Oxidized sulfur-rich arc magmas formed porphyry Cu deposits by 1.88 Ga

Most known porphyry Cu deposits formed in the Phanerozoic and are exclusively associated with moderately oxidized, sulfur-rich, hydrous arc-related magmas derived from partial melting of the asthenospheric mantle metasomatized by slab-derived fluids. Yet, whether similar metallogenic processes also operated in the Precambrian remains obscure. Here we address the issue by investigating the origin, fO2, and S contents of calc-alkaline plutonic rocks associated with the Haib porphyry Cu deposit in the Paleoproterozoic Richtersveld Magmatic Arc (southern Namibia), an interpreted mature island-arc setting. We show that the ca. 1886–1881 Ma ore-forming magmas, originated from a mantle-dominated source with minor crustal contributions, were relatively oxidized (1‒2 log units above the fayalite-magnetite-quartz redox buffer) and sulfur-rich. These results indicate that moderately oxidized, sulfur-rich arc magma associated with porphyry Cu mineralization already existed in the late Paleoproterozoic, probably as a result of recycling of sulfate-rich seawater or sediments from the subducted oceanic lithosphere at that time.


Supplementary Note 1: the Haib porphyry Cu deposit
The Haib porphyry Cu deposit is located in the western Richtersveld Magmatic Arc of southern Namibia, where the arc assemblage mainly comprises the greenschist facies Orange River Volcanic Group and underlying younger Vioolsdrif Plutonic Suite (Fig.   1). In Haib, the Vioolsdrif Plutonic Suite is remarkably undeformed compared to the localized intense deformation of the Orange River Volcanic Group along the NWtrending shear zones (Fig. 1), which may reflect rheological contrasts between strong crystalline plutonic rocks and the weaker phyllosilicate-rich volcanic rocks. General lack of deformation in the volcanic rocks within Haib and near contacts with the granodiorite intrusions may indicate local induration by contact metamorphism prior to the regional deformation. Metamorphism to greenschist grade occurs during the ~1.1 Ga Namaqua Orogeny 1 and has variably affected the rocks in Haib.
The granodiorite porphyry locally contains xenoliths of plagioclase-phyric andesite porphyry and is intruded with a sharp contact by leucocratic granodiorite porphyry dikes, as observed in drill core ( Supplementary Fig. 2c-e). Granodiorite in the batholith locally contains quartz monzonite enclaves ( Supplementary Fig. 2f). Aplite dikes cut the granodiorite porphyry and intrude equigranular granodiorite of the batholith (Supplementary Fig. 2h and i). The dike is poorly mineralized and is crosscut by late quartz-ankerite veins, which is interpreted to have formed later than the main porphyry Cu-mineralization stage.
Note that the plagioclase-phyric andesite porphyry, granodiorite porphyry, and leucocratic granodiorite porphyry are formerly referred to as feldspar porphyry, quartz-feldspar porphyry, and quartz-feldspar porphyry II in refs. 2,3 , and reports from the Deep Resources, Inc. A phase of quartz biotite porphyry is also identified, but its geochronology and lithogeochemical composition are comparable to the granodiorite porphyry 3 . We therefore argue that petrographic differences may reflect various degrees of alteration. A mineralized dioritic feldspar porphyry dike containing granodiorite porphyry xenolith is identified ( Supplementary Fig. 1b), which is interpreted to be a breccia 3 and is not studied here.

Alteration
Alteration at Haib has been described in detail in refs. 2,3 , and the noted alteration assemblages are comparable to Phanerozoic porphyry Cu systems 4, 5 , which include early potassic alteration with local overprinting by albite, chlorite, and sericite alteration types ( Supplementary Fig. 1c). Pervasive biotite and minor K-feldspar alteration are commonly associated with epidote, anhydrite, muscovite, and minor tourmaline, titanite, and rutile ( Supplementary Fig. 4a-f). Epidote is rarely associated with potassic alteration and mineralization in typical Phanerozoic porphyry Cu systems, and its presence in Haib may reflect the hydrothermal fluid being more calcic. Local K-feldspar selvages are seen around veins, and most of this alteration has associated minor molybdenite mineralization.
Albite alteration is rare and was only identified in the granodiorite porphyry at a depth of ~780 m in drill hole TCDH-10 ( Supplementary Fig. 1c). It is characterized by replacement of K-feldspar with albite and is associated with low-Cu grades.
Chlorite alteration is mainly observed in the plagioclase-phyric andesite porphyry Sericite alteration is locally seen in drill core and mainly replaces plagioclase and biotite ( Supplementary Fig. 4h). At the property surface, which is considered as the shallow levels of the system, the potassic alteration is locally overprinted by sericite alteration whereby sericite completely replaces plagioclase.
Regional metamorphism during Namaqua Orogeny 1 is mainly characterized by the presence of epidote, sericite, and minor chlorite, and has variably affected all of the rock types in the area.
The sinuous quartz A veins, with granular textures, contain relatively minor amounts of sulfide minerals, mainly chalcopyrite, and are in equilibrium with potassic alteration in the wall rock, but also locally have narrow K-feldspar alteration halos ( Supplementary Fig. 4f).
Quartz ± molybdenite ± chalcopyrite ± pyrite B veins are rarely found and pyritic D To evaluate the relative timing of crystallization of titanite/zircon-hosted apatite to volatile saturation in parent melt for Haib, we plot apatite halogen compositions in Supplementary Fig. 9. We restrict the plotting to analyses containing Cl content above the detection limit (190 ppm), and these analyses are mainly for samples HB-30 (diorite) and HB-51 (leucocratic granodiorite porphyry). One single analysis is also available for granodiorite (sample HB-34) but cannot form a trend in itself, and is therefore not included. The c-axis of most of the apatite grains are perpendicular to the electron beam, so that beam damages are interpreted to have been minimized (see the Methods section in the main text).
The apatite XF/XOH ratio decreases with decreasing XCl/XOH ratio, which is consistent with the modeling result for apatite crystallized in a volatile-undersaturated environment in ref. 6 , suggesting that the apatites crystallized earlier than magmatic degassing. This is also consistent with the fact that apatite from individual samples yielded limited compositional variability (Supplementary Data 9) and lacks zoning ( Supplementary Fig. 9).

Supplementary Note 3: P-T correction of magmatic fO2 from apatite S µ-XANES data
In silicate glasses, it has been demonstrated that a decrease in temperature of 100 °C and a decrease in pressure of 300 MPa may result in an approximate deviation of ΔFMQ + 0.5 and -0.2, respectively 9, 10 . Given that apatite may crystallize as a nearliquidus phase, we use the model apatite saturation temperature (AST) 11