Introduction

The Tethyan orogenic belt, extending for 12,000 km from the Pyrenees through the Alps and the Turkish–Iranian plateau, across Pakistan and the Himalayan–Tibetan Plateau, and into the Indochina Peninsula (Fig. 1a), is one of the world’s most important porphyry Cu belts1,2,3,4. This orogen formed through successive stages of continental breakup, ocean basin formation, and eventual collision between Gondwana-derived continents and Eurasia (Supplementary Fig. 1 and Supplementary Material). The Tethys realm can be divided chronologically into the Proto-, Paleo-, and Neo-Tethys oceans, formed during the Early Paleozoic, Late Paleozoic, and Mesozoic, respectively5. Of particular interest is the Neo-Tethys Ocean, which opened in the Early Permian and underwent protracted subduction during the Mesozoic, culminating in extensive Cenozoic continental collisions that led to the formation of porphyry Cu deposits in regions such as the Gangdese and Yulong belts (Tibet), the Kerman belt (Iran), and the Anatolides belt (Turkey)1,2,5,6,7,8 (Fig. 1a and Supplementary Material).

Fig. 1: Spatial and temporal distribution of Tethyan porphyry Cu deposits.
figure 1

a Topographic map (www.mapswire.com) of the Tethyan domain highlighting Tethyan sutures3,4, porphyry deposit locations, and names of giant deposits (Cu > 2 Mt). b Temporal distribution of porphyry Cu deposits (polygons) across different metallogenic belts, illustrating their association with oceanic subduction (blue arrows) and continental collision (gray shadows). Detailed information about the deposits is available in the Supplementary Materials and Supplementary Data 1.

Full size image

Despite a long Mesozoic subduction history, significantly fewer and smaller porphyry Cu deposits formed during subduction-related magmatism than during Cenozoic collisional or post-collisional magmatic stages in the Tethyan belt (Figs. 1b, 2c). The reason for this disparity remains elusive. A key control on porphyry copper formation is the oxidation state of the magmatic system, i.e., magmatic oxygen fugacity, commonly expressed as ∆FMQ (the log of oxygen fugacity relative to the fayalite–magnetite–quartz buffer). Oxidized arc magmas (∆FMQ ≈ +1 to +2) favor the mobilization of chalcophile elements, enabling their transport to the upper crust9,10. Detrital zircon provides a valuable record of magmatic redox conditions over geologic timescales. Global geological events–particularly oceanic anoxic events and seawater incursion events–can influence the geochemical composition of subducted sediments11,12, altering mantle wedge oxygen fugacity and affecting porphyry mineralization potential.

Fig. 2: Redox variations and associated geological events in the Tethyan domain.
figure 2

a Global sea-level curve43 and average surface temperature trends42; b Number of hydrocarbon source rocks38 and the proportion of global oil and gas reserves39,40. c Ore-forming ages and Cu–Au resources of Tethyan porphyry Cu deposits (Supplementary Data 1) alongside major oceanic anoxic events 26. d Zircon ∆FMQ variation over time, presented as binned averages with a bin size of 5 Myr. Error bars represent ±2 standard errors of the mean. e Degree of restriction in sedimentary basins, with regions of potential anoxia highlighted in red and areas of well-oxygenated waters shown in blue44.

Full size image

Here, we apply a multidisciplinary approach to track the oxygen fugacity evolution of Tethyan magmas from the Early Jurassic to the Miocene. Using the novel zircon oxybarometer, we reconstruct ∆FMQ through time and integrate these data with sedimentary rocks, global sea-level and temperature changes, whole-rock geochemical data, and the timing of major geological events. Our findings illuminate the mechanisms responsible for redox variations in the mantle wedge and their implications for subduction- vs. collision-related porphyry Cu deposits.

