotla ignimbrite (Martiny et al., 2013) normalised to abundances in average CI chondrites as reported by McDonough and Sun (1995).  As are typical of zircons from Cu-unmineralised igneous complexes, these zircon REE patterns feature large negative Eu anomalies and relatively small positive Ce anomalies. As indicated by the dashed line, Ce* is the expected abundance of Ce if it were all Ce3+, like the adjacent, entirely trivalent REEs; Ce* = √(LaN X PrN).  (C) Steep, hook-shaped REE patterns with a minimum around Er and absence of a significant negative Eu anomaly are typical features of felsic arc magmas parental to major magmatic-hydrothermal copper ore deposits, such as these examples from the giant Tampakan porphyry copper-gold deposit (4.3 Ma) and superposed high-sulfidation epithermal Cu-Au deposit (3.3 Ma).  These features produce a relatively large value of the pattern shape-factor (EuN/Eu*)/YbN.  (D) Zircons from porphyry-copper ore-stage magma and high-sulfidation ore-stage magma at Tampakan illustrate the features typical of zircons from copper-ore-forming arc intrusions globally, showing relatively large values of the cerium anomaly, Ce/Ce*, and small negative europium anomalies.  (E) Whole-rock REE patterns of tonalitic intrusions that generated the Batu Hijau porphyry Cu-Au deposit, Sumbawa (Fiorentini and Garwin, 2010) share the distinguishing features in panel C. (F) Zircons from Batu Hijau demonstrate that zircon crystallising from Eu-undepleted silicate melts inherits that feature of the melt.  NMORB and chondrite normalisations yield REE patterns of nearly identical shape, but chondrite normalisation increases the numbers along the ordinate axis by about 10-fold.

The late Eocene-early Oligocene Tilzapotla igneous complex in the Sierra Madre del Sur, Guerrero, Mexico, illustrates the REE patterns (from Mori et al., 2012) typical of Cu-infertile felsic arc igneous complexes worldwide (Fig. 1A).  The Tilzapotla igneous suite represents low-pressure magmatic differentiation from a parental arc-basaltic magma to andesitic, dacitic, and rhyolitic residual melt compositions (Mori et al, 2012), with little crustal assimilation, during crustal extension between major strike-slip shear zones that localise a pair of long belts of magmatic centres (Martini et al., 2009).  No economically significant magmatic-hydrothermal copper mineralisation has been reported in the WNW-ESE-trending, 360-km-long string of 39-32 Ma igneous complexes that includes the Tilzapotla diorites, granodiorites, and related volcanics, although small porphyry copper ore deposits developed at 36-32 Ma in Michoacán state several hundred kilometres northwest of Tilzapotla (Potra and Macfarlane, 2014).  The strong negative Eu anomaly and flattish pattern in the Dy-Lu interval yield low values of the pattern shape-factor (EuN/Eu*)/YbN. The strong Eu depletion and weak depletion of middle and heavy REEs are inherited by zircon crystallising from such melts, as shown in Figure 1B. 

In contrast to the whole-rock REE patterns in Figure 1A, Figures 1C and 1E illustrate the distinctive features of whole-rock REE patterns typical of Cu-ore-forming arc magmas.  The most conspicuous feature is lack of a significant negative Eu anomaly.  In addition, REE patterns of Cu-ore-productive magmas generally have steeper, spoon-profile or hook-shaped patterns with a minimum in the Ho-Tm interval and a rise toward Lu.  To numerically characterise these features, Loucks (2000) formulated the pattern shape-factor (EuN/Eu*)/YbN, which has much higher values in panels C and E than in A.

Fig. 2.  Arc igneous mafic-to-felsic differentiation series parental to major porphyry-type and high-sulfidation-epithermal-type Cu(-Au) deposits are compared with barren reference suites from unmineralised time and space intervals at well explored convergent plate margins.  (A.)  Circles represent the mafic-to-felsic differentiation series in compressive (solid orange dots) and non-compressive (unfilled orange circles) segments of the Sangihe intra-oceanic arc from southern Mindanao through northern Sulawesi. The syn-orogenic mafic-to-felsic magmatic differentiation series in the Tampakan complex follows a trend of rising (EuN/Eu*)/YbN, whereas farther south in non-compressive parts of the Sangihe arc, the differentiation series from the same type of tholeiitic basaltic parent magma follows a trend of declining (EuN/Eu*)/YbN as wt % SiO2 increases.  Pink triangles represent the Pliocene-Pleistocene mafic-to-felsic differentiation series in a strongly compressive setting on Sumbawa island (solid pink triangles; data from Fiorentini and Garwin, 2010, and references therein) as compared with the differentiation series from the same type of tholeiitic basaltic parent magma in weakly to non-compressive intervals above the same subduction zone (unfilled pink triangles) flanking it to the west and east along the Sunda-Banda arc. (B.)  Continent-margin arc igneous complexes that host major magmatic-hydrothermal copper ore deposits developed in settings that experienced orogenic deformation in the lead-up to the ore-forming magmatism (Sillitoe, 1998; Loucks & Ballard, 2002), and all follow trends of rising (EuN/Eu*)/YbN with increasing wt % SiO2 up to ~70–72 wt%. In contrast, unmineralised reference suites in the 23-0 Ma age range in the Central Volcanic Zone and in the Southern Volcanic Zone of the Andes and in the 10-0 Ma age range in the North Honshu arc developed in relatively non-compressive settings and follow mafic-to-felsic magmatic differentiation trends of declining (EuN/Eu*)/YbN with increasing SiO2.  Fresh or little-altered (weak propylitic or less) samples of ore-stage intrusions parental to major porphyry and high-sulfidation Cu(-Au) deposits of Phanerozoic age on five continents are shown by dark grey diamonds and inverted triangles (the ARCGOLDCOPPER compilation of ore-forming intrusion compositions described by Loucks, 2014).

