1 Introduction

It is generally considered that the granitoids related to tin deposits are derived from supracrustal remelting, whereas the magmas associated with porphyry Cu systems are the products of partial melting of mantle wedge metasomatized by dehydrated fluid from subducted oceanic slab (Richards, 2003). Both copper and tin resources explored in the Circum Pacific metallogenic belts are predominant over the world. The copper reserve of the porphyry-skarn copper deposits in these belts reaches 86% of the world reserves (Sillitoe, 2012; Mao et al., 2014), whereas the tin reserve of the granite-related tin deposits in the belts is also more than 80% (Lehmann, 1990), mainly contributed by five largest world-class tin deposits (i.e. Gejiu, Dachang and Dulong in China, Llallagua in Bolivia and San Rafael in Peru). The eastern continental margins of the Pacific Ocean host porphyry-skarn copper deposits, whereas numerous tin deposits occur in the western continental margin. Both porphyry-skarn copper and granite-related tin deposits of these belts have been well studied, particularly those in the eastern continental margins which are referred to as "cradles of metallogeny". In this paper, we provide an overview of the research progress achieved regarding copper and tin deposits in East China, in comparison to those in the eastern margins of Pacific Ocean. We also attempt to identify the major scientific problems remained.

2 Research progress of the Cu and Sn deposits in the continental margins of the eastern Pacific Ocean

The Cordillera-Andean giant metallogenic belt extends for ~10, 000km along the eastern Pacific Ocean and has been central to various models on ore genesis (Fig. 1). Lindgern (1913) developed a theory of magmatic deposit, and argued that most of these deposits were derived from the magmatic-hydrothermal fluids, and proposed the criteria to identify the physico-chemical conditions of ore formation. He also proposed a classification scheme of mineral deposits with sub-division into magmatic exsolution, contact metamorphism, pegmatite, and hydrothermal (high T-P, moderate T-P and low T-P). The classification scheme has been a milestone for modern economic geology.

Granite-related Sn and Cu deposits are commonly of hydrothermal origin and associated with reduced ilmenite-series or S-type and oxidized magnetite-series or I-type granites, respectively. Application of plate tectonics since 1950s-1970s provided insights into the linkage between ore formation and tectonics. Sillitoe(1972a, 1976) and Mitchell and Garson (1981) proposed that the Cu and Sn ore belts in convergent continental margins are mutually parallel based on research on the Andean orogeny in South America. These studies also show that the Cu and Sn mineralization are associated with a variety of granitic rocks that have close spatial-temporal relationship with tectonic events.

2.1 Progress of research on the continental arc in South and North America as the cradle of the theory of porphyry Cu formation

Globally, porphyry Cu deposits are mainly distributed in the continental margin and island arc of the Circum Pacific Rim, Central Asian orogenic belt, and Tethyan. Sillitoe (2013) compiled a spatial distribution map of the main porphyry Cu deposits with proved reserves, which shows that these porphyry Cu deposits are not distributed homogenously but concentrated in certain regions. Cooke et al. (2005) examined ages and copper reserves of the 25 largest porphyry Cu deposits in the world, and found that most of these formed during Cenozoic, and are clustered within central Andean in South America, Southwest Cordillera of North America, and Northwest America such as Alaska. The belt contains three largest porphyry Cu deposits over the world, such as El Teniente, Chuquicamata, and Rio Blanco-Los Bronces in South America. The central Andean and Southwest Cordillera of North America are two largest Cu deposit clusters, whereas the latter produced several models on modern porphyry formation (Leveille and Stegen, 2012). The central Andean in South America is also a window to the processes in active continental margins (Lehmann, 2004). It contains five N-S trending belts from west coast to east inland (Fig. 2): ① Cretaceous IOCG (Iron-oxide-Copper-Gold deposit); ② Paleocene-Early Eocene (66~52Ma) porphyry Cu; ③ Late Eocene-Early Oligocene (42~31Ma) porphyry Cu; ④ Middle Miocene-Early Pliocene (16~5Ma) porphyry Cu ore belts (Deckart et al., 2014); ⑤ granite-related tin-Ag (24~12Ma, Lehmann, 2004). The Northwest Cordillera of North America mainly contains porphyry Cu deposits in continental margin of Late Cretaceous-Early Cenozoic (Leveille and Stegen, 2012; Barra and Valencia, 2014). These deposits are located in Alaska, Arizona, and New Mexico in America and Yukon in Canada. It is worth noting that the porphyry Cu deposits in southwest America not only formed in continental margin but also extends up to ~1000km from California, Arizona, New Mexico, to Utah and North Mexico, somewhere to Texas. This interesting phenomenon attracted Livingston (1973) and Lowell (1974) to argue against the linkage between subducted slab and porphyry Cu deposit and to support a linkage between intraplate extensional environment and porphyry Cu deposit formation. They reckoned that the magma derived from the upper mantle or the lower crust will rise up and stall, and give rise to Cu mineralization along zones of crustal weakness, and this model is suitable for the porphyry Cu mineralization in the Southwest North America. Furthermore, Davies (1989) demonstrated that the ages of the mineral deposits of western North America show a decrease trend from east to west, ranging from 200Ma, 150Ma, to 80~40Ma, and he correlated the metallogeny with terrane accretion. Although the mechanism by which cratonic structures may be propagated laterally into accreted terrane remains unclear, it is evident that the deep-seated structures in the craton have obviously controlled mantle upwelling and the NW-trending mantle-crust reaction where the crustal weakness provided sites for porphyry intrusion and mineralization. After the 1980s, increasing evidences support the porphyry Cu formation to be associated with subduction. For example, Titley and Beane (1981) proposed rapid initial subduction of slab and subsequent slow down during Late Cretaceous, followed by obliquely subduction (Titley, 1993). The North American block rotated dextrally with respect to the Pacific plate, forming an extensional setting for porphyry Cu mineralization. Murphy (2001) speculated that the porphyry Cu formation in southwest America was possibly associated with low-angle plate subduction, although as evidenced only by the sparsity of volcanic rocks, widespread intracontinental deformation, and thick-skinned tectonics. In reality, the tectonic setting of the widespread Late Cretaceous porphyry Cu formation is still controversial and requires further investigation.