Results

Redox variation from Mesozoic to Cenozoic

Detrital zircon oxybarometry reveals an overall increase in ∆FMQ from the Early Jurassic through the Miocene in the Tethyan belt (Fig. 2d). Although data density during the Jurassic to Early Cretaceous is relatively low and thus more variable, Mesozoic ∆FMQ values are dominantly below +1. By contrast, Cenozoic ∆FMQ values commonly exceed +1, consistent with a transition to more oxidized magmatic conditions (Fig. 2d).

Discussion

Explaining the scarcity of Mesozoic subduction-related porphyry Cu deposits

The scarcity of Mesozoic subduction-related porphyry deposits in the Tethyan belt contrasts with the abundance of Cenozoic collision-related porphyry deposits (Fig. 2c). The porphyry Cu deposits are formed in the upper crust, usually at a depth of 1 to 5 km13, and globally occur mainly in the Phanerozoic, particularly in the Cenozoic14. While some workers have suggested that extensive erosion may have removed older subduction-related porphyry Cu deposits15,16, there is no credible evidence for widespread pre-Cenozoic porphyry Cu mineralization before the continental collision in the Tethyan belt.

The fertile intrusions for porphyry Cu formation worldwide are oxidized with ∆FMQ + 1 to +29,17,18. By contrast, the notably low ∆FMQ (<+1) during the Jurassic and Cretaceous suggests that arc magmas were relatively reduced, contrasting to high ∆FMQ (>+1) during the Cenozoic (Fig. 2d). The V/Sc ratios are also widely used as a redox proxy in igneous rocks, with higher ratios indicating more oxidized magmas19. The oxidation states of mid-oceanic ridge basalts (MORBs) close to ∆FMQ ~ 0 have V/Sc ratios of 6.74 ± 1.11. We compiled V/Sc data from Tethyan magmatic rocks (Supplementary Data 3). The result shows the lower V/Sc ratios of Mesozoic rocks (median = 6.9, n = 631) are like MORBs, while the Cenozoic rocks have higher V/Sc ratios (median = 8.4, n = 786) (Supplementary Fig. 2). The V/Sc results are consistent with our zircon ΔFMQ results, both indicated the Cenozoic ƒO2 is higher than the Mesozoic. Under low ƒO2, sulfur tends to form sulfides. The chalcophile elements (e.g., Cu, Au) have high partition coefficients between sulfide and silicate melts10. As a result, the chalcophile elements were sequestered in the lower crust via sulfide accumulation under low ƒO2, resulting in rare mineralization in the upper crust9,10,20. This process parallels the anoxic conditions inferred for the Paleo-Tethys Ocean basin during the Permian when the basin was in a restricted environment near the equator21,22. The correlation between redox state and porphyry Cu deposit frequency (Fig. 2c, d) supports redox-driven control over the spatiotemporal distribution of Tethyan porphyry Cu mineralization.

The Andes porphyry copper belt is a critical part of the Circum-Pacific metallogenic domain, where porphyry deposits are primarily related to oceanic subduction23. Andes porphyry copper belt is south-north trending, always in an open environment, and hence less prone to anoxia. This is consistent with the high ∆FMQ (>+1) in the Andes belt, which may display a tendency to increase in ∆FMQ from +1.0 around the equator to +2.0 at high latitude24. This is different from the east-west trending Tethyan metallogenic domain.

Mechanisms of redox state variation

Oceanic subduction-related arc magmas are commonly oxidized due to slab-derived fluids25, but the Neo-Tethys domain appears to have bucked this trend during the Mesozoic, exhibiting lower ∆FMQ (Fig. 2d). To elucidate this discrepancy, we compiled data on sedimentary rocks, global sea-level and temperature changes, whole-rock geochemical data, and major global geological events (Figs. 2–4).

Fig. 3: Representative stratigraphic columns from the Tethyan domain.
figure 3

The locations of these stratigraphic columns are indicated in Supplementary Fig. 1c. Stratigraphic data are modified from the following sources: Columns 174, 275, 3–512, and 676.