Figure 1 shows that zircons crystallising from Cu-fertile melts inherit the melts’ absence of a substantial negative Eu anomaly. In addition, the zircons from Cu-ore-productive intrusions typically have a large Ce spike, usually Ce/Ce* ≈ 75 to 750, but with large grain-to-grain variability within samples, due at least in part to analytical imprecision in measuring La, which is typically near the detection limit of LA-ICPMS and SIMS ion probe analysis.  The small Eu anomaly is the more consistent feature of zircons from Cu-ore-productive magmas.

In order to elucidate the petrogenetic origin of the features in Figure 1 that are diagnostic of Cu metallogenic fertility in magmas and zircons, Figure 2 shows how the REE pattern shape-factor (EuN/Eu*)/YbN varies over the course of magmatic differentiation from parental mafic magmas to residual ore-stage felsic magmas, as compared with mafic-to-felsic magmatic differentiation trends in well explored arc segments that are apparently devoid of significant Cu mineralisation in the time interval considered.  Figure 2A compares whole-rock compositions of Cu-ore-bearing and unmineralised segments of two intra-oceanic arcs.  The 7-0 Ma Tampakan igneous complex in south-central Mindanao is at the north end of the Sangihe arc where it has collided face-to-face with the Halmahera arc by subduction-elimination of the intervening Molucca Sea plate (Rohrlach and Loucks, 2005).  The Tampakan igneous complex hosts giant high-sulfidation and porphyry Cu-Au deposits with ore reserves of 2500 million tonnes containing 12.5 million tonnes Cu and 500 tonnes Au (Rohrlach and Loucks, 2005). 

Farther south, the intra-oceanic Sangihe arc overlies the westwardly subducting Molucca Sea Plate in a non-compressive crustal stress regime.  In the Sunda-Banda arc of Indonesia, northward subduction of the Scott Plateau prong of northwestern Australia is producing orogenic deformation and uplift of Sumba and Sumbawa islands and incipient subduction reversal by southward subduction of the Banda Sea lithosphere along the Flores Thrust north of Sumbawa, Flores, and Wetar (Harris et al, 2009). The late Neogene Batu Hijau porphyry Cu deposit (1840 million tonnes ore with 8.1 million tonnes Cu and 644 tonnes Au) and Elang porphyry Cu-Au prospect on Sumbawa developed during orogeny associated with arrival of the Scott Plateau prong of the Australian continent at the Sunda Trench.  Figure 3B shows mafic-to-felsic differentiation series in major Cu-ore-bearing igneous complexes in continent-margin arcs, in comparison with reference suites from well explored continental-arc segments that are apparently devoid of economically significant Cu mineralisation in the time interval considered.  

In Figure 2A and B, convergence of magmatic differentiation trends at the low-SiO2 end of the series implies that all these ore-forming and barren suites share a common type of parental, mantle-derived arc-basaltic magma.  The divergence of Cu-ore-productive and Cu-infertile magmatic differentiation trends from the same type of mafic parent magma implies that copper metallogenic fertility is an emergent property of an atypical magmatic differentiation process, not a primary property acquired by different melting processes or by melting different kinds of source material.  Cu-ore-productive magma series undergo magmatic differentiation mainly at high pressures in the range 0.6–1.2 GPa in compressive, orogenic stress regimes, whereas the Cu-infertile suites evolve from basaltic parental magmas mainly at upper-crustal pressures in tectonic settings that permit easy magma ascent from the mantle by dyke propagation against low resistance by horizontal tectonic stress (Loucks and Ballard, 2002; Rohrlach and Loucks, 2005; Loucks, 2014). In strongly compressive stress regimes, mantle-derived magmas tend to pond near the Moho where basaltic magmas lose about half their buoyancy at the density discontinuity from mantle peridotite to lower-crustal gabbroic or mafic granulitic country rocks, and where the uppermost lithospheric mantle is usually the stiffest and strongest part of the arc lithosphere (Bürgmann and Dresen, 2008) and can laterally transmit tectonic stress to provide relatively high resistance to ascent of magma by hydraulic fracturing. 