2.1.1 Ore-forming environment

Sillitoe(1972a, b) initially noticed the link between porphyry Cu formation and plate subduction, and proposed a tectonic-metallogenic model within subduction-related setting. He suggested that the causative magma is the product of remelting of oceanic crust during subduction. Sillitoe (1998) further argued that the porphyry Cu deposits preferentially formed in a compressional environment under which the crustal thickening was synchronous with the formation of the world-class deposits in central North Chile and Northwest Arizona. Meanwhile, the strike-slip within the Paleocene-Eocene arc in South America is favorable for the formation of giant deposits, such as the Cujaone and Toquepala, whereas the small deposits preferentially occur in the extensional environment in north Chile. Sillitoe (1998) concluded that ① compressional regime can effectively prevent a long-distance transportation of magma through upper crust to form volcanic rocks, thus the shallow-emplaced magma chamber would be relatively large; ② the shallow emplacement of magma chamber under a compressional context is not easy to give rise to the volcanic eruption but help fractional crystallization of a magma chamber that is in favor of volatile saturation and exsolution; ③ it is difficult for a steep tensional fracture to form under compression, which effectively reduces the number of apophyses on top of the magma chamber and helps concentrate magmatic-hydrothermal fluid. Based on the research of the Paleocene-Oligocene igneous rocks in South Peru-North Chile, James and Sacks (1999) concluded that the porphyry Cu formation is related to the change of subduction slab angle from normal, flat, to normal, which is characterized by the following seven factors: ① a wide alkalic magmatic belt; ② compressional plane shortage; ③ dehydration of lithosphere above the flat slab; ④ cooling of lithospheric convection causes abnormal low-temperature geothermal flow; ⑤ back to normal subduction and reacts with hot asthenosphere and dehydrated mantle to cause wet melting; ⑥ extensive melting of crust induced by penetration of mantle melt into the continental lithosphere; ⑦ the lithospheric thinning and weakness favoring circulation of heat flow and resulting in significant shortage and rise of continents. Cooke et al. (2005) reviewed the characteristics of the 25 largest Cu and Au-rich porphyry deposits in the world, and found that six of 7 giant copper and nine of 13 giant gold deposits are estimated to occur in the low-angle subduction of aseismic ridges, seamount chains, or oceanic plateaus which resulted in crustal thickening, rapid uplift, and exhumation. The oceanic subduction in such locations better explains why multistage porphyry Cu mineralization has been recognized in central South America during the Cenozoic. Others also noticed a connection between mid-ocean ridge subduction and porphyry Cu formation (Uyeda and Miyashiro, 1974; Dickinson and Snyder, 1979; Thorkelson, 1996). Rosenbaum et al. (2005) argued that the porphyry Cu deposit clusters in southern Peru and southern Chile were produced from Nazca and Juan Fermand mid-ocean ridge subduction. Recently, Richards and Holm (2013) proposed that the formation of giant and supergiant porphyry Cu deposits is genetically associated with the subduction of major transform structures on the seafloor. The ultramafic-mafic oceanic crust and lithosphere with oceanic transforms and fracture zones undergo metasomatism (e.g., serpentinization), if subducted at convergent plate boundaries (e.g., island arc, or continent), and the hydrous lithospheric-scale structural weakness might cause a vertical tear of the slab aiding in further mantle flow and slab melting. This would trigger the formation of Cu-bearing melts.