Full size image
Fig. 4: Whole-rock elemental ratios and Monte Carlo isotopic simulation for mafic rocks in the Tethyan domain.
figure 4

a Ba/La versus Th/Yb; b Lu/Hf versus Th/La; c143Nd/144Nd versus initial 87Sr/86Sr in arc settings; d143Nd/144Nd versus initial 87Sr/86Sr in collisional settings. Data for the mafic rocks are collected from the database GEOROC and listed in Supplementary Data 3. The crosses with different colors represent random mixing of pelagic or terrigenous sediments with the depleted mantle at variable proportions from a Monte Carlo simulation. Details of the end-members are listed in Supplementary Data 4.

Full size image

Subduction of reduced, organic-rich sediments in the Mesozoic

Multiple Jurassic and Cretaceous oceanic anoxic events are recognized in the Tethyan belt, including the early Toarcian (~183 Ma), Callovian (~166 Ma), early Aptian (~120 Ma), early Albian (~111 Ma), late Albin (~102 Ma and ~100 Ma), Cenomanian–Turonian (~93 Ma), and late Coniacian to Santonian (~86 Ma)26,27,28,29,30,31 (Fig. 2c). These events coincided with massive deposition of organic-rich black shales in marine and terrestrial settings26. The compiled stratigraphic columns in the Tethyan domain show abundant organic matter-rich sediments in the Mesozoic (Fig. 3). Subduction of such reduced black shales with organic carbon would release CH4-rich fluids into the mantle wedge, and CH4 acts as a reducing agent, consuming Fe3+ and O2 via: CH4 + Fe2O3 + O2 → FeO+CO2 + H2, decreasing the oxidation state of mantle wedge and related arc magmas32,33,34. The temporal correlation of low ∆FMQ values with oceanic anoxic events suggests that organic-rich subducted slabs played a pivotal role in producing relatively reduced magmas during the Mesozoic (Fig. 2c, d).

This is further supported by whole-rock elemental and isotopic data for mafic rocks in the Tethyan domain (Supplementary Data 3). The mantle can be modified by the subducting slab and overlying sediments. Different sediment melt (pelagic or terrigenous) and fluid display distinct elemental and isotopic signatures. As a result, some specific geochemical signatures can be used to decipher the contribution of fluid and sediment melt to a depleted mantle wedge35,36. Slab fluid has higher concentrations of fluid-mobile elements such as large ion lithophile elements (e.g., Ba, Rb, Sr, K) but lower concentrations of light rare earth elements, thorium, and high field strength elements (e.g., Nb, Ta, Zr, and Hf) compared to sediment melt35,36. Furthermore, owing to the ‘zircon effect’, melted terrigenous sediments will have lower Lu/Hf ratios due to the abundant detrital zircons that enrich Hf compared to melted pelagic sediments that lack detrital zircons37. The high Ba/La, Lu/Hf ratios (Fig. 4a, b), Sr isotopic ratios and low Nd isotopic ratios (Fig. 4c) of Tethyan mafic rocks produced during Mesozoic oceanic subduction suggest more pelagic sediment melts (less than 4%) and/or slab fluid component was added into the mantle magmatic source reservoir. The enriched isotopic features are mainly attributed to mantle source metasomatism rather than crustal contamination, as there are no negative or positive covariation patterns between whole-rock 87Sr/86Sr(i), 143Nd/144Nd ratios and SiO2 contents (Supplementary Fig. 3).

Influence of high sea-level and warm climates

Frequent seawater incursion events during the Mesozoic, coupled with elevated global sea levels and warmer temperatures (Fig. 2a), facilitated extensive organic carbon burial and the formation of hydrocarbon source rocks11,12,38,39,40. Rising sea levels played a critical role in controlling organic facies deposition41. During the Jurassic and Cretaceous, higher global temperatures42 and sea-level rises43 (Fig. 2a) led to increased nutrient influx from continental erosion into the oceans. This influx enhanced marine primary productivity, creating optimal conditions for significant organic matter burial in continental margin basins. Consequently, these conditions drove the formation of extensive hydrocarbon source rocks, establishing the Jurassic and Cretaceous as the world’s most prolific periods for oil and gas generation (Fig. 2b).