Magma chambers embedded in hot country rocks near the Moho cool very slowly and tend to last long enough to experience intermittent replenishment from the mantle during continuous fractional crystallisation of cumulates on the chamber floor, so after multiple replenishment-and-crystallisation cycles, the residual felsic melts have inherited H2O, SO3, Cl, and other cumulate-incompatible chemical components from all prior batches of magma that have entered the chamber.  Tapping the top of the deep chamber during its replenishment from the mantle can lead to a succession of epizonal intrusions and volcanics that provide a time series monitoring chemical evolution of the deep chamber over millions of years of its lifespan. Such time series in ore-bearing igneous complexes indicate that, in the long-lived, Moho-level, intermittently replenished, intermittently tapped, continuously crystallising magma chamber, the rising accumulation of dissolved H2O in residual melts of successive cycles causes hornblende to arrive earlier in the paragenetic sequence and crystallise in greater abundance in successive cycles.  Hornblende crystallisation depletes the melt in yttrium and lanthanide elements heavier than Tb, so hornblende saturation early in the paragenetic sequence imposes on residual melts the “Nike swoosh”-shaped REE patterns illustrated in Figures 1C and E.  REE patterns of that shape tell us that the magmas were unusually H2O-rich. Conversely, crystallisation of anhydrous pyroxenes until dacitic stages of magmatic differentiation is responsible for flatter REE patterns in the Dy to Lu interval (e.g., Fig. 1A), which tell us the magma series was relatively H2O-poor.

In long-lived, Moho-vicinity magma chambers, rising accumulation of dissolved H2O in residual melts of successive cycles causes plagioclase production to arrive later in the mineral saturation sequence in successive cycles. At the usual oxidation states of granitoid arc magmas, ~25-50% of the Eu in the melt is divalent (Wilke & Behrens, 1999). Eu2+ and Sr2+ partition strongly into plagioclase; early and prolific production of plagioclase from relatively shallow and/or H2O-poor melts is responsible for the pronounced negative Eu anomaly in Figure 1A and B. To the degree that high pressure and high dissolved H2O in the melt delay plagioclase saturation to late, dacitic stages of magmatic differentiation, Eu2+ and Sr2+ accumulate in residual melts until plagioclase eventually saturates.  A plagioclase-undersaturated, Sr- and Eu-undepleted melt departing a Moho-vicinity magma chamber becomes plagioclase-saturated as a consequence of decompression during ascent, and arrives in an epizonal, subvolcanic magma chamber as a plagioclase porphyry with undepleted Eu and Sr, and high bulk Sr/Y and (Eu/Eu*)/Yb, as demonstrated by Rohrlach and Loucks (2005) over the 7-Myr duration of multi-cycle magmatism in the Tampakan igneous complex, Mindanao. These processes account for the magmatic differentiation trends shown in Figure 2, and the features of ore-stage magmas and zircons in Figure 1.

Insights from Figures 1 and 2 prompted formulation of the zircon REE-pattern shape-factors plotted as coordinates in Figure 3, which shows 702 new analyses of zircons done for this study, together with a compilation of 2760 published zircon analyses by ion probe and LA-ICPMS, done in many laboratories around the world.  The objective of the compilation of published zircon analyses was to test for consistency of signal by varied analytical methods and laboratories and consistency of signal among porphyry ore deposits that developed in intra-oceanic arcs, at continent margins above subducting oceanic lithosphere, and in Tethyan “collisional” settings (Yunnan and Tibet) above deeply subducting continental lithosphere, wherein Cenozoic deep subduction of continental crust and its sedimentary cover provided aqueous fluids to flux mantle-wedge melting during metamorphic dehydration of subducted continental sediments (e.g., Guo et al., 2013; Lu et al., 2015).  The comparison of results by different analytical methods shows that fast, inexpensive analysis by LA-ICPMS using atomic- to molecular-scale cold photo-evaporation by 193nm-wavelength far-ultraviolet excimer laser (introduced for LA-ICPMS by Loucks et al., 1996, who calculated that 193nm photon quantum energies match many inter-atomic bond energies in minerals)—but not chunky thermal-expansion sputtering by longer-wavelength ultraviolet and infrared lasers—typically yields analytical detection limits and precision equivalent to slow, expensive SHRIMP and Cameca ion-probe analyses, which also ablate at the atomic- to molecular scale.  The 702 new analyses plotted in Figure 3 were done using the 193nm excimer LA-ICPMS facility at Curtin University.

Fig. 3.  Porphyry- and high-sulfidation copper-ore-forming arc magmas, all of which have “adakitic” or “appinitic” whole-rock REE patterns and Sr/Y ratios, produce zircons having REE compositions that are usefully distinct from zircons in most Cu-unmineralised igneous suites from convergent and divergent plate margins and intra-plate tectonic settings.  Zircons from Cu-(±Au±Mo) ore-forming igneous suites are represented by orange and reddish symbols.  More than 3450 zircon analyses by ion-probe and LA-ICPMS from 43 Phanerozoic igneous suites are plotted, of which 2760 analyses from 30 suites were compiled from published literature; samples from another 13 suites in the USA, Philippines, and Indonesia were collected and analysed by us using the ArF LA-ICPMS facility at Curtin University. (A) The Ce/Ce* ratio is widely used by zircon geochemists to represents the amplitude of the Ce spike (Fig. 1).  The rare-earth element Y has the same ionic charge (+3) and nearly identical ionic radius and zircon/melt partition coefficient as the lanthanide element Ho, but Y is much more abundant and more precisely analysable than Ho. Because Ce* is evaluated by interpolation between La and Pr [Ce* = √(LaN X PrN)], both of which are typically near or below the detection limits by ion-probe and LA-ICPMS analysis, there is substantial analytical imprecision and data scatter in panel A, and many analysed zircons could not be plotted because La was below the detection limit.  Only convergent-plate-margin igneous suites are plotted in panel A.   (B) Nd abundance in zircon is typically well above the analytical detection limit, so the ratio CeN/NdN is used to represent the amplitude of the Ce spike in zircon. The REE pattern shape-factor (CeN/NdN)/Y is measurable with relatively high analytical precision and obviously provides a more useful data-sorting efficacy than the shape-factor (CeN/Ce*)/YN in panel A.  All new and published zircon analyses that pass the purity screens (eliminating analyses contaminated by phosphate and other mineral micro-inclusions) are plotted in panel B, including igneous suites from intra-plate hotspots and rifts and mid-ocean-ridge settings, because potential users of detrital zircons as mineral exploration “pathfinders” would be interested to know whether zircons from convergent-plate-margin igneous suites are chemically distinguishable from zircons in intra-plate and divergent-plate-margin igneous suites.

