porphyry deposits and oxidized magmas_2015

Ore Geology Reviews 65 (2015) 97–131

Contents lists available at ScienceDirect

Ore Geology Reviews

j ourna l homepage: www.e lsev ie r .com/ locate /oregeorev

Review

Porphyry deposits and oxidized magmas

Weidong Sun a,⁎, Rui-fang Huang b,c, He Li a, Yong-bin Hu a,c, Chan-chan Zhang a,c, Sai-jun Sun a,c,Li-peng Zhang a,c, Xing Ding b, Cong-ying Li a, Robert E. Zartman a, Ming-xing Ling b

a CAS Key Laboratory of Mineralogy and Metallogeny, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, 511 Kehua Street, Wushan, Guangzhou 510640, Chinab State Key Laboratory of Isotope Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, 511 Kehua Street, Wushan, Guangzhou 510640, Chinac University of the Chinese Academy of Sciences, Beijing 100049, China

⁎ Corresponding author.E-mail address: [email protected] (W. Sun).

http://dx.doi.org/10.1016/j.oregeorev.2014.09.0040169-1368/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t
a r t i c l e i n f o
Article history:Received 17 June 2014Received in revised form 19 August 2014Accepted 2 September 2014Available online 16 September 2014

Keywords:Porphyry depositOxidized magmasOxygen fugacityAdakiteSlab meltsArc magmasPlate subductionRidge subduction

Porphyry deposits supply most of the world's Cu andMo resources. Over 90% of the porphyry deposits are foundat convergent margins, especially above active subduction zones, with much fewer occurrences at post-collisional or other tectonic settings. Porphyry Cu–(Mo)–(Au) deposits are essentially magmatic–hydrothermalsystems, which are generally initiated by injection of oxidizedmagmas saturatedwithmetal-rich aqueous fluids,i.e., the parental magmas need to be water rich and oxidized with most of the sulfur appearing as sulfate in themagma. Sulfur is themost important geosolvent that controls the behavior of Cu and other chalcophile elements,due to high partition coefficients of chalcophile elements between sulfide and silicate melts. Small amount ofresidual sulfides can hold a large amount of Cu. Therefore, it is essential to eliminate residual sulfides to gethigh Cu contents inmagmas for the formation of porphyry deposits. Sulfate (SO4

2−) is over 10 timesmore solublethan sulfide (S2−), and thus the solubility of sulfur depends strongly on sulfur speciation, which in turn dependson oxygen fugacities. The magic number of oxygen fugacity is log fO2 N FMQ+ 2 (i.e.,ΔFMQ+ 2), where FMQ isthe fayalite–magnetite–quartz oxygen buffer. Most of the sulfur inmagmas is present as sulfate at oxygen fugac-ities higher than ΔFMQ + 2. Correspondingly, the solubility of sulfur increases from ~1000 ppm up to N1 wt.%.Oxidation promotes the destruction of sulfides in the magma source, and thereby increases initial chalcophileelement concentrations, forming sulfur-undersaturated magmas that can further assimilate sulfides during as-cent. Copper, Mo and Au act as incompatible elements in sulfide undersaturated magmas, leading to highchalcophile element concentrations in evolved magmas. The final porphyry mineralization is controlled bysulfate reduction, which is usually initiated by magnetite crystallization, accompanied by decreasing pH andcorrespondingly increasing oxidation potential of sulfate. Hematite forms once sulfate reduction lowers the pHsufficiently, driving the oxidation potential of sulfate up to the hematite–magnetite oxygen fugacity (HM) buffer,which is ~ΔFMQ+ 4. Given that ferrous iron is the most important reductant that is responsible for sulfate re-duction during porphyry mineralization, the highest oxygen fugacity favorable for porphyry mineralization isthe HM buffer. In addition to the oxidation of ferrous iron during the crystallization of magnetite and hematite,reducing wallrock may also contribute to sulfate reduction and mineralization. Nevertheless, porphyry depositsare usually mineralized in the whole upper portion of the pluton, whereas interactions with country rocks aregenerally restricted at the interface, therefore assimilation of reducing sediments is not likely to be a decisive con-trolling process. Degassing of oxidized gases has also been proposed as amajor process that is responsible for sul-fate reduction. Degassing, however, is not likely to be a main process in porphyry mineralization that occurs at2–4 km depths in the upper crust. Sulfide formed during sulfate reduction is efficiently scavenged by aqueousfluids, which transports metals to shallower depths, i.e., the top of the porphyry and superjacent wallrock. Ac-cording to traditional views, sulfide saturation and segregation during magma evolution is not favorable forthe formation of porphyry Cu ± Au ± Mo deposits. This is the main difference between porphyry deposits andNi–Cu sulfide deposits. Nevertheless, in places with thick sections of reducing sediments, e.g., the westernNorth America, sulfide saturation and segregationmay occur during evolution of themagma, forming Cu-rich cu-mulates at the base of plutons. These Cu-rich sulfides may evolve into porphyry mineralization or even controlthe ore-forming process. Their contribution depends heavily on subsequent oxidation, i.e., a major contributioncan be expected onlywhen the sulfide cumulates are oxidized to sulfate, liberating the chalcophile elements. Sul-fate reduction and ferrous Fe oxidation form H+, which dramatically lowers the pH values of ore-forming fluidsand causes pervasive alteration zones in porphyry Cu deposits. The amount of H+ released duringmineralizationand the alkali content in the porphyry together control the intensity of alterations. In principle, H2 and methane

http://crossmark.crossref.org/dialog/?doi=10.1016/j.oregeorev.2014.09.004&domain=pdf

http://dx.doi.org/10.1016/j.oregeorev.2014.09.004

mailto:[email protected]

http://dx.doi.org/10.1016/j.oregeorev.2014.09.004

http://www.sciencedirect.com/science/journal/01691368

98 W. Sun et al. / Ore Geology Reviews 65 (2015) 97–131

form during the final mineralization process of porphyry deposits, but are mostly oxidized by ferric Fe duringsubsequent processes. Some of the reduced gases, however,may survive the highly oxidizing environment to es-cape from the system, or even to get trapped in fluid inclusions. Therefore, small amount of reduced gases influidinclusions cannot argue against the oxidized feature of the magmas. Reduced magmas are not favorable for por-phyry mineralization. Reduced porphyry deposits so far reported are just mineralization that has either been re-duced in host rock away from the causative porphyry or through assimilation of reducing components duringemplacement.

© 2014 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 982. Brief introduction of major oxygen buffers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

2.1. Fayalite–magnetite–quartz (FMQ) oxygen buffer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1022.2. Hematite–magnetite (HM) oxygen buffer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1022.3. Ni–NiO(NNO) oxygen buffer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1022.4. Pyrite + pyrrhotite +magnetite (PPM) oxygen buffer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

3. The association of porphyry deposits with oxidized magmas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1033.1. Large porphyry deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

3.1.1. Porphyry Cu and Au deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1043.1.2. Porphyry Cu–Mo deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1063.1.3. Porphyry Mo deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

3.2. Linkage between oxidized magmas and porphyry deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1083.2.1. Sulfur oxidation, sulfide under saturation and residual sulfide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1083.2.2. Sulfate reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1103.2.3. Hematite–magnetite intergrowth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

3.3. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1124. The association of porphyry deposits with reduced magmas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

4.1. Reduced magmas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1134.1.1. Evidence for reduced magmas of Catface porphyry deposit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1144.1.2. Evidence for reduced magma in the Baogutu porphyry deposit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

4.2. Source of copper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1154.3. Formation of reduced porphyry deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

4.3.1. The formation of the Catface porphyry deposit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1164.3.2. The formation of the Baogutu porphyry deposit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

4.4. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1175. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

5.1. The oxygen fugacities at convergent margins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1175.2. The difference between porphyry and epithermal in terms of oxygen fugacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

5.2.1. Magnetite crisis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1175.2.2. Oxygen fugacity and open systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

5.3. Adakite, slab melting, ridge subduction and porphyry Cu deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1215.3.1. Adakite and porphyry Cu deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1215.3.2. Ridge subduction and porphyry Cu deposits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

5.4. Alterations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1236. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

1. Introduction

Porphyry deposits are hosts to one of the most important economicmineral associations (Cooke et al., 2005; Halter et al., 2005; Heinrichet al., 2004; Mutschler et al., 2010; Sillitoe, 2010), accounting for ~80%Cu and ~95%Mo of the world's total reserves. It is also an important re-source of Au, Ag, Zn, Sn andW.Most porphyry deposits are found aboveactive subduction zones (Fig. 1) (e.g., Chiaradia, 2014; Chiaradia et al.,2012; Gonzalez-Partida et al., 2003; Hedenquist et al., 1998; Kesler,1997; Lee, 2014; Richards, 1999, 2013; Sillitoe, 2010; Sun et al., 2011;Wilkinson, 2013), with a few occurrences at post-collisional or othertectonic settings (Sillitoe, 2010), e.g., porphyry Mo deposits in the east-ern Qinling orogenic belt (Chen, 2013; Li et al., 2012a; N. Li et al., 2013)and, arguably porphyry Cu–Mo deposits in Gangdese belt on the southTibetan Plateau (Hou et al., 2009; Qu et al., 2004; Xiao et al., 2012)and some porphyry Cu deposits in Iran (Calagari, 2003; Castillo, 2006;Haschke et al., 2010; Shafiei et al., 2009).

The consensus is that most of the porphyry Cu ± Mo ± Au systemsare initiated by injection of oxidized adakitic magma saturated withaqueous fluids that are S- and metal-rich, i.e., the parental magmasmust be water rich and oxidized (e.g., Ballard et al., 2002; Burnhamand Ohmoto, 1980; Garrido et al., 2002; Imai, 2002; Liang et al., 2006;Mungall, 2002; Sillitoe, 2010; Stern et al., 2007; Sun et al., 2013b). It is,however, still controversial as regards to: why high oxygen fugacity isfavorable for the mineralization of porphyry deposits, how oxidizedthe magma could be, whether adakitic magma is essential for porphyrymineralization orwhether the porphyry deposits can be associatedwithnormal arc rocks (Fig. 2), and why the pure porphyry Mo deposits arealso closely associated with highly oxidized magmas.

Copper, Au and Mo are chalcophile elements, which are stronglycontrolled by the behavior and speciation of sulfur. Therefore, the lessthe quantity of residual sulfide, the higher the initial Cu contents in pri-marymagmas (Fig. 3) (Lee et al., 2012; Sun et al., 2004a, 2013b). Exper-iments show that sulfate is much more soluble than sulfide in magmas

Fig. 1.Worldwide distribution of porphyry Cu deposits. Note, most of the porphyry deposits are distributed along convergent margins. Porphyry Mo deposits are not shown.Modified after Sun et al. (2013b). Data sources: Mutschler et al. (2010).

99W. Sun et al. / Ore Geology Reviews 65 (2015) 97–131

(Beermannet al., 2011; Jugo, 2009). Therefore, considerablymore sulfuris removed in the form of sulfate under higher oxygen fugacity (Leeet al., 2012; Liang et al., 2009; Sun et al., 2013b). Meanwhile, sulfide iskept undersaturated during the evolution of oxidized magmas, suchthat no sulfide segregation is expected (Sun et al., 2012a; Sun et al.,2013b). Therefore, Cu, Mo and Au act as moderately incompatible ele-ments (McDonough and Sun, 1995; Sun et al., 2003a,b,c), which becomeenriched during the early stage of magma evolution, and suddenly gointo magmatic fluids during magnetite crystallization (Sun et al.,2004a, 2013a). It has been proposed that oxygen fugacity of logfO2 N FMQ + 2 (i.e., ΔFMQ + 2, where FMQ is the fayalite–magnetite–quartz oxygen buffer) is themagic number for porphyry mineralization(e.g., Mungall, 2002; Sun et al., 2013b). Sulfate is the dominant speciesat log fO2 N FMQ + 2, which is much more soluble than sulfide inmagmas (Jugo, 2009). Therefore, residual sulfides are more efficientlydestroyed at oxygen fugacities higher than FMQ + 2, releasing theirchalcophile elements (Sun et al., 2013b). Others have argued that SO2

is the main sulfur species dissolved in porphyry magmas (Richards,2014; Smith et al., 2012), not sulfate nor sulfide. It was further arguedthat variations in oxidation state over typical ranges for arc magmas(ΔFMQ = 0 to +2) have no major effects on the potential of magmasto form porphyry Cu–(Mo) deposits during plate subduction(Richards, 2011b). The implication is that ΔFMQ + 2 is not of impor-tance for porphyry Cumineralization. Instead, water is themost impor-tant factor that controls porphyrymineralization (Richards, 2011a). Thequestion then becomes that most arc plutons are water saturated andhighly oxidized, why do only a very small fraction of special magmas(mostly adakitic characterized by high Sr/Y and high Sr) in arc settingsform porphyry deposits. For example, there are essentially no porphyrydeposits in Japan (Fig. 1). Note that Japan arc is associated with olderand presumably wetter subducting plate, such that arc rocks are pre-sumably wetter than subduction related rocks along the easternmarginof the Pacific Ocean (Sun et al., 2012a,b, 2013a,b).

A recent study has suggested that the source region for arc magmasis probably neither unusually oxidized nor enriched in economic ele-ments of interest, such as Cu, as has been shown by studies of primitivearc magmas (Lee et al., 2010, 2012). This is seemingly consistent withthe moderate incompatibility of Cu during mantle magmatism (Sun

et al., 2003a,b), which leads to Cu depletion in the mantle wedge.Also, the moderate mobility of Cu during plate subduction (J.L. Li et al.,2013) can compensate for the depletion caused by previous melting.The implication thus is that normal arc rocks, i.e., peridotitic melts, arenot favorable for porphyry Cu mineralization.

Still other authors have argued that, instead of high oxidation, sul-fide saturation of the magma and consequent pre-enrichment throughsulfide accumulation are the most important step for porphyry ore de-posits (Chiaradia, 2014; Lee, 2014; Wilkinson, 2013). This is probablytrue along the western margin of the North American continent andother places where reduced sediments are well developed (see morein Section 4). However, this hypothesis raises several other questions.Why are porphyry deposits usually associated with oxidized magmas,which is a situation so different from Ni–Cu sulfide deposits that clearlyexperienced sulfide saturation? How does porphyry remain oxidizedafter collecting sulfide from the pre-enriched sulfide accumulates?Why are the grades of porphyry deposits much lower than sulfide satu-rated Ni–Cu sulfide deposits? Why is there no Ni in porphyry deposits?

Moreover, some porphyry deposits are apparently associated withreduced magmas at ΔFMQ − 0.5 to −3 (Cao et al., 2014; Rowins,2000; Smith et al., 2012). Although reduced porphyry deposits arerare, with tonnages much smaller than oxidizing porphyry deposits(Cao et al., 2014), the mechanism needs to be clarified.

This contribution focuses on the controlling factors of porphyrymin-eralization. Major controversies to be discussed here are: (1) Why domost of the porphyry deposits associated with oxidized magmas occurat convergent margins? (2)What is themost favorable oxygen fugacityrange for porphyry mineralization? (3) What is the genetic connectionbetween high oxygen fugacity magmas and porphyry deposits?(4) Why are some porphyry deposits associated with reducingmagmas? (5) What are the connections between porphyry depositsand reduced magmas, if there is any?

2. Brief introduction of major oxygen buffers

Oxygen fugacity (fO2) is an important geological parameter affectingthe stability ofminerals, the evolution ofmagmas, and ore-forming pro-cesses. It is an equivalent of the partial pressure of oxygen in a particular

image of Fig.�1

Fig. 2. Twodifferentmodels for porphyry Cu±Au±Modeposits. A. Porphyry deposits are formed in normal arc rocks (after Richards, 2011a). According to thismodel, even the formationof giant porphyry deposits is nothing special but optimization of normal ore-forming processes, controlled by distinct tectonic configurations, reactive host rocks, or focusedfluid flow thathavehelped to enhance the overall process (Richards, 2013). B. Porphyry deposits are associatedwith slabmelts (modified afterWilkinson, 2013),whichhavehigh initial Cu contents (Sunet al., 2011).


environment (atmosphere, magmas, rocks, etc.) corrected for the non-ideal character of the gas. Oxygen fugacity is very important to the be-haviors of elements, but it is not a very precise term because in somecases, oxidation–reduction reactions do not involve any oxygen andthe oxygen fugacity changes with pressure and temperature (Sunet al., 2014a,b). “Redox state” is a better term to describe the relativeproportions of an element among its different oxidation states. Never-theless, oxygen fugacity is popularly used in Earth sciences.

Oxygen fugacity is usually notated as variations relative to a certainoxygen buffer. Oxygen buffer refers to an assemblage ofminerals or com-pounds that constrains oxygen fugacity as a function of temperature and

pressure (Fig. 4), i.e., the oxygen fugacity of equilibration at a fixed pres-sure is defined by one of the curves in oxygen fugacity versus tempera-ture diagrams. The concept of oxygen fugacity and oxygen fugacitybuffer have been well developed and widely used in high pressureexperiments.

Oxygen fugacity is of critical importance for the behaviors of ele-ments. Therefore, it is crucial to know the oxygen fugacity of a geochem-ical process and to control oxygen fugacity during experiments. Thereare several ways to control oxygen fugacity of experiments in the labo-ratory. Gas-mixing techniques have been commonly used for control-ling oxygen fugacity of experiments conducted at high temperature

image of Fig.�2

Fig. 3. Copper contents during partial melting of mantle peridotite under different oxygen fugacities modeled by Lee et al. (2012). (A) Variation of aggregated melt and residual mantlecomposition with degree of partial melting (F) at 2GPa, 1350 °C and fO2 at FMQ + 0, assuming initial S content of 200 ppm, i.e. 0.06 wt.% of sulfide. (B) Copper contents in primarymelt as a function of F and log fO2 (ΔFMQ); same P–T conditions as in (A). (C) Cu content of aggregate liquids versus fO2 for different F. Gray field refers to Cu in primitive MORB andarc magmas (Lee et al., 2012).

Fig. 4. Comparison of log fO2–T relationships for the Ni–NiO, Co–CoO, FMQ, MnO–Mn3O4,and HM oxygen buffers at 1 bar.The log fO2 of Co–CoO, FMQ and HM are taken from Chou (1978). The log fO2 of Ni–NiOand MnO–Mn3O4 are taken from Huebner and Sato (1970).


(above 1300 K) and 1 atm pressure (Huebner, 1987). The gas mixturesmostly used are CO2–CO and CO2–H2. At higher pressure, oxygen fugac-ity can be controlled by the oxygen buffer technique (Eugster, 1957)and the Shawmembrane (Shaw, 1963), and measured with the hydro-gen fugacity sensor technique (Chou, 1978). This oxygen buffer tech-nique was first developed by Eugster in 1957 to prevent oxidation ofiron during growth of hydrous ferrous silicates and then later to deter-mine the stability of annite (Eugster, 1957; Eugster and Wones, 1962;Wones and Eugster, 1965). The oxygen buffers commonly used inhigh pressure and high temperature experiments are fayalite–magnetite–quartz (FMQ), Co–CoO, Ni–NiO (NNO), hematite–magnetite(HM), andMnO–Mn3O4. These buffers are also useful as notations in themeasurement of the oxygen fugacity of natural rocks and magmas. Acomparison of the log fO2–T relationship for these buffers is shown inFig. 4. In experiments using the conventional double-capsule method,the mixture of buffer materials and water is loaded in the outer capsuleand the sample is sealed in the inner capsule. Hydrogen gas formsduring the reaction between buffermaterials andwater in the outer cap-sule, then diffuses through the inner capsule wall and equilibrates withthe hydrogen in the inner capsule. Capsuleswith high hydrogen diffusiv-ity are usually taken as the inner capsule (e.g., Pt), whereas thatwith lowhydrogen diffusivity is used as the outer capsule (e.g., Ag). The

image of Fig.�3

image of Fig.�4

Fig. 5. Pressure dependence of log fO2 and log fH2 of FMQ oxygen buffer. The log fO2 of Ni–NiO buffer is taken from Chou (1978). Method for calibrating log fH2 of FMQ buffer is thesame as that of the HM buffer.


prerequisite for using the double-capsule technique is that the buffermaterials themselves do not diffuse into the inner capsule. Therefore,they cannot be used for system with haplogranitic melts and H2SO4

solutions because Co and Ni diffuse into the inner capsule, resulting information of nickel sulfide or cobalt sulfide (Keppler, 2010).

The other way to regulate the oxygen fugacity of sample assem-blages is to control the hydrogen fugacity using the Shaw membrane,which introduces variable amount of hydrogen gas into the system(Shaw, 1963). The advantage of this technique is that the hydrogen fu-gacity can be varied independently and continuously. Moreover, theAg–AgCl–H2O–HCl acid buffer can be used as a fH2 sensor by loadingthe acid-buffer assemblage in a small Pt capsule and putting the capsulein a larger capsule containing the sample assemblages (Chou, 1978). Inthis case, the measured concentration of HCl in the sensor indicates thefH2 of the outer sample system.

2.1. Fayalite–magnetite–quartz (FMQ) oxygen buffer

The FMQ oxygen buffer is widely used in studies of natural samples,as well as for experiments. Given that the oxygen fugacity of most igne-ous rocks plot within several log units of the FMQ buffer, the oxygen fu-gacities of natural samples are often noted as log unit variations fromthe FMQ buffer, i.e., in ΔFMQ units. For example, ΔFMQ+ 2 means ox-ygen fugacity 2 log units higher than the values defined by the FMQbuffer. The FMQ equilibrium has been determined experimentally bymany researchers, e.g., Chou (1978), Myers and Eugster (1983) andEugster and Wones (1962). The oxygen fugacity is controlled throughreaction (1):

3Fe2SiO4 þ O2 ¼ 3SiO2 þ 2Fe3O4: ð1Þ

As shown in Fig. 5, the pressure dependence of log fO2 for this bufferis not significant, but log fH2 changes greatly from 1 bar to 10 kbar.

2.2. Hematite–magnetite (HM) oxygen buffer

The HM (hematite + magnetite) equilibrium has been determinedby several researchers (e.g., Chou, 1978; Eugster and Wones, 1962;Hemingway, 1990; Myers and Eugster, 1983). Oxygen fugacity is con-trolled through reaction (2) or (3):

4Fe3O4 þ O2 ¼ 6Fe2O3 ð2Þ

2Fe3O4 þ H2O ¼ 3Fe2O3 þ H2: ð3Þ

Eq. (3) better describes the system with water present. As shown inFig. 6, the pressure dependence of log fO2 is quite small, while log fH2 in-creases rapidly at low pressures but less at higher pressures. For highpressure experiments, the HMbuffer is more oxidized than the intrinsicoxygen fugacity of the hydrothermal vessel. Thus, hematite was con-sumed quite fast, e.g., in some experiments around 3.73 mg Fe2O3 wasconsumed in the buffer assemblage every hour under steady-state con-ditions, and 100mg of Fe2O3 lasted only 26.8 h (Chou, 1986). Therefore,the HM buffer is not suitable for experiments conducted in hydrother-mal vessels with water as the pressure medium. In natural systems,the oxygen fugacity of the HM buffer is usually much higher than thatof magmas. Most of the porphyry deposits, however, reach the HMbuffer value during mineralization.

2.3. Ni–NiO(NNO) oxygen buffer

The Ni–NiO oxygen buffer has been experimentally determined byHuebner and Sato (1970) and O'Neil and Pownceby (1993). The oxygenfugacity is controlled through reactions (4) and (5):

Ni þ 1=2O2 ¼ NiO ð4Þ

Ni þ H2O ¼ NiO þ H2: ð5Þ

As shown in Fig. 7, the pressure dependence of log fO2 is not signif-icant, but log fH2 increases rapidly at low pressures but less at higherpressures. The Ni–NiO buffer is commonly used in experiments, bring-ing with it distinct advantages. First, the oxygen fugacity of Ni–NiO isquite close to the intrinsic oxygen fugacity of the hydrothermal vessel(NiNiO + 0.5). Thus, even quite small amounts of Ni and NiO powdercould last for a long time during experiments, e.g., around 200 mg ofNi and NiO is enough for around a 10-day experiment. Moreover,oxygen fugacity controlled by the Ni–NiO buffer is quite importantfor the oxidation states of many elements, e.g., sulfur in the fluids ismostly as H2S at the oxygen fugacity of NNO, but SO2 and SO3 appearwhen the oxygen fugacity increases by 0.5 log unit (Binder andKeppler, 2011).

2.4. Pyrite + pyrrhotite + magnetite (PPM) oxygen buffer

The PPM (pyrite + pyrrhotite + magnetite) equilibrium controlsfS2 and fO2 through reactions (6)–(8):

2Fe1−xS þ ð1−2xÞS2 ¼ 2ð1−xÞFeS2 ð6Þ

6Fe1−xS þ 4O2 ¼ 2ð1−xÞFe3O4 þ 3S2 ð7Þ

3FeS2 þ 2O2 ¼ Fe3O4 þ 3S2: ð8Þ

image of Fig.�5

Fig. 6. Pressure dependence of log fO2 and log fH2 for the HM buffer. The log fO2 ofHM buffer is taken from Chou (1978). The log fH2 is calculated using log K = log fH2 +1/2log fO2, where K is the equilibrium constant of H2O = H2 + 1/2O2 determined bySupcrt92 with the database DPRONS2003.

Fig. 7. Pressure dependence of log fO2 and log fH2 (Ni–NiO oxygen buffer). The logfO2 ofNi–NiO buffer is taken from Huebner and Sato (1970). Method for calibrating log fH2 ofNi–NiO buffer is the same as that of HM buffer.


This buffer has been used in hydrothermal experiments, e.g., Spryand Scott (1986) and Crerar et al. (1978). The fO2 of the PPM bufferhas been determined by Kishima (1989) and Shi (1992), and their re-sults are in general agreement. Fig. 8. shows a comparison of the fO2–Trelationship of the PPM buffer with those of the NNO and HM oxygenbuffers. The log fO2 of the PPM oxygen buffer is located between thoseof NNO and HMbuffers, e.g., at 700 °C, the log fO2 of PPM oxygen bufferis around 1 log10 unit lower than that of theHMbuffer, but 3.3 log10 unithigher than that of theNNObuffer. As shown in Fig. 8b, log fO2 decreasesslightly with increasing pressure (Kishima, 1989; Shi, 1992).