The degree of restriction in sedimentary basins–classified as open or restricted–serves as a proxy for oceanic anoxia44. Restricted ocean basins, often enclosed by surrounding land, are more likely to develop anoxic conditions. In contrast, open marine basins are less prone to anoxia44. Although this classification does not take into account detailed geochemical constraints or ocean dynamics (e.g., upwelling, surface currents, or salinity), certain trends can be inferred from these maps. During the Jurassic, the Neo-Tethys Ocean was predominantly in an open state, resulting in limited anoxia (Fig. 2e). However, as the oceanic subduction progressed, the Neo-Tethys basins gradually closed from the Jurassic to the Cretaceous, becoming increasingly restricted and anoxic (Fig. 2e). These periods of basin closure, accompanied late Mesozoic subduction, preceded Cenozoic continent-continent collision, align with widespread oceanic anoxic events and lower zircon ∆FMQ values, and thus are consistent with an anoxic environment (Fig. 2c–e).

In summary, the Neo-Tethys Ocean evolved into a warm, high sea-level, progressively closing restricted ocean basin during the Jurassic and Cretaceous. This environment fostered the deposition of abundant reduced organic matter, the subduction of which would undoubtedly decrease the oxygen fugacity in the mantle wedge (Fig. 2).

Transition to oxidized magmas in the Cenozoic

Cenozoic magmas in the Tethyan belt are predominantly oxidized with ∆FMQ > +1 (Fig. 2d). Several mechanisms have been proposed to explain this elevated Cenozoic oxidation state:

(1) Subduction of oxidized continental sediments. After the Indian, Arabian, and African continents collided with Eurasia, the subduction of oxidized continental sediments (e.g., carbonates and evaporitic sulfates, both are oxidants) of the passive continental margins likely elevated the oxidation state of the mantle wedge and lower crust by sediment dehydration and releasing oxidized fluids2,45,46,47,48,49 (Fig. 3).

(2) Injection of oxidized ultrapotassic rocks into the lower crust. Although ultrapotassic magmas are not commonly associated with porphyry systems in oceanic subduction zones (e.g., Andes belt), they are frequently observed in spatial and temporal association with porphyry Cu mineralization in collisional settings50,51. As mantle-derived ultrapotassic magmas ascend from a paleo depth of ~56 km to ~15 km, their ΔFMQ increases from +0.8 to +3.052. Besides, the stability field of sulfide shifts significantly towards more oxidized conditions with increasing pressure53. At the depth of the lower crust (1–2 GPa), oxidizing sulfides to sulfates requires higher oxygen fugacity (ΔFMQ > 3). Therefore, their capacity to significantly oxidize magma in the lower crust remains limited.

(3) Auto-oxidation during amphibole and/or garnet fractionation. Garnet fractionation, favored by crust thickening induced by continental collision, preferentially removes Fe2+ from magma, thus auto-oxidizing the residual melt54,55,56. Amphibole, another common Fe-bearing mineral in hydrous melt, also contributes to magma oxidation57. Its Fe3+/ƩFe ratio decreases during magma evolution, enriching Fe3+ in the residual melt58. However, these auto-oxidation processes induced by garnet and amphibole fractionation under high pressure and hydrous conditions primarily occur during crustal differentiation rather than in the mantle source.

Overall, the tectonic switch from subduction of reduced pelagic to oxidized terrigenous sediments emerges as a dominant factor in increasing the oxidation state of magmas. Whole-rock elemental and isotopic data further support this inference. The high Th/Yb, Th/La, and Sr isotopic ratios, and low Lu/Hf and Nd isotopic ratios of mafic rocks emplaced during continental collision indicate increased involvement of terrigenous sediments in the mantle source (Fig. 4). Monte Carlo isotopic mixing model between the depleted mantle and terrigenous sediments shows about 0.5–4% of terrigenous sediment melts were added into the mantle source (Fig. 4d). Moreover, the light Mg-Ca isotopic compositions of Tethyan post-collisional potassic-ultrapotassic rocks are interpreted as evidence for CO2-related metasomatism in the mantle source replacing the earlier CH4-related metasomatism59,60,61.