Not all of the unmineralised suites plotted in Figure 3 are necessarily Cu-infertile; for example, the Mt. St. Helens and Mt. Pinatubo zircons are from recent (<50 ka) adakitic eruptions on stratovolcanoes not eroded to depths sufficient to expose coeval ore mineralisation, if any.  The 1991 Pinatubo adakitic dacite (zircons plotted) is chemically indistinguishable from the dacite porphyry stock that produced the 2.5 Ma Dizon porphyry Cu-Au deposit that is exposed in the eroded south flank of the Pinatubo edifice (Imai, 2005).  The appinitic/adakitic Criffell-Dalbeattie granodioritic to granitic pluton of Caledonian age in SW Scotland (zircons plotted) contains no reported Cu mineralization, but near-90° vertical-to-horizontal rotation of the plutonic-to-volcanic igneous complex and host strata has preserved the structurally shallower, contemporaneous and co-magmatic Black Stockarton Moor subvolcanic intrusive complex exposed 3-5 km farther west, which contains porphyry-Cu mineralisation in granodiorite porphyry stocks, dykes, and sills (Leake and Cooper, 1983). The Criffell-Dalbeattie complex is one of a number of adakitic/appinitic Caledonian volcanic-plutonic complexes in western Scotland that have whole-rock chemistry matching the characteristics of Cu-ore-productive arc magmas, such as in Figure 2.

In an extended treatment of the petrogenetic basis of our zircon indicators of Cu metallogenic fertility (Loucks et al., in prep.), we demonstrate that Eu2+/Eu3+ and Ce4+/Ce3+ in zircon and the closely related parameters Eu/Eu* and Ce/Ce* in zircon are not meaningful proxies for the oxidation state of the zircon’s parent melt, in contrast to claims by Trail et al. (2011, 2012, 2015) and by Dilles et al. (2015).  We demonstrate that variations of those parameters in and among zircon populations are due almost entirely to variation of magmatic hydration state and crystallisation temperature.

Fig. 4.  Sulphur contents of hydrous silicate liquids with compositions in the basaltic andesite to rhyolite range (~56–70 wt % SiO2) that were experimentally saturated with pyrrhotite and/or anhydrite over a range of experimentally imposed oxygen fugacity are shown. The oxidation state of the experimental melt is represented as the difference in log units from the nickel metal + nickel oxide reference buffer (NNO). Pressures in the range 200–400 MPa correspond to lithostatic pressures at depths of ~7–15 km.  The experiments demonstrate that the solubility of sulfur as sulfate in the melt is about an order of magnitude greater at log ƒO2 > NNO+1.5 (limited by igneous anhydrite saturation) than at log ƒO2 < NNO mainly as FeS and H2S species in the melt (limited by pyrrhotite saturation).  The usual range of oxidation state of porphyry- and high-sulfidation Cu-ore-forming arc magmas is from Burnham and Ohmoto, (1980).

Figure 4 shows why a valid indicator of magmatic oxidation state, applicable to detrital zircons, would be useful to exploration programs for porphyry- and high-sulfidation-epithermal Cu-Au deposits.  As demonstrated experimentally by Scaillet et al. (1998), sulfate (S6+O3) dissolved in silicate melt partitions into exsolving hydrothermal fluid as the more volatile sulfite (S4+O2).  At subsolidus temperatures within a pluton’s crystalline carapace or country rocks, the fluid’s SO2 undergoes hydrolysis according to the reaction 4SO2 + 4H2O --> 3H2SO4 + H2S, which supplies nearly all the reduced sulfur that is available to precipitate Cu as sulfide minerals.  So the sulfate content of the silicate melt largely determines whether there is enough sulfide sulfur available to achieve efficient precipitation of the available Cu in the hydrothermal fluid.