2.1.2 Origin of the ore-related magma

Porphyry Cu deposits are genetically related to calc-alkaline to mildle alkaline intermediate-felsic magmas. The rock types range from quartz diorite to granite. The Cu-fertile porphyries in the continental arc are dominantly calc-alkalic and subordinately high-K calc-alkalic, and are characterized by granodiorite and quartz monzonite (Singer et al., 2005). In contrast, the Cu-fertile porphyries in island arc are typically calc-alkalic and characterized by quartz diorite and minor granite diorite and quartz monzonite (Misra, 2000). The calc-alkalic granite porphyry is mainly associated with Cu-Mo mineralization, whereas the alkalic granite is related to Cu-Au mineralization (Sillitoe, 1997; Müeller and Groves, 2000).

Several investigations have been conducted on the magma origin and evolution associated with subduction settings. The calc-alkalic Cu-fertile porphyry magma is thought to be the product of direct melting of subducted oceanic slab (Sillitoe, 1972a, b; Burnham, 1979). Dehydration of the subducted slab metasomatized the mantle wedge which has been partially melted to form arc magmatism (Tatsumi, 1986, 1989; Peacock, 1993; Schmidt and Poli, 1998; Bourdon et al., 2003; Grove et al., 2006, 2012). The product of partial melting of the dehydrated peridotitic asthenosphere is high-Mg basalt or picrite with high water content (up to 7.5%; Sobolev and Chaussidon, 1996; Moore and Carmichael, 1998; Ulmer, 2001; Pichavant et al., 2002; Cervantes and Wallace, 2003; Grove et al., 2006, 2012; Wallace, 2005; Kelley et al., 2010; Zimmer et al., 2010). The magma is relatively more oxidized than the typical asthenospheric melt (FMQ +1~3, relative to the fayalite-magnetite-quartz buffer; Ballhaus, 1993; Brandon and Draper, 1996; Parkinson and Arculus, 1999; de Hoog et al., 2004). Sun et al. (2015) suggested that the solubility of SO₄^2- is one order of magnitude higher than S^2- in magma, and therefore the sulfur solubility largely depends on magmatic redox state. In addition, previous studies showed that the magma derived from mature subduction zone to the overriding plate is commonly oxidized (Jugo et al., 2005, 2010; Klimm et al., 2012). When the redox state reaches FMQ +1~2, the solubility of metal complexes with sulfide increases rapidly in the silicate melt.

The metal-fertile basaltic melt formed from the mature subduction zone is commonly hot, hydrous, and highly oxidized (Herzberg et al., 1983). The density of such magma is lower than the mantle, and will tend to rise from the asthenospheric mantle and penetrate the mantle lithosphere. However, the buoyancy is too low to drive magma to the surface and would therefore stall at the level of neutral buoyancy (Glazner and Ussler, 1988). The interaction between the hydrous basalt flux and the surrounding rocks will generate andesites with uniform composition, involving a process of crustal melting, assimilation, storage of homogenized magma (MASH; Hildreth and Moorbath, 1988). Hildreth and Moorbath (1998) proposed a contribution of crustal material to the formation of porphyry Cu deposit in central Chile of South America, and argued that the calc-alkalic magma was formed during the MASH process of basaltic magma at the base of the lower crust where the basaltic magma was the product of partial melting of the mantle wedge. The MASH process has been widely accepted (Richards, 2003, 2013; Richards and Holm, 2013; Wilkinson, 2013). The hybrid and evolving magmas generate basaltic andesite and andesite. The densities of such magmas are too low to rise through the lower continental crust and would stall in the middle to upper crust to form large batholiths (Cruden, 1998; Cobbing, 1999; de Saint-Blanquat et al., 2001; Klepeis et al., 2003; Lipman, 2007; Richards and Mumin, 2013). The volatile-rich buoyant magma will rise up along fractures and crystallizes as stocks and dikes, or erupt to the surface to form a volcanic sequence of andesite-dacite-rhyolite composition (Walker, 1989; Jaupart and Allègre, 1991; Carrigan et al., 1992; Eichelberger, 1995; Legros and Kelfoun, 2000; Huppert and Woods, 2002).

Wilkinson (2013) highlights four key stages in the formation of porphyry Cu deposits of the volcanic arc setting: ① cyclic enrichment of magmas with metal and water in the deep crust; ② saturation of the magma with sulfide concentrates metal into a smaller volumes of material and releases; ③ efficient transfer of metals into hydrothermal fluids that are exsolved from the magmas; ④ trigger the precipitation of ore minerals in a local zone. He also argues that sulfide saturation is a most critical step in porphyry Cu formation, and can thus be used in evaluating the magma fertility.

2.1.3 Source of the ore-forming materials

Dehydration of the subducted oceanic lithosphere is a critical factor affecting magma fertility in volcanic arcs. The dehydration will supply a large amount of volatile (H₂O, S, and Cl) and large-ion lithophile elements (LILEs) into the mantle wedge (Tatsumi, 1986; Davidson, 1996; de Hoog et al., 2004). The H₂O addition is thought to elevate oxygen fugacity of the magmas which are derived from the mantle wedge melting (Richards, 2003). In moderately oxidized magma, sulfur is mainly dissolved as sulfate (~1.5% S; Jugo et al., 2001) which make Cu and Au to be incompatible (Hamlyn et al., 1985; Bornhorst and Rose, 1986; Richards, 1995, 2013). The above explains why the arc magmatism is conducive for the enrichment of chalcophile and siderophile elements (e.g., Cu and Au). Campos et al. (2002) conducted microthermometry and electron microprobe analysis of melt inclusion trapped in the quartz phenocrysts of the Llamo porphyry intrusion at the Zaldivar porphyry Cu deposit. The result showed that the Cu contents in the same inclusions range between 0.03% and 0.57% with 0.1% in average, which are at least one order of magnitude higher than what should be expected for calc-alkaline intrusions (Einaudi, 1977). Although the sulfur content in the melt is below the detection limit, Campos et al. (2002) argued that the copper is sourced from the magma.

The rapid development of microanalytical techniques, such as PIXE, SXRF, and LA-ICP-MS, allows the quantification of major, trace, and rare elements and isotopes at higher resolution (μm-scale). Wallace and Edmonds (2011) compared the sulfur concentration of magmatic sulfide and silicate melt inclusion, and suggested that the sulfur in the intermediate-felsic magma is derived from the underlying mafic magma. Audétat and Simon (2012) further conducted experimental studies of the melt partition coefficients between sulfide and silicate melt and showed that copper strongly partitions into sulfide liquid relative to silicate melt, whereas molybdenite partitions into the silicate melt. Another application is testing the partition behavior of elements in different phases. Zajacz et al. (2008) reported the fluid/melt partition coefficients of some economically important metals, in which the partition coefficient of copper is highly variable but can be up to 2700. Pettke et al. (2010) measured lead isotope ratios in single fluid inclusions from the Bingham Canyon Cu-Mo-Au deposit using LA-MC-ICP-MS and argued that the anciently metasomatized subcontinental lithospheric mantle was involved in the mineralization.

2.2 High ore grade and special mineralization style in the Bolivia Sn-Ag ore belt provide new insights for exploring Sn deposits

The Sn ore belt that developed within the volcanoplutonic sequences is well known for its high ore grade. The Bolivia Sn ore belt in the central eastern Andean extends up to 900km from North Argentina, and the whole Bolivia, to southeastern Peru. A series of world-class vein-type and porphyry Sn deposits also have been recognized (Fig. 2; Lehmann et al., 2000). All of the vein-type Sn deposits are of high grade and tonnage, including the Chojila, Oruto, Llallagua, Potosi, Cerro Rico, Tasna, Choroque in Bolivia and San Rafael in Peru (Dietrich et al., 2000). For example, the total Sn reserve in the Llallagua deposit is up to 1Mt (Ahlfeld and Schneider-Scherbina, 1964; Redwood and Rice, 1997); the San Rafael deposit contains Sn reserve of 0.73Mt with a grade of 5.7% and 22, 000t Cu at a grade of 0.16% (Mlynarczyk et al., 2003); and the Cerro Rico de Potosi was most important Ag deposit in the early 20^th century with ~50, 000t Ag produce (Zartman and Cunningham, 1995; Lehmann, 2004).

The exposed strata in the Bolivia Sn ore belt contains Paleozoic mudstone and clastic rocks and was metamorphosed to sub-greenschist facies during Early Devonian-Late Carboniferous, and was then unconformably overlain by Carboniferous-Early Permian clastic and carbonate rocks. The younger Middle Permian-Triassic red beds are intercalated with alkalic volcanic rocks, which unconformably underlie a ~1km Cretaceous clastic and carbonate rocks and the Miocene-Pliocene ignimbrite and red beds. The multistage magmatism emplaced during the Late Devonian-Late Tertiary, including Late Devonian-Early Carboniferous granite associated with minor Sn-W, the Permian-Triassic granite batholith hosting W-Cu-Sn-Mo, the Late Cretaceous granodiorite stock associated with minor Ag and base metal, and Middle-Late Tertiary granite associated with Sn-W-Ag ore systems (Kontak and Clark, 1988). The Permian Sn mineralization is unfortunately not economic (Clark et al., 1983). The Cenozoic Sn mineralization can be divided into two parts geographically: Oligocene-Miocene Sn-W in the north and Late Miocene Sn-W-Bi-Ag-Pb-Zn in the south. The north part includes vein-type (e.g., San Rafael, Illimani, Quimsa Cruz, and Santa Vera Cruz) and subvolcanic-associated (Llallagua) Sn deposits, whereas the south part includes Cerro Rico, Quechisla, Chorolque, Tasna, and Chocaya. The mineralization ages of these two parts are 24~21Ma and 16~12Ma, respectively (Lehmann, 2004).

2.2.1 The alteration and mineralization styles are typical to establish ore genetic model

Two types of Sn mineralization were identified: porphyry-type and granite-related vein-type, both of which are associated with the late Tertiary stocks of intermediate composition. Sillitoe et al. (1975) examined several of these stocks, including those at the major Llallagua, Potosí, Oruro, and Chorolque mineralized centers, and proposed a porphyry Sn deposit model (Fig. 3). These deposits show characteristics of stockwork, disseminated, and breccia cassiterite mineralization similar with those of porphyry Cu deposit. The host rocks include the Paleozoic sedimentary, intermediate intrusion and the coeval rhyolite, dacite, and andesite. The orebody is manifested as an inverted cone with Sn grade of 0.2%~0.3% and is cut by Sn-Ag vein swarm to form bonanza ores with Sn grade of 1%~5%. The K-feldspar alteration is weak, the quartz-sericite alteration is pervasive, and the tourmaline prevails in depth; whereas the silicification, kaolinization, and illitization develop close to the surface. In contrast, vein-type Sn mineralization is commonly associated with volcanic edifice and subvolcanic dikes of the dacitic, rhyolitic-dacitic, and quartz latitic composition. The orebody is cylindrical with a diameter of ~1km. Similar to porphyry Cu deposits, many of the hydrothermal breccia pipes were discovered in the ore district. The lode Sn deposit is argued to be triggered by one of the late events in a volcanic center. The vein-type Sn, Sn-Ag, and Sn-polymetallic mineralization overprinted either on the porphyry deposit or the Paleozoic-Mesozoic sedimentary rocks and formed swarms of orebodies with thickness of tens of meters and length of several kilometers.

Exploration activity introduced by Sillitoe (1998) shows that the porphyry Sn deposit also has lithocap like porphyry Cu deposit. The lithocap of porphyry Cu deposit is composed of advanced argillic alteration, whereas that of porphyry Sn deposit (such as the Pulacayou deposit) commonly occurs on top of the alteration center and comprises vuggy residue quartz and quartz-alunite. Based on a systematic study of the lithocaps in the south Bolivia Sn ore belt, Sillitoe (1998) proposed a comprehensive alteration model involving the advanced argillic alteration on the top as a lithocap, to the quartz-sericitization zone and tourmalinization towards the depth. The chemical zoning ranges from Ag, Sb±Sn, Sn and base metal, to deep W±Bi±Sn (Fig. 3). The advance argillic lithocap represents a shallow epithermal part of the Sn-Ag system and indicates a high potential of orebody in depth. Therefore, the lithocap and alteration zoning can be useful exploration indicators in volcanic districts.

2.2.2 Ore genetic models

The porphyry and vein-type mineralization are genetically associated with granitic rocks, although intrusions have not been found in some areas, whilst the rhyolitic porphyry dikes can be extensively observed. These dikes might play a role as channels in the fluid uprising.

Geochemical studies in the Bolivia Sn ore belt reveal that the granite and cogenetic volcanic rocks are peraluminous which can be ascribed to be reduced S-type or ilmenite-series granite (Pichavant et al., 1988; Ericksen et al., 1990; Lehmann et al., 1990; Morgan et al., 1998). Lehmann (1990) argued that the Sn-bearing granite was produced from remelting of Paleoproterozoic rocks which experienced fractional crystallization. Dietrich et al. (2000) showed that the Bolivia Sn-bearing granites are rhyolitic dacite and dacite in composition but only experienced moderate fractional crystallization. The melt inclusions in quartz phenocrysts are highly fractionated rhyolite, and distinct from the Sn-bearing granite. The difference is difficult to be explained by fractional crystallization in a single system but can be attributed to magma recharge into the chamber. The magma mixing process resulted in the formation of the granite porphyry with moderate fractionation, whereas the magmatic-hydrothermal fluid exsolved from the highly fractionated Sn-bearing granite went through the volcanic pipe to the shallow crustal level and generated Sn deposit. The argument attracts our attention on the metal fertility of the moderately fractionated granite since the highly fractionated granite has not been exposed.

Wagner et al. (2009) conducted a comprehensive stable isotope and fluid inclusion investigation of the San Rafael giant Sn deposit and identified that the fluid responsible for the mineralization was hot, acidic, hypersaline, and reduced. Based on those data, it is thought that the ore-forming fluid was exsolved from the magma and then interacted with the wall rock to form extensive quartz-sericitization and tourmalinization. The oxidation, dilution, cooling, and neutralization of the acidic ore fluids during their mixing with hot groundwater of meteoric origin triggered large-scale precipitation of cassiterite. Tourmalinization is typical in the Bolivia Sn deposits. Lehmann et al. (2000) measured the chemical composition of the melt inclusions in the quartz phenocrysts, and noticed that boron abundance is positively correlated with cesium, rubidium, and arsenic. He thus argued that the B, As, Cs, Rb are incompatible elements, their enrichments can be attributed to magmatic fractionation. As the mudstone of the Early Paleozoic strata is rich in B, he further suggested that the metal-fertile porphyry was derived from remelting of the Paleozoic rocks.

2.2.3 Geodynamic setting of the Sn deposit

The formation of the Sn deposit is commonly restricted to specific tectonic settings. Lehmann et al. (1990) concluded three types of tectonic setting for ore formation, these are post-collision (e.g., Carboniferous-Permian Sn ore belt in the West Europe, the Permian-Triassic Sn ore belt in southeast Asia), intracontinental arc of the continental margin (e.g., the Tertiary Sn ore belt in Bolivia and the Cretaceous-Triassic Sn ore belt in Thailand), and intracontinental extension and rifting (e.g., Permian-Triassic Sn prospects in Bolivia, Cretaceous Sn-W deposits in South China, Cretaceous Sn deposit in Nigeria, and Precambrian Rondonia Sn deposit in Brazil). Sillitoe (1976) proposed a plate subduction model for porphyry deposits that can explaine why these deposits are arranged in parallel and linear belts. Based on the published deep-crust structural data (James and Sacks, 1999), Lehmann (2004) proposed that the Cenozoic porphyry Cu deposits in Chile and the Sn ore belt in Bolivia formed in the continental margin and back arc, respectively, and argued that their formation is affected by the plate subduction angle. The transition between normal angle and flat angle subduction favored the porphyry Cu formation, whereas the transition between flat subduction to normal angle resulted in asthenospheric upwelling and crustal anatexis to form S-type granite in the back-arc. The Sn-bearing granite was generated by fractional crystallization.

3 Current research progress of the Sn and Cu deposits in East China and the adjacent regions

The East China and the adjacent areas yield extensive mineralization at the Mesozoic rather than the Cenozoic, with ages ranging from 170Ma to 80Ma (Mao et al., 2011; Hu and Zhou, 2012; Zhou et al., 2015; Pirajno and Zhou, 2015). Particularly, the distribution of deposits extends ~1000km from eastern coast to inland, which contains W, Sn, Mo, Bi, Cu, Pb, Zn, Ag, Au, Sb, U and Fe deposits (Zhou et al., 2002; Mao et al., 2003, 2011, 2018; Xie et al., 2012; Zeng et al., 2013; Ouyang et al., 2013; Wang et al., 2017; Goldfarb et al., 2014; Chen et al., 2017; Hu et al., 2008, 2017a, b). About ~1200km east of the continental margin, a cluster of porphyry Cu deposits were newly discovered and recognized in the Zhashan district, South Qinling, which was dated to be 148~145Ma (Xie et al., 2015, 2017). Since these metallogenic events in East China do not show the regional paralleling belts compared to the continental margin that occurs in continental margins of South America, it has traditionally been regarded as a typical intraplate setting.

3.1 Major porphyry/skarn copper provinces in East China and adjacent regions

Until now, no world-class porphyry Cu deposit has been discovered in the continental margin of the west Pacific Circum belt. The Dexing porphyry Cu deposit and the Luoboling-Zijinshan porphyry-epithermal system contains reserves of >10Mt Cu with 50t Au and >5Mt Cu with >400t Au, respectively.

Based on a synthesis of the published precise ages of the mineralization and related granitoids, two episodes of copper mineralization can be recognized: Middle-Late Jurassic-Early Cretaceous and Late Cretaceous. In the first period the porphyry Cu deposits are lineated in three parallel metallogenic belts: the Qinhang, Middle-Lower Yangtze River Valley, and Nengjiang-Northeast Taihang (Fig. 1; Fig. 4). The mineralization ages are concentrated at 170~137Ma (Mao et al., 2014; Zhou et al., 2015). Recent industrial exploration and scientific research have revealed the presence of the Middle-Late Jurassic (170~154Ma) porphyry Cu-Au-Mo deposits at the continental margin of South China (Mao et al., 2017). In addition to the Dexing, Yinshan, Yongping, Longtougang in Jiangxi Province, Tongcun and Linghou in Zhejiang Province, Gutian in Fujian Province, and the Xinliaodong, Zhongqiuyang, Jilongshan and Honggoushan in eastern Guangdong Province, and the Qiguling, Potoumian and Didougang in western Guangdong Province, and the Yushui fracture-controlled high grade copper deposit recognized during our field trip in March, 2018. In addition, both the Dingjiashan and Fengyan stratiform Pb-Zn deposits where Proterozoic strata was exposed in the center and surrounded by the Jurassic volcanic rocks in the Wuyishan region are identified to be Late Mesozoic porphyry Mo-skarn Cu-Pb-Zn-vein Pb-Zn system, rather than exhalative ores previously proposed. Xiao Xiaoniu (unpublished data) obtained rock-forming age of 158~155Ma for the granitic porphyry associated with porphyry Mo mineralization beneath the Fengyan Pb-Zn deposit (personal communication). On the whole, those Jurassic deposits are situated in a series of NE-trending Middle-Late Jurassic volcanic basins which were covered by the Late Cretaceous andesitic volcanic rocks. Due to the volcanic cover, these deposits are only exposed in the southern and western parts of the southeast coast belt, and some uplift areas in the Cretaceous volcanic rocks.

Second stage of copper mineralization appeared as porphyry Cu-epithermal Cu-Au and/or Au deposits which developed in Cretaceous extensional volcanic basins (Fig. 5). For instance, the Xiaoxinancha porphyry Cu-Au ore district in the Yanji basin, eastern Jinlin Province (Sun et al., 2013; Han et al., 2013); the Wangjiazhuang and Qibaoshan porphyry copper deposits occur in the Zouping and Jiaolai basins, respectively in eastern Shandong Province (Xu et al., 2000; Wang et al., 2015); the Luoboling porphyry Cu-Mo, Zijinshan high sulfidation epithermal Cu-Au and Yueyang low sulfidation epithermal Au-Ag deposits in the Shanghang basin, western Fujian Province, and coupling porphyry Cu and granite-related Sn deposits in the Yangchun basin, western Guangdong Province. In addition, along the continental margin low sulfidation epithermal Au±Ag deposits are extensively distributed at the margins of Cretaceous volcanic and fault basins, such as the Zhilingtou Au deposit and Baishan Au deposits in Zhejiang and Jinlin provinces, respectively. The mineralization ages range from 100Ma to 80Ma (Mao et al., 2011, 2013; Jiang et al., 2013; Sun et al., 2013; Zhong et al., 2014).

3.1.1 Magma source remains controversial

Many studies have been conducted on the magmatic rocks associated with the porphyry/skarn copper deposits in East China. The characteristics of these rocks can be summarized as follows: ① granodiorite porphyry and subordinate monzonitic granite porphyry and quartz diorite porphyry in which the monzonitic granite porphyry is closely associated with the Cu-Mo mineralization; ② the intrusive rocks are calc-alkalic with wide ranges of major elements, e.g., SO₂ content of 55%~70%. They are also enriched in LILEs and LREEs but depleted in HREEs and high-field-strength elements (HFSE), with the Nb, Ta, and Ti negative anomaly which is typical of arc magmatism (Liu et al., 2012; Zhou et al., 2015); ③ wide ranges of Sr-Nd-Hf isotopes; their initial ⁸⁷Sr/⁸⁶Sr, ε_Nd(t), and zircon ε_Hf(t) and δ¹⁸O values are 0.705~0.709, -12~-3, and -11~0, and 6‰~9‰, respectively (Li et al., 2013) for the ore-related intrusive rocks in the Middle-Lower Yangtze River Valley. The isotopes of the Dexing ore-related intrusive rocks are characterized by mantle values, such as initial ⁸⁷Sr/⁸⁶Sr ratios of 0.705~0.706, ε_Nd(t) of -0.9~0.6, ε_Hf(t) of 2.0~7.0, and δ¹⁸O of 5‰~6‰ (Liu et al., 2012; Hou et al., 2013). These ore-bearing intrusive rocks are characterized by high Sr and low Y abundances, which are similar to the adakitic rocks (Zhang et al., 2001). However, these rocks have high K, low Mg, and Sr-Nd isotopic enrichment, inconsistent with those from the adakitic magma derived from partial melting of subducted oceanic crust. The genesis of the intrusive rocks associated with these porphyry Cu deposits in East China remains controversial. The main viewpoints include: ① partial melting of subducted oceanic plate (Ling et al., 2009); ② partial melting of thickened lower crust (Zhang et al., 2001); ③ mixture of enriched mantle and lower crust or assimilation between the partial melting products of enriched mantle and lower crust (Zhou et al., 2015); ④ remelting of Neoproterozoic mafic arc cumulate (Liu et al., 2012; Hou et al., 2013).

3.1.2 The tectonic setting remains equivocal

Zhen (1999), Rui et al. (1984), and Huang et al. (2001) argued that porphyry Cu deposits not only form during orogeny but also within continental rifting. Rui and Zhang (1986) proposed that the porphyry copper deposits in East China formed in the post-collision. In past ten years, the new concept of adakite and related porphyry Cu deposits has strongly impacted on the geochemistry society in China. Zhang et al.(2001, 2004) suggested a new term of 'continental adakite (C-type)' which is thought to be the product of partial melting of the lower crustal granulite induced by thickened continental crust during underplating of basaltic magma. The C-type adakite is also characterized by high Sr and low Y (Yb) abundances. The transition of amphibole to the garnet during dehydration provides enough water to leach metal from the mantle and mafic rocks. The process explains the close relationship of adakite to mineralization. Hou et al. (2015) emphasized that the ore sources are not related to subduction slab but instead are derived from juvenile crust or mantle-crust interaction triggered by asthenospheric upwelling, and the ore-bearing magma of the Mujicun deposits is derived from remelting of lower crust mafic rocks (Hou et al., 2015). Zhou et al. (2015) worked on the porphyry-skarn Cu polymetallic ore belt in the Middle-Lower Yangtze River Valley, and correlated it with typical intracontinental porphyry Cu ore belt. The ore-forming magma was derived from a mixture of mafic magma (from enriched mantle) and felsic magma (from partial melting of thickened lower crust). The H₂O, S, Cu, and Au are mainly from the enriched mantle.

Qi (1990) and Qi et al. (2000) suggested that the Paleo-pacific plate subduction and strike-slip fault zones are genetically linked to the Yanshanian metallogeny in East China. In past twenty years, many researchers have focused on the tectonic, structural, petrology, and mineral deposits, which have identified two significant tectonic transitions in East China (Wan and Zhu, 2002; Wan, 2004; Wan et al., 2008; Mao et al., 2003, 2005, 2008, 2011, 2013, 2014; Zhao et al., 2004; Dong et al., 2007; Zhang et al., 2009; Zhou et al., 2006; Li and Li, 2007). The NW-trending oblique subduction of Paleo-Pacific oceanic plate under Eurasian or eastern Asian continental started at ~170Ma, and the transpressional regime changed to extensional regime at ca. 135Ma. We proposed that the ore-forming geodynamic settings witnessed a transition from oblique transpression convergent margin during 170~135Ma to strike-slip margin at 135~80Ma, which are related to the markedly low-angle subduction and post-subduction of Izanagi or Paleo-pacific plate, respectively (Mao et al., 2003, 2005, 2011, 2013). The ore-related porphyries in the compressional environment are mainly derived from teared subducted slab in the first period, whereas those in the extensional basins are sourced from stagnant subducted slab or metasomatized subcontinental lithosphere mantle later. Ling et al. (2009) and Sun et al.(2010, 2015) proposed that the porphyry Cu deposits in the Middle-Lower Yangtze River Valley are closely related to mid-ocean ridge subduction. The mid-ocean ridge subduction supplied energy and material. In particular, the newly recognized Middle-Late Jurassic porphyry Cu ore belt and calc-alkaline magmatic arc in the southeastern coastal area provides an excellent evidence for better understanding of the subduction of Pacific oceanic plate beneath the Eurasia continent (Mao et al., 2017).

3.2 Major tin provinces in East China and adjacent regions

The Sn deposits in East China are mainly distributed in South China and southern part of the Great Khingan Range (Fig. 4). South China is the most important Sn province in China and also over the world. It was thought that W deposits are mainly distributed in the eastern part of South China, whereas Sn deposits are in the western part (Chen et al., 1989; Hua et al., 2005). However, the precise isotopic ages suggest an obvious difference in the mineralization ages of the W-Sn deposits in the eastern section of South China (Nanling Range) and Sn deposits in the west. The Nanling Range is dominated by W deposits with subordinate Sn deposits, such as Furong and Xitian. The dense W-Sn deposits in the Nanling Range are mainly distributed in Hunan, Jiangxi, Guangdong and Guangxi Provinces, along a group of paralleling NE-trending faults, particularly at the intersections of regional NE- and EW-trending faults. Large-scale W-Sn mineralization in the Nanling Range mainly formed during 160~150Ma (Mao et al., 2004, 2007, 2008; Hua et al., 2005; Peng et al., 2006; Hu et al., 2012a). The dominant mineralization styles are composed of skarn and quartz-vein types. Those W-Sn deposits are distributed in a scope of NE-trending oval at surface (Mao et al., 2013) rather than EW-trending one as previous described. All the Sn deposits in western South China developed along the margins of the Youjiang basin, including the world-class Gejiu and Dachang Sn deposits in China, as well as the Tinh Nuc and Nui Pao Sn deposits in Vietnam, and are dated to be 98~76Ma (Cheng and Mao, 2010). In fact, these tin deposits form part of a 10, 000km long Sn ore belt along the continental margins from eastern Vietnam, southeast coast of China (including Guangdong, Fujian, Zhejiang, and Shandong provinces), Korean Peninsula to Sikhote Alin, Far East of Russia (Fig. 5). The Sn mineralization in the regional belt was dated at ca. 100~76Ma (Cheng et al., 2016). The predominant mineralization styles are skarn-type and stratiform sulfide type with subordinate porphyry and quartz-vein type. In the belt there are also some tungsten deposits, such as large-scale Sangdong skarn tungsten deposit in South Korea and the Damingshan quartz vein type in southwestern Guangxi Province, China. They are dated to be about 87Ma and 95Ma (Seo et al., 2017; Li et al., 2008), respectively.

The southern part of the Great Khingan Range is another important Sn metallogenic province, where a series of skarn, cassiterite-sulfide, porphyry, and greisen-type Sn deposits have been explored (Wang et al., 2006), and the large-scale Huanggang skarn Sn-Fe deposit has been well studied (Zhou et al., 2012). Recently, two large Sn deposits have been discovered, namely the Weilasituo porphyry Sn deposit and the Baiyinchagan cassiterite-sulfide Sn-Ag deposit. The Weilasituo porphyry Sn deposit is characterized by disseminated-veinlet Nb-Ta-Sn, greisen-type, and quartz-vein type Sn mineralization around the cupola of the deep-seated granite pluton. The tin deposit is spatial-temporally and genetically associated