The link between subduction- and collision-related porphyry Cu deposits

The temporal distribution of Mesozoic subduction-related versus Cenozoic collision-related porphyry Cu deposits in the Tethys domain is highly uneven, with the majority of large deposits forming during the Cenozoic. Previous studies proposed a genetic link between these two types of porphyry Cu deposits. Magmas from earlier oceanic subduction emplaced metal-rich sulfide cumulates in the lower crust and these were suggested to have later remelted during continental collision to provide abundant metals for collision-related porphyry Cu deposits62,63.

A critical aspect of this model requires the transition from reduced to oxidized conditions, a feature which is documented with the evidence presented in this study. Thus, a reduced mantle wedge and associated arc magmas are present during oceanic subduction, which is followed by an increased oxidation state during continental collision. The initial reduced mantle wedge is largely attributed to the subduction of organic-rich sediments. Similar conditions have been observed in Japan33,64 and the Paleo-Tethys Ocean basins21,22. Low oxygen fugacity traps chalcophile elements in the lower crust during magma fractionation, resulting in the emplacement of barren arc magmas in the upper crust (Fig. 5a).

Fig. 5: Genetic model linking oceanic subduction and continental collision to porphyry copper formation.
figure 5

a Subduction of reduced, organic matter-rich sediments decreases oxygen fugacity, causing chalcophile elements to accumulate in the lower crust and inhibiting porphyry mineralization. b Subduction of oxidized continental sediments increases oxygen fugacity, converting residual sulfides in the lower crust to sulfates and releasing metals, thereby enabling porphyry mineralization.

Full size image

For these metal-rich sulfide cumulates to be remobilized, the oxidation state must increase. Subduction of oxidized continental sediments (e.g., carbonates and evaporitic sulfates) during collision releases oxidized fluids to the mantle wedge and lower crust, converting residual sulfides in the lower crust to sulfates, thereby liberating metals for porphyry mineralization (Fig. 5b). This process underscores the interplay between subduction and collision in the formation of porphyry Cu deposits.

Implications for mineral exploration

Our findings highlight the crucial role of the nature of subducted sediments in modulating the redox state and controlling the formation of porphyry Cu deposits. Given the importance of oxidation and hydration state for magmas to form porphyry Cu deposits, the key for Cu exploration is to identify areas or periods that may host oxidized and hydrous shallowly emplaced igneous complexes. Magmatic zircons are directly from specific magmatic units and reflect the nature of specific parent magma. However, they require exposure to the surface, limiting application in covered areas. Zircon is a common accessory mineral in felsic igneous provinces. Owing to its mechanical and chemical robustness, zircon can not only provide reliable geochronologic data but also contains metallogenic information, such as oxidation and hydration states17,18, which is becoming a cost-competitive exploration tool for porphyry Cu deposits. Temporally, detrital zircons can be used to trace the evolution of redox and hydration state for a long period of geological history. This study provides an example of the application of detrital zircons in the Tethyan domain. Higher oxidation and hydration state revealed by zircon Ce-U-Ti (Fig. 2) and the ratio of europium anomaly to ytterbium (Supplementary Fig. 4) in Cenozoic intrusions implies higher Cu prospectivity. Spatially, detrital zircons can provide a much wider footprint than other more conventional geochemical exploration media, particularly when applied in paleo-watersheds to identify potential fertile units upstream from the sampling site18,65,66. Moreover, this method can also effectively guide exploration for W-Sn deposits. Economic W-Sn deposits demonstrate preferential association with reduced magmatic systems, as elevated oxygen fugacity triggers premature crystallization of cassiterite and scheelite, thereby hindering efficient transport of Sn-W elements from deep magmatic sources through crustal pathways67. The reduced Mesozoic granitoids in the Neo-Tethyan belt are potential targets for tungsten-tin deposits.