Fig. 5. Nernst partition coefficients of Ce and U and other trace elements between zircon and synthetic andesitic melt were measured by secondary-ion mass spectrometry (SIMS, ion probe) and by LA-ICPMS in Burnham and Berry’s (2012) experiments at 1300°C and a 14 log units variation of oxygen fugacity.  Experimental oxygen fugacity is represented here as log units difference from that of the fayalite-magnetite-quartz reference buffer. The oxidation state of cerium varies from entirely Ce3+ in melt at the lowest ƒO2 to increased (but still small) amounts of Ce4+ in the melt at the highest ƒO2, over which range the Ce content of zircon rises ~100-fold as the proportion of Ce4+ in the melt increases, despite the melt’s supply of Nb5+ and P5+ that could provide local charge-balance for Ce3+ in zircon at low ƒO2.  Uranium shows the opposite trend, being entirely U4+ in the melt at the lowest ƒO2, but with increasing amounts of U5+ and U6+ in the melt as ƒO2 is increased.  The data show a very strong preference for U4+ in Zr4+ lattice sites, and the zircon/melt partition coefficient of U declines ~20-fold as the melt’s supply of U4+ diminishes with rising ƒO2, despite the availability of the entire series of REE3+ in the melt that could provide local charge balance for U5+ and U6+ on the Zr4+ sublattice at high ƒO2.

Burnham and Berry’s (2012) experimental study of trace-element partitioning between zircon and haplo-andesitic melt over a large range of ƒO2 offers inspiration as to how a new zircon oxybarometer might be formulated using the Ce/U ratio in zircon.  Ce and U occur in natural zircons at concentration levels that are easy to measure precisely, and their zircon/melt partition coefficients vary oppositely as ƒO2 varies within the range relevant to terrestrial magmas, as we illustrate in Figure 5.  However, the Ce/U ratio in zircon must also vary with Ce/U ratio in its parent melt, which varies slightly among mantle-derived primary melts as percentage of partial melting varies (Ce/U ≈ 30-40 in most primary arc basalts) and varies more strongly with stage of magmatic differentiation, as illustrated in Figure 6.  In arc-magma mafic-to-felsic differentiation series, aluminous augite and hornblende accept modest amounts of light REEs in Ca2+ lattice sites (DCe < 1), but augite and hornblende have even lower uptake of U4+ (DU << 1); apatite typically saturates at a basaltic-andesite stage, and during continued magmatic differentiation apatite selectively sequesters the melt’s Ce relative to U (DCeapatite/melt ≈ 26 and DUapatite/melt ≈ 1.8 in El Chichón apatite phenocryst/rhyolitic groundmass pairs, for example; Luhr et al., 1994).

Fig. 6.  The whole-rock ratio Ce/U decreases with increasing SiO2 over the mafic-to-felsic calc-alkalic (to adakitic) magmatic differentiation series (A) and arc-tholeiitic differentiation series (B) in continent-margin arcs as well as in intra-oceanic-arc calc-alkalic series (C) and oceanic-arc tholeiitic series (D).  The plotted Andean data may be found on the GEOROC website (http://georoc.mpch-mainz.gwdg.de/georoc/). Adak data are mainly from Romick et al. (1992) and Kay and Kay (1994).  Okmok data are mainly from Miller et al. (1992) and Kay and Kay (1994).

Fig. 7.  The whole-rock ratio U/Ti increases with increasing SiO2 over the mafic-to-felsic calc-alkalic magmatic differentiation series (A) and arc-tholeiitic differentiation series (B) in continent-margin arcs as well as in intra-oceanic-arc calc-alkalic series (C) and oceanic-arc tholeiitic series (D).  The plotted Andean data may be found on the GEOROC website (http://georoc.mpch-mainz.gwdg.de/georoc/). Adak data are mainly from Romick et al. (1992) and Kay and Kay (1994).  Okmok data are mainly from Miller et al. (1992) and Kay and Kay (1994).

In order to damp the component of variance in zircon’s Ce/U ratio that is due to stage of magmatic differentiation, we employ the zircon’s U/Ti ratio as a differentiation index of its parent melt.  Figure 7 illustrates the typical variation of whole-rock U/Ti ratio in calc-alkalic and tholeiitic differentiation series of continent-margin and intra-oceanic arc magmas.  The parameter (U/Ti) in zircon affords an appropriate weighting of the parent melt’s differentiation index. 

Fig. 8. The average value of the parameter Ce/√((U X Ti) for the analysed zircon population in each rock sample is plotted against oxygen fugacity of the parent magma, as provided by independent constraints from laboratory crystallisation experiments and natural mineral assemblages.  The least-squares fit to the data is ΔFMQ = 2.555 + 4.317 X log [Ce/√(U X Ti)], with a correlation coefficient R = 0.95 and a standard deviation of 0.51 log unit ƒO2.  The biotite-hornblende granodiorite samples from the Caledonian Criffell-Dalbeattie pluton in SW Scotland have poorly constrained ƒO2; Stephens et al. (1985) state “biotite mineral separates from the granodiorites have high Fe3+/Fe2+ ratios lying close to the magnetite-haematite buffer estimated by Wones and Eugster (1965).”  No biotite analyses are reported.  Because the tilted Criffell-Dalbeattie pluton is structurally overlain by the coeval Black Stockarton Moor granodiorite-porphyry stock-dyke-sill swarm that generated porphyry copper mineralisation (Leake and Cooper, 1983), the data are of interest, even though the crude ƒO2 constraint lies well above the accepted range of calc-alkalic biotite-hornblende granitoid magmas (Carmichael, 1991). The Criffell samples are plotted at the ƒO2 of the hematite-magnetite buffer, but parenthesised and excluded from the regression fit. 