At high temperature, the assemblage PPM is replaced by magnetiteand pyrrhotite (MPo) because pyrite is not stable (Sun et al., 2013a,b;Tomkins, 2010). Consequently, SO2 becomes the dominant species,and fH2, fH2O, and fH2S decrease abruptly. The breakdown temperatureof pyrite increases with increasing pressures, e.g., at 2 kbar, the break-down temperature is 745 °C, and increases to around 800 °C to10 kbar (Shi, 1992).

3. The association of porphyry deposits with oxidized magmas

It has long been proposed that most porphyry deposits are closelyassociated with oxidized magmas (Burnham and Ohmoto, 1980;Candela, 1992; Hedenquist and Lowenstern, 1994; Liang et al., 2006;Liang et al., 2009; Mungall, 2002; Sillitoe, 2010; Sun et al., 2012a,2013b), also known as the magnetite-series magmas (Ishihara andTerashima, 1989) (Fig. 9). These porphyry Cu–(Mo)–(Au) deposits are

almost always attributable to the injection of an oxidized adakiticmagma that is saturated with S- and metal-rich aqueous fluids,i.e., the parental magmas are water rich and oxidized (Ballard et al.,2002; Mungall, 2002; Sillitoe, 2010). Several small “reducing porphyrydeposits” constitute the rare exceptions (Cao et al., 2014; Rowins,2000; Sillitoe, 1999) (see detailed discussion in Section 4 below). Inter-growths of magnetite and hematite indicate that oxygen fugacities ofporphyry deposits often reach the values defined by the HM buffer(Fig. 10). In this section, we review the oxygen fugacities of a numberof famous deposits and then discuss the genetic connections betweenoxidizedmagmas and porphyry deposits. Themain questions discussedin this section include that: whether oxygen fugacity is a controllingfactor that dictates the distribution of porphyry deposits. What is thegenetic link between high oxygen fugacity magmas and porphyrydeposits?What is themost favorable oxygen fugacity range for porphyrymineralization? And how do oxygen fugacities change during porphyrymineralization?

3.1. Large porphyry deposits

The close association between highly oxidizedmagmas and porphy-ry deposits may best be illustrated by a FeO versus fO2 diagram (Fig. 9).The highly oxidized nature of porphyry magmas have been reported inessentially all types of porphyries (Mungall, 2002; Sillitoe, 2010; Sunet al., 2013b; Vila et al., 1991), including porphyry Cu and Cu–Audeposits (Sillitoe, 2010), porphyry Au deposits (Vila and Sillitoe, 1991;Vila et al., 1991), to porphyry Cu–Mo deposits (Cuadra and Rojas,2001; Lynch and Ortega, 1997; Stern et al., 2007) and pure porphyry

image of Fig.�6

image of Fig.�7

Fig. 8. Comparison of the log fO2 of PPM (pyrite–pyrrhotite–magnetite) buffer with thoseof Ni–NiO and hematite–magnetite buffer (A) and pressure dependence of the logfO2 ofPPM buffer.The log fO2 of PPM buffer is calibrated from Shi (1992).


Modeposits (Seedorff and Einaudi, 2004a,b). They are found in differenttectonic settings, ranging from continental arc (Sillitoe, 2010; Sternet al., 2007, 2011) to island arc (Hedenquist et al., 1998; Imai and

Fig. 9. Schematic plot of Fe content inmagmas versus oxidation state (fO2) for calc-alkalineto alkaline magmas associated with porphyry Cu, Cu–Mo and Mo deposits and W, Sn de-posits. The approximate boundary between magnetite- and ilmenite-series magmas(Ishihara, 1977) is also shown. Nevertheless, theHMbuffer is the upper limit of oxygen fu-gacity favorable for porphyry deposits.Modified after Thompson et al. (1999).

Ohno, 2005; Imai et al., 1993; Sun et al., 2013b), and fromactive subduc-tion zones to post-collisional zones (Hou et al., 2004, 2007b, 2009; Liet al., 2012a; Xiao et al., 2012).

One common indicator of the high oxygen fugacity of porphyry de-posits is sulfate, i.e., both magmatic and hydrothermal sulfate,e.g., hydrothermal anhydrite and magmatic anhydrite and gypsum areabundant in essentially all large porphyry deposits (Cooke et al., 2011;Zhang, Ling et al., 2013; Halter et al., 2005; Imai et al., 2007; Kavalieriset al., 2011; Li et al., 2008; Liang et al., 2009; Stern et al., 2007; Vilaet al., 1991). In addition to sulfate, hypogene hematite and specularitehave also been reported in many porphyry deposits (Baker et al.,1997; Hedenquist et al., 1998; Imai, 2001; Imai et al., 2007; Li et al.,2008; Seedorff and Einaudi, 2004b; Sillitoe, 2010; Spry et al., 1996;Vila and Sillitoe, 1991; Vila et al., 1991). The hematite–magnetite(HM) oxygen buffer has been taken as the upper limit of oxygen fugac-ities that are favorable for porphyry mineralization (Sun et al., 2013b).Previous authors also have proposed that the oxygen fugacities of differ-ent porphyry deposits are slightly different, in the order of: porphyry Cuand porphyry Au deposits N porphyry Cu–Mo deposits N porphyry Modeposits (Fig. 9) (Thompson et al., 1999).

3.1.1. Porphyry Cu and Au depositsPorphyry Cu and Au deposits almost do have systematically higher

oxygen fugacities and FeO contents compared to other porphyry de-posits (i.e., Cu–Mo and Mo). The best examples are the Cenozoic por-phyry Cu and Au deposits occurring in the southwestern Pacificislands, and to a less extent, the Paleozoic porphyry Cu–Au deposits inthe Central Asian Orogenic Belt.

From this one could surmise that porphyry Cu–Au deposits are asso-ciated with island arcs, whereas porphyry Cu–Mo deposits are associat-ed with continental arcs. For example, Cenozoic porphyry depositslocated in island arcs in the southwestern Pacific are all Cu–Au deposits,e.g., Grasberg and Batu Hijau in Indonesia; Panguna and Ok Tedi inPapua New Guinea; Lepanto–Far South East, Tampakan, Atlas andSipilay in the Philippines (Cooke et al., 2005). Some researchers haveeven argued that Cu comes from the mantle, whereas Mo comes fromthe continental crust (Mao et al., 2011). This is not the case, however, al-though porphyry Cu (±Au and ±Mo) deposits indeed do have higherεNd isotopic values (Hou et al., 2007a), implying less contribution fromthe continental crust. This observation, however, does not necessarilysupport amantle origin for Cu, either, because both the depletedmantleand the oceanic crust have high εNd values, whereas enriched mantleand the continental crust have low εNd values. As argued below(Section 3.2.1), a strong case can be made that most of the Cu comesfrom subducted oceanic crust. The distribution of porphyry depositsdoes not support the premise that the Cu comes from the mantle.Moreover, porphyry Cu–Au deposits are not always formed in islandarc environments, either. For example, there are also many porphyryCu–Au deposits (without economic levels of Mo) in the westernAmerican continents, including some of the world's top 25 largest por-phyry deposits, e.g., Cerro Casale in Chile; Minas Conga in Peru; Bajode la Alumbrera in Argentina; Pebble Copper in the USA; and Prosperityin Canada. Here we give brief introductions to three famous porphyryCu–Au deposits.

3.1.1.1. Grasberg, Indonesia.Grasberg in Irian Jaya, Indonesia is one of thelargest high-grade hypogene porphyry Cu–Au deposit (38.32 Mt Cu @1.12% and 3662 t Au @ 1.07 g/t) in the world, ranking no. 1 among por-phyry Au deposits and among the top 10 porphyry Cu deposits (Cookeet al., 2005). It was formed ~3 Ma ago. The presence of anhydrite andhematite indicates very high oxygen fugacities (Cooke et al., 2005;Mathur et al., 2000), near the HM buffer. This deposit is related to theclosure of a small backarc basin between Indonesia and the SouthChina Sea. As is characteristic of Cenozoic porphyry deposits in south-western Pacific, it contains no Mo.

image of Fig.�8

image of Fig.�9

Fig. 10. Images ofmagnetite–hematite intergrowths frommajor porphyry deposits in China, indicating that the oxygen fugacity reached the HMbuffer (Zhang, Ling et al., 2013). A. Dexing(Zhang, Ling et al., 2013), B. Xiongcun, C. Duobuza, D. Yulong (Sun et al., 2013a,b), E. Qulong, F. Zijinshan (Sun et al., 2013a,b).


3.1.1.2 . Far Southeast–Lepanto, Philippines. The Far Southeast–Lepantoporphyry and epithermal Cu–Au deposits, Philippines, are parts of alarge epithermal–porphyry Cu–Au body with a total of 5.48 Mt Cu @0.8 wt.% and 973 t Au @ 1.42 g/t (Table 1; Cooke et al., 2005;Hedenquist et al., 1998). Sulfates, including anhydrite and alunite (likelyrelated to the epithermal mineralization) are abundant in the minerali-zation system, and hematite formed during chlorite alteration, indicat-ing high oxygen fugacity up to the HM buffer (Hedenquist et al., 1998).

3.1.1.3. Santo Tomas II (Philex), Philippines. The Santo Tomas II (Philex)porphyry deposit, Philippines, has total reserves of 1.2 Mt Cu @0.33 wt.% and 233 t Au @ 0.64 g/t (Table 1). It is also associated witha highly oxidized magma, as indicated by anhydrite veinlets and a

Table 1Age, grade and tonnages of major porphyry deposits discussed in the text.

Deposit Province Ref

Cu–Au Grasberg Irian Jaya Cooke et al. (2005), Mutschler et al. (2010)Lepanto-FarSouth East

N. Luzon Cooke et al. (2005), Mutschler et al. (2010)

Santo Tomas II Philippines Mutschler et al. (2010)Cu–Mo El Teniente Central Chile Cooke et al. (2005), Mutschler et al. (2010)

Chuquicamata Nothern Chile Cooke et al. (2005), Mutschler et al. (2010)Dexing China Mutschler et al. (2010), Zhang, Ling et al. (201

Zhang, Sun et al. (2013)Qulong China Xiao et al. (2012)Yulong China Mutschler et al. (2010), Liang et al. (2006)Sar-Cheshmeh Iran Cooke et al. (2005), Mutschler et al. (2010)Batu Hijau Indonesia Mutschler et al. (2010)Panguna Bougainville Cooke et al. (2005)Ok Tedi PNG Mutschler et al. (2010)Tampakan Philippines Cooke et al. (2005)Atlas Philippines Cooke et al. (2005)

Mo Henderson United States Mutschler et al. (2010)Reduced Catface Canada Mutschler et al. (2010)Reduced Baogutu China Cao et al. (2014)

magnetite–titanohematite assemblage, which indicates oxygen fu-gacities near the HM buffer at nearly magmatic temperature (Imai,2001).

In summary, essentially, most if not all, porphyry Cu–Au deposits arehighly oxidized (Figs. 9, 10). In addition to those mentioned above, he-matite flakes in quartz veinlets have been reported in the Waisoi por-phyry Cu deposit (Namosi district), Viti Levu, Fiji (Imai et al., 2007).The Tongshankou porphyry skarn Cu–Au deposit, in the Lower YangtzeRiver belt, central eastern China, also has primary hematite next to sul-fides (Li et al., 2008). The euhedral characteristics of the hematite implya hydrothermal origin.

Porphyry Au deposits have oxygen fugacities similar to, if not higherthan, porphyry Cu deposits. For example, Marte, a large porphyry Au

Age Tonnage Augrade

Au Cugrade

Cu Mograde

Mo

(Ma) (Mt) (g/t) (t) (wt.%) (Mt) (wt.%) (Mt)

4–3 3409 1.07 3662 1.12 38.31.5–1.2 685 1.42 973 0.8 5.48

1 364 0.64 233 0.33 1.27.1–4.6 11,845 0.0035 437 0.92 109 0.02 2.5033.6 15,052 0.04 301 0.71 106 0.024 1.81

3), 170–148 1500 0.18 19 0.45 8.4 0.01 0.29

14 0.5 10.4 0.03 0.5Paleogene 850 0.84 7.14 0.03 0.1512 1200 0.27 324 1.20 14.4 0.03 0.365 1644 0.35 572 0.44 7.263.5 1415 0.57 799 0.46 6.511.2–1.1 700 0.64 446 0.64 4.483.3–2.2 1400 0.24 336 0.55 7.7061 1380 0.24 331 0.50 6.9030–27 727 0.17 1.24Middle Eocene 308 0.37 1.14 0.01 0.02

0.1 14 0.28 0.63 0.011 0.018

image of Fig.�10


deposit located in theMaricunga belt of the AndeanCordillera, northernChile, contains several percents of hematite, magnetite, anhydrite andgypsum (Vila et al., 1991), clearly reaching the HM buffer (Figs. 10, 11).

It has also been argued that iron oxides–copper–gold (known asIOCG) deposits are the low S version of porphyry Cu–Au deposits, con-trolled by secular changes in oceanic sulfate content and the geothermalgradients at the end of the Precambrian (Richards and Mumin, 2013).This is still elusive. Nevertheless, IOCG deposits are all highly oxidized,up to the HM buffer.

3.1.2. Porphyry Cu–Mo depositsIn general, porphyry Cu–Mo deposits have FeO contents systemati-

cally lower than porphyry Cu, but higher than pure porphyry Mo de-posits, with oxygen fugacities also falling between them (Fig. 8). Manyporphyry Cu–Mo deposits have high sulfate contents, with much lessmagnetite. No hypogene hematite has yet been reported in some ofthe supergiant porphyry Cu–Mo deposits, partially due to severe super-gene oxidation, which makes it difficult to identify primary hematite.

Porphyry Cu–Mo deposits have Cu and Au grades comparable tothose of porphyry Cu ± Au deposits, with low Mo grades, rangingfrom 0.01 to 0.03 wt.% (Table 1) (Cooke et al., 2005). Some of the super-giant deposits may have large Mo reserves of several million tons(Cooke et al., 2005). The Cu–Au and Mo mineralizations that occur inthe same ore bodies usually do not happen at the same time, indicatingdifferent sources and/or different mineralization processes. Consistent-ly, the Mo/Cu ratios of these deposits are generally more than 20 timeshigher than the primitive mantle ratio value (McDonough and Sun,1995), suggesting that Mo has been added to the porphyry depositsfrom different sources.

Oneof themost important geologic processes that enrichesMo is theoxidation–reduction cycle operating during chemical weathering at theEarth's surface, i.e., Mo is mobilized during oxidation and then enrichedin organic-rich sediments due to reduction (Li et al., 2012a,b). The

Fig. 11. Stability domains of the trisulfur ion S3−, sulfate, and sulfide in an aqueous solution,as a function of oxygen fugacity (log10 fO2) and acidity (pH=−log10mH+, inmol per kg)at 350 °C and 0.5 GPa illustrating sulfate reduction (modified after Pokrovski andDubrovinsky, 2011; Sun et al., 2013a,b). Also shown is the oxygen fugacity of the majormineral buffers (HM, thick horizontal orange line; NNO and FMQ, horizontal dashedlines) and theneutrality point of purewater (the vertical dashed line). The orangefield be-tween HM and ΔFMQ+2 is the optimum condition for porphyry Cu, Au, Mo mineraliza-tion (Sun et al., 2013b). Lines E1, E2, E3 and E4, show trajectories for sulfur reduction (Sunet al., 2013b). Stability domains of trisulfur ion S3− at total dissolved sulfur concentrationsof 1 wt.% and 0.1 wt.% are shown.

involvement of such reducing agents as organic matter seemingly ex-plains the system's lower oxygen fugacity. The most famous porphyryCu–Mo deposits are the Cenozoic ones located along the eastern Pacificmargin.

3.1.2.1. El Teniente. The supergiant El Teniente deposit in Chile is thelargest porphyry Cu–Mo–Au deposit in the world, with total reservesof ~109 Mt Cu @0.92%, 2.5 Mt Mo @ 0.02% and 437 t Au @0.035 g/t(Table 1; Cooke et al., 2005;Mutschler et al., 2010). The deposit spatiallycorresponds to the Juan Fernandez Ridge (Cooke et al., 2005), and thushas been attributed to slab melting during ridge subduction (Sun et al.,2010). Others argued that ridge subduction is not responsible for mak-ing the supergiant porphyry deposit based on the inferred migrationhistory of the arc (Kay et al., 2005). El Teniente is a nested porphyry sys-tem that, according to different authors, was active for ~1 Ma (Bakeret al., 2011; Cannell et al., 2003) up to more than 7 Ma (14.2–6.5 Ma)(Barra, 2011; Stern et al., 2011; Vry et al., 2010). High Fe2O3/FeO valuesof 1 to 3 (Garrido et al., 2002), quartz anhydrite veins and anhydrite-cemented breccias as well as gypsum clearly indicate high oxygen fu-gacities (Klemm et al., 2007; Vry et al., 2010). Although it is not spelledout, the high Fe2O3/FeO ratios of 1 to 3 (Garrido et al., 2002) indicateabundant hematite, reaching the HM buffer. Hydrothermal rutile alsoindicates high oxygen fugacity (NNO + 1.3) (Rabbia et al., 2009),which is however much lower than the HM buffer, though. Neverthe-less, the oxygen fugacity undoubtedly fluctuated during porphyry min-eralization (Liang et al., 2009; Sun et al., 2013b), and the hydrothermalrutile only records the oxygen fugacities under which it crystallized.

Most of the Cu in El Tenientewas emplaced during the latemagmat-ic stage (Cannell et al., 2005). A small sulfur- (N3 wt.%) and copper-rich(N0.5 wt.%) fine-grained igneous rock known as “Porphyry A” stock(b1 km3, 6.09 ± 0.18 Ma) (Stern et al., 2011) has abundant igneous an-hydrite, with varied textures ranging from interstitial to poikilitic andcorresponding modal abundances from 10 to 20%, respectively. Theseigneous anhydrite grains with planar crystal boundaries, occur alongwith fresh and unaltered biotite, feldspars, quartz, and Fe-oxides(Stern et al., 2007). Because the anhydrite-rich stock is isotopically sim-ilar to all the other igneous rocks of the late Miocene deposit (Bakeret al., 2011; Stern et al., 2007) formed at a time of regional compressivedeformation, it has beenproposed that anoxidized parentmagma in thelarge productive magma chamber was undergoing igneous fraction-ation at that time. During a period of recharge by mantle-derivedmafic magmas into the base of the chamber and associated volatiletransfer and concentration near its roof, the opportunity arose to pro-duce the Cu- and S-rich magmas that formed the anhydrite-bearing in-trusive rocks (Stern et al., 2007).

3.1.2.2. Chuquicamata. Chuquicamata is another supergiant porphyrydeposit in Chile, with a total reserve of ~106 Mt Cu @ 0.71%, 1.81 MtMo @ 0.024% and 301 t Au @ 0.04 g/t (Table 1. Cooke et al., 2005;Mutschler et al., 2010). Similar to El Teniente, the formation of theChuquicamata porphyry Cu deposits also lasted for several millionyears, between 36 and 31 Ma (Ballard et al., 2001; Ossandon et al.,2001). It is closely associated with highly oxidized adakite (Oyarzunet al., 2001, 2002). Because quantitative oxygen barometers based onFe–Ti oxides are prone to resetting, and primary whole rock Fe(III)/Fe(II) ratios and anhydrite, if originally present, are unlikely to havesurvived the hydrothermal alteration and surficial weathering of thisdeposit, the zircon Ce4+/Ce3+ ratio has been used to indicate its highoxygen fugacity (Fig. 12) (Ballard et al., 2002).

3.1.2.3. Dexing. The Dexing porphyry Cu deposit is located in southeast-ern China, with total reserves of 8.4Mt Cu @ 0.45%, 0.29MtMo @ 0.01%and 19 t Au@ 0.18 g/t (Table 1. Zhang, Ling et al., 2013; Zhu et al., 1983).The deposit consists of a cluster of three porphyries, with Tongchang asthe largest in the center, Fujiawu the second in the southeast andZhushahong, a small deposit in the northwest (Li et al., 2007; X.F. Li

image of Fig.�11

Fig. 12. Ce4+/Ce3+ and (Eu/Eu*)N ratios of zircon grains for Dexing and Shapinggouporphyry deposits as well as ore-bearing and ore-barren porphyries from Chile. HighCe4+/Ce3+ signifies high oxygen fugacity. Data of ore-bearing and ore-barren samples inChile from Ballard et al. (2002), Dexing from Zhang, Ling et al. (2013) and Shapinggoufrom Zhang, Li et al. (2013).


et al., 2013; Zhou et al., 2013). The oreminerals comprise pyrite, chalco-pyrite, molybdenite, minor tetrahedrite, bornite and chalcocite. Gangueminerals include quartz, muscovite, chlorite, calcite, minor epidote andanhydrite. The porphyry is an adakite formed at ~170 Ma (Zhang, Linget al., 2013;Wang et al., 2006a,b). Another poly-metal deposit, Yinshan,formed roughly at the same time and may be paragenetically related(Wang et al., 2013). The Dexing porphyry deposit is closely associatedwith highly oxidized magmas (Zhang, Ling et al., 2013; Li and Sasaki,2007), e.g., it is famous for abundant specular hematite, which mostlyformed during late stage hydrothermal alteration (Fig. 13). Inter-growths of magnetite and hematite indicate its oxygen fugacity reachedthe magnetite–hematite buffer (Fig. 11). Hematite is also found in fluidinclusions in Dexing (Liu et al., 2011).

3.1.2.4. Qulong. The Qulong porphyry Cu–Mo deposit is now the largestporphyry-type deposit in China, with reserves of 10.4 Mt Cu @ 0.5%and 0.5 Mt Mo @ 0.03% (Table 1. Xiao et al., 2012). It is located in theGangdese orogenic belt in southern Tibet. In addition to abundant hy-drothermal anhydrite of up to 10% or more, magmatic anhydrite isalso reported in unaltered granodiorite porphyry. These anhydrite oc-currences indicate that the Qulong magmatic–hydrothermal systemwas highly oxidized and sulfur-rich, with abundant sulfates (Xiaoet al., 2012; Yang et al., 2009) (Fig. 14). Given that it formed at~14 Ma, which post-date the initial collision between Indian and

Fig. 13. An image of specularite from the Tongchang deposit in the Dexing porphyry Cudeposits. Specularite cutting across carbonate veins, indicating that it was formed verylate, likely through hydrothermal alteration.

Eurasian continents, Qulong is taken as a typical post-collisional por-phyry deposit. Nevertheless, the Indian plate is still subducting north-ward. It is not clear whether the portion of the subducting slabunderneath Qulong was continental or oceanic. Intergrowths of hema-tite and magnetite (Fig. 10) indicate the oxygen fugacity of Qulongreached the HM buffer. Adakites found in the Qulong deposit are mix-tures of continental and slab melts (Sun et al., 2012a).

3.1.2.5. Yulong. The Yulong porphyry Cu–Au deposit belt is distributedalong the northwestern extension of the Red River–Ailao Shan fault sys-tem, at the eastern margin of the Tibetan Plateau (Hou et al., 2007b;Jiang et al., 2006; Liang et al., 2006), and covers an area of ~300 kmlong and ~20 km wide. Five major porphyries, which contain most ofthe Cu reserves so far discovered in the belt, are located in a narrow,elongated domain of approximately 50 km long and 10 km wide andare closely associated with Cenozoic high potassium intrusive rocks(Liang et al., 2006, 2007). The Yulong porphyry is the largest one inthe Yulong copper deposit belt, with reserves of N7.14 Mt of Cu @0.84% and 0.15 Mt of Mo @ 0.028% (Table 1; Liang et al., 2006;Mutschler et al., 2010). All five known porphyry deposits all togethercontain a total ofmore than 8 million tons of Cu andMo reserves. ZirconU–Pb dating shows that the formation ages of deposits within theYulong ore belt range from 41.2 to 36.9 Ma, extending over a period ofsimilar to 4.3 Ma, with formation ages decreasing systematically fromnorthwest to southeast (Liang et al., 2006). Zircon grains from theYulong ore-bearing porphyries have higher Ce4+/Ce3+ values thanthose from barren porphyries in the region (Liang et al., 2006). Abun-dant magnetite and hematite suggest that the ore-bearing porphyriesare highly oxidized, reaching the HM buffer (Figs. 10, 11) (Sun et al.,2013b).

Fig. 14. Images of magmatic anhydrite from the Qulong giant porphyry deposit in theGangdese porphyry belt, south Tibetan Plateau. Bi = biotite; Anh = anhydrite; Kf =K-feldspar; Py = pyrite; Cpy = chalcopyrite; Ap = apatite.

image of Fig.�12

image of Fig.�13

image of Fig.�14


3.1.2.6. Sar-Cheshmeh. The Sar-Cheshmeh is one of the top 20 largestporphyry Cu–Au–Mo deposit in the world with reserves of 14.4 Mt Cu@ 1.2%, 0.36 Mt Mo @ 0.03%, and 324 t Au @ 0.27 g/t (Cooke et al.,2005). It is located in southwestern Iran and is associated with severalintrusive pulses of Miocene stocks (~12.2 Ma) (Mutschler et al., 2010),ranging in composition from diorite through granodiorite to quartz–monzonite (Hezarkhani, 2006). Molybdenum enrichment and deposi-tion took place before Cu. Anhydrite is a popular mineral identified inmost veins (Hezarkhani, 2006). Hematite has been reported, but it oc-curs as a secondary supergene ore forming mineral (Shahabpour,1991). The Sar-Cheshmeh and several other giant Miocene porphyrydeposits of Iran and Pakistan are located in a region undergoing conti-nent–continent collision. The geodynamic and architectural controlson porphyry formation in this complex tectonic zone are unclear(Cooke et al., 2005). The ore-forming magmas in this section of theTethys margin have high Sr contents (N500 ppm) and Sr/Y (N50)(Haschke et al., 2010), which are clearly adakitic (Defant andDrummond, 1990). Previous studies suggested that these magmas area product of continental arc-style magmatism (Cooke et al., 2005).They were attributed to either the attempted subduction of the Arabianplate beneath the Eurasian plate, or the change from subduction ofoceanic to continental crust (Cooke et al., 2005), or partial melting of afertile copper- and sulfur-enriched arc crustal keel (Haschke et al.,2010). Further studies are needed on the formation of Sar-Cheshmehand other Cenozoic porphyry deposits in the belt.