Methods

Materials

The zircon age and trace elements are compiled from the global detrital zircon database68 and our newly acquired data in Tibet. The complete database includes 13981 detrital zircons (see details in Supplementary Data 2).

Zircon U–Pb dating and trace element analyses

Zircon grains were separated from two rivers in Tibet and mounted in epoxy resin for age and trace element determinations. Zircon U–Pb dating and trace element analyses were conducted simultaneously by laser ablation–inductively coupled plasma–mass spectrometry (LA–ICP–MS) in the Mineral and Fluid Inclusion Microanalysis Laboratory, Institute of Geology, Chinese Academy of Geological Sciences, Beijing, China. The NWR 193UC laser ablation system (Elemental Scientific Lasers, USA) was equipped with a Coherent Excistar 200 excimer laser and a two-volume ablation cell. The laser ablation system was coupled to an Agilent 7900 ICP–MS instrument (Agilent, USA). Zircons were mounted in epoxy resin discs, polished to expose grain interiors, cleaned ultrasonically in ultrapure water, and then cleaned again before analysis using analytical-grade methanol. Pre-ablation was conducted before each spot analysis using five laser shots (~0.3 μm in depth) to remove potential surface contamination. The analyses used a 30 μm laser beam diameter with a laser frequency of 8 Hz and fluence of 2 J/cm2. The Iolite software package was used for data reduction69. Zircons 91500 and GJ-1 were the primary and secondary reference materials, respectively. The exponential function was used to correct for down-hole fractionation70. NIST 610 and 91Zr were used to calibrate the trace element concentrations as an external reference material and internal standard, respectively. Zircon age and trace element data are listed in Supplementary Data 2.

Calculation of oxygen fugacity

Data points outside the Tethys domain were excluded. Zircon ages younger than 200 Ma are used for the following discussion to minimize the influence of the Paleo-Tethys Oceanic subduction because the closure of the Paleo-Tethys Ocean occurred in the Late Permian–Late Triassic5. Zircons with a Th/U ratio of <0.1 were used to exclude metamorphic zircon71. Moreover, zircons derived from S-type granites were screened out using zircon P and REE contents72. After screening, 3010 zircon grains were used for reconstructing the Tethyan oxygen fugacity variation (Supplementary Data 2). The novel magmatic oxybarometer using ratios of Ce, U, and Ti in zircon is independent of temperature and pressure, with a standard error of ±0.6 log unit ƒO273. The ∆FMQ values can be calculated by the equation:

$$\Delta {{{\rm{FMQ}}}}=3.998(\pm 0.124)\times \log [{{{\rm{Ce}}}}/ \surd \, ({{{{\rm{U}}}}}_{{{{\rm{i}}}}}\times {{{{\rm{T}}}}}_{{{{\rm{i}}}}})]+2.284(\pm 0.101)$$
(1)

where Ui denotes age-corrected initial U content73. The ∆FMQ values are plotted as binned averages (bin size = 5 Myr).

Monte Carlo isotopic simulation

The Sr-Nd concentrations and isotopic ratios of the sediment–depleted mantle mixture can be calculated using the following equations:

$${Cx}={Cs}\times F+{Cm}\times (1-F)$$
(2)
$${Rx}={Rs}\times F\times {Cs}/{Cx}+{Rm}\times (1-F)\times {Cm}/{Cx}$$
(3)

where x is Sr or Nd, Cx, Cs, and Cm are element concentrations of x after mixing, sediment, and depleted mantle, respectively. Rx, Rs, and Rm are isotopic ratios of x after mixing, sediment, and depleted mantle, respectively, and F is the proportion of sediment-derived melt.

To obtain as many as possible mixture scenarios, random two pelagic or terrigenous sediments were selected and mixed to create a new pelagic or terrigenous sediment end-member. Then the new sediment end-member was mixed with the depleted mantle in random proportions. 30,000 pelagic sediment–mantle mixtures and 30,000 terrigenous sediment–mantle mixtures were generated using this method. The simulation results are shown in Fig. 4c, d.