Figure 8 shows the average value of the parameter (Ce/U) X√(U/Ti) (which can be rearranged as Ce/√(U X Ti)) in each analysed zircon population in 69 samples from rock units for which independent constraints are available on the parent melt’s oxidation state. The ƒO2constraints are of many types and vary substantially in precision. The ƒO2 values in our calibration dataset have reported 1s uncertainties in the range ±0.2–0.4 log units in the best cases (crystallisation experiments at controlled ƒO2 and multiple magnetite-ilmenite phenocryst pairs in volcanics) or greater uncertainties in the cases of single values applied to all data from the moon, kimberlites, ocean-ridge basalts, and titanite+magnetite assemblages. In addition to uncertainties in ƒO2 values in the calibration dataset, some scatter arises from chemical heterogeneities in analysed zircons.  Diffusion of high-valence ions through the silicate melt boundary layer adjacent to growing crystal surface is slower than the rate of crystal growth in fast-cooling subvolcanic plutons, leading to varying degrees of disequilibrium crystal/melt partitioning in oscillatory- and sector-zoned zircon crystals, and differential uptake of trace elements on different crystal faces in a single growth zone (Chamberlain et al., 2014). Therefore, the ƒO2 calibration in Figure 8 is more reliably applied to an average of several analyses per crystal than to a single spot analysis. Most of the plotted values are averages of 5–50 analyses per sample.

Despite significant uncertainties in some of the ƒO2 constraints in our calibration dataset, the regression fit’s standard error of ±0.5 log unit ƒO2 in recovery of dataset values in Figure 8 assures us that the formulation our new oxybarometer is theoretically sound, so its reliability could be refined in future by re-calibration using an expanded dataset relying heavily on natural zircon-bearing volcanic phenocryst and melt-inclusion assemblages having multiple, corroborative ƒO2 indicators.


We thank Paul Agnew and Alan Kobussen of Rio Tinto Exploration for financial support and for authorisation to publicise research results. Additional financial support was provided by the ARC Centre of Excellence for Core to Crust Fluid Systems (CCFS). Noreen Evans and Chris Kirkland of Curtin University assisted with LA-ICPMS data acquisition and reduction.  We thank Steve Garwin, Bruce Rohrlach, Doone Wyborn, Alfred Anderson, Nelia Dunbar, and Michael Palin for providing rock samples, and Henrietta Cathey for providing unpublished zircon analyses.

References Cited


Botcharnikov, R.E., Linnen, R.L., Wilke, M., Holtz, F., Jugo, P.J., Berndt, J., 2011. High gold concentrations in sulphide-bearing magma under oxidizing conditions. Nature Geoscience 4: 112–115.

Bürgmann, R., and Dresen, G., 2008.  Rheology of the lower crust and upper mantle:evidence from rock mechanics, geodesy, and field observations.  Annual Reviews of Earth and Planetary Sciences, 36: 531–67. doi: 10.1146/annurev.earth.36.031207.124326.

Burnham, A.D., and Berry, A.J. (2012) An experimental study of trace element partitioningbetween zircon and melt as a function of oxygen fugacity.  Geochimica et Cosmochimica Acta, 95: 196-212.

Carmichael, I.S.E., 1991. The redox states of basic and silicic magmas: a reflection of their source regions? Contributions to Mineralogy and Petrology 106: 129-141

Carroll, M.R., and Rutherford, M.J., 1987. The stability of igneous anhydrite: experimental results and implications for sulfur behaviour in the 1982 El Chichón trachyandesite and other evolved magmas.  Journal of Petrology, 28: 781-801.

Chamberlain, K.J., Wilson, C.J.N., Wooden, J.L., Charlier, B.L.A., and Ireland, T.R., 2014.  New perspectives on the Bishop Tuff from zircon textures, ages, and trace elements.  Journal of Petrology, 55: 395-426. doi:10.1093/petrology/egt072

Finney, B., Turner, S., Hawkesworth, C. Larsen, J., Nye, C. George, R. Bindeman, I. and Eichelberger, J., 2008. Magmatic differentiation at an island-arc caldera: Okmok volcano, Aleutian Islands, Alaska.  Journal of Petrology, 49: 857-884.

Fiorentini, M.L. and Garwin, S.L., 2010.  Evidence of a mantle contribution in the genesis of magmatic rocks from the Neogene Batu Hijau district in the Sunda arc, southwestern Sumbawa, Indonesia.  Contributions to Mineralogy and Petrology, 159: 819-837.

Ghiorso, M.S., Evans, B.W., 2008. Thermodynamics of rhombohedral oxide solid solutions and a revision of the Fe–Ti two oxide geothermometer and oxygen-barometer. American Journal of Science, 308: 957–1039.

Harris, R., Vorkink, M.W., Prasetyadl, C., Zobell, E., Roosmawati, N. and Apthorpe, M., 2009.  Transition from subduction to arc-continent collision: geologic and neotectonic evolution of Savu Island, Indonesia.  Geosphere,  5: 152–171; doi: 10.1130/GES00209.1

Hora, J.M., Singer, B.S., Wörner, G., Beard, B.L., Jicha, B.R., and Johnson, C.M., 2009.  Shallow and deep crustal control on differentiation of calc-alkaline and tholeiitic magma.  Earth and Planetary Science Letters, 285: 75-86.

Imai, A., 2005, Evolution of hydrothermal system at the Dizon porphyry Cu-Au deposit, Zambales, Philippines. Resource Geology, 55: 73-90.

Ishihara, S., 1977. The magnetite-series and ilmenite-series granitic rocks. Mining Geology, 27: 293–305.

Kay, S. M. and Kay, R. W., 1982.  Tectonic controls on tholeiitic and calc-alkaline magmatism in the Aleutian arc.  Journal of Geophysical Research, 87: 4051-4072.

Kay, S. M. and Kay, R. W., 1994. Aleutian magmas in space and time. In: Plafker, G. and Berg, H. C. (eds) The Geology of Alaska. The Geology of America, Volume G-1. Boulder, CO: Geological Society of America, pp. 687-722.

Lang, J.R. and Titley, S.R., 1998, Isotopic and geochemical characteristics of Laramide magmatic systems in Arizona and implications for the genesis of porphyry copper deposits.  Economic Geology, 93: 138-170.

Leake, R.C. and Cooper, C., 1983, The Black Stockarton Moor subvolcanic complex, Galloway.  Journal of the Geological Society London, 140: 665-676.

Loucks, R. R. 2000. Report 1: Development of petrochemical discriminants of metallogenically fertile calc-alkalic igneous suites and their application to regional assessments of gold and copper prospectivity of Neogene and Quaternary volcano-plutonic centers througout Chile and in southwestern Bolivia. Unpublished report for industry-sponsored research project: Predictive Guides to Copper and Gold Mineralisation at Circum-Pacific Convergent Plate Margins, 73 p.

Loucks, R.R., 2012a, Chemical characteristics, geodynamic settings, and petrogenesis of copper-ore-forming arc magmas.  CET Quarterly News, issue 19.

Loucks, R.R., 2012b, Chemical characteristics, geodynamic settings, and petrogenesis of gold-ore-forming arc magmas.  CET Quarterly News, issue 20.

Loucks, R.R., 2014.  Distinctive composition of copper-ore-forming arc magmas: Australian Journal of Earth Sciences, 61: 5-16, doi: 10.1080/08120099.2013.865676.

Loucks, R. R. and Ballard,  J. R., 2002. Report 2A: Correlated variation of tectonic stress with chemical composition of arc volcanism: Unpublished report for industry-sponsored research project. Predictive Guides to Copper and Gold Mineralisation at Circum-Pacific Convergent Plate Margins, 99 p.

Loucks, R.R., Eggins, S.M., Shelley, J.M.G., Kinsley, L.P.J.  and Ware, N.G., 1996.  Development of the inductively-coupled-plasma mass-spectrometry ultraviolet laser trace-element micro-analyzer (ICP–MS -ULTEMA).  Research School of Earth Sciences Annual Report 1995

Lu, Y.-J., Loucks, R.R., Fiorentini, M.L., Yang, Z.M. and Zeng-Qian Hou, Z.-Q., 2015. Fluid flux melting generated postcollisional high Sr/Y copper ore–forming water-rich magmas in Tibet.  Geology, 43: 583-586. doi:10.1130/G36734.1

Luhr J. F. (1990) Experimental phase relations of water- and sulfur-saturated arc magmas and the 1982 eruptions of El Chichón Volcano. Journal of Petrology, 31: 1071–1114.

Luhr, J.F., Carmichael, I.S.E., and Varekamp, J.C., 1994.  The 1982 eruptions of El Chichón volcano, Chiapas, Mexico: mineralogy and petrology of the anhydrite-bearing pumices.  Journal of Volcanology and Geothermal Research, 23: 69-108.

Martini, M., Ferrari, L., López-Martínez, M., Cerca-Martínez, M., Valencia, V., and Serrano-Durán, L., 2009. Cretaceous- Eocene magmatism and Laramide deformation in Southwestern Mexico: No role for terrane accretion, in Kay, S.M., Ramos, V.A., and Dickinson, W.R., eds.,Backbone of the Americas: Shallow subduction, plateau uplift, and ridge and terrane collision. Geological Society of America Memoir, 204: 151–182.

Martiny, B.M., Morán-Zenteno, D.J., Solari, L.,  López-Martínez, M., de Silva, S.L., Flores-Huerta, D.,  Zúñiga-Lagunes, L. and Luna-González, L., 2013. Caldera formation and progressive batholith construction: Geochronological, petrographic and stratigraphic constraints from the Coxcatlán-Tilzapotla area, Sierra Madre del Sur, Mexico.  Revista Mexicana de Ciencias Geológicas, 30: 247-267.

McDonough, W.F. and Sun, S.-s., 1995.  The composition of the Earth.  Chemical Geology, 120: 223-253.

Miller, D.M., Langmuir, C.H., Goldstein, S.L., and Franks, A.L., 1992.  The importance of parental magma composition to calc-alkaline and tholeiitic evolution: evidence from Umnak Island in the Aleutians.  Journal of Geophysical Research, 97: 321-343.

Mori, L., Morán-Zenteno, D.J., Martiny, B.M., González-Torres, E.A., Chapela-Lara, M.,  Díaz-Bravo, B.A. and Roberge, J., 2012.  Thermomechanical maturation of the continental crust and its effects on the late Eocene–early Oligocene volcanic record of the Sierra Madre del Sur Province, southern Mexico.  International Geology Review, 54: 1475–1496.

Nieto-Samaniego, A.F., Alaniz-Álvarez, S.A., Silva-Romo, G., Eguiza-Castro, M.H., and Mendoza-Rosales, C.C., 2006, Latest Cretaceous to Miocene deformation events in the eastern Sierra Madre del Sur, Mexico, inferred from the geometry and age of major structures: Geological Society of America Bulletin, 118: 238–252, doi: 10.1130/B25730.1.

Parat, F., Holtz, F., Feig, S., 2008. Pre-eruptive conditions of the Huerto Andesite (Fish Canyon system, San Juan volcanic field, Colorado): Influence of volatiles (C–O–H–S) on phase equilibria and mineral composition. Journal of Petrology, 49: 911–935

Potra, A. and Macfarlane, A.W., 2014. Lead isotope studies of the Guerrero composite terrane, west-central Mexico: implications for ore genesis.  Mineralium Deposita, 49: 101-117, doi: 10.1007/s00126-013-0477-0.

Rohrlach, B. D., 2002. Tectonic evolution, petrochemistry, geochronology and palaeohydrology of the Tampakan porphyry and high sulphidation epithermal Cu–Au deposit Mindanao, Philippines, Unpub. PhD thesis, Australian National University, Canberra, Australia, 499 p.

Rohrlach, B.D. and Loucks, R.R., 2005. Multi-million-year cyclic ramp-up of volatiles in a lower-crustal magma reservoir trapped below the Tampakan copper–gold deposit by Mio–Pliocene crustal compression in the southern Philippines. In: Porter T. M. ed. Super Porphyry Copper & Gold Deposits—A Global Perspective, v.2, pp. 369– 407. PCG Publishing, Adelaide.

Romick, J.D., Kay, S.M., and Kay, R.W., 1992.  The influence of amphibole fractionation on the evolution of calc-alkaline andesite and dacite tephra from the central Aleutians, Alaska.  Contributions to Mineralogy and Petrology, 112: 108-118.

Scaillet, B., Clemente, B., Evans, B.W., and Pichavant, M., 1998. Redox control of sulfur degassing in silicic magmas. Journal of Geophysical Research, 103: 23,937−23,949.

Shannon, R.D., 1976. Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides.  Acta Crystallographica, A32: 751-767. 

Sillitoe, R. H., 1998. Major regional factors favoring large size, high hypogene grade, elevated gold content and supergene oxidation and enrichment of porphyry copper deposits. In: Porter T. M. ed. Porphyry and Hydrothermal Copper and Gold Deposits: A Global Perspective, pp. 21–34. Conference Proceedings Perth, 1998. Australian Mineral Foundation, Glenside, South Australia.

Singer, B.S., Jicha, B.R., Harper, M.A., Naranjo, J.A., Lara, L.E., and Moreno-Roa, H., 2008.  Eruptive history, geochronology, and magmatic evolution of the Puyehue-Cordón Caulle volcanic complex, Chile.  Geological Society of America Bulletin, 120: 599-618.

Stephens, W.E., Whitley, J.E., Thirlwall, M.F., and Halliday, A.N., 1985. The Criffell zoned pluton: correlated behaviour of rare earth element abundances with isotopic systems. Contributions to Mineralogy and Petrology, 89: 226-238. 

Sun, S.-s. and McDonough, W.F., 1989. Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes. In: A.D. Saunders and M.J. Norry (Editors). Magmatism in the Ocean Basins. The Geological Society, London. pp. 313-345.

Trail, D., Watson, E.B., and Tailby, N.D., 2011, The oxidation state of Hadean magmas and implications for early Earth’s atmosphere.  Nature, 480: 79-82, doi:10.1038/nature10655

Trail, D., Watson, E.B., and Tailby, N.D., 2012.  Ce and Eu anomalies in zircon as proxies for the oxidation state of magmas.  Geochimica et Cosmochimica Acta, 97: 70–87.

Trail, D., Tailby, N.D., Lanzirotti, A., Newville, M., Thomas, J.B., and Watson, E.B., 2015.  Redox evolution of silicic magmas: Insights from XANES measurements of Ce valence in Bishop Tuff zircons.  Chemical Geology, 402: 77–88.  doi.org/10.1016/j.chemgeo.2015.02.033.

Wilke, M., and Behrens, H., 1999.  The dependence of the partitioning of iron and europium between plagioclase and hydrous tonalitic melt on oxygen fugacity.  Contributions to Mineralogy and Petrology, 137: 102-114.

Wones, D.R., and Eugster, H.P., 1965). Stability of biotite: experiment, theory, and application. American Mineralogist, 50: 1228- 1272.