3.1.3. Porphyry Mo depositsPorphyry Mo deposits generally have higher SiO2 and lower FeO

contents. It also has been proposed that porphyryModeposits have sys-tematically lower oxygen fugacities than porphyry Cu (Au) and Cu–Modeposits (Fig. 9). There are threemajor porphyryMo deposit belts in theworld, the Henderson–Climax (Klemm et al., 2008; Pettke et al., 2010;Seedorff and Einaudi, 2004a,b; Singer, 2008), Qinling–Dabie (Chen,2013; Chen et al., 2000; Li et al., 2012a; Mao et al., 2008; Stein et al.,1997; Yang et al., 2013; Zeng et al., 2013; Zhang, Li et al., 2013) andXing'an–Mongolia belts (Wu et al., 2011; Y. Zhang et al., 2013; Zenget al., 2013). The oxygen fugacity can also reach the HM buffer duringsulfate reduction and mineralization (Fig. 10), likely because of thelower FeO contents.

3.1.3.1. Henderson.Henderson porphyryMo deposit is located in the Cli-max–Henderson Mo belt, Colorado. It consists of 12 Oligocene rhyoliticstocks in three centers, Henderson (oldest), Seriate, and Vasquez(deepest and youngest), at depths of ~1 km below the surface at theRed Mountain (Seedorff and Einaudi, 2004a), with a reserve of1.24 Mt Mo @ 0.17 wt.% (Table 1). Previous studies proposed that, inthe Henderson porphyry Mo deposit, Fe was leached at high to moder-ately high temperatures and then fixed in the rock at lower tempera-tures, first mainly as magnetite and then as pyrite and minorpyrrhotite and specular hematite (Seedorff and Einaudi, 2004a). Otherporphyry deposits in the Climax–Henderson belt also contain hematite,indicating these deposits reached the HM buffer.

3.1.3.2. Shapinggou. The Shapinggou porphyry Mo deposit is the largestClimax-type Mo deposit in the world, with total proven Mo reservesof over 2.2 million metric tons @ 0.17 wt.% (Zhang, Li et al., 2013). It islocated in the western Dabie Mountains, along the east extension ofthe East Qinling Mo mineralization belt (Chen, 2013; Chen et al., 2013;Gao et al., 2010; H.Y. Li et al., 2012; Han et al., 2013; N. Li et al., 2013;Stein et al., 1997; Yang et al., 2013; Zeng et al., 2013; Zhu et al., 2010),to the north of the Triassic suture between the north and south Chinablocks (Sun et al., 2002). Both Re–Os isochron and U–Pb zircon datinggive the same age of ~111 Ma (Zhang, Li et al., 2013), which assigns itto the third mineralization pulse of the East Qinling Mo belt (Li et al.,2012a; Mao et al., 2008, 2011). High zircon Ce4+/Ce3+ ratios indicate

very high oxygen fugacity, comparable to those of Chuquicamata(Fig. 12) (Zhang, Li et al., 2013).

3.2. Linkage between oxidized magmas and porphyry deposits

Copper, Au and Mo are all chalcophile elements, the behaviors ofwhich aremainly controlled by reduced sulfur, e.g., sulfide, hydrosulfidecomplexes (Sun et al., 2004a), or polysulfide (e.g., S22−, S3−) complexes(Sun et al., 2013b). Although most geologists agree that porphyry de-posits are closely associated with oxidized magmas (Sillitoe, 2010;Sun et al., 2013b), it is still hotly debated as regarding to why oxidizedmagmas favor porphyry mineralization and how oxidized the ore-forming magmas should be. Major disagreements include: (1) What isthe most favorable oxygen fugacity for porphyry mineralization?Someworkers proposed thatΔFMQ+2 is amagic number for porphyrymineralization (Mungall, 2002), whereas ΔFMQ + 2 to +4 is the mostfavorable range of oxygen fugacity for porphyry deposits (Sun et al.,2013b). Yet others claimed that variations in oxidation state over typicalranges for arc magmas formed during plate subduction (ΔFMQ = 0to +2) have no major effects on the potential to form porphyry Cu ±Au ± Mo deposits during plate subduction (Richards, 2011b). It iseven argued that some porphyry deposits formed in reduced magmas(Cao et al., 2014; Rowins, 2000; Smith et al., 2012). (2) What is themain sulfur species in porphyry? Some workers concluded that sulfateis the predominant sulfur species in ore-forming porphyries (Cookeet al., 2011; Field et al., 2005; Liang et al., 2009; Sotnikov et al., 2004;Sun et al., 2013a,b), whereas others argue that the main sulfur speciesis SO2 dissolved in porphyry magmas (Richards, 2014; Smith et al.,2012), i.e., neither sulfate nor sulfide. (3) Is sulfide saturation duringmagma evolution important for porphyry mineralization? Someworkers proposed that sulfide undersaturation is important for porphy-ry mineralization, i.e., no residual sulfide remains in the source (Li et al.,2012b; Liang et al., 2009; Sun et al., 2013b), whereas others argued thatsulfide saturation and accumulation are of critical importance for por-phyry mineralization (Wilkinson, 2013).

In general, the porphyry mineralization process consists of 3 majorphases: (1) source, i.e., extraction of chalcophile elements from thesource; (2) transportation and concentration of ore-forming elements;and (3) deposition and fixation of the ore forming elements intoorebodies. The relationship between oxidized magmas and porphyrydeposits during these three steps are discussed below.

3.2.1. Sulfur oxidation, sulfide under saturation and residual sulfideCopper, Au and Mo are all chalcophile elements, with high partition

coefficients between sulfide and melts, e.g., DCu = 1334 ± 210 (Pattenet al., 2013), DAu = 4500–11,200 (Mungall and Brenan, 2014), andDMo = 0.15–5.15 (Li and Audetat, 2013), respectively. Some experi-ments suggest that DMo increases with decreasing oxygen fugacities(Li and Audetat, 2013). Therefore, the behaviors of Cu and Au arestrongly controlled by sulfide, whereasMo is much less sensitive to sul-fide, especially at high oxygen fugacities. During mantle melting, Cu aswell as Au and Mo are all moderately incompatible elements, with par-tition coefficients ranging between those of Yb and Ce (McDonough andSun, 1995; Sun et al., 2003a,b, 2004a). The estimated Cu abundance inthe primitive mantle is 30 ppm (McDonough and Sun, 1995), whereasthe Cu concentrations in MORB so far published range from 70 to150 ppm (Hofmann, 1988; Sun et al., 2003b).When there is no residualsulfide in themantle source, Cu and also other chalcophile elements be-comehighly incompatible (Lee et al., 2012; Liu et al., 2014). Partialmelt-ing of mantle peridotite may form melts with high Cu contents up to350 ppm at high oxygen fugacities, at very low degree of partial melting(Fig. 3) (Lee et al., 2012). Nevertheless, most arc magmas formed atmuch N10% partial melting, resulting in much lower Cu contents(e.g., 100 ppm), which is not favorable than reducing magmas for por-phyry mineralization. In contrast to mantle peridotite, MORB has sulfurabundances of about 1000 ppm, the effects of oxygen fugacity on

Fig. 15. Sulfur speciation curve vs. fO2 for basaltic glasses based on the S6+/ΣS estimates ofXANES (in bold), EPMA (dashed line) (Jugo, 2009; Jugo et al., 2005, 2010). Also shown arefields of different tectonic settings, (A) Japan arc and Mexican arc; (B) MORB, OIB andmantle wedge, and arc. MORB = mid-ocean ridge basalt; IAB = Island arc basalt;BABB = Back arc basin basalt; OIB = Oceanic island basalt.


residual sulfides, and, consequently, the Cu content in the melts, be-comes much more obvious when considering partial melting ofsubducted oceanic slabs.

The S abundance in the primitive mantle is estimated to be about200–250 ppm (McDonough and Sun, 1995), and about 150 ppm in de-pleted mantle (Lorand, 1990; Mavrogenes and O'Neill, 1999). Mainly,the sulfur speciation inmagmas is controlled by oxygen fugacity. Sulfideis the predominant sulfur species at oxygen fugacities lower than theFMQ buffer. Sulfate proportions start to climb up above the FMQ buffer.Most of the sulfur in the magmas is present as sulfate at ΔFMQ + 2(Fig. 11) (Jugo, 2009; Jugo et al., 2005, 2010). Experiments show thatsulfate ismuchmore soluble than sulfide inmagmas. The sulfur contentat sulfide saturation (SCSS) increases solely as a function of increasingfO2 (Jugo, 2009):

SCSS ¼ S2–h i

1þ exp 2:23ΔFMQ–2:89ð Þð Þ ð9Þ

and

ΔFMQC ¼ 1:29þ 0:45 ln S6þh i

– ln S2–h i� �

ð10Þ

where ΔFMQC refers to the critical fO2 for simultaneous saturation ofsulfide and sulfate (Jugo, 2009).

The predicted SCSS content at sulfide saturation in basalts rangesfrom 1300 ppm at ΔFMQ − 1 and 1500 ppm at ΔFMQ + 0.5, to7500 ppm at ΔFMQ + 2 and 1.4 wt.% at ΔFMQ + 2.3 (Fig. 15) (Jugo,2009). At oxygen fugacities of ΔFMQ 0 to +2.5, the more oxidizingthe system, the higher the S contents in the magmas, and thus high ox-ygen fugacity is the most efficient way to eliminate residual sulfides(Lee et al., 2012; Sun et al., 2013b). The solubility of sulfide in magmasis independent of oxygen fugacity below the FMQ buffer (Mavrogenesand O'Neill, 1999), implying that most of the sulfur is removed in theform of sulfate under high oxygen fugacity (Jugo, 2009), andwith its re-moval releasing much more of the chalcophile elements (Lee et al.,2012; Liang et al., 2009; Sun et al., 2004b).

The average S concentration in MORB is ~1000 ppm (O'Neill andMavrogenes, 2002), while laboratory experimental measurementshave determined the S content at sulfide saturation in basalts to be1300 ppmat oxygen fugacities ofΔFMQ0 (Jugo, 2009). The small differ-ence between the average MORB and experimental values is likely dueto slight differences in oxygen fugacities and pressure effects, i.e., sulfideis less soluble under higher pressures (Mavrogenes and O'Neill, 1999).Given that most MORB forms through ~10% partial melting of thedepleted MORB mantle, it extracts about 100 ppm of sulfur fromthemantle source, leaving behind ~150 ppm of sulfur as residual sul-fide. In case all the sulfur is present as sulfide, ~25% partial melting isneeded to eliminate all the residual sulfides from the mantle sourceat ΔFMQ 0 (Fig. 3) (Lee et al., 2012). Such large degrees of partialmelting would dilute the released Cu and produce picrite or evenkomatiite, not basalt.

For the mantle wedge above subducting slabs, the estimated S con-tent at sulfide saturation ranges from 1500 ppm (at ΔFMQ + 0.4) to4500 ppm (at ΔFMQ + 1.7) (Jugo, 2009). As shown in Fig. 16, thehighest oxygen fugacity in arc magmas may be NΔFMQ + 3, whichallows up to 1.4 wt.% of S at sulfide saturation (Jugo, 2009). At oxygenfugacities below ΔFMQ + 1, 10% or less partial melting of the mantlewedge allows preservation of residual sulfides, assuming that the Sabundances in the mantle wedge is also 250 ppm. At higher oxygen fu-gacity, no residual sulfides can be expected unless there is additionalsulfur added from the subducting slab, which in actuality is probablythe case. During metamorphism of the subducting slab, S is mobilefrom metamorphic rocks ranging in grade from blueschist to amphibo-lite (Sun et al., 2013a; Tomkins, 2010), such that they can release S andchalcophile elements to the mantle wedge (J.L. Li et al., 2013).

Nevertheless, normal arc rocks are not favorable materials for por-phyry Cu mineralization. Arc magmas are probably neither unusuallyoxidized nor enriched in economic elements of interest, such as Cu(Lee et al., 2010; Lee et al., 2012). Instead, most porphyry Cu depositsare closely associated with adakite (Oyarzun et al., 2001; Sajona andMaury, 1998; Sun et al., 2010, 2012a; Thieblemont et al., 1997), or socalled high Sr/Y porphyries (Chiaradia et al., 2012; Richards, 2011a).Most of the ore forming adakites were formed through slab melting(Sun et al., 2011, 2012a), deriving from the subducting oceanic crustitself rather than the overlying mantle wedge. MORB has higher Cucontents (70–150 ppm) (Hofmann, 1988; Sun et al., 2003b) thanother portions of the subducted oceanic slab, and thus is likely themain contributor to mineralization. Therefore, partial melting ofsubducted oceanic slabs forms magmas with high Cu contents (Sunet al., 2011, 2012b), which plausibly explains the close associationsbetween adakites, especially formed during ridge subductions, and por-phyry Cu deposits. These papers, however, did not consider the effectsof oxygen fugacity and dramatically underestimated the contributionsof slab melting.

image of Fig.�15

Fig. 16.Oxygen fugacity of different tectonic settings. Note the range of oxygen fugacity forcontinental peridotite ismuch less than that of Bryant et al. (2007), after removing perido-tite samples from arc settings. The oxygen fugacities of convergent margin magmas aresystematically higher than those of intraplate settings. Magmas from intraplate settingswithout influence from plate subduction are too reduced for mineralization (Sun et al.,2013b).Modified after Bryant et al. (2007).


Slabmelts formed at oxygen fugacities higher thanΔFMQ+2 are fa-vorable for porphyry Cumineralization.MORB has an average S contentof ~1000 ppm (O'Neill and Mavrogenes, 2002), from which it is easy toget the 22% and 10% partial melting needed to eliminate residual sul-fides at ΔFMQ + 1.7 and ΔFMQ + 2, respectively. Correspondingly,the Cu contents of suchmeltswould be 450 and 1000 ppm, respectively.Slab melts formed at ΔFMQ + 2 would have Cu contents more thantwice as high as those formed atΔFMQ+1.7, and thus are considerablymore favorable for Cu porphyry mineralization than mantle wedgemelts. Only four times greater enrichments of Cu are needed to reacheconomic porphyry deposit proportions of ~4000 ppm. This ought tobe easily achieved throughmagma evolution plus hydrothermal pro-cesses. At oxygen fugacities of ΔFMQ+ 2.3 or higher, only 5% partialmelting is needed to remove all the residual sulfides, forming meltswith initial Cu contents of ~2000 ppm. In contrast, unrealistic N50%partial melting is needed to eliminate all residual sulfides at oxygenfugacities lower than ΔFMQ + 1, corresponding to a Cu content ofless than 200 ppm in the melt. Residual sulfide would exist untilthe entire slab melted, in which case sulfide would remain the pre-dominant sulfur species in both melt and slab (under reducing con-ditions of bΔFMQ 0).

In order to form supergiant porphyry deposits, e.g., El Teniente, up to600 km3 of magma with a Cu content of 100 ppm would be needed toget the necessary amount of metal (Stern et al., 2011). Note that most

porphyry deposits have very small surface exposures, ranging in areafrom less than 1 km2 (mostly) to ~5 km2 (in rare cases). Thus, themagma chamber would need to be unrealistically thick, e.g., thickerthan the continental crust. Alternatively, the magma must scavengelaterally for distances over 10 km and concentrates the Cu at one rel-atively small spot. None of these scenarios are practical. For adakiticmelts with initial Cu contents of ~1000 ppm, the amount of magmaneeded measures only 60 km3 for the supergiant El Teniente. Amagma chamber extending to a depth of several kilometers issufficient.

Slab melts need to pass through the mantle lithosphere before theyeventually are emplaced in the upper crust, and thusmay lose or collectchalcophile elements by interactions with themantle. Sulfide saturatedmelts would lose some chalcophile elements. For slab melts formed atmelting degrees higher than that require for SCSS, theywould be under-saturated in sulfide and thus could assimilate more sulfides andchalcophile elements from the surroundingmantlewhile ascending, de-pending on the partitioning relationships between melt and residualsulfide in the mantle. For example, 10% partial melting of MORB witha S content of 1000 ppm atΔFMQ N 2.3, would form amelt with a S con-tent of ~1 wt.%, which is only half of its SCSS. Thismelt thenwould havethe capacity to assimilate large amounts of sulfide. This plausibly ex-plains the common association between high-Mg adakites and porphy-ry Cu deposits.

Meanwhile, sulfide is kept undersaturated during the evolutionof the oxidized magma, such that no sulfide segregation occurs,and Cu, Mo, and Au act as moderately incompatible elements.They would become enriched at an early stage of magma evolution,and dissolve into fluids following reduction accompanying magne-tite crystallization (Sun et al., 2004a, 2013b), or other reductionprocesses.

It has long been proposed that oxygen fugacities higher thanΔFMQ+2 have reached the magic number for porphyry mineralization(Mungall, 2002; Sun et al., 2013b). As discussed above, this is exactlythe point atwhich residual sulfide is eliminated by ~10% partial meltingof a subducted slab, forming melts with initial Cu contents up to1000 ppm. In contrast, at oxygen fugacity even only slightly lower,e.g., ΔFMQ + 1.7, the highest Cu contents that can be reached throughslab melting is ~450 ppm.

The systematically low oxygen fugacities in Japan arc volcanic rocks(bΔFMQ + 2) compared to those from the western American conti-nents (up to ΔFMQ + 3) (Fig. 16), is an important factor controllingthe distribution of Cu porphyry deposits, i.e., abundant porphyry Cu de-posits are located in the western American continents, whereas noneoccur in Japan (Fig. 1) (Sun et al., 2012b, 2013b).

3.2.2. Sulfate reductionAt oxygen fugacities higher than ΔFMQ + 2, most of the sulfur in

magmas is present as sulfate, and suchmagmas usually have higher con-tents of total S and chalcophile elements. Copper however, is achalcophile element, and its final deposition during mineralizationshould be controlled mainly by the behavior of reduced sulfur (Lianget al., 2009; Sun et al., 2004a, 2013b). Therefore, the final stage of miner-alization inevitably requires the reduction of sulfate (S6+: HSO4

−/SO42−)

in the oxidized source magmas to sulfides (S2−: H2S/HS−/S2−) orpolysulfides (e.g., S22−, S3−) (Sun et al., 2013b).

In principle, there are several ways to lower the oxygen fugacitiesduring magma evolution. Among them are degassing of oxidized vola-tile species (i.e. CO2, SO3), assimilation of reducing country rocks(Ishihara andMatsuhisa, 1999; Smith et al., 2012), and reduction of sul-fate by other elements, e.g., fractional crystallization of magnetite andeven hematite (Liang et al., 2009; Sun et al., 2004a, 2013b).

It has also been proposed that degassing of SO2 may lower the oxy-gen fugacity (Kelley and Cottrell, 2012). As illustrated in Eq. (11),degassing of SO2 indeed consumes oxygen, such that it may lower theoxygen fugacity of the system.However, degassing of SO2 cannot reduce

image of Fig.�16

Fig. 18. Ferric/ferrous Fe ratios and estimated oxygen fugacities of Manus glasses (Sunet al., 2004a). Oxygen fugacity of Manus glasses was calculated using the following equa-

tion: ln XFe2O3XFeO

� �¼ a lnƒO2 þ b

T þ cþ ΣdiXi, where, a = 0.196, b = 1.1492 × 104, c =

−6.675, dAl2O3 = −2.243, dFeO = −1.828, dCaO = 3.201, dNa2O = 5.854, dK2O =6.215 (Kress and Carmichael, 1991).


sulfate to sulfide in themagmas. Instead, it drives the reaction (Eq. (11))to the right, which leads to the oxidation of sulfide.

2H2S þ 3O2 ¼ 2H2O þ 2SO2 ð11Þ

Degassing of CO2 may reduce sulfate to sulfide, through a processsimilar to that described by Eq. (12).

2C þ SO3 þ H2O ¼ 2CO2 þ H2S ð12Þ

As shown in Fig. 17, however, the C–CO2 buffer is lower than the SSObuffer (Mungall, 2002), therefore there is little C in the system at SSObuffer. Moreover, degassing of any type is not likely to be a key processduring porphyry mineralization, which is usually occurring at depths of~2–4 km (Sillitoe, 2010).

Assimilation of reducing sediments may lead to sulfate reduction asproposed by previous authors (Shen and Pan, 2013; Smith et al., 2012)(also see Section 4). Nevertheless, assimilation occurring during theemplacement of porphyries would result in heterogeneous reduction,because reduction is mainly focused along the porphyry–country rockinterface. This kind of reduction process has been reported for Baogutu(Shen and Pan, 2013), which was classified as a reduced porphyry Cudeposit associated with ilmenite-series magmas (Cao et al., 2014) (seedetailed discussion in Section 4.3). Assimilation duringmagma chamberdevelopmentmay be better mixed andmore homogenized. However, itwould result in early sulfide saturation and segregation from the melt.Because sulfide is denser than silicate melt, it would tend to sink,which is more likely to occur during the mineralization of Cu–Nideposits, rather than porphyry deposits.

Elements that can exist in variable oxidation states and that are pres-ent in sufficient abundances to affect the redox state of the silicate Earthare C, H, S, and Fe (Mungall, 2002). In highly oxidized porphyrymagmas, sulfate, CO2 and H2O are the dominant species, which are al-ready oxidized and thus cannot reduce sulfate to sulfide, leaving ferrousiron as the only probable reducing agent (Sun et al., 2013b). Iron is amajor element, with about 20 to 30% of the total Fe in a typical porphyryoccurring as ferrous iron, corresponding to an oxygen fugacity of lessthan ΔFMQ+ 2 (Fig. 18). Magnetite contains 66.7% ferric iron, whereashematite contains 100% ferric iron. Therefore, the crystallization ofmag-netite and hematite tends to lower the proportions of ferric iron in themagma, which apparently would lower the oxygen fugacity. In thecase of the Manus deposit in volcanic rocks, the Fe3+/(total Fe) didnot changemuch duringmagnetite crystallization (Fig. 18). This was at-tributed to sulfate buffering (Sun et al., 2004a), i.e., the oxidation of

Fig. 17. Sulfate reduction and oxygen buffers. CCO= carbon dioxide–carbon oxide buffer;SSO = sulfide–sulfur oxide buffer.After Mungall (2002) and Sun et al. (2014a,b).

ferrous iron is coupled with the reduction of sulfate to sulfide(Eqs. (13), (14)). Magnetite crystallization at Manus and many otherarc volcanic rock series is coupled with dramatic decreases in Cu andAu (Sun et al., 2004a), which is an important factor in understandingAu and Cu mineralization (Sun et al., 2004a). The reduction in theamount of Cu and Au during magma evolution is well-known in arcmagmas (Moss et al., 2001; Sun et al., 2011; Togashi and Terashima,1997), and has been referred to as the magnetite crisis (Jenner et al.,2010). Indeed, magnetite crystallization/alteration is often taken asthe controlling process for porphyry Cu and Au mineralization by caus-ing sulfate reduction and presumably also oxygen fugacity fluctuations(Liang et al., 2009).

There are several ways to describe the oxidation of ferrous Fe. Inmagma system, this is better described as oxidation of FeO by sulfateor another oxidant (Eq. (13)) (Sun et al., 2004a):

SO2−4 þ 8FeO ¼ 4Fe2O3 þ S

2−: ð13Þ

Eq. (13) is simplified, Fe2O3 represents the ferric Fe in Fe3O4. This re-action can be more precisely described by Eq. (14):

SO2−4 þ 12FeO ¼ 4Fe3O4 þ S

2−: ð14Þ

These reactions have no effects on the pH value of the system. This isone reason why the oxygen fugacity of the Manus magma did notchange much (Sun et al., 2004a), or even decrease slightly during thecrystallization of magnetite and sulfate reduction (Jenner et al., 2010)(see also Section 5).

3.2.3. Hematite–magnetite intergrowthInterestingly, in addition to abundant magnetite in porphyry de-

posits (Astudillo et al., 2010; Audetat et al., 2004; Baker et al., 1997;

image of Fig.�17

image of Fig.�18


Dilles et al., 2011; Imai, 2005; Liu et al., 2011; Yang et al., 2002), hema-tite and specularite (a hydrothermal variety of hematite) are commonin porphyry copper deposits (Figs. 10, 13) (Zhang, Ling et al., 2013;Sillitoe, 2010; Sun et al., 2013b). Asmentioned previously, primitive he-matite–magnetite intergrowths have been reported in many porphyryCu deposits all over the world (Fig. 10), including nearly all major por-phyry deposits in China (Sun et al., 2013b), as well as some depositsin South America (Ballard et al., 2002; Patricio and Gonzalo, 2001; Vilaand Sillitoe, 1991) and the Southwest Pacific islands (Hedenquistet al., 1998; Imai et al., 2007), and Mongolia (Khashgerel et al., 2008).Most of the hematite and/or specularite formed during the late stagesof porphyry mineralization (Sillitoe, 2010; Sun et al., 2013b).

The occurrence of primitive hematite in closely association withmagnetite strongly suggests a very high oxygen fugacity (Fig. 11),such that reaching the HM oxygen fugacity buffer (equivalent to~ΔFMQ + 4) during porphyry Cu mineralization. This suggests that, incontrast to hydrothermal systems in the Manus backarc basin (Jenneret al., 2010; Sun et al., 2004a), the oxygen fugacity of porphyriesincreases during magnetite crystallization (Sun et al., 2013b).

The further oxidation of magnetite is controlled by the pH values ofthe system (Sun et al., 2013b). In contrast to themagma system, ferrousFe is mainly present as Fe2+ in aqueous fluids. In this case, the oxidationof ferrous Fe is better described by Eq. (15):

SO2−4 þ 12Fe

2þ þ 12H2O ¼ 4Fe3O4 þ S2− þ 24H

þ: ð15Þ

As shown in Eqs. (2) and (1), HM and FMQ oxygen fugacity buffersdo not change with pH. The oxidation potential of sulfate, however, de-pends strongly on pH (Figs. 11). The oxidation of ferrous iron by sulfateleads to a decreases in pH in ore-forming fluids within the porphyrysystem (Eq. (15)), which drives up the oxidation potential of sulfate.Consequently, the fO2 increases during the reduction of sulfate(Fig. 11) (Pokrovski and Dubrovinsky, 2011; Sun et al., 2013b). It is es-timated that the amount of H+, released during the reduction of sulfateand oxidation of ferrous iron,may lower thepHdown to+2, dependingon temperatures, and if there are no pH buffers in the system to signif-icantly elevate the oxygen fugacity (Sun et al., 2014a,b) and oxidizemagnetite or ferrous Fe in hydrothermal fluids to hematite and/orspecularite (Eq. (16)).

SO2−4 þ 8Fe

2þ þ 8H2O ¼ 4Fe2O3 þ S2− þ 16H

þ ð16Þ

Or further oxidation of magnetite (Eq. (17)):

SO2−4 þ 8Fe3O4 ¼ 12Fe2O3 þ S

2−: ð17Þ

The formation of hydrothermal hematite will further lower the pH,whereas further oxidation of magnetite will not change the pH. All thesulfate reduction reactions provide S2−, which promotes mineraliza-tion, forming chalcopyrite (Eq. (18)) and more pyrite (Eq. (19))(Heinrich, 1990) and releasing H2. Hydrogen may further react withCO2 and/or O2 (Eqs. (20), (21)) or escape from the porphyry systemduring degassing or diffusive loss (Sun et al., 2014a,b).

Cuþ þ Fe

2þ þ 2H2S ¼ CuFeS2 þ 3Hþ þ §H2 ð18Þ

Fe2þ þ 2H2S ¼ FeS2 þ 2H

þ þ H2 ð19Þ

2H2 þ O2 ¼ 2H2O ð20Þ

4H2 þ CO2 ¼ CH4 þ 2H2O ð21Þ

The porphyry mineralization process requires a continuous reduc-tion of sulfate to sulfide, coupled with consumption of oxygen andeven reduction of CO2 to CH4. Given that porphyry magmas are highlyoxidized, reaction (21) is not likely to occur. No significant amounts ofmethane have been detected in most cases, except in reduced porphy-ries (see details in Section 4).

Based on high pressure experiments, it was proposed that the tri-sulfur ion S3− is an important sulfur species (Fig. 11) at pressures andtemperatures (Pokrovski and Dubrovinsky, 2011) that are commonfor porphyry systems (Hedenquist and Lowenstern, 1994; Seo et al.,2009; Sillitoe, 2010; Sun et al., 2014a,b). It has been further proposedthat in general geological processes where the tri-sulfur ion may havebeen involved need to be reconsidered (Pokrovski and Dubrovinsky,2011), and this would require the investigation of different reactionpathways (Sun et al., 2013b). In fluids, the reactions can be describedby Eqs. (22) and (23)

6SO2−4 þ 52H2O þ 57Fe

2þ ¼ 2S−3 þ 19Fe3O4 þ 104H

þ ð22Þ

2S−3 þ 20H2O þ 15Fe

2þ ¼ 6S2− þ 5Fe3O4 þ 40H

þ: ð23Þ

Both of these reactions (Eqs. (22), (23)) also would lower the pH ofthefluids, which in turnwould elevate the oxidation potential of sulfate.Such conditions will promote the formation of hematite and would befavorable for porphyry mineralization. The oxidation potential of theSO4

2−–S3− reaction also rises with decreasing pH (Fig. 11), so that mag-netite may be further oxidized to hematite by SO4

2−, releasing OH−

(Eq. (24)). This in turn increases the pH andpromotesmagnetite forma-tion (Sun et al., 2013b), which may explain magnetite and hematite in-tergrowths (Fig. 10). In addition to direct oxidation of magnetite,hematite and specularite may form directly from fluids (Eq. (25)), re-leasing H+, which lowers pH and promotes further oxidation of ferrousFe. Pure hydrothermal hematite and specularite can often be deter-mined to have formed during the late stages of porphyrymineralization(Sillitoe, 2010), e.g., the specularite occurring in late stage veins atDexing (Fig. 13).

38Fe3O4 þ 6SO2−4 þ 5H2O ¼ 57Fe2O3 þ 2S

−3 þ 10OH

− ð24Þ

38Fe2þ þ 6SO

2−4 þ 33H2O ¼ 19Fe2O3 þ 2S

−3 þ 66H

þ ð25Þ

In magmas, in contrast to fluids, FeO is more abundant than Fe2+. IfS3− is indeed stable as proposed (Pokrovski and Dubrovinsky, 2011),then the reduction of sulfate in magmas first leads to an increase inpH (Eq. (26)), which is compensated by further reduction of S3−

(Eq. (27)). Therefore, S3−may indeed causemore fluctuations in oxygenfugacities through a different reaction pathway, but it has no major in-fluences on the final results (Sun et al., 2014a,b).

6SO2−4 þ 5H2O þ 57FeO ¼ 2S

−3 þ 19Fe3O4 þ 10OH

− ð26Þ

2S−3 þ 5H2O þ 15FeO ¼ 6S

2− þ 5Fe3O4 þ 10Hþ ð27Þ

For large porphyry deposits, the reduction of sulfate inevitably re-sults in the oxidation of magnetite to hematite as the pH is lowered(Figs. 10, 11). Therefore, the hematite–magnetite intergrowths maybe taken as a possible ore indicator in future prospecting for copperdeposits.

3.3. Summary

Sulfur is one of themost important geosolvents that controls the be-havior of copper and other chalcophile elements, therefore it is essential

Fig. 19. Diagram of log fO2 versus temperature and 1000/T illustrating the oxygenfugacities of (A) oxidized porphyry deposits and (B) reduced deposits, Catface andBaogutu. The systematic difference in oxygen fugacities is likely not primary. Baogutu ishosted in ilmenite-series diorites, containing native antimony (An and Zhu, 2010),hypogenepyrrhotite, and methane-rich fluid inclusions. Previous studies suggested thatassimilation occurring during the emplacement of porphyries resulted in reduction,which ismainly focused along the porphyry–country rock interface (Shen and Pan, 2013).Modified after Oyarzun et al. (2001) and Cao et al. (2014) and Smith et al. (2012),respectively.


to the understanding ofmineralization processes of copper and a varietyof other metal resources. Sulfate is the dominant sulfur mineral speciesin porphyries associated with large copper deposits. This is because theoxidation of sulfide to sulfate during partial melting is essential for theefficient extraction of chalcophile elements out of the source region,especially from subducted oceanic slabs, which have more than5 times more sulfur than the mantle. In contrast, metals of porphyryCu deposits are hosted in sulfides, which require low oxygen fugacityfor their stabilization during the final stage of mineralization. Therefore,the key process of porphyrymineralization is oxidation and reduction ofsulfur, controlled by ferrous/ferric Fe and pH values. Sulfate reduction inhydrothermal fluids lowers the pH and consequently elevates the oxy-gen fugacity of the system up to the HMbuffer. The acid released duringsulfate reduction causes alterations. In contrast, sulfate reduction inmagmas does not change the pH, such that the oxygen fugacity is slight-ly reduced. All types of porphyry deposits, ranging from Cu + Au,Cu + Mo, and Mo deposits, may be oxidized sufficiently to reach theredox states of the HM buffer.

4. The association of porphyry deposits with reduced magmas

Although most of the porphyry deposits are closely associated withoxidized magmas, a small group of porphyry deposits are reported asapparently related to reduced magmas with oxygen fugacities rangingfrom ΔFMQ − 0.5 to −3 (Fig. 19) (Cao et al., 2014; Rowins, 2000;Smith et al., 2012). Reduced porphyry deposits so far reported are: 17Mile Hill, Western Australia; the Minãs de San Anton, Mexico (Rowins,2000); the Baogutu, northwestern China (Cao et al., 2014; Shen andPan, 2013; Shen et al., 2010a, 2012); North Fork Deposit, west CentralCascades, western North America (Smithson and Rowins, 2005);Catface; Vancouver Island, British Columbia (Smith et al., 2012); and afew other small porphyry deposits (Rowins, 2000). Three criteria havebeen proposed for the classification of reduced porphyry Cu deposits:(1) lack of primary hematite and sulfate minerals, (2) rich in hypogenepyrrhotite and CH4, and (3) oxides belongs to the ilmenite-series,reduced, I-type granitoids (Rowins, 2000).

Instead of primary hematite, magnetite, and sulfate (e.g., anhydrite,gypsum) that are frequently seen in oxidized porphyries, these“reduced” porphyry Cu–Au deposits contain abundant hypogene pyr-rhotite, and commonly have carbonic-rich ore fluids with substantialproportions of CH4, clearly indicating a reducing condition (Cao et al.,2014; Rowins, 2000; Smith et al., 2012). More importantly, they are as-sociated with ilmenite-bearing, reduced I-type granitoids (Cao et al.,2014; Rowins, 2000), in direct contrast to oxidized porphyries that arerelated to a magnetite-series mineralogy (Ishihara and Sasaki, 1991;Thompson et al., 1999). For example, ilmenite is an end-member ofthe hematite–ilmenite solid solutions with a very low hematite compo-nent, varying from XHem (0.01–0.1), for the Catface phase and XHem

(0.04–0.09) for the Hecate Bay phase (Smith et al., 2012). Magnetite isnearly pure Fe3O4 with less than 2% Fe2TiO4 component for Catface(Smith et al., 2012) and Baogutu (An and Zhu, 2010; Cao et al., 2014;Shen and Pan, 2013; Shen et al., 2010a,b). Ilmenite is more abundantthan magnetite in the reduced porphyries (Cao et al., 2014; Smithet al., 2012), e.g., all intrusive phases at Catface have accessory FeTioxides with ilmenite/magnetite ratios of ~9:1 (Smith et al., 2012), indi-cating a relatively reduced oxidation state, i.e., belonging to theilmenite-series (Ishihara, 2004). Experiments with apatite have shownthat SO3 contents reflect the oxygen fugacity of the magmas (Penget al., 1997). The uniform and low SO3 contents in apatite from reducedmagma deposits were used to argue that the low oxygen fugacity ofthese porphyries is primary (Fig. 20) (Cao et al., 2014; Smith et al.,2012).

The reduced porphyry deposits range in age from the Late Archeanto the Oligocene (Rowins, 2000). Although the tonnages of these re-duced porphyry deposits are generallymuch smaller than oxidized por-phyry deposits, the mechanism that controls these mineralization

process needs to be clarified. Some major questions to be answeredare: How reduced magmas formed in an arc environment, where oxi-dizedmagmas are common, e.g., in western North America? In contrast,why are there no porphyry deposits in Japan, where the ilmenite-seriesis well-developed (Ishihara and Murakami, 2006)? How do Cu and Auget enriched in reducedmagmas?What's the relation between oxidizedand reduced porphyries?

4.1. Reduced magmas

Arc magmas usually have high oxygen fugacities (Fig. 16) (Arculus,1994; Ballhaus, 1993; Carmichael, 1991; Kelley and Cottrell, 2009; Sunet al., 2012a, 2013a,b), a condition that has been attributed to a variety

image of Fig.�19

Fig. 20. Diagram of Cl versus SO3 in apatite indicating the oxygen fugacities of Catface andBaogutu. The SO3 contents of Catface apatite are all lower than 0.15 wt.%, corresponding tooxygen fugacities below the NNO buffer. In contrast, the SO3 contents of Baogutu apatiterange from less than 0.1 wt.% to much higher than 0.2 wt.%, corresponding to variedoxygen fugacities, ranging from below the NNO buffer to close to the NNO +1. HDP =hornblende diorite porphyry; GDP = Granodiorite porphyry; D = diorite.Modified after Cao et al. (2014) and Smith et al. (2012).


of mechanisms. The addition of subduction released oxidizing compo-nents to the mantle wedge or directly to arc magmas is proposed asthe most straightforward way, e.g., water (Kelley and Cottrell, 2009),oxidizing fluids and/or melts (Brandon and Draper, 1996), fluids withoxidizing components, e.g., hematite, sulfate (X.M. Sun et al., 2007),slab melts with high Fe3+/Fe2+ratios (Mungall, 2002), as well as otheroxidized components, e.g., carbonates, or other oxidized sediments.Arc magmas may also get oxidized during their evolution and ascentdue to the following mechanisms (Ballhaus, 1993; Lee et al., 2005,2010): (1) Changing oxygen buffers from graphite CO2 (CCO) equilibriain the mantle source to Fe3+–Fe2+ equilibria after graphite has beeneliminated by partial melting (Ballhaus, 1993); (2) Fractional crystalli-zation of olivine and other minerals that favors ferrous iron over ferriciron (Carmichael, 1991); (3) Assimilation of oxidizing country rocks(Lee et al., 2005); (4) Degassing of reduced volatile species (e.g., H2,H2S and CH4) (Ballhaus, 1993; Lee et al., 2005); (5) Changing pH(e.g., lowering pH in a sulfate–sulfide system) (Sun et al., 2013a,b);(6) Magma recharging which preferentially enriches ferric Fe (Leeet al., in press).

There are also several mechanisms that, in principle, may lower theoxygen fugacity of magmas: (1) Addition of reducing sediments fromplate subduction (Takagi, 2004); (2) Assimilation of reducing countryrocks (Ishihara and Matsuhisa, 1999; Smith et al., 2012); (3) Degassingof oxidized volatile species (i.e. CO2, SO3); (4) Addition of subduction re-leased reducing fluids (Song et al., 2009); and (5) Fractional crystalliza-tion of magnetite and even hematite (Liang et al., 2009; Sun et al.,2004a, 2013a,b). As discussed previously (Section 3.2.2), the magnetitecrystallization reduces the oxygen fugacity only in magmatic systems(reactions (13) and (14)), and may only lower the oxygen fugacity tothe FMQoxygen buffer. Herewe discuss two reduced porphyry depositsin detail.

4.1.1. Evidence for reduced magmas of Catface porphyry depositCatface is the largest reduced porphyry Cu deposit so far reported,

emplaced at 40.4–41 Ma, with an indicated reserve of 56.9 Mt @ 0.4%Cu and an additional inferred resource of 262.4 million tons @ 0.38%Cu. Three potential mechanisms for the formation of Catface reducingmagmas had been discussed before degassingwas assigned as the caus-ative process: (1) upwelling of the reducing asthenospheric mantle in-duced by the opening of a slab window during ridge subduction,

(2) evolvement of subducted reducing sediments, and (3) assimilationof reduced sediments duringmagma emplacement (Smith et al., 2012).

Previous studies suggested that a slab window introduces hot, up-welling asthenospheric mantle in the subduction zone environment,forming non-arc-like alkalic and adakitic magmatism in the volcanicarc (Abratis and Worner, 2001; Groome and Thorkelson, 2009; H. Liet al., 2011, 2012; Kinoshita, 1997). The Kula–Farallon ridge in thenortheastern Pacific began descending under theNorth American conti-nent in Alaska in the Late Cretaceous (Madsen et al., 2006; Scharmanet al., 2012), and then between 62 and 11 Ma migrated southward(Cole and Stewart, 2009; Cole et al., 2006; Liu et al., 2008), forming asemi-continuous forearc magmatic belt from Alaska to Oregon(Madsen et al., 2006). Ridge subduction forms oxidized adakites byslab melting (Defant and Drummond, 1990; Ling et al., 2009, 2013),followed by reduced A-type granites (Abratis and Worner, 2001; H. Liet al., 2012; Thorkelson and Breitsprecher, 2005; Yogodzinski et al.,1994). Interestingly, the Catface porphyry deposit is closely associatedwithMtWashington,which is an adakite (Smith et al., 2012) and seem-ingly would support the slab window model. Nevertheless, adakitesformed by slab melting have much higher initial Cu contents thanmantle-derived melts, and thus are favorable for porphyry mineraliza-tion (Sun et al., 2010, 2011), suggesting that the reduced Catface por-phyry could be a country rock hosting the ore deposit, rather than thecausative porphyry.

Subduction of reducing sediments also reduces the oxygen fugacity ofarc magmas, as seen in the Japan arc (Takagi, 2004). Carbon rich sedi-ments in the descending plate react with H2O to produce CH4 and CO2

(Ballhaus, 1993; Takagi, 2004), or directly releases CH4 by devolatilization(Song et al., 2009). This process, however, was excluded from furtherconsideration based on the low 87Sr/86Sr of Catface (Smith et al.,2012). Methane, however, may decouple from Sr, such that Sr isotopescannot give a conclusive answer. Nevertheless, the oceanic platesubducting underneath Japan is Cretaceous in age, which has experi-enced eight Ocean Anoxic Events (Jenkyns, 2010), and thus containsabundant organic rich sediments. In contrast, the oceanic platessubducting underneath the western North American continent areyounger and have experienced only oneOceanAnoxic Event.Moreover,ocean ridges are young and generally sediment-poor. Therefore, sub-duction of reducing sediments is not a plausible explanation for thelow fO2 shown by the Catface porphyry.

The third mechanism proposed for the reduced condition of theCatface porphyrywas by assimilation of reduced sediments. The Catfaceporphyrymay have ascended through and interactedwith graphite-richreducing country rock known to exist in the region (Smith et al., 2012).The reduction of ferric Fe through reaction with reducing materials,e.g., graphite, in sediments was proposed as a potential factor responsi-ble for the low fO2 of the parental Catface magma. This mechanismwasexcluded based on Sr isotopes (see details below) (Smith et al., 2012).

The authors then proposed that the reduced Catface porphyry wasdue to degassing of SO2 from the magma. This argument is based onthe low S content in the rock's apatite and a recent work by Kelleyand Cottrell (2012), which proposed that fractional crystallizationcoupled with degassing of S in a sub volcanic arc magma chamber canresult in a reduction of N2 log fO2 units in themagma. This calculated re-duction was based on the rock's ferric iron content and S in melt inclu-sions, and would hold true for magmas varying in composition frombasaltic andesite to dacite (Kelley and Cottrell, 2012). Degassing ofSO2 indeed has a significant influence on the oxygen fugacity in volcanicsystems. It is, however, not likely to be amajor process in plutons,whichshould experience much less degassing.

Moreover, apatite from the Catface has a homogenous S contentthroughout the crystal (Smith et al., 2012), which argues againstmajor degassing induced S loss during magma evolution. Experimentson felsic liquids showed that the SO3 content in apatite increasewith in-creasing fO2 from 0.04wt.% at the FMQ buffer, up to N1.0wt.% at the HMbuffer (Peng et al., 1997). As pointed out by Smith et al. (2012), apatite

image of Fig.�20


crystallizes below 950 °C in silicic calc-alkaline magmas (Green andWatson, 1982). If the magma at Catface was reduced during degassingof SO3, there should be compositional zonation in the apatite, which isnot the case (Fig. 20) (Smith et al., 2012).

4.1.2. Evidence for reduced magma in the Baogutu porphyry depositBaogutu is located inWest Junggar, Central Asian Orogenic Belt, and

is the second largest reduced porphyry Cu deposit so far reported,with controlled reserves of 0.63 Mt Cu @ 0.28 wt.%, 14 t Au @0.1 ppm, 1.8 × 104 t Mo @ 0.011 wt.% and 390 t Ag @ 1.8 ppm (Caoet al., 2014). It contains native antimony (An and Zhu, 2010), abundanthypogenepyrrhotite, andmethane-richfluid inclusions. The low apatiteSO3 content, whole rock Fe2O3/FeO, and fluid compositions indicate alow fO2 of ~NNO for the magma and NNO–NNO − 2 for associatedhydrothermal fluid (Cao et al., 2014; Shen et al., 2010b).

Based on H–O and sulfide S–Pb isotope data, it has been argued thatthe methane-rich ore-forming fluids were derived from a deep mantlesource with little contamination from penetrated sediments (Caoet al., 2014). The assimilation model, however, does not contradictwith a deep mantle source for the S, Pb, nor Cu. In fact, most Cu shouldhave come from slabmelts (Sun et al., 2011, 2012b), as also should havecome the S and Pb, which are present in abundances similar to mantlederived magmas. The H isotopes of fluid inclusions from Baogutu havevery low δD values (Shen et al., 2012; Zhang et al., 2010), which aremuch lower than for either magmatic fluids or metamorphic fluids(Fig. 21). This was interpreted as the main evidence for degassing(Cao et al., 2014). Degassing, however, should result in heavier hydro-gen and oxygen isotopes in the residualmagmas. The authors used pub-lished heavy H values from circum-Pacific volcanic gases (Giggenbach,1992b) to argue that degassing should leave behind fluids with evenlighter H than contained in the magmatic and metamorphic fluids(Cao et al., 2014). The heavy H of circum-Pacific volcanic gases, howev-er, was originally explained as having arisen from the addition of seawa-ter to the gases (Giggenbach, 1992a). Basic principles of isotopegeochemistry dictate instead that in degassing, i.e., evaporating, the re-maining liquid becomes isotopically heavier in H and O. In addition,

Fig. 21. δD versus δ18O diagram and calculated isotopic composition of waters in hydro-thermal fluids derived from measured isotopic composition of quartz and its fluid inclu-sion for the Baogutu deposit from Cao et al. (2014). Reference lines and boxes are asfollow: meteoric water line (Craig, 1961), felsic magmatic water (Taylor, 1992), residualwater in intrusion after degassing and crystallization (Taylor, 1974), low-salinity vapordischarges from high-temperature volcanic fumaroles (Giggenbach, 1992b), primarymagmatic water area, metamorphic water area and multiple formation water area(Hoefs, 2004). H–O isotopic data are from (Zhang et al., 2010) and (Shen et al., 2012).Low-salinity vapor shows strong seawater signals, which cannot represent “degassing”fluids.

degassing cannot explain the decoupling between δD and O isotopecompositions. Moreover, the degassing of porphyries should be muchless pronounced than the degassing of volcanic rocks.

The systematically lighter H coupled with a slightly lighter O canbest be explained by the addition ofmeteoriticwater. Baogutu is locatedin the center of the Eurasian continent. Meteoriticwater there should bevery light in H and O isotope composition. Addition of such watershould dramatically reduce the δD of the magma with a much lessereffect on O becausemagmas have low hydrogen contents and abundantoxygen contents.

4.2. Source of copper

For oxidized magmas, as discussed in Section 3, excess sulfur is pre-cipitated in the form of sulfate, leaving behind less residual sulfide, suchthat Cu, Au and other chalcophile elements become enriched inmagmas(Lee et al., 2012; Sun et al., 2012a, 2013a,b). In contrast, reducedmagmas usually retain residual sulfides, such that they have low prima-ry Cu contents (Fig. 3) (Lee et al., 2012). Then, how do reducedmagmasget enriched in Cu?

As pointed out by previous authors, porphyry deposits associatedwith reducedmagmas are generally small (Cao et al., 2014). Thiswas at-tributed to the originally low Cu and Au contents of the magmas, lessmagmatic fluids released due to deep emplacement, or tectonic settingsthat are not favorable for porphyry mineralization (Cao et al., 2014).They also argued that significantly lower sulfur solubility in reducedmelt (Jugo, 2009; Jugo et al., 2005, 2010) keeps S2− as the dominantsulfur species, which potentially isolates sulfides from the magma dur-ing its migration to the site of final emplacement, thus producing onlyrelatively small chalcophile endowments (Cao et al., 2014). None ofthese arguments are convincing to us. First, lower Cu and Au contentswould first result in lower grade, not necessarily smaller tonnages.The grades of porphyry Cu and Au deposits associated with reducedmagmas are comparable to, if not higher than, those associated withoxidized magmas (Cooke et al., 2005; Rowins, 2000). Second, there isno systematic difference in terms of emplacement depths and tectonicsettings between reduced and oxidized porphyry deposits. More impor-tantly, compared to oxidized magmas, reduced magmas indeed havelower sulfur contents with S2− occurring as the dominant sulfur species(Jugo et al., 2010), but this does not necessarily mean higher S2− in re-duced magmas as claimed by Cao et al. (2014). Given that the solubilityof sulfide is independent of oxygen fugacity, but increaseswith decreas-ing pressure under reducing conditions (Mavrogenes and O'Neill,1999), the high proportions of S2− cannot “isolate sulfides from themagma during migration” as proposed.

Based on a synthesis of theoretical, experimental, and field data, ithas been proposed that Cu and Au can be transported via the vaporphase to distal sites as far as several kilometers away from the causativeporphyry due to fluid boiling or immiscible phase separation. Conse-quently, the source porphyry becomes a low-grade sub-economic Cu–Au core or failed porphyry Cu system (Rowins, 2000), which is moreor less similar to an epithermal ore system. Experiments find that themain transporting agents of Cu at the porphyry level are brines andthat models based on transporting copper in the vapor phase are incor-rect (Lerchbaumer and Audetat, 2012). It is further demonstrated, usingexperimental studies, that brine–vapor separation in porphyry depositsdoes not cause selective Cu transfer to the vapor, but is more likely todestabilize Cu complexes and promote copper ore deposition duringde-compression and unmixing of the two fluid phases. In contrast, Au maybe selectively transferred into the vapor phase, allowing its transportfrom the deeper porphyry copper deposits to form shallowerepithermal gold deposits (Seo and Heinrich, 2013). This explains theAu-rich capping feature of many reduced porphyry deposits. Mean-while, Cu may be transported to distal sites through normal fluids.Such a reduced porphyry Cu–Aumineralization model does not contra-dict the current understanding of porphyry Cu–Au formation. In fact,

image of Fig.�21


the recognition of reduced porphyry Cu–Au systems encourages asearch for distal sites that are favorable for focusing and precipitatingAu and Cu-rich vapors (Rowins, 2000). This implies that such porphyrydeposits themselves are not genetically related to reduced magmas,i.e., they are not the causative magma, but the host rock.

4.3. Formation of reduced porphyry deposits

Porphyry deposits related to reduced magmas have two other dis-tinct characteristics. They are associatedwith carbonic-rich ore formingfluids (Cao et al., 2014; Rowins, 2000; Smith et al., 2012), and multipleintrusive events (Cao et al., 2014; Rowins, 2000; Smith et al., 2012). Inaddition to the transportation of Cu and Au from the causal porphyryto distal sites of deposition (Rowins, 2000), we propose two otherways that may form reduced porphyry deposits: (1) by reduction ofoxidized magmas during their ascent in the crust, and (2) by using re-duced magmatic rocks that are present only as country rocks to hostore deposits originating from underlying oxidized magmas.

4.3.1. The formation of the Catface porphyry depositAs discussed above (Section 3.2.2), assimilation of reduced sedi-

ments may have reduced the oxygen fugacity of the Catface porphyry,a possibility that was excluded based on Sr isotopes. The 87Sr/86Srratio of the Catface intrusions is 0.704, whereas those of the graphite-rich metasediments of the Pacific Rim terrane range from 0.706 to0.708 (Smith et al., 2012). We find such small differences in Sr isotopesto be negligible, and thus do not argue strongly against the assimilationmodel. Graphite and especially methane need not be consideredcoupled with silicate minerals, such that the reducing action of thegraphite and methane is not proportionally related to the assimilationof silicates nor sulfides, thereby decoupling them from the Sr, Pb, andS isotopes. Relationship between Sr and C aside, the amount of C neededis very small. The total iron (expressed as Fe2O3) content ranges from2.25 wt.% to 5.53 wt.% for the Catface porphyry. One carbon atom re-duces 4 ferric Fe atoms (Eq. (28)), assuming all the Fe in the parentalCatface magmas was ferric Fe, and all the ferric Fe was reduced bygraphite, then only 0.04 to 0.1 wt.% of graphite is required, which mayonly have had a limited influence on Sr isotopes. More interestingly,the Catface intrusions have fO2 values nearly identical to the C–CO–CO2 buffer at similar pressure and temperature conditions (Smithet al., 2012), whichwould support the assimilation of graphite-rich sed-iments.

C þ 2Fe2O3 ¼ 4FeO þ CO2 ð28Þ

Reduced sediments usually also contain methane (Cao et al., 2014),which is a more efficient reductant, i.e., 1 methane molecule reduces 8ferric Fe atoms (Eq. (29)), such that only ~0.02 to 0.05 wt.% of methaneis needed. Therefore, the parental Catface magmas may have acquired alow fO2 during their ascent through the crust. The homogenous SO3 inapatite may simply be due to an early reduction of the magmas, beforeapatite crystallization.

CH4 þ 4Fe2O3 ¼ 8FeO þ CO2 þ 2H2O ð29Þ

Considering that the Catface porphyry is not adakitic, it may well besimply the country rock that hosts the deposit. The Catface porphyrydeposit is closely associated with a nearby pluton on Mt Washington(35–41Ma), which is an adakite (Smith et al., 2012). Themineralization(40.9 Ma) and the emplacement age of the Catface porphyry (40.4–41 Ma) (Smith et al., 2012) are both within the age range of the MtWashington adakite. Adakites along the eastern Pacific rim are mostlyformed by slab melting (Liu et al., 2010; Sun et al., 2012a,b), which arelikely to have high initial Cu contents, and are thus favorable for porphy-rymineralization (Sun et al., 2010, 2011). In contrast, the Cu contents ofthe asthenosphere (~30 ppm) (McDonough and Sun, 1995) and thecontinental crust (~27 ppm) (Rudnick and Gao, 2003) are much lower

than MORB (~100 ppm) (Sun et al., 2003b), such that melt derivedfrom them should have much lower Cu contents and be less favorablefor porphyry Cumineralization. Therefore, the reduced Catface porphy-ry is likely to be a country rock that only hosts the deposit.

Based on the above discussion, we propose the following model forthe formation of the Catface porphyry. Intrusions in the VancouverIsland range in age from 51 to 35.5 Ma (Madsen et al., 2006), spanningthe time during which the Kula–Farallon ridge collided with the conti-nent. Accompanying the oblique subduction of the ridge (Madsenet al., 2006), its two limbs separated, with a slab window opened in be-tween. The first limb formed early through partial melting of the hotsubducting plate. The resulting adakites may then have been reducedthrough reaction with the thick (maximum thickness of 4 km)carbon-rich Cretaceous Nanaimo Group sediments (Madsen et al.,2006) as they migrated upward. Consequently, sulfate is reduced tosulfide, leaving behind Cu-rich sulfide accumulations in the lowercrust. This is followed by emplacement of the Catface porphyry at 41–40.4 Ma (Smith et al., 2012), which took place at a time that the slabwindow was open. The mantle derived parent magmas of the Catfaceporphyry are expected to have been more reduced (near the FMQoxygen fugacity buffer) and drier than the adakite, and to have alower initial Cu contents. Nevertheless, it became wetter and evenmore reduced (ΔFMQ − 0.3 to −3) with a higher Cu content afterassimilating the Cu-rich sulfide accumulations in the lower crust.Mean-while, ridge subduction induced compression and consequent upliftand erosion occurred, which favored the exposure of the porphyry de-posit. Oxidized Mt Washington adakite was emplaced at 41–35.3 Maduring the subduction of the west limb of the ridge, bringing more oreforming fluids into the Catface porphyry. We infer that large propor-tions of these adakites are still buried due to the dramatically lessenedcompression, uplifting and erosion following ridge subduction.

4.3.2. The formation of the Baogutu porphyry depositBased on previously obtained fluid inclusion H–O isotope data and

sulfide S–Pb isotope data, it was proposed that the methane-rich oreforming fluids in the Baogutu porphyry deposit were derived from adeepmantle sourcewith little contamination from sedimentary compo-nents (Cao et al., 2014). As discussed above (Section 4.1.2, Fig. 21), theauthors' understanding about H–O isotopes is in error, whereas onlysmall amount of C is enough to lower the oxygen fugacity of theBaogutulow Femagmas. Therefore, S–Pb isotopes cannot place any decisive con-straints on the role of assimilation in forming the deposit. As pointed outby the authors, detailed studies are needed to clarify the origin of theCH4.

More importantly, there are three intrusive phases at Baogutu,which from old to young are: (1) themain diorite phase, (2) dikes of di-orite porphyry and granodiorite porphyry intruding the early dioritestock, and (3) dikes of hornblende diorite porphyry intruding all thethree phases. All the samples with the exception of two were collectedfrom the diorite, which are not porphyry at all. The reduced dioritemagma may have no relation with the porphyry mineralization, exceptto act as a host rock.

We propose that the reduced features of the Baogutu porphyry de-posit are secondary and occurred during emplacement, thus having nomajor influence on the mineralization process. A similar model hasbeen proposed based on geochemical and mineralogical studies byShen and Pan (2013), in which mineral composition data suggest thatthe primary magma of the Baogutu porphyry deposit is oxidized. Theheterogeneous and reduced characteristics of the deposit are attributedto significant country-rock contamination after emplacement (Shenand Pan, 2013) in an arc setting (Shen et al., 2013a,b).

Similar to Catface, there are also adakites in Baogutu. The Baogutuadakites have been attributed to the mixing of ~95% slab melt with~5% sediment-derived melt in the Late Carboniferous close to asubducting spreading ridge (Tang et al., 2010). Therefore, more work


on porphyry dykes are needed to clarify the origin of the Baogutu por-phyry deposit.

4.4. Summary

Reducedmagmas are not favorable for porphyrymineralization. Thereduced porphyry deposits so far reported are either simply distal hostrocks located as far as several kilometers away from or above the caus-ative porphyry buried underneath, or originally oxidized magmas thatwere reduced through assimilation of reducing components. Degassingof SO2 is not likely to be a main process for forming porphyry depositsbecause it cannot explain the features of reduced porphyries. Carefulstudies are needed to identify the causative porphyries. Oxidizedadakitic (slabmelts) rocks are overwhelmingly the most favorable can-didates for porphyry mineralizations.

5. Discussion

5.1. The oxygen fugacities at convergent margins

The oxygen fugacity of the mantle and volcanic arc has been understudy for a long time (Ballhaus, 1993; Carmichael, 1991; Cottrell andKelley, 2013; Kelley and Cottrell, 2009; Lee et al., 2005, 2010;Parkinson and Arculus, 1999). The consensus is that the oxygen fugacityof arc magmas is systematically higher than that of MORB (Fig. 16)(Carmichael, 1991; Kelley and Cottrell, 2009; Parkinson and Arculus,1999; Sun et al., 2012a, 2013a,b). Nevertheless, it is still hotly debatedas regards to how arc magmas get oxidized. As mentioned inSection 3, a variety of mechanisms have been proposed.

The most straightforward way to elevate the oxygen fugacity of arcmagmas is by the addition of oxidizing components to the mantlewedge or directly to the magmas. Water is the most abundant volatilecomponent released during plate subduction. It has been proposedthat H2O reacts with FeO, forming Fe2O3 with the release of H2

(Eq. (30)) (Brandon and Draper, 1996; Kelley and Cottrell, 2009).Most of the ferric Fe will be transferred into the melt because ferric Feis more incompatible than ferrous Fe (Lee et al., in press). This reaction,however, is not controlled by water. Instead, it is controlled by ferrousFe, which is more stable under high pressure. Recent experimentsshow that water and H2 can coexist as two immiscible phases. Thisimmiscibility implies that water is stable in the mantle under highpressure (Bali et al., 2013).

More accurately, the ferrous Fe in fluids reacts with water, releasingH2 and H+ (Eq. (31)). This is supported by the abundance of oxidizedcomponents, e.g.magnetite–hematite and sulfates, found in the subduc-tion released fluids of ultrahigh pressure quartz veins, which may ele-vate the oxygen fugacity of the mantle wedge (X.M. Sun et al., 2007).It has also been proposed that slab melts are the most efficient way totransfer high Fe3+/Fe2+ ratios responsible for high oxygen fugacity inadakites (Mungall, 2002) and arc magmas (Brandon and Draper, 1996).

H2O þ 2FeO ¼ Fe2O3 þ H2 ð30Þ

3H2O þ 2Fe2þ ¼ Fe2O3 þ H2 þ 4H

þ ð31Þ

The addition of other oxidized materials, e.g., Fe3+, C4+, and S6+,from subducted sediments and oceanic crust to the mantle wedgemay elevate the redox states of the mantle as well (Fig. 18) (Evanset al., 2012). Arc magmas may also get oxidized during their evolutionand ascent by processes accompanying melting, crystallization, assimi-lation, degassing, etc (Ballhaus, 1993; Lee et al., 2005, 2010). For exam-ple, graphite and diamond are stable in the deepmantle. Duringmantlemelting, CO2 is incompatible whereas C is compatible, and graphitemaybe consumed through oxidation melting (Stagno et al., 2013). Most ofthe C in magma occurs as CO2, therefore the oxygen buffer changes

from graphite–CO2 (CCO) equilibria in the mantle source to Fe3+–

Fe2+ equilibria as the magma ascends (Ballhaus, 1993). Consistently,it becomes more oxidizing with decreasing depths and pressures(Stagno and Frost, 2010; Stagno et al., 2013). Moreover, most mantleminerals, such as olivine, favor ferrous Fe over ferric Fe, and fractionalcrystallization of these minerals elevates the Fe3+/Fe2+ (Carmichael,1991). Given that ferric Fe is highly incompatible, magma rechargingwill further enrich the ferric Fe (Lee et al., in press), resulting in higheroxygen fugacities. In addition, assimilation of oxidizing country rocks(Lee et al., 2005), degassing of reduced volatile species (e.g., H2, H2Sand CH4) (Ballhaus, 1993; Lee et al., 2005) and lowering of pH valuesin a sulfate–sulfide dominated system (Sun et al., 2013a,b) may also el-evate the oxygen fugacity of themagmas. Based on similar Zn/FeT ratiosof mantle peridotite and primitive arc magmas, it has been argued thatprimary arc magmas are not necessarily oxidized (Lee et al., 2010). In-creasing Zn/FeT ratios with decreasing MgO content argues that thehigh oxygen fugacity of arc magmas is acquired through magma evolu-tion (Lee et al., 2010), indicating that continuing chemical evolutionduring magma transport and emplacement may have major effects onthe oxidized characteristics of arc magmas. This may equally explainthe diversity in oxygen fugacities at convergent margins. Nevertheless,high oxygen fugacities developed duringmagma evolutionmay providelittle or no contribution to eliminating residual sulfides, and thus littlecontribution to porphyry mineralization. This decoupling may partiallyexplain why most arc magmas are highly oxidized, whereas only asmall portion of them form porphyry Cu deposits.

5.2. The difference between porphyry and epithermal in terms of oxygenfugacity

Most Cuporphyry deposits form at depths of 2–4 km, and are usuallyassociated with epithermal deposits at shallow depths if not removedby erosion (Figs. 22d, 23a, b) (Cooke et al., 2011; Hedenquist et al.,1998; Heinrich, 2005; Heinrich et al., 2004; Hollings et al., 2005;Sillitoe, 2010). This seemingly implies that porphyry and epithermal de-posits are closely related. Many epithermal deposits, however, are notlinked to porphyry deposits. For example, epithermal deposits formedin themid-ocean ridge and backarc basins usually do not have any asso-ciated porphyry deposit. This is probably mainly because the crust inbackarc basins is too thin. Regardless, the oxygen fugacities ofepithermal deposits have a much larger ranges than porphyry deposits.

5.2.1. Magnetite crisisThe magnetite crisis (Jenner et al., 2010) refers to the dramatic de-

creases in Cu and Au during magnetite crystallization in arc volcanicrocks (Sun et al., 2004a), which is a common phenomenon (Mosset al., 2001; Sun et al., 2011; Togashi and Terashima, 1997). Thismagne-tite crisis is taken as a main process that leads to the formation of ore-forming fluids, responsible for the Au and Cu mineralization of bothepithermal and porphyry deposits (Liang et al., 2009; Sun et al., 2004a,2013a,b) (see also Section 3): Sulfate is reduced to sulfide during mag-netite crystallization, forming hydrosulfide complexes, which scavengechalcophile elements into fluids and subsequently transport the metalsto favorable places for forming deposits (Sun et al., 2004a).

This mineralization model has been challenged by some laterstudies. Although using the same set of samples (Jenner et al., 2012),and essentially repeating the results of the previous study (Sun et al.,2004a), it was proposed that magnetite fractionation triggers sulfidesaturation (Jenner et al., 2010). This conclusion was drawn basedmain-ly on the behavior of Se, which was assumed as a proxy that follows Sclosely during magmatic evolution except that it is not lost duringlow-pressure (near sea-floor) degassing (Jenner et al., 2010). Fluid ex-traction, however, is different from degassing, and thus Se provides noconstraints on the process. In fact, as shown in the supplementary infor-mation, sulfide is undersaturated in Manus glasses (Fig. 24) (Sun et al.,2004a).

Fig. 22. Different models for the mineralization of porphyry deposits. After (A) Lee (2014): Mantle-derived magmas (red) intrude into the cold upper plate (black) of subduction zones,generating crystallizing magma chambers. Thick continental arcs are more favorable for porphyry mineralization because of accumulation of copper-rich sulfide cumulates; (B) afterWilkinson (2013), which also proposes that sulfide saturation and accumulation are very important for the formation of porphyry Cu deposits; (C) after Richards (2011b), illustratingporphyry and epithermalmineralizations in arc and postcollisional settings; (D) after Richards (2013), highlighting features or processes thatmay result in supercharging of these systemsto form giant deposits.


image of Fig.�22

Fig. 22 (continued).


There are indeed sulfide inclusions in mineral phenocryst fromManus glasses (Fig. 25) (Sun et al., 2004a). The sulfide phase containshigh Au and Ag contents in addition to Cu in sulfides, but no otherchalcophile elements, such as Ni, Re, and Pt (Jenner et al., 2010). Thiswas used to argue that the sulfide phase is crystalline rather than an im-miscible sulfide melt (Jenner et al., 2010).

Crystallization of sulfide cannot explain this phenomenon, becauseNi and Pt are even more chalcophile than Cu, with partition coefficientsbetween sulfide and silicate melts of several hundreds and more than20 thousand, respectively (Table 2). We propose that these sulfidesformed directly from magmatic fluids. As shown in Fig. 25, sulfides inphenocrysts are associated with fluid inclusions (Sun et al., 2004a).It is very likely that the preferential enrichments of Cu and Au over

Ni and Pt are controlled by sulfide complexes in fluid. This may ex-plain why there is no Ni and Pt in most porphyry and epithermaldeposits.

The lack of Re in these sulfides can be plausibly interpreted by thehigh oxygen fugacity, under which most of the Re is Re6+, and thus be-haves as lithophile rather than chalcophile elements. As shown inFig. 26, in contrast to Cu and Au (Sun et al., 2004a) and also V, Co, Pt,and Se (Jenner et al., 2010), Re concentrations inManus glasses keep in-creasing when magnetite starts to crystallize, and then decrease gradu-ally (Sun et al., 2003a). This can best be explained by the reduction ofRe6+ to Re4+ during magnetite crystallization. Correspondingly, Rechanges from incompatible (Sun et al., 2003a,b,c), to compatible(Mallmann and O'Neill, 2007).

image of Fig.�22

Fig. 23. Cartoons illustrate the relationship between some epithermal deposits with porphyry deposits underneath. Modified after: (A) Hedenquist and Lowenstern (1994) and(B) Heinrich (2005). Note, not all epithermal deposits are associated with porphyry deposits.


5.2.2. Oxygen fugacity and open systemsIn contrast to the dramatically elevated oxygen fugacities during

magnetite crisis, the oxygen fugacity of the Manus magmas did notchange much, or even lessened slightly (Jenner et al., 2010) during thecrystallization of magnetite and sulfate reduction (Sun et al., 2004a).This can be interpreted as a magmatic process, i.e., the magnetite crys-tallization and sulfate reduction recorded in glasses occurred duringmagma evolution. As mentioned in Section 3.2.2, magnetite crystalliza-tion in magmas reduces sulfate without changing the pH values. Thesolubility of sulfides in fluids is very high, with partition coefficients

of ~500 (Keppler, 2010). Sulfides formed through sulfate reductionare released into magmatic fluids, where they scavenge chalcophileelements (Jenner et al., 2010; Sun et al., 2004a). Meanwhile, thereactions of Eqs. (16) and (17) are driven to the right, promotingsulfate reduction.

Hydrothermal magnetite and even hematite may also form in fluidsreleased from arc volcanic magmas, e.g., Manus glasses, releasing H+.Nevertheless, volcanic magma systems are much more open than por-phyry systems, so that the formation of hydrothermal iron oxides doesnot necessarily affect the oxygen fugacity of the arc magmas.

image of Fig.�23

Fig. 24. Sulfide contents in Manus submarine volcanic glasses, showing that sulfide is farbelow saturation (after Sun et al., 2004a). Therefore, sulfide segregation induced by sulfidesaturation as proposed by previous authors (Jenner et al., 2010) is not a feasibleway to ex-plain the magnetite crisis. Filled and open circles represent volcanic glasses analyzed bydifferent authors, filled triangles represent melt inclusions.


Given that volcanic systems are farmore open, the initial Cu, Au con-tents are not of critical importance to porphyry mineralization. Metalscan be leached out as far as there is enoughwater circulation. Therefore,it does not require very high oxygen fugacities to eliminate residual

Fig. 25. Images of sulfide inclusion in an olivine phenocryst of Manus glass under(A) reflected light and (B) and transparent light. Only one big sulfide globule is clearlyidentified under reflected light, which seemingly indicates sulfide saturation. This grainis actually associated with fluid inclusions as shown under transparent light. In addition,there are several sulfide grains, which are all associated with fluid inclusions (Sun et al.,2004a). The unique composition of sulfides inManus glasses (lowNi, Pt, etc.)may be plau-sibly interpreted by sulfides crystallized from magmatic fluids.

sulfides. This explainswhy epithermal deposits are distributed in a vari-ety of tectonic settings, ranging frommid-ocean ridges and backarc ba-sins to arcs.

5.3. Adakite, slab melting, ridge subduction and porphyry Cu deposits

Adakite was initially named for rocks formed through partial melt-ing of subducted young oceanic crust (b25 Ma, represented by mid-ocean ridge basalt, MORB) (Defant and Drummond, 1990; Kay, 1978).In contrast to most other rock types, adakite is defined by geochemicalcompositions (e.g., SiO2 ≥ 56 wt.%, Al2O3 ≥ 15 wt.%, Y ≤ 18 ppm,Yb ≤ 1.9 ppm and Sr ≥ 400 ppm) without detailed petrographic con-straints. Therefore, (1) both eruptive and intrusive rocks can be classi-fied as adakites; and (2) adakites may be produced simply by partialmelting of mafic rocks in the presence of garnet and absence of plagio-clase. Different mechanisms have been proposed to produce adakites,e.g., partial melting of the lower continental crust (Chung et al., 2003;Gao et al., 2004; Guo et al., 2006; Xu et al., 2002, 2006; Zhang et al.,2001b) or underplated new crust (Hou et al., 2009; Martin, 1999), orby fractional crystallization of normal arc magmas (Castillo, 2006;Macpherson et al., 2006; Richards and Kerrich, 2007). Given that theoceanic crust is very different from continental crust, slab melts can bedistinguished from lower continental crust melts using geochemicalcriteria (Ling et al., 2011; Liu et al., 2010; Sun et al., 2012a).

5.3.1. Adakite and porphyry Cu depositsMost porphyry Cu deposits are associated with adakites (Oyarzun

et al., 2001; Sajona and Maury, 1998; Sun et al., 2011, 2012a,b, 2013a,b; Thieblemont et al., 1997), but the association is not true vice versa.Many adakites, e.g., those from the Dabie Mountains, are not mineral-ized at all (Huang et al., 2008; Ling et al., 2013; Liu et al., 2012; Wanget al., 2007a). Partial melting of thickened eclogitic lower continentalcrust (Wang et al., 2006a,b, 2007a,b; Zhang et al., 2001a) and fractionalcrystallization of garnet (Macpherson et al., 2006) or amphibole(Richards and Kerrich, 2007) may also form high Sr/Y magmas. Thelower continental crust melts have much lower Cu abundance andlower oxygen fugacity than subducting slabs, such that they are not fa-vorable for forming porphyry Cu deposits. Fractional crystallization ofgarnet and/or amphibole plays no positive role in Cumineralization, ei-ther (Sun et al., 2011, 2012a).

Adakite formation through slab melting, on the other hand, doesfavor porphyry Cu mineralization (Mungall, 2002; Sajona and Maury,1998; Sun et al., 2011, 2012a; Thieblemont et al., 1997). Several differ-ent explanations for the fertility of adakitic slab melts have beenproposed:

Oxidized. It has been argued that slab melts might be unusually oxi-dized and rich in sulfur (Oyarzun et al., 2001), due to high Fe3+ con-tent from oxidative sea-floor alteration. As a consequence, thiselevated fO2 state causes oxidation of chalcophile metal-bearing sul-fide phases in themantlewedge, releasingmetals to the silicatemeltphase (Mungall, 2002). However, although oxidation is crucial toporphyry mineralization, adakites are not systematically more oxi-dized than normal arc magmas (Ballhaus, 1993; X.M. Sun et al.,2007; Sun et al., 2013a,b). Normal arc melts are not systematicallymore enriched in Cu than MORB, either (Lee et al., 2012).Water: It has been proposed that slab melts might be unusuallywater rich (Sajona and Maury, 1998). High water contents inmagmas may suppress the crystallization of plagioclase, and pro-mote the formation of amphibole, resulting in high Sr/Y signatures(Richards, 2011a, 2012). The problem with this model is that nostudies have ever demonstrated that adakites are more hydrousthan normal arc rocks. Moreover, the crystallization of amphiboleplays no role in Cu mineralization. Copper is incompatible in mostmajor silicate minerals, but could be compatible in amphibole, de-pending on the composition of the magmas and amphibole (Sun

image of Fig.�24

image of Fig.�25

Table 2Summary of sulfide/silicate partition coefficients.

Cu Au Ag Ni Pt Re Mo References

212–278 – – 240–308 – – – Rajamani and Naldrett (1978)1383 1.5E+4–1.9E+4 – 500–900 – – – Peach et al. (1990)250–970 – – 540–4400 – – – Gaetani and Grove (1997)330–1070 7E+3–1.3E+4 300–1680 210–1270 – – 0.20–14.48 Li and Audetat (2012, 2013)1040–1624 – 853–1528 646–965 – – – Patten et al. (2013)1050–1850 4.1E+03–1.12E+04 – – 2.41E+05–4.19E+05 361–960 – Mungall and Brenan (2014)

Fig.whmuformwasMo


et al., 2012b, 2014a). Therefore, none of the arguments concerningwater abundance can be plausibly interpreted to explain an associa-tion between adakites and porphyry deposits. Because of these un-certainties, it has been asserted that there is no single, obviousreasonwhy slabmelts have an unusually high potential to form por-phyry deposits (Richards, 2013).Felsic: Adakite is formed by partial melting of basaltic rocks, suchthat they should logically be more felsic than peridotite melts(Sajona andMaury, 1998). It has been further proposed that adakitesmight more readily crystallize as intrusive plutons because of theirviscous felsic nature, leading to the generation of a more efficientcrustal magmatic–hydrothermal systems (Sajona and Maury,1998). In fact, adakites associated with porphyry Cu deposits aremostly intermediate in composition, not felsic.Compression. Adakites are often generated by flat subduction ofyoung oceanic crust with associated compressional stress in theupper plate. Such an environment should be favorable for trapping

26. Diagrams of SiO2 versus Cu and Re. Note, in contrast to Cu, Re keeps increasingen Cu drops suddenly at the point when magnetite starts to crystallize. Rhenium isch less chalcophile than Cu under high oxygen fugacities. It is present mainly in theof Re6+ and acts as an incompatible element before magnetite crystallization andgradually reduced to Re4+ during magnetite crystallization.

dified after Sun et al. (2003a, 2004a).

magma in a non-erupting, closed-system pluton where sulfurmight precipitate as hydrothermal sulfides and sulfates instead ofbeing degassed as SO2 (Oyarzun et al., 2001, 2002). In addition, com-pression also results in uplifting and erosions, which are later favor-able for the exposure of porphyry deposits. The question is againwhynormal arc rocks in compressed environments do not formpor-phyry deposits.High Cu contents. Oceanic crust has a much higher Cu abundance(~100 ppm) (Sun et al., 2003a,b) than the mantle (30 ppm)(McDonough and Sun, 1995) or the continental crust (~27 ppm)(Rudnick and Gao, 2003). It has been proposed that partial meltingof the subducted oceanic crust forms adakites with systematicallyhigher Cu initial contents—favorable for porphyry Cu mineralization(Sun et al., 2011, 2012a,b). Previous studies, however, did not quan-titatively model the influence of oxygen fugacity, although its posi-tive effect has been emphasized (Sun et al., 2011, 2013a,b).

The effects of oxygen fugacity on normal arc magmas have beennicely modeled (Fig. 3) (Lee et al., 2012). Normal arc rocks form by par-tial melting of peridotite from the mantle wedge. The primitive mantlecontains ~250 ppmof S (McDonough and Sun, 1995),whereas depletedmantle peridotite contains ~150 ppm of S (O'Neill and Mavrogenes,2002). Partial melting of mantle peridotite can easily eliminate residualsulfide even under reducing conditions, e.g., by ~20% partial melting atΔFMQ 0 (Fig. 3) (Lee et al., 2012). In contrast, oceanic crust has sulfurabundances of over 1000 ppm (O'Neill and Mavrogenes, 2002), whichis about 4 times greater than the primitive mantle value (McDonoughand Sun, 1995). As discussed in Section 3, oxygen fugacities higherthan ΔFMQ+ 2 are of critical importance to eliminate residual sulfidesduring slab melting. However, at oxygen fugacities higher than ΔFMQ+ 2, ~5–10% of partial melting is enough to eliminate residual sulfidefrom the subducted oceanic crust, forming adakitic melts with high Cuand S contents up to 2000 ppmand percent levels, respectively. Such el-evated contents in the magma, of course, are consistent with the highCu, S contents present in ore bearing porphyries.

5.3.2. Ridge subduction and porphyry Cu depositsMore than half of the world porphyry Cu deposits are located along

the western coasts of the North and South American continents com-prising the eastern Pacificmargin. The total resources there are estimat-ed to be N1.8 billion tons, accounting for about 60% of the world's totalCu resource estimation (Mutschler et al., 2010). Twenty out of theworld's top 25 giant porphyry Cu deposits are located there. In contrast,there are essentially no porphyry Cu deposits located along the north-western Pacific margin, e.g., Japan.

Many large porphyry Cu–Audeposits are connected to subduction ofspreading and aseismic ridges (Fig. 27) (Cooke et al., 2005; Sun et al.,2010). As discussed above, slab melting is usually involved in creatingmagmas with high Cu contents and adequately high oxygen fugacity,which are the twomain controlling factors for porphyry Cu mineraliza-tion (Section 3). Subduction of young ridges, both spreading andaseismic, in particular, produces adakites with high oxygen fugacity,making it the best geological process for porphyry Cu deposits (Sunet al., 2010, 2013a,b). There are several subducting ridges along theeast Pacific margin, e.g., in Chile and Peru in the South America. These

image of Fig.�26

Fig. 27.Many large porphyry Cu deposits are closely associated with ridge subduction, because subduction of young ridges is the most favorable geologic process for slab melting in thePhenozoic, forming highly oxidized melt with high initial Cu contents.Modified after Sun et al. (2010).


ridges were mostly younger than 25 Ma when they began to subduct,and are closely associated with large porphyry Cu–Au deposits (Fig. 27).

There are also several ridges (most of which are aseismic, i.e., islandchains) on thewestern Pacific plate (Fig. 28).Most of these ridges, how-ever, are older than 100Ma (W.D. Sun et al., 2007), and are not likely toform adakites. In addition, the oxygen fugacities in Japan, Izu–Bonin–Mariana, and other arcs along the northwestern Pacificmargins, are sys-tematically lower than ΔFMQ+ 2 (Fig. 16a). For these reasons, we findit not surprising that no economically viable porphyry deposits associat-ed with volcanic arcs are known in the northwestern Pacific region.

There are porphyry ore deposits located at the southwestern Pacificmargin, which, however, are much less developed in terms of tonnageand the number of deposits. It is noteworthy that these deposits are as-sociated with the closure of backarc basins younger than 25 Ma(Fig. 29).

Considering the geotherm's concave downward shape of subductingslabs, ridge subduction is the most favorable tectonic setting for slab

melting in the Phanerozoic. Geochemical signatures of ridge subductionare important exploration targets for large porphyry Cu–Au deposits.

5.4. Alterations

Porphyry deposits have very well developed alteration zones thattypically affect several cubic kilometers of rock (Lowell and Guilbert,1970; Sillitoe, 2010; Titley, 1981) (Fig. 30), which is of critical impor-tance to understanding porphyrymineralization processes and improv-ing their exploration.

Alteration is mainly controlled by pH values of the ore-forming fluids(Sillitoe, 2010). The amount of H+ released during mineralization(e.g., Eqs. (15), (16), (18), (19), (22), (23), (25)) together with alkali con-tents in the porphyry together control advanced argillic lithocapformation and alteration (Sillitoe, 2010). For example, sericitic andadvanced argillic alteration are much less well developed in porphyryCu deposits associated with alkaline than with calc-alkaline intrusions

image of Fig.�27

Fig. 28. Distributions of aseismic ridges (island chains) in the Western Pacific. These aseismic ridges are all much older than 25 Ma, such that do not get melted during subduction.Modified after W.D. Sun et al. (2007).

Fig. 29. Ages of backarc basins in the southwestern Pacific. Subduction of young backarc basin crust forms adakite, which is favorable for porphyry mineralization.


image of Fig.�28

image of Fig.�29


(Lang et al., 1995; Sillitoe and Burrows, 2002), reflecting control of theK+/H+ ratio bymagma chemistry (Sillitoe, 2010). As summarized recent-ly (Sun et al., 2014b), the main alteration reactions are Eqs. (32)–(37):

2K Mg; Feð Þ3AlSi3O10 OHð Þ2Biotite

þ4Hþ

¼ Al Mg; Feð Þ5AlSi3O10 OHð Þ8chlorite

þ Mg; Feð Þ2þ þ 2Kþ þ 3SiO2 ð32Þ

Fig. 30. Alteration patterns of porphyry deposits. After: (A) Lowell a

3KAlSi3O8Potassium feldspar

þ2Hþ ¼ KAl2Si3AlO10 OHð Þ2sericite

þ2Kþ þ 6SiO2 ð33Þ

3NaAlSi3O8Sodiumfeldspar

þKþ þ 2Hþ ¼ KAl2 AlSi3O10½ � OHð Þ2sericite

þ6SiO2 þ 3Naþ ð34Þ

2KAl3Si3O10 OHð Þ2Sericite

þ2Hþ þ 3H2O ¼ 3Al2Si2O5 OHð Þ4kaolinite

þ2Kþ ð35Þ

nd Guilbert (1970); (B) Sillitoe (2010) and (C), Richards (2011b).

image of Fig.�30

Fig. 30 (continued).


KFe3AlSi3O10 OHð Þ2Biotite

þ1=2O2 ¼ KAlSi3O8potassium feldspar

þFe3O4 þH2O ð36Þ

CaAl2Si2O8Anorthite

þ2KClþ 4SiO2 ¼ 2KAlSi3O8potassium feldspar

þCaCl2: ð37Þ

The alteration zone in porphyry Cu deposits starts from barren, earlysodic–calcic upward through potentially ore-grade potassic, chlorite–sericite, and sericitic, to advanced argillic, and finally the lithocap(Sillitoe, 2010). In general, the alteration–mineralization zones becomeprogressively younger upward (Fig. 30), consequently the shallower al-teration–mineralization zones overprint deeper ones.

Sodic–calcic alteration is typically sulfide andmetal poor (except forFe asmagnetite) but can hostmineralization in Au-rich porphyry Cu de-posits. It is commonly located in the immediate wallrocks of the por-phyry intrusion, or are found as a centrally located zone of someporphyry Cu stocks (Sillitoe, 2010).

Potassic alteration is located in the center and deeper portions of theporphyry. Dominant mineral changes from biotite in relatively maficporphyry intrusions and host rocks, to K-feldspar in more felsic, grano-dioritic to quartz monzonitic settings (Sillitoe, 2010). Quartz-K ± Na-feldspar overprints may destroy the more typical potassic assemblages.The chalcopyrite ± bornite ore in many porphyry Cu deposits is largelyconfined to potassic zones. Potassic-alteredwallrocks may attain N1 kmthickness. The potassic alteration generally becomes less intense fromthe older to younger porphyry phases (Sillitoe, 2010). This is likelydue to a lowering of the pH as mineralization continues.

Chlorite–sericite alteration produces pale-green rocks and is wide-spread in the shallower parts of some porphyry Cu deposits,overprinting preexisting potassic assemblages. The alteration is charac-terized by transformation of mafic minerals to chlorite, plagioclase to

sericite (fine-grained muscovite) and/or illite, and magnetite to hema-tite (martite and/or specularite), along with deposition of pyrite andchalcopyrite (Sillitoe, 2010).

Sericitic alteration in porphyry Cu deposits normally overprints thepotassic and chlorite–sericite assemblages. It may be subdivided intotwo different types—a less common, early greenish to greenish-gray incolor alteration, and a far more common white alteration. The sericiticalteration is commonly pyrite dominated, implying effective removalof the Cu (±Au) present in the former chlorite–sericite and/or potassicassemblages. It may also constitute ore with Cu either in the form ofchalcopyrite or as high sulfidation-state assemblages (Sillitoe, 2010).

The lower portion of argillic lithocapsmay overprint the upper partsof the porphyry Cu deposits, whereas the sericitic alteration transformsupwardly to quartz–pyrophyllite. The advanced argillic alteration pref-erentially affects lithologic units with low acid-buffering capacities(Sillitoe, 2010).

6. Conclusion

The key processes of porphyry mineralization are oxidation and re-duction of sulfur. Most of the porphyry deposits are closely associatedwith oxidized magmas, with sulfate as the dominant sulfur mineralspecies. Sulfur is one of the most important geosolvents that controlsthe behaviors of copper and other chalcophile elements, thereforeknowledge of its geochemical behavior is essential to the understandingof mineralization processes for copper and a variety of other metal re-sources. Given that for most chalcophile elements, the partition coeffi-cient between sulfide and melt is very high, elimination of residualsulfide is essential for the extraction of chalcophile elements from thesource and thus the formation of porphyry deposits. The solubility ofsulfur depends strongly on sulfur speciation, which in turn dependson oxygen fugacities. Sulfate is over 10 times more soluble than sulfide.

image of Fig.�30


At oxygen fugacities higher than ΔFMQ+ 2, most of the sulfur in meltsis present as sulfate, such that the solubility of sulfur increases from~1300 ppm to 2 wt.%. Consequently, slab melts becomes sulfur under-saturated at N5% partial melting, with Cu contents of over 1000 ppm.Therefore,ΔFMQ+2 is often considered themagic number for porphy-ry mineralization.

Metals of porphyry Cu deposits are hosted in sulfides, which requirereduction of sulfate to sulfide during the final stage of mineralization. Inprinciple, sulfate can also be reduced by assimilation of reducing sedi-ments or degassing of oxidizing gases. Porphyry deposits are usuallymineralized throughout the whole pluton, whereas interactions withcountry rocks occurmainly at the interface, such that assimilation of re-ducing sediments is not likely to be the controlling process. Degassing isnot themain process in deeply emplaced porphyry bodies, such that it isnot likely to be amajor process for sulfate reduction, either. Ferrous ironis themost important reductant that is responsible for sulfate reductionduring porphyry mineralization. The highest oxygen fugacity favorablefor porphyrymineralization is the HMbuffer. Otherwise, there is no fer-rous Fe in the system. The reduction of sulfate and oxidation of ferrousFe lower the pH value. This, in turn, elevates the oxidation potential ofsulfate, driving the oxygen fugacities up to the HM buffer. Thesebehaviors explain the popular occurrence of hypogene magnetite andhematite (and specularite) in porphyry deposits.

Low pH fluids cause formation of pervasive alteration zones in por-phyry Cu deposits, starting frombarren, early sodic–calcic alteration up-ward through potentially ore-grade potassic, chlorite–sericite, andsericitic alterations, to advanced argillic alterations, and finally formingthe lithocaps. The amount of H+ released duringmineralization and thealkali content in the porphyry together control advanced argilliclithocap formation and alterations.

Hydrogen andmethane formduring the finalmineralization processof porphyry deposits. Most of the hydrogen and methane should havebeen oxidized by ferric Fe. In special cases, some of the reduced gasesmay escape from the system, and even get trapped in fluid inclusions.Therefore, small amount of reduced gases in fluid inclusions cannotargue against the oxidized feature of the magmas.

Reduced magmas are not favorable for porphyry mineralization.There are indeed several small porphyry deposits that appear to be re-lated to reduced porphyries. In our opinion, the reduced porphyry de-posits so far reported are either host rocks at distal sites as far asseveral kilometers away from or above the causative porphyry lyingdeep underneath, or a consequence of fluid transportation. Some ofthe reduced porphyries were originally associated with oxidizedmagmas but were reduced through assimilation of reducing compo-nents during emplacement. Degassing of SO2 is not likely to be a mainprocess for porphyry deposit formation, and it doesn't reduce sulfateto sulfide, either. Therefore, degassing of SO2 cannot explain the featuresof reduced porphyries. Given that a large portion ofmetals in a porphyrydeposit are hosted inwall rocks, attention is needed to identify the caus-ative porphyries.

Acknowledgments

We would like to thank Professor Cin-ty A. Lee for constructivediscussions. This study is supported by the National Natural ScienceFoundation of China (nos. 41090374, 41121002, 41172080). This is con-tribution no. IS-1939 from GIGCAS.

References

Abratis, M., Worner, G., 2001. Ridge collision, slab-window formation, and the flux ofPacific asthenosphere into the Caribbean realm. Geology 29 (2), 127–130.

An, F., Zhu, Y.F., 2010. Native antimony in the Baogutu gold deposit (west Junggar, NWChina): its occurrence and origin. Ore Geol. Rev. 37 (3–4), 214–223.

Arculus, R.J., 1994. Aspects of magma genesis in arcs. Lithos 33 (1–3), 189–208.

Astudillo, N., Roperch, P., Townley, B., Arriagada, C., Chauvin, A., 2010. Magnetic polarityzonation within the El Teniente copper–molybdenum porphyry deposit, centralChile. Mineral. Deposita 45 (1), 23–41.

Audetat, A., Pettke, T., Dolejs, D., 2004. Magmatic anhydrite and calcite in the ore-formingquartz–monzodiorite magma at Santa Rita, New Mexico (USA): genetic constraintson porphyry–Cu mineralization. Lithos 72 (3–4), 147–161.

Baker, T., Ash, C.H., Thompson, J.F.H., 1997. Geological setting and characteristics of theRed Chris porphyry copper–gold deposit, northwestern British Columbia. Explor.Min. Geol. 6 (4), 297–316.

Baker,M.J., Cooke, D.R., Hollings, P.N., 2011. LA-ICP-MSU–Pb zircon ages and Nd isotope datafor the Teniente Plutonic Complex: implications for the evolution of the El TenientePorphyry Cu–Mo Deposit. Let's Talk Ore Deposits vols. I and Ii, (853–855 pp.).

Bali, E., Audetat, A., Keppler, H., 2013.Water and hydrogen are immiscible in Earth's man-tle. Nature 495 (7440), 220–222.

Ballard, J.R., Palin, J.M., Williams, I.S., Campbell, I.H., Faunes, A., 2001. Two ages of porphy-ry intrusion resolved for the super-giant Chuquicamata copper deposit of northernChile by ELA-ICP-MS and SHRIMP. Geology 29 (5), 383–386.

Ballard, J.R., Palin, J.M., Campbell, I.H., 2002. Relative oxidation states of magmas inferredfrom Ce(IV)/Ce(III) in zircon: application to porphyry copper deposits of northernChile. Contrib. Mineral. Petrol. 144 (3), 347–364.

Ballhaus, C., 1993. Oxidation states of the lithospheric and asthenospheric upper mantle.Contrib. Mineral. Petrol. 114, 331–348.

Barra, F., 2011. Assessing the longevity of porphyry Cu–Mo deposits: examples from theChilean Andes. Let's Talk Ore Deposits vols. I and Ii, (145–147 pp.).

Beermann, O., Botcharnikov, R.E., Holtz, F., Diedrich, O., Nowak, M., 2011. Temperature de-pendence of sulfide and sulfate solubility in olivine-saturated basaltic magmas.Geochim. Cosmochim. Acta 75 (23), 7612–7631.

Binder, B., Keppler, H., 2011. The oxidation state of sulfur inmagmatic fluids. Earth Planet.Sci. Lett. 301, 190–198.

Brandon, A.D., Draper, D.S., 1996. Constraints on the origin of the oxidation state ofmantle overlying subduction zones: an example from Simcoe, Washington, USA.Geochim. Cosmochim. Acta 60 (10), 1739–1749.

Bryant, J.A., Yogodzinski, G.M., Churikova, T.G., 2007. Melt–mantle interactions beneaththe Kamchatka arc: evidence from ultramafic xenoliths from Shiveluch volcano.Geochem. Geophys. Geosyst. 8 (4), Q04007. http://dx.doi.org/10.1029/2006GC001443.

Burnham, C.W., Ohmoto, H., 1980. Late-stage processes of felsic magmatism. Min. Geol.Spec. Issue 8, 1–12.

Calagari, A.A., 2003. Stable isotope (S, O, H and C) studies of the phyllic and potassic–phyllic alteration zones of the porphyry copper deposit at Sungun, East Azarbaidjan,Iran. J. Asian Earth Sci. 21 (7), 767–780.

Candela, P.A., 1992. Controls on ore metal ratios in granite-related ore systems — an ex-perimental and computational approach. Trans. R. Soc. Edinb. Earth Sci. 83, 317–326.

Cannell, J., Cooke, D.R., Stein, H.J., Markey, R., 2003. New paragenetically constrained Re–Os molybdenite ages for El Teniente Cu–Mo porphyry deposit, Central Chile. MineralExploration and Sustainable Development vols. 1 and 2, (255–258 pp.).

Cannell, J., Cooke, D.R.,Walshe, J.L., Stein, H., 2005. Geology, mineralization, alteration, andstructural evolution of the El Teniente porphyry Cu–Mo deposit. Econ. Geol. 100 (5),979–1003.

Cao, M., Qin, K., Li, G., Jin, L., Evans, N.J., Yang, X., 2014. Baogutu: an example of reducedporphyry Cu deposit in western Junggar. Ore Geol. Rev. 56, 159–180.

Carmichael, I.S.E., 1991. The redox states of basic and silicicmagmas— a reflection of theirsource regions. Contrib. Mineral. Petrol. 106 (2), 129–141.

Castillo, P.R., 2006. An overview of adakite petrogenesis. Chin. Sci. Bull. 51 (3), 258–268.Chen, Y.J., 2013. The development of continental collisionmetallogeny and its application.

Acta Petrol. Sin. 29 (1), 1–17.Chen, Y.J., Li, C., Zhang, J., Li, Z., Wang, H.H., 2000. Sr and O isotopic characteristics of por-

phyries in the Qinling molybdenum deposit belt and their implication to geneticmechanism and type. Sci. China. Ser. D Earth Sci. 43, 82–94.

Chen, W., Xu, Z.W., Lu, X.C., Yang, X.N., Li, H.C., Qu, W.J., Chen, J.Q., Wang, H., Wang, S.H.,2013. Petrogenesis of the Bao'anzhai granite and associated Momineralization, west-ern Dabie orogen, east-central China: constraints from zircon U–Pb and molybdeniteRe–Os dating, whole-rock geochemistry, and Sr–Nd–Pb–Hf isotopes. Int. Geol. Rev. 55(10), 1220–1238.

Chiaradia, M., 2014. Copper enrichment in arc magmas controlled by overriding platethickness. Nature Geosci. 7, 43–46.

Chiaradia, M., Ulianov, A., Kouzmanov, K., Beate, B., 2012. Why large porphyry Cu depositslike high Sr/Y magmas? Sci. Rep. 2. http://dx.doi.org/10.1038/srep00685.

Chou, I.M., 1978. Calibration of oxygen buffers at elevated P and T using the hydrogen fu-gacity sensor. Am. Mineral. 63, 690–703.

Chou, I.M., 1986. Permeability of precious metals to hydrogen at 2 kbar total pressure andelevated temperatures. Am. J. Sci. 286, 638–658.

Chung, S.L., Liu, D.Y., Ji, J.Q., Chu, M.F., Lee, H.Y.,Wen, D.J., Lo, C.H., Lee, T.Y., Qian, Q., Zhang,Q., 2003. Adakites from continental collision zones: melting of thickened lower crustbeneath southern Tibet. Geology 31 (11), 1021–1024.

Cole, R.B., Stewart, B.W., 2009. Continental margin volcanism at sites of spreading ridgesubduction: examples from southern Alaska and western California. Tectonophysics464 (1–4), 118–136.

Cole, R.B., Nelson, S.W., Layer, P.W., Oswald, P.J., 2006. Eocene volcanism above a de-pleted mantle slab window in southern Alaska. Geol. Soc. Am. Bull. 118 (1–2),140–158.

Cooke, D.R., Hollings, P., Walsh, J.L., 2005. Giant porphyry deposits: characteristics, distri-bution, and tectonic controls. Econ. Geol. 100 (5), 801–818.

Cooke, D.R., Deyell, C.L., Waters, P.J., Gonzales, R.I., Zaw, K., 2011. Evidence for magmatic–hydrothermal fluids and ore-forming processes in epithermal and porphyry depositsof the Baguio District, Philippines. Econ. Geol. 106 (8), 1399–1424.

http://refhub.elsevier.com/S0169-1368(14)00205-4/rf0005





































http://dx.doi.org/10.1029/2006GC001443

http://dx.doi.org/10.1029/2006GC001443






























http://dx.doi.org/10.1038/srep00685



















Cottrell, E., Kelley, K.A., 2013. Redox heterogeneity in mid-ocean ridge basalts as a func-tion of mantle source. Science 340 (6138), 1314–1317.

Craig, H., 1961. Isotopic variations in meteoric waters. Science 133, 1702–1703.Crerar, D.A., Susak, N.J., Borcsik, M., Schwartz, S., 1978. Solubility of the buffer assemblage

pyrite–pyrrhotite–magnetite in NaCl solutions from 200 to 350 °C. Geochim.Cosmochim. Acta 42, 1427–1437.

Cuadra, P., Rojas, G., 2001. Oxide mineralization at the Radomiro Tomic porphyry copperdeposit, northern Chile. Econ. Geol. Bull. Soc. Econ. Geol. 96 (2), 387–400.

Defant, M.J., Drummond, M.S., 1990. Derivation of some modern arc magmas by meltingof young subducted lithosphere. Nature 347, 662–665.

Dilles, J.H., Tomlinson, A.J., Garcia, M., Alcota, H., 2011. The geology of the Fortuna Grano-diorite Complex, Chuquicamata district, Northern Chile: relation to porphyry copperdeposits. Let's Talk Ore Deposits. vols I and Ii, (398–400 pp.).

Eugster, H.P., 1957. Heterogeneous reactions involving oxidation and reduction at highpressures and temperatures. J. Chem. Phys. 26, 1160.

Eugster, H.P., Wones, D.R., 1962. Stability relations of the ferruginous biotite, annite. J.Petrol. 3, 82–125.

Evans, K.A., Elburg, M.A., Kamenetsky, V.S., 2012. Oxidation state of subarc mantle. Geol-ogy 40 (9), 783–786.

Field, C.W., Zhang, L., Dilles, J.H., Rye, R.O., Reed, M.H., 2005. Sulfur and oxygen isotopicrecord in sulfate and sulfide minerals of early, deep, pre-Main Stage porphyry Cu–Mo and late Main Stage base-metal mineral deposits, Butte district, Montana.Chem. Geol. 215 (1–4), 61–93.

Gaetani, G.A., Grove, T.L., 1997. Partitioning of moderately siderophile elements amongolivine, silicate melt, and sulfide melt: constraints on core formation in the Earthand Mars. Geochim. Cosmochim. Acta 61 (9), 1829–1846.

Gao, S., Rudnick, R.L., Yuan, H.L., Liu, X.M., Liu, Y.S., Xu, W.L., Ling, W.L., Ayers, J., Wang, X.C.,Wang, Q.H., 2004. Recycling lower continental crust in the North China craton. Nature432 (7019), 892–897.

Gao, Y.L., Zhang, J.M., Ye, H.S., Meng, F., Zhou, K., Gao, Y., 2010. Geological characteristicsand molybdenite Re–Os isotopic dating of Shiyaogou porphyry molybdenum depositin the East Qinling. Acta Petrol. Sin. 26 (3), 729–739.

Garrido, I., Cembrano, J., Sina, A., Stedman, P., Yanez, G., 2002. Highmagma oxidation stateand bulk crustal shortening: key factors in the genesis of Andean porphyry copperdeposits, central Chile (31–34° S). Rev. Geol. Chile 29 (1), 43–54.

Giggenbach, W.F., 1992a. Isotopic shifts in waters from geothermal and volcanic systemsalong convergent plate boundaries and their origin. Earth Planet. Sci. Lett. 113 (4),495–510.

Giggenbach, W.F., 1992b. SEG distinguished lecture — magma degassing and mineral de-position in hydrothermal systems along convergent plate boundaries. Econ. Geol.Bull. Soc. Econ. Geol. 87 (7), 1927–1944.

Gonzalez-Partida, E., Levresse, G., Carrillo-Chavez, A., Cheilletz, A., Gasquet, D., Jones, D.,2003. Paleocene adakite Au–Fe bearing rocks, Mezcala, Mexico: evidence from geo-chemical characteristics. J. Geochem. Explor. 80 (1), 25–40.

Green, T.H., Watson, E.B., 1982. Crystallization of apatite in natural magmas under high-pressure, hydrous conditions, with particular reference to orogenic rock series.Contrib. Mineral. Petrol. 79 (1), 96–105.

Groome,W.G., Thorkelson, D.J., 2009. The three-dimensional thermo-mechanical signatureof ridge subduction and slab window migration. Tectonophysics 464 (1–4), 70–83.

Guo, F., Fan,W., Li, C., Zhang, S.-H., Zhao, Y., Song, B., Yang, Z., Hu, J.-M.,Wu, H., 2006. Geo-chemistry of lateMesozoic adakites from the Sulu belt, eastern China:magma genesisand implications for crustal recycling beneath continental collisional orogens:Carboniferous granitic plutons from the northern margin of the North China Block:implications for a late Palaeozoic active continental margin. Geol. Mag. 143 (1), 1–13.

Halter, W., Heinrich, C., Pettke, T., 2005. Magma evolution and the formation of porphyryCu–Au ore fluids: evidence from silicate and sulfide melt inclusions. Mineral.Deposita 39 (8), 845–863.

Han, Y.G., Zhang, S.H., Pirajno, F., Zhou, X.W., Zhao, G.C., Qu, W.J., Liu, S.H., Zhang, J.M.,Liang, H.B., Yang, K., 2013. U–Pb and Re–Os isotopic systematics and zircon Ce4+/Ce3+ ratios in the Shiyaogou Mo deposit in eastern Qinling, central China: insightsinto the oxidation state of granitoids and Mo (Au) mineralization. Ore Geol. Rev.55, 29–47.

Haschke, M., Ahmadian, J., Murata, M., McDonald, I., 2010. Copper mineralizationprevented by arc-root delamination during Alpine–Himalayan collision in centralIran. Econ. Geol. 105 (4), 855–865.

Hedenquist, J.W., Lowenstern, J.B., 1994. The role of magmas in the formation of hydro-thermal ore-deposits. Nature 370 (6490), 519–527.

Hedenquist, J.W., Arribas, A., Reynolds, T.J., 1998. Evolution of an intrusion-centeredhydrothermal system: Far Southeast–Lepanto porphyry and epithermal Cu–Audeposits, Philippines. Econ. Geol. Bull. Soc. Econ. Geol. 93 (4), 373–404.

Heinrich, C.A., 1990. The chemistry of hydrothermal tin (–tungsten) ore deposition. Econ.Geol. 85, 457–481.

Heinrich, C.A., 2005. The physical and chemical evolution of low-salinity magmatic fluidsat the porphyry to epithermal transition: a thermodynamic study. Mineral. Deposita39, 864–889.

Heinrich, C.A., Driesner, T., Stefansson, A., Seward, T.M., 2004. Magmatic vapor contractionand the transport of gold from the porphyry environment to epithermal ore deposits.Geology 32 (9), 761–764.

Hemingway, B.S., 1990. Thermodynamic properties for bunsenite, NiO, magnetite, Fe3O4,and hematite, Fe2O3 with comments on selected oxygen buffer reactions. Am. Miner-al. 75, 781–790.

Hezarkhani, A., 2006. Hydrothermal evolution of the Sar–Cheshmeh porphyry Cu–Mo de-posit, Iran: evidence from fluid inclusions. J. Asian Earth Sci. 28 (4–6), 409–422.

Hoefs, J. (Ed.), 2004. Stable Isotope Geochemistry. Springer, Germany.Hofmann, A.W., 1988. Chemical differentiation of the Earth: the relationship between

mantle, continental crust, and oceanic crust. Earth Planet. Sci. Lett. 90, 297–314.

Hollings, P., Cooke, D., Clark, A., 2005. Regional geochemistry of tertiary igneous rocks incentral Chile: implications for the geodynamic environment of giant porphyry copperand epithermal gold mineralization. Econ. Geol. 100 (5), 887–904.

Hou, Z.Q., Gao, Y.F., Qu, X.M., Rui, Z.Y., Mo, X.X., 2004. Origin of adakitic intrusives gener-ated during mid-Miocene east–west extension in southern Tibet. Earth Planet. Sci.Lett. 220 (1–2), 139–155.

Hou, Z.Q., Pan, X.F., Yang, Z.M., Qu, X.M., 2007a. Porphyry Cu–(Mo–Au) deposits no relat-ed to oceanic-slab subduction: examples from Chinese porphyry deposits in conti-nental settings. Geoscience 21 (2), 332–351 (in Chinese with English abstract).

Hou, Z.Q., Xie, Y.L., Xu, W.Y., Li, Y.Q., Zhu, X.K., Khin, Z., Beaudoin, G., Rui, Z.Y., Wei, H.A.,Ciren, L., 2007b. Yulong deposit, eastern Tibet: a high-sulfidation Cu–Au porphyrycopper deposit in the eastern Indo-Asian collision zone. Int. Geol. Rev. 49 (3),235–258.

Hou, Z.Q., Yang, Z.M., Qu, X.M., Meng, X.J., Li, Z.Q., Beaudoin, G., Rui, Z.Y., Gao, Y.F., Zaw, K.,2009. The Miocene Gangdese porphyry copper belt generated during post-collisionalextension in the Tibetan Orogen. Ore Geol. Rev. 36 (1–3), 25–51.

Huang, F., Li, S.G., Dong, F., He, Y.S., Chen, F.K., 2008. High-Mg adakitic rocks in the Dabieorogen, central China: implications for foundering mechanism of lower continentalcrust. Chem. Geol. 255 (1–2), 1–13.

Huebner, J.S., 1987. Use of gas mixtures at low pressure to specify oxygen and other fu-gacities of furnace atmospheres. Hydrothermal Experimental Techniques, pp. 20–60.

Huebner, J.S., Sato, M., 1970. The oxygen fugacity–temperature relationships of manga-nese oxide and nickel oxide buffers. Am. Mineral. 55, 934–952.

Imai, A., 2001. Generation and evolution of ore fluids for porphyry Cu–Au mineralizationof the Santo Tomas II (Philex) deposit, Philippines. Resour. Geol. 51 (2), 71–96.

Imai, A., 2002. Metallogenesis of porphyry Cu deposits of the western Luzon arc,Philippines: K–Ar ages, SO3 contents of microphenocrystic apatite and significanceof intrusive rocks. Resour. Geol. 52 (2), 147–161.

Imai, A., 2005. Evolution of hydrothermal system at the Dizon porphyry Cu–Au deposit,Zambales, Philippines. Resour. Geol. 55 (2), 73–90.

Imai, A., Ohno, S., 2005. Primary ore mineral assemblage and fluid inclusion study of theBatu Hijau porphyry Cu–Au deposit, Sumbawa, Indonesia. Resour. Geol. 55 (3),239–248.

Imai, A., Listanco, E.L., Fujii, T., 1993. Petrologic and sulfur isotopic significance of highlyoxidized and sulfur-rich magma of Mt-Pinatubo, Philippines. Geology 21 (8),699–702.

Imai, A., Ohbuchi, Y., Tanaka, T., Morita, S., Yasunaga, K., 2007. Characteristics of porphyryCu mineralization at Waisoi (Namosi district), Viti Levu, Fiji. Resour. Geol. 57 (4),374–385.

Ishihara, S., 1977. The magnetite-series and ilmenite-series granitic rock. Min. Geol. 27,930–941.

Ishihara, S., 2004. The redox state of granitoids relative to tectonic setting and earth his-tory: the magnetite–ilmenite series 30 years later. Trans. R. Soc. Edinb. Earth Sci. 95,23–33.

Ishihara, S., Matsuhisa, Y., 1999. Oxygen isotopic constraints on the geneses of the Mio-cene Outer Zone granitoids in Japan. Lithos 46 (3), 523–534.

Ishihara, S., Murakami, H., 2006. Fractionated ilmenite-series granites in Southwest Japan:source magma for REE–Sn–W mineralizations. Resour. Geol. 56 (3), 245–256.

Ishihara, S., Sasaki, A., 1991. Ore-deposits related to granitic magmatism in Japan — amagmatic viewpoint. Episodes 14 (3), 286–292.

Ishihara, S., Terashima, S., 1989. Carbon contents of the magnetite-series and ilmenite-series granitoids in Japan. Geochem. J. 23 (1), 25–36.

Jenkyns, H.C., 2010. Geochemistry of oceanic anoxic events. Geochem. Geophys. Geosyst.11, Q03004. http://dx.doi.org/10.1029/2009GC002788.

Jenner, F.E., O'Neill, H.S.C., Arculus, R.J., Mavrogenes, J.A., 2010. The magnetite crisis in theevolution of arc-related magmas and the initial concentration of Au, Ag and Cu.J. Petrol. 51 (12), 2445–2464.

Jenner, F.E., O'neill, H.S.C., Arculus, R.J., Mavrogenes, J.A., 2012. The magnetite crisis in theevolution of arc-related magmas and the initial concentration of Au, Ag and Cu (vol51, pg 2445, 2010). J. Petrol. 53 (5), 1089-1089.

Jiang, Y.H., Jiang, S.Y., Ling, H.F., Dai, B.Z., 2006. Low-degree melting of a metasomatizedlithospheric mantle for the origin of Cenozoic Yulong monzogranite-porphyry, eastTibet: geochemical and Sr–Nd–Pb–Hf isotopic constraints. Earth Planet. Sci. Lett.241 (3–4), 617–633.

Jugo, P.J., 2009. Sulfur content at sulfide saturation in oxidized magmas. Geology 37 (5),415–418.

Jugo, P.J., Luth, R.W., Richards, J.P., 2005. Experimental data on the speciation of sulfur as afunction of oxygen fugacity in basaltic melts. Geochim. Cosmochim. Acta 69 (2),497–503.

Jugo, P.J., Wilke, M., Botcharnikov, R.E., 2010. Sulfur K-edge XANES analysis of natural andsynthetic basaltic glasses: implications for S speciation and S content as function ofoxygen fugacity. Geochim. Cosmochim. Acta 74 (20), 5926–5938.

Kavalieris, I., Rinchin, O., Gombojav, J., Gombosuren, N., Crane, D., Orssich, C., Bat-Erdene,K., 2011. Characteristics of the Oyu Tolgoi porphyry deposits, South Gobi, Mongolia.Let's Talk Ore Deposits vols. I and Ii, (342–344 pp.).

Kay, R.W., 1978. Aleutian magnesian andesite: melts from subducted Pacific Ocean crust.J. Volcanol. Geotherm. Res. 4, 117–132.

Kay, S.M., Godoy, E., Kurtz, A., 2005. Episodic arc migration, crustal thickening, subductionerosion, and magmatism in the south-central Andes. Geol. Soc. Am. Bull. 117 (1–2),67–88.

Kelley, K.A., Cottrell, E., 2009. Water and the oxidation state of subduction zone magmas.Science 325 (5940), 605–607.

Kelley, K.A., Cottrell, E., 2012. The influence of magmatic differentiation on the oxidationstate of Fe in a basaltic arc magma. Earth Planet. Sci. Lett. 329, 109–121.

Keppler, H., 2010. The distribution of sulfur between haplogranitic melts and aqueousfluids. Geochim. Cosmochim. Acta 74 (2), 645–660.















































































































































http://dx.doi.org/10.1029/2009GC002788

































Kesler, S.E., 1997. Metallogenic evolution of convergent margins: selected ore depositmodels. Ore Geol. Rev. 12 (3), 153–171.

Khashgerel, B.E., Kavalieris, I., Hayashi, K., 2008. Mineralogy, textures, and whole-rockgeochemistry of advanced argillic alteration: Hugo Dummett porphyry Cu-Au depos-it, Oyu Tolgoi mineral district. Mongolia. Mineral. Deposita 43, 913–932.

Kinoshita, O., 1997. Slab window-related magmatism caused by the Kula–Pacific ridgesubduction beneath the Eurasia continent in the Cretaceous. Episodes 20 (3),185–187.

Kishima, N., 1989. A thermodynamic study on the pyrite–pyrrhotite–magnetite–watersystem at 300–500 °C with relevance to the fugacity/concentration quotient of aque-ous H2S. Geochim. Cosmochim. Acta 53, 2143–2155.

Klemm, L.M., Pettke, T., Heinbicth, C.A., Campos, E., 2007. Hydrothermal evolution of theEl Teniente deposit, Chile: Porphyry Cu–Mo ore deposition from low-salinity mag-matic fluids. Econ. Geol. 102 (6), 1021–1045.

Klemm, L.M., Pettke, T., Heinrich, C.A., 2008. Fluid and source magma evolution ofthe Questa porphyry Mo deposit, New Mexico, USA. Mineral. Deposita 43 (5),533–552.

Kress, V.C., Carmichael, I.S.E., 1991. The compressibility of silicate liquids containing Fe2O3

and the effect of composition, temperature, oxygen fugacity and pressure on theirredox states. Contrib. Mineral. Petrol. 108, 82–92.

Lang, J.R., Lueck, B., Mortensen, J.K., Russell, J.K., Stanley, C.R., Thompson, J.F.H., 1995.Triassic–Jurassic silica-undersaturated and silica-saturated alkalic intrusions in thecordillera of British-Columbia — implications for arc magmatism. Geology 23 (5),451–454.

Lee, C.-T.A., 2014. Copper conundrums. Nat. Geosci. 7, 10–11.Lee, C.T.A., Leeman, W.P., Canil, D., Li, Z.X.A., 2005. Similar V/Sc systematics in MORB and

arc basalts: implications for the oxygen fugacities of their mantle source regions. J.Petrol. 46 (11), 2313–2336.

Lee, C.T.A., Luffi, P., Le Roux, V., Dasgupta, R., Albarede, F., Leeman, W.P., 2010. The redoxstate of arc mantle using Zn/Fe systematics. Nature 468 (7324), 681–685.

Lee, C.T.A., Luffi, P., Chin, E.J., Bouchet, R., Dasgupta, R., Morton, D.M., Le Roux, V., Yin, Q.Z.,Jin, D., 2012. Copper systematics in arc magmas and implications for crust–mantledifferentiation. Science 336 (6077), 64–68.

Lee, C.-T.A., Lee, T.C., Wu, C.-T., 2014. Modeling the compositional evolution of recharging,evacuating, and fractionating (REFC) magma chambers: implications for differentia-tion of arc magmas. Geochim. Cosmochim. Acta http://dx.doi.org/10.1016/j.gca.2013.08.009 (in press).

Lerchbaumer, L., Audetat, A., 2012. High Cu concentrations in vapor-type fluid inclusions:an artifact? Geochim. Cosmochim. Acta 88, 255–274.

Li, Y., Audetat, A., 2012. Partitioning of V, Mn, Co, Ni, Cu, Zn, As, Mo, Ag, Sn, Sb, W, Au, Pb,and Bi between sulfide phases and hydrous basanite melt at uppermantle conditions.Earth Planet. Sci. Lett. 355, 327–340.

Li, Y., Audetat, A., 2013. Gold solubility and partitioning between sulfide liquid,monosulfide solid solution and hydrous mantle melts: implications for the formationof Au-richmagmas and crust–mantle differentiation. Geochim. Cosmochim. Acta 118,247–262.

Li, X.F., Sasaki, M., 2007. Hydrothermal alteration and mineralization of middle JurassicDexing porphyry Cu–Mo deposit, southeast China. Resour. Geol. 57 (4), 409–426.

Li, X.F., Watanabe, Y., Mao, J.W., Liu, S.X., Yi, X.K., 2007. Sensitive high-resolution ion mi-croprobe U–Pb Zircon and 40Ar–39Ar muscovite ages of the Yinshan deposit in thenortheast Jiangxi province, South China. Resour. Geol. 57 (3), 325–337.

Li, J.W., Zhao, X.F., Zhou, M.F., Vasconcelos, P., Ma, C.Q., Deng, X.D., de Souza, Z.S., Zhao, Y.X., Wu, G., 2008. Origin of the Tongshankou porphyry-skarn Cu–Mo deposit, easternYangtze craton, Eastern China: geochronological, geochemical, and Sr–Nd–Hf isotopicconstraints. Mineral. Deposita 43 (3), 315–336.

Li, H., Zhang, H., Ling, M.X., Wang, F.Y., Ding, X., Zhou, J.B., Yang, X.Y., Tu, X.L., Sun, W.D.,2011. Geochemical and zircon U–Pb study of the Huangmeijian A-type granite: impli-cations for geological evolution of the Lower Yangtze River belt. Int. Geol. Rev. 53(5–6), 499–525.

Li, C.Y., Wang, F.Y., Hao, X.L., Zhang, H., Ling, M.X., Zhou, J.B., Li, Y.L., Fan, W.M., Sun, W.D.,2012a. Formation of the world's largest molybdenum metallogenic belt: a plate-tectonic perspective on the Qinling molybdenum deposits. Int. Geol. Rev. 54 (9),1093–1112.

Li, C.Y., Zhang, H., Wang, F.Y., Liu, J.Q., Sun, Y.L., Hao, X.L., Li, Y.L., Sun,W.D., 2012b. The for-mation of the Dabaoshan porphyry molybdenum deposit induced by slab rollback.Lithos 150, 101–110.

Li, H., Ling, M.X., Li, C.Y., Zhang, H., Ding, X., Yang, X.Y., Fan,W.M., Li, Y.L., Sun,W.D., 2012c.A-type granite belts of two chemical subgroups in central eastern China: indication ofridge subduction. Lithos 150, 26–36.

Li, H.Y., Wang, X.X., Ye, H.S., Yang, L., 2012d. Emplacement ages and petrogenesis of themolybdenum-bearing granites in the Jinduicheng area of East Qinling, China: con-straints from zircon U–Pb ages and Hf isotopes. Acta Geol. Sin. Eng. Ed. 86 (3),661–679.

Li, J.L., Gao, J., John, T., Klemd, R., Su, W., 2013a. Fluid-mediated metal transport in subduc-tion zones and its link to arc-related giant ore deposits: constraints from a sulfide-bearing HP vein in lawsonite eclogite (Tianshan, China). Geochim. Cosmochim. Acta120, 326–362.

Li, N., Chen, Y.J., Pirajno, F., Ni, Z.Y., 2013b. Timing of the Yuchiling giant porphyry Mo sys-tem, and implications for ore genesis. Mineral. Deposita 48 (4), 505–524.

Li, X.F., Hu, R.Z., Rusk, B., Xiao, R., Wang, C.Y., Yang, F., 2013c. U–Pb and Ar–Ar geochronol-ogy of the Fujiawu porphyry Cu–Mo deposit, Dexing district, Southeast China: impli-cations for magmatism, hydrothermal alteration, and mineralization. J. Asian EarthSci. 74, 330–342.

Liang, H.Y., Campbell, I.H., Allen, C., Sun, W.D., Liu, C.Q., Yu, H.X., Xie, Y.W., Zhang, Y.Q.,2006. Zircon Ce4+/Ce3+ ratios and ages for Yulong ore-bearing porphyries in easternTibet. Mineral. Deposita 41 (2), 152–159.

Liang, H.Y., Campbell, I.H., Allen, C.M., Sun, W.D., Yu, H.X., Xie, Y.W., Zhang, Y.Q.,2007. The age of the potassic alkaline igneous rocks along the Ailao Shan–RedRiver shear zone: implications for the onset age of left-lateral shearing. J. Geol.115 (2), 231–242.

Liang, H.Y., Sun, W.D., Su, W.C., Zartman, R.E., 2009. Porphyry copper–gold mineralizationat Yulong, China, promoted by decreasing redox potential during magnetite alter-ation. Econ. Geol. 104 (4), 587–596.

Ling, M.X., Wang, F.Y., Ding, X., Hu, Y.H., Zhou, J.B., Zartman, R.E., Yang, X.Y., Sun, W.D.,2009. Cretaceous ridge subduction along the Lower Yangtse River Belt, EasternChina. Econ. Geol. 104 (2), 303–321.

Ling, M.X., Wang, F.Y., Ding, X., Zhou, J.B., Sun, W.D., 2011. Different origins of adakitesfrom the Dabie Mountains and the Lower Yangtze River Belt, eastern China:geochemical constraints. Int. Geol. Rev. 53 (5–6), 727–740.

Ling, M.X., Li, Y., Ding, X., Teng, F.Z., Yang, X.Y., Fan,W.M., Xu, Y.G., Sun,W.D., 2013. Destruc-tion of the North China Craton induced by ridge subductions. J. Geol. 121 (2), 197–213.

Liu, L.J., Spasojevic, S., Gurnis, M., 2008. Reconstructing Farallon plate subduction beneathNorth America back to the Late Cretaceous. Science 322 (5903), 934–938.

Liu, S.A., Li, S.G., He, Y.S., Huang, F., 2010. Geochemical contrasts between early Cretaceousore-bearing and ore-barren high-Mg adakites in central-eastern China: implicationsfor petrogenesis and Cu–Au mineralization. Geochim. Cosmochim. Acta 74,7160–7178.

Liu, X., Fan, H.R., Hu, F.F., Hu, B.G., Zhu, X.Y., 2011. SEM-EDS investigation of daughtermin-erals of fluid inclusions at the Dexing porphyry Cu–Mo deposit, Jiangxi Province,China. Acta Petrol. Sin. 27 (5), 1397–1409.

Liu, S.A., Li, S.G., Guo, S.S., Hou, Z.H., He, Y.S., 2012. The Cretaceous adakitic–basaltic–granitic magma sequence on south-eastern margin of the North China Craton: impli-cations for lithospheric thinning mechanism. Lithos 134, 163–178.

Liu, X., Xiong, X., Audétat, A., Li, Y., Song, M., Li, L., Sun, W., Ding, X., 2014. Partitioning ofcopper between olivine, orthopyroxene, clinopyroxene, spinel, garnet and silicatemelts at upper mantle conditions. Geochim. Cosmochim. Acta 125, 1–22.

Lorand, J.P., 1990. Are spinel lherzolite xenoliths representative of the abundance of sulfurin the upper mantle. Geochim. Cosmochim. Acta 54 (5), 1487–1492.

Lowell, J.D., Guilbert, J.M., 1970. Lateral and vertical alteration–mineralization zoning inporphyry ore deposits. Econ. Geol. 65 (4) (373-).

Lynch, G., Ortega, S., 1997. Hydrothermal alteration and tourmaline–albite equilibria atthe Coxheath porphyry Cu–Mo–Au deposit, Nova Scotia. Can. Mineral. 35, 79–94.

Macpherson, C.G., Dreher, S.T., Thirlwall, M.F., 2006. Adakites without slab melting: highpressure differentiation of island arc magma, Mindanao, the Philippines. Earth Planet.Sci. Lett. 243 (3–4), 581–593.

Madsen, J.K., Thorkelson, D.J., Friedman, R.M., Marshall, D.D., 2006. Cenozoic to Recentplate configurations in the Pacific Basin: ridge subduction and slab windowmagmatism in western North America. Geosphere 2 (1), 11–34.

Mallmann, G., O'Neill, H.S.C., 2007. The effect of oxygen fugacity on the partitioning of Rebetween crystals and silicate melt duringmantlemelting. Geochim. Cosmochim. Acta71 (11), 2837–2857.

Mao, J.W., Xie, G.Q., Bierlein, F., Qu, W.J., Du, A.D., Ye, H.S., Pirajno, F., Li, H.M., Guo, B.J., Li,Y.F., Yang, Z.Q., 2008. Tectonic implications from Re–Os dating of Mesozoic molybde-num deposits in the East Qinling–Dabie orogenic belt. Geochim. Cosmochim. Acta 72(18), 4607–4626.

Mao, J.W., Pirajno, F., Xiang, J.F., Gao, J.J., Ye, H.S., Li, Y.F., Guo, B.J., 2011. Mesozoic molyb-denum deposits in the east Qinling–Dabie orogenic belt: characteristics and tectonicsettings. Ore Geol. Rev. 43 (1), 264–293.

Martin, H., 1999. Adakitic magmas: modern analogues of Archaean granitoids. Lithos 46(3), 411–429.

Mathur, R., Ruiz, J., Titley, S., Gibbins, S., Margotomo, W., 2000. Different crustal sourcesfor Au-rich and Au-poor ores of the Grasberg Cu–Au porphyry deposit. Earth Planet.Sci. Lett. 183 (1–2), 7–14.

Mavrogenes, J.A., O'Neill, H.S.C., 1999. The relative effects of pressure, temperature andoxygen fugacity on the solubility of sulfide in mafic magmas. Geochim. Cosmochim.Acta 63 (7–8), 1173–1180.

McDonough, W.F., Sun, S.S., 1995. The composition of the Earth. Chem. Geol. 120 (3–4),223–253.

Moss, R., Scott, S.D., Binns, R.A., 2001. Gold content of eastern Manus basin volcanic rocks:implications for enrichment in associated hydrothermal precipitates. Econ. Geol. Bull.Soc. Econ. Geol. 96 (1), 91–107.

Mungall, J.E., 2002. Roasting themantle: slabmelting and the genesis of major Au and Au-rich Cu deposits. Geology 30 (10), 915–918.

Mungall, J.E., Brenan, J.M., 2014. Partitioning of platinum-group elements and Au betweensulfide liquid and basalt and the origins of mantle–crust fractionation of thechalcophile elements. Geochim. Cosmochim. Acta 125, 265–289.

Mutschler, F.E., Ludington, S., Bookstrom, A.A., 2010. Giant porphyryrelated metal campsof the world—a database. USGS Open-File Report: 99-556. (6 pp. (http://geopubs.wr.usgs.gov/open-file/of99-556/)).

Myers, J., Eugster, H., 1983. The system Fe–Si–O: oxygen buffer calibrations to 1500 K.Contrib. Mineral. Petrol. 82, 75–90.

O'Neil, H.St.C., Pownceby, M.I., 1993. Thermodynamic data from redox reactions at hightemperatures. Ι. An experimental and theoretical assessment. Contrib. Mineral. Petrol.114, 296–314.

O'Neill, H.S.C., Mavrogenes, J.A., 2002. The sulfide capacity and the sulfur content at sul-fide saturation of silicate melts at 1400 °C and 1 bar. J. Petrol. 43 (6), 1049–1087.

Ossandon, G., Freraut, R., Gustafson, L.B., Lindsay, D.D., Zentilli, M., 2001. Geology of theChuquicamata mine: a progress report. Econ. Geol. Bull. Soc. Econ. Geol. 96 (2),249–270.

Oyarzun, R., Marquez, A., Lillo, J., Lopez, I., Rivera, S., 2001. Giant versus small porphyrycopper deposits of Cenozoic age in northern Chile: adakitic versus normal calc-alkaline magmatism. Mineral. Deposita 36 (8), 794–798.


































http://dx.doi.org/10.1016/j.gca.2013.08.009

http://dx.doi.org/10.1016/j.gca.2013.08.009






















































































































http://geopubs.wr.usgs.gov/open-file/of99-556/

http://geopubs.wr.usgs.gov/open-file/of99-556/















Oyarzun, R., Marquez, A., Lillo, J., Lopez, I., Rivera, S., 2002. Reply to discussion on “Giantversus small porphyry copper deposits of Cenozoic age in northern Chile: adakiticversus normal calc-alkaline magmatism” by Oyarzun R, Marquez A, Lillo J, Lopez I,Rivera S (Mineralium Deposita 36: 794–798, 2001). Mineral. Deposita 37 (8),795–799.

Parkinson, I.J., Arculus, R.J., 1999. The redox state of subduction zones: insights from arc-peridotites. Chem. Geol. 160 (4), 409–423.

Patricio, C.C., Gonzalo, R.S., 2001. Oxide mineralization at the Radomiro Tomic porphyrycopper deposit, northern Chile. Econ. Geol. 96, 387–400.

Patten, C., Barnes, S.J., Mathez, E.A., Jenner, F.E., 2013. Partition coefficients of chalcophileelements between sulfide and silicate melts and the early crystallization history ofsulfide liquid: LA-ICP-MS analysis of MORB sulfide droplets. Chem. Geol. 358,170–188.

Peach, C.L., Mathez, E.A., Keays, R.R., 1990. Sulfide melt silicate melt distribution coeffi-cients for noble-metals and other chalcophile elements as deduced from MORB —

implications for partial melting. Geochim. Cosmochim. Acta 54 (12), 3379–3389.Peng, G.Y., Luhr, J.F., McGee, J.J., 1997. Factors controlling sulfur concentrations in volcanic

apatite. Am. Mineral. 82 (11–12), 1210–1224.Pettke, T., Oberli, F., Heinrich, C.A., 2010. The magma and metal source of giant porphyry-

type ore deposits, based on lead isotope microanalysis of individual fluid inclusions.Earth Planet. Sci. Lett. 296 (3–4), 267–277.

Pokrovski, G.S., Dubrovinsky, L.S., 2011. The S3− ion is stable in geological fluids at elevatedtemperatures and pressures. Science 331 (6020), 1052–1054.

Qu, X., Hou, Z., Li, Y., 2004. Melt components derived from a subducted slab in late oro-genic ore-bearing porphyries in the Gangdese copper belt, southern Tibetan plateau.Lithos 74 (3–4), 131–148.

Rabbia, O.M., Hernandez, L.B., French, D.H., King, R.W., Ayers, J.C., 2009. The El Tenienteporphyry Cu–Mo deposit from a hydrothermal rutile perspective. Mineral. Deposita44 (8), 849–866.

Rajamani, V., Naldrett, A.J., 1978. Partitioning of Fe, Co, Ni, and Cu between sulfide liquidand basaltic melts and composition of Ni–Cu sulfide deposits. Econ. Geol. 73 (1),82–93.

Richards, J.P., 1999. Tectonomagmatic controls on localization of porphyry copper de-posits, Chile. Mineral Deposits: Processes to Processing vols. 1 and 2, (425–428 pp.).

Richards, J.P., 2011a. High Sr/Y arc magmas and porphyry Cu ± Mo ± Au deposits: justadd water. Econ. Geol. 106 (7), 1075–1081.

Richards, J.P., 2011b. Magmatic to hydrothermal metal fluxes in convergent and collidedmargins. Ore Geol. Rev. 40 (1), 1–26.

Richards, J.P., 2012. Discussion of Sun et al. (2011): ‘The genetic association of adakitesand Cu–Au ore deposits’. Int. Geol. Rev. 368–369.

Richards, J.P., 2013. Giant ore deposits formed by optimal alignments and combinations ofgeological processes. Nat. Geosci. 6 (11), 911–916.

Richards, J.P., 2014. Discussion of : the link between reduced porphyry copper depositsand oxidized magmas. Geochim. Cosmochim. Acta 126, 643–645.

Richards, J.P., Kerrich, R., 2007. Adakite-like rocks: their diverse origins and questionablerole in metallogenesis. Econ. Geol. 102 (4), 537–576.

Richards, J.P., Mumin, A.H., 2013. Magmatic–hydrothermal processes within an evolvingEarth: iron oxide–copper–gold and porphyry Cu ± Mo ± Au deposits. Geology 41(7), 767–770.

Rowins, S.M., 2000. Reduced porphyry copper–gold deposits: a new variation on an oldtheme. Geology 28 (6), 491–494.

Rudnick, R.L., Gao, S., 2003. Composition of the continental crust. In: Heinrich, D.H.,Turekian, K.K. (Eds.), Treatise on Geochemistry. Pergamon, Oxford, pp. 1–64.

Sajona, F.G., Maury, R.C., 1998. Association of adakites with gold and copper mineraliza-tion in the Philippines. C. R. Acad. Sci. IIA Earth Planet. Sci. 326 (1), 27–34.

Scharman, M.R., Pavlis, T.L., Ruppert, N., 2012. Crustal stabilization through the processesof ridge subduction: examples from the Chugach metamorphic complex, southernAlaska. Earth Planet. Sci. Lett. 329, 122–132.

Seedorff, E., Einaudi, M.T., 2004a. Henderson porphyry molybdenum system, Colorado: I.Sequence and abundance of hydrothermal mineral assemblages, flow paths ofevolving fluids, and evolutionary style. Econ. Geol. Bull. Soc. Econ. Geol. 99 (1), 3–37.

Seedorff, E., Einaudi, M.T., 2004b. Henderson porphyry molybdenum system. Colorado: II.Decoupling of introduction and deposition of metals during geochemical evolution ofhydrothermal fluids. Econ. Geol. Bull. Soc. Econ. Geol. 99 (1), 39–72.

Seo, J.H., Heinrich, C.A., 2013. Selective copper diffusion into quartz-hosted vaporinclusions: evidence from other host minerals, driving forces, and consequences forCu–Au ore formation. Geochim. Cosmochim. Acta 113, 60–69.

Seo, J.H., Guillong, M., Heinrich, C.A., 2009. The role of sulfur in the formation of mag-matic–hydrothermal copper–gold deposits. Earth Planet. Sci. Lett. 282 (1–4),323–328.

Shafiei, B., Haschke, M., Shahabpour, J., 2009. Recycling of orogenic arc crust triggers por-phyry Cu mineralization in Kerman Cenozoic arc rocks, southeastern Iran. Mineral.Deposita 44 (3), 265–283.

Shahabpour, J., 1991. Some secondary ore formation features of the Sar–Cheshmeh por-phyry copper–molybdenum deposit, Kerman, Iran. Mineral. Deposita 26 (4),275–280.

Shaw, H.T., 1963. Hydrogen–water vapor mixtures: control of hydrothermal atmospheresby hydrogen osmosis. Science 139, 1220–1222.

Shen, P., Pan, H.D., 2013. Country-rock contamination of magmas associated with theBaogutu porphyry Cu deposit, Xinjiang, China. Lithos 177, 451–469.

Shen, P., Shen, Y.C., Pan, H.D., Wang, J.B., Zhang, R., Zhang, Y.X., 2010a. Baogutu porphyryCu–Mo–Au deposit, West Junggar, northwest China: petrology, alteration, andminer-alization. Econ. Geol. 105 (5), 947–970.

Shen, P., Shen, Y.C., Wang, J.B., Zhu, H.P., Wang, L.J., Meng, L., 2010b. Methane-rich fluidevolution of the Baogutu porphyry Cu–Mo–Au deposit, Xinjiang, NW China. Chem.Geol. 275 (1–2), 78–98.

Shen, P., Shen, Y., Pan, H., Li, X.H., Dong, L., Wang, J., Zhu, H., Dai, H., Guan, W., 2012.Geochronology and isotope geochemistry of the Baogutu porphyry copper depositin the West Junggar region, Xinjiang, China. J. Asian Earth Sci. 49, 99–115.

Shen, P., Pan, H.D., Xiao, W.J., Chen, X.H., Eleonorad, S., Shen, Y.C., 2013a. Twogeodynamic–metallogenic events in the Balkhash (Kazakhstan) and the WestJunggar (China): Carboniferous porphyry Cu and Permian greisenW–Momineraliza-tion. Int. Geol. Rev. 55 (13), 1660–1687.

Shen, P., Pan, H.D., Xiao, W.J., Li, X.H., Dai, H.W., Zhu, H.P., 2013b. Early Carboniferousintra-oceanic arc and back-arc basin system in the West Junggar, NW China. Int.Geol. Rev. 55 (16), 1991–2007.

Shi, P.F., 1992. Fluid fugacities and phase equilibria in the Fe–Si–O–H–S system. Am. Min-eral. 77, 1050–1066.

Sillitoe, R.H., 1999. Styles of high-sulphidation gold, silver and copper mineralisation inporphyry and epithermal environments. Pacrim'99: International Congress onEarth Science, Exploration and Mining Around the Pacific Rim, Proceedings. 99(29–44 pp.).

Sillitoe, R.H., 2010. Porphyry copper systems. Econ. Geol. 105 (1), 3–41.Sillitoe, R.H., Burrows, D.R., 2002. New field evidence bearing on the origin of the El Laco

magnetite deposit, northern Chile. Econ. Geol. Bull. Soc. Econ. Geol. 97 (5),1101–1109.

Singer, D.A., 2008. Mineral deposit densities for estimating mineral resources. Math.Geosci. 40 (1), 33–46.

Smith, C.M., Canil, D., Rowins, S.M., Friedman, R., 2012. Reduced graniticmagmas in an arcsetting: the Catface porphyry Cu–Mo deposit of the Paleogene Cascade Arc. Lithos154, 361–373.

Smithson, D.M., Rowins, S.M., 2005. Reduced I-type magmatism and porphyry Cu–Aumineralization in the west Central Cascades, WA: the ca. 37 Ma North Fork Deposit.Geochim. Cosmochim. Acta 69 (10), A566-A566.

Song, S.G., Su, L., Niu, Y.L., Lai, Y., Zhang, L.F., 2009. CH4 inclusions in orogenic harzburgite:evidence for reduced slab fluids and implication for redox melting in mantle wedge.Geochim. Cosmochim. Acta 73 (6), 1737–1754.

Sotnikov, V.I., Ponomarchuk, V.A., Pertseva, A.P., Berzina, A.P., Berzina, A.N., Girnon, V.O.,2004. Evolution of sulfur isotopes in porphyry Cu–Mo ore-magmatic systems ofSiberia and Mongolia. Geol. Geofiz. 45 (8), 963–974.

Spry, P.G., Scott, S.D., 1986. The stability of zincian spinels in sulfide systems and their po-tential as exploration guides for metamorphosedmassive sulfide deposits. Econ. Geol.81, 1446–1461.

Spry, P.G., Paredes, M.M., Foster, F., Truckle, J.S., Chadwick, T.H., 1996. Evidence for a ge-netic link between gold–silver telluride and porphyry molybdenum mineralizationat the Golden Sunlight deposit, Whitehall, Montana: fluid inclusion and stable isotopestudies. Econ. Geol. Bull. Soc. Econ. Geol. 91 (3), 507–526.

Stagno, V., Frost, D.J., 2010. Carbon speciation in the asthenosphere: experimental mea-surements of the redox conditions at which carbonate-bearing melts coexist withgraphite or diamond in peridotite assemblages. Earth Planet. Sci. Lett. 300 (1–2),72–84.

Stagno, V., Ojwang, D.O., McCammon, C.A., Frost, D.J., 2013. The oxidation state of themantle and the extraction of carbon from Earth's interior. Nature 493 (7430)(84–+).

Stein, H.J., Markey, R.J., Morgan, J.W., Du, A., Sun, Y., 1997. Highly precise and accurate Re–Os ages for molybdenite from the East Qinling molybdenum belt, Shaanxi Province,China. Econ. Geol. Bull. Soc. Econ. Geol. 92 (7–8), 827–835.

Stern, C.R., Funk, J.A., Skewes, M.A., Arevalo, A., 2007. Magmatic anhydrite in plutonicrocks at the El Teniente Cu–Mo deposit chile, and the role of sulfur- and copper-rich magmas in its formation. Econ. Geol. 102 (7), 1335–1344.

Stern, C.R., Skewes, M.A., Arevalo, A., 2011. Magmatic evolution of the Giant El TenienteCu–Mo Deposit, Central Chile. J. Petrol. 52 (7–8), 1591–1617.

Sun, W.D., Williams, I.S., Li, S.G., 2002. Carboniferous and triassic eclogites in the westernDabie Mountains, east-central China: evidence for protracted convergence of theNorth and South China Blocks. J. Metamorph. Geol. 20 (9), 873–886.

Sun, W.D., Arculus, R.J., Bennett, V.C., Eggins, S.M., Binns, R.A., 2003a. Evidence for rheni-um enrichment in themantle wedge from submarine arc-like volcanic glasses (PapuaNew Guinea). Geology 31 (10), 845–848.

Sun, W.D., Bennett, V.C., Eggins, S.M., Arculus, R.J., Perfit, M.R., 2003b. Rhenium systemat-ics in submarine MORB and back-arc basin glasses: laser ablation ICP-MS results.Chem. Geol. 196 (1–4), 259–281.

Sun, W.D., Bennett, V.C., Eggins, S.M., Kamenetsky, V.S., Arculus, R.J., 2003c. Enhancedmantle-to-crust rhenium transfer in undegassed arc magmas. Nature 422 (6929),294–297.

Sun,W.D., Arculus, R.J., Kamenetsky, V.S., Binns, R.A., 2004a. Release of gold-bearing fluidsin convergent margin magmas prompted by magnetite crystallization. Nature 431(7011), 975–978.

Sun,W.D., Bennett, V.C., Kamenetsky, V.S., 2004b. Themechanism of Re enrichment in arcmagmas: evidence from Lau Basin basaltic glasses and primitive melt inclusions.Earth Planet. Sci. Lett. 222 (1), 101–114.

Sun, W.D., Ding, X., Hu, Y.H., Li, X.H., 2007. The golden transformation of the Cretaceousplate subduction in the west Pacific. Earth Planet. Sci. Lett. 262 (3–4), 533–542.

Sun, X.M., Tang, Q., Sun, W.D., Xu, L., Zhai, W., Liang, J.L., Liang, Y.H., Shen, K., Zhang, Z.M.,Zhou, B., Wang, F.Y., 2007. Monazite, iron oxide and barite exsolutions in apatite ag-gregates from CCSD drillhole eclogites and their geological implications. Geochim.Cosmochim. Acta 71 (11), 2896–2905.

Sun, W.D., Ling, M.X., Yang, X.Y., Fan, W.M., Ding, X., Liang, H.Y., 2010. Ridge subductionand porphyry copper–gold mineralization: an overview. Sci. China Ser. D Earth Sci.53 (4), 475–484.

Sun,W.D., Zhang, H., Ling, M.X., Ding, X., Chung, S.L., Zhou, J.B., Yang, X.Y., Fan,W.M., 2011.The genetic association of adakites and Cu–Au ore deposits. Int. Geol. Rev. 53 (5–6),691–703.













































































































































































Sun, W.D., Ling, M.X., Chung, S.L., Ding, X., Yang, X.Y., Liang, H.Y., Fan, W.M., Goldfarb, R.,Yin, Q.Z., 2012a. Geochemical constraints on adakites of different origins and coppermineralization. J. Geol. 120 (1), 105–120.

Sun, W.D., Ling, M.X., Ding, X., Chung, S.L., Yang, X.Y., Fan, W.M., 2012b. The genetic asso-ciation of adakites and Cu–Au ore deposits: a reply. Int. Geol. Rev. 54 (3), 370–372.

Sun,W.D., Li, S., Yang, X.Y., Ling, M.X., Ding, X., Duan, L.A., Zhan, M.Z., Zhang, H., Fan,W.M.,2013a. Large-scale gold mineralization in eastern China induced by an Early Creta-ceous clockwise change in Pacific plate motions. Int. Geol. Rev. 55 (3), 311–321.

Sun, W.D., Liang, H.Y., Ling, M.X., Zhan, M.Z., Ding, X., Zhang, H., Yang, X.Y., Li, Y.L., Ireland,T.R.,Wei, Q.R., Fan,W.M., 2013b. The link between reduced porphyry copper depositsand oxidized magmas. Geochim. Cosmochim. Acta 103, 263–275.

Sun, W.D., Huang, R.F., Liang, H.Y., Ling, M.X., Li, C.Y., Ding, X., Zhang, H., Yang, X.Y.,Ireland, T., Fan, W.M., 2014a. Magnetite–hematite, oxygen fugacity, adakite and por-phyry copper deposits: reply to Richards. Geochim. Cosmochim. Acta 126, 646–649.

Sun, W.D., et al., 2014b. The genetic association between magnetite–hematite and por-phyry copper deposits: reply to Pokrovski. Geochim. Cosmochim. Acta 126, 639–642.

Takagi, T., 2004. Origin of magnetite- and ilmenite-series granitic rocks in the Japan Arc.Am. J. Sci. 304 (2), 169–202.

Tang, G.J., Wang, Q., Wyman, D.A., Li, Z.X., Zhao, Z.H., Jia, X.H., Jiang, Z.Q., 2010. Ridge sub-duction and crustal growth in the Central Asian Orogenic Belt: evidence from LateCarboniferous adakites and high-Mg diorites in the western Junggar region, northernXinjiang (west China). Chem. Geol. 277 (3–4), 281–300.

Taylor, H.P., 1974. The application of oxygen and hydrogen isotope studies to problems ofhydrothermal alteration and ore deposition. Econ. Geol. 69, 843–883.

Taylor, B.E., 1992. Degassing of H2O from rhyolite magma during eruption and shallowinrusion, and the isotopic composition of magmatic water in hydrothermal systems.Geol. Surv. Jpn. Rep. 279, 190–194.

Thieblemont, D., Stein, G., Lescuyer, J.L., 1997. Epithermal and porphyry deposits: theadakite connection. C.R. Acad. Sci., Ser. IIa: Sci. Terre Planets 325 (2), 103–109.

Thompson, J.F.H., Sillitoe, R.H., Baker, T., Lang, J.R., Mortensen, J.K., 1999. Intrusion-relategold deposits associated with tungsten–tin provinces. Mineral. Deposita 34, 323–334.

Thorkelson, D.J., Breitsprecher, K., 2005. Partial melting of slab window margins: genesisof adakitic and non-adakitic magmas. Lithos 79 (1–2), 25–41.

Titley, S.R., 1981. Porphyry copper. Am. Sci. 69 (6), 632–638.Togashi, S., Terashima, S., 1997. The behavior of gold in unaltered island arc tholeiitic

rocks from Izu-Oshima, Fuji, and Osoreyama volcanic areas, Japan. Geochim.Cosmochim. Acta 61 (3), 543–554.

Tomkins, A.G., 2010. Windows of metamorphic sulfur liberation in the crust: implicationsfor gold deposit genesis. Geochim. Cosmochim. Acta 74 (11), 3246–3259.

Vila, T., Sillitoe, R.H., 1991. Gold-rich porphyry systems in the Maricunga Belt, NorthernChile. Econ. Geol. Bull. Soc. Econ. Geol. 86 (6), 1238–1260.

Vila, T., Sillitoe, R.H., Betzhold, J., Viteri, E., 1991. The porphyry gold deposit at Marte,Northern Chile. Econ. Geol. Bull. Soc. Econ. Geol. 86 (6), 1271–1286.

Vry, V.H., Wilkinson, J.J., Seguel, J., Millan, J., 2010. Multistage intrusion, brecciation, andveining at El Teniente, evolution of a nested porphyry system. Econ. Geol. 105 (1),119–153.

Wang, Q., Wyman, D.A., Xu, J.F., Zhao, Z.H., Jian, P., Xiong, X.L., Bao, Z.W., Li, C.F., Bai, Z.H.,2006a. Petrogenesis of Cretaceous adakitic and shoshonitic igneous rocks in theLuzong area, Anhui Province (eastern China): implications for geodynamics andCu–Au mineralization. Lithos 89 (3–4), 424–446.

Wang, Q., Xu, J.F., Jian, P., Bao, Z.W., Zhao, Z.H., Li, C.F., Xiong, X.L., Ma, J.L., 2006b. Petro-genesis of adakitic porphyries in an extensional tectonic setting, Dexing, SouthChina: implications for the genesis of porphyry copper mineralization. J. Petrol. 47(1), 119–144.

Wang, Q., Wyman, D.A., Xu, J., Jian, P., Zhao, Z., Li, C., Xu, W., Ma, J., He, B., 2007a. Early Cre-taceous adakitic granites in the Northern Dabie Complex, central China: implicationsfor partial melting and delamination of thickened lower crust. Geochim. Cosmochim.Acta 71 (10), 2609–2636.

Wang, Q., Wyman, D.A., Xu, J.F., Zhao, Z.H., Jian, P., Zi, F., 2007b. Partial melting of thick-ened or delaminated lower crust in the middle of eastern China: implications forCu–Au mineralization. J. Geol. 115 (2), 149–161.

Wang, G.G., Ni, P., Wang, R.C., Zhao, K.D., Chen, H., Ding, J.Y., Zhao, C., Cai, Y.T., 2013. Geo-logical, fluid inclusion and isotopic studies of the Yinshan Cu–Au–Pb–Zn–Ag deposit,

South China: implications for ore genesis and exploration. J. Asian Earth Sci. 74,343–360.

Wilkinson, J.J., 2013. Triggers for the formation of porphyry ore deposits inmagmatic arcs.Nat. Geosci. 6 (11), 917–925.

Wones, D.R., Eugster, H.P., 1965. Stability of biotite: experiement, theory, and application.Am. Mineral. 50, 1228–1272.

Wu, H.Y., Zhang, L.C., Wan, B., Chen, Z.G., Zhang, X.J., Xiang, P., 2011. Geochronological andgeochemical constraints on Aolunhua porphyry Mo–Cu deposit, northeast China, andits tectonic significance. Ore Geol. Rev. 43 (1), 78–91.

Xiao, B., Qin, K.Z., Li, G.M., Li, J.X., Xia, D.X., Chen, L., Zhao, J.X., 2012. Highly oxidizedmagma and fluid evolution of Miocene Qulong giant porphyry Cu–Mo deposit, South-ern Tibet, China. Resour. Geol. 62 (1), 4–18.

Xu, J.F., Shinjo, R., Defant, M.J., Wang, Q.A., Rapp, R.P., 2002. Origin of Mesozoic adakitic in-trusive rocks in the Ningzhen area of east China: Partial melting of delaminated lowercontinental crust? Geology 30 (12), 1111–1114.

Xu, W.L., Wang, Q.H., Wang, D.Y., Guo, J.H., Pei, F.P., 2006. Mesozoic adakitic rocks fromthe Xuzhou–Suzhou area, eastern China: evidence for partial melting of delaminatedlower continental crust. J. Asian Earth Sci. 27 (4), 454–464.

Yang, X.Y., Wang, K.R., Yang, X.M., Sun, L.G., 2002. Characteristics of mineralization andgold occurrence in Shaxi porphyry copper–gold deposit, central Anhui, China.Neues Jahrb. Mineral. Abh. 177 (3), 293–320.

Yang, Z.M., Hou, Z.Q., White, N.C., Chang, Z.S., Li, Z.Q., Song, Y.C., 2009. Geology of the post-collisional porphyry copper–molybdenum deposit at Qulong, Tibet. Ore Geol. Rev. 36(1–3), 133–159.

Yang, L., Chen, F.K., Liu, B.X., Hu, Z.P., Qi, Y., Wu, J.D., He, J.F., Siebel,W., 2013. Geochemistryand Sr–Nd–Pb–Hf isotopic composition of the Donggou Mo-bearing granite porphy-ry, Qinling orogenic belt, central China. Int. Geol. Rev. 55 (10), 1261–1279.

Yogodzinski, G.M., Volynets, O.N., Koloskov, A.V., Seliverstov, N.I., Matvenkov, V.V., 1994.Magnesian andesites and the subduction component in a strongly calc-alkaline seriesat Piip Volcano, Far Western Aleutians. J. Petrol. 35 (1), 163–204.

Zeng, Q.D., Liu, J.M., Qin, K.Z., Fan, H.R., Chu, S.X., Wang, Y.B., Zhou, L.L., 2013. Types, char-acteristics, and time–space distribution of molybdenum deposits in China. Int. Geol.Rev. 55 (11), 1311–1358.

Zhang, Q., Wang, Y., Wang, Y.L., 2001a. Preliminary study on the components of the lowercrust in east China Plateau during Yanshanian Period: constraints on Sr and Nd isoto-pic compositions of adakite-like rocks. Acta Petrol. Sin. 17 (4), 505–513.

Zhang, Q., Wang, Y., Qian, Q., Yang, J.H., Wang, Y.L., Zhao, T.P., Guo, G.J., 2001b. The char-acteristics and tectonic–metallogenic significances of the adakites in Yanshan periodfrom eastern China. Acta Petrol. Sin. 17 (2), 236–244 (in Chinese with Englishabstract).

Zhang, Z.X., Yang, F.Q., Yan, S.H., Zhang, R., Chai, F.M., Liu, F.L., Geng, X.X., 2010. Sources ofore-forming fluids and materials of the Baogutu porphyry copper deposit in Xinjiang:constrains from sulfur–hydrogen–oxygen isotopes geochemistry. Acta Geol. Sin. 26,707–716 (in Chinese with English abstract).

Zhang, H., Ling,M.X., Liu, Y.L., Tu, X.L.,Wang, F.Y., Li, C.Y., Liang, H.Y., Yang, X.Y., Arndt, N.T.,Sun, W.D., 2013. High oxygen fugacity and slab melting linked to Cu mineralization:evidence from Dexing porphyry copper deposits, southeastern China. J. Geol. 121 (3),289–305.

Zhang, Y., Sun, J.G., Chen, Y.J., Zhao, K.Q., Gu, A.L., 2013. Re–Os and U–Pb geochronology ofporphyry Mo deposits in central Jilin Province: Mo ore-forming stages in northeastChina. Int. Geol. Rev. 55 (14), 1763–1785.

Zhang, H., Li, C.Y., Yang, X.Y., Sun, Y.L., Deng, J.H., Liang, H.Y.,Wang, R.L., Wang, B.H.,Wang,Y.X., Sun, W.D., 2013. Shapinggou: the largest Climax-type porphyry Mo deposit inChina. Int. Geol. Rev. 56, 313–331.

Zhou, Q., Jiang, Y.H., Zhang, H.H., Liao, S.Y., Jin, G.D., Zhao, P., Jia, R.Y., Liu, Z., 2013. Mantleorigin of the Dexing porphyry copper deposit, SE China. Int. Geol. Rev. 55 (3),337–349.

Zhu, X., Huang, C., Rui, Z., Zhou, Y., Zhu, X., Hu, C., Mei, Z., 1983. The Geology of DexingPorphyry Copper Ore Field. Geological Publishing House, Beijing (in Chinese).

Zhu, L.M., Zhang, G.W., Guo, B., Lee, B., Gong, H.J., Wang, F., 2010. Geochemistry of theJinduicheng Mo-bearing porphyry and deposit, and its implications for thegeodynamic setting in East Qinling, PR China. Chem. Erde 70 (2), 159–174.























































































































porphyry deposits and oxidized magmas_2015

Documents

porphyry cu deposits

porphyry mineralization

porphyry depositoxidized

porphyry depositsare

c tporphyry deposits

reviewporphyry deposits

porphyry minerthe ore

curich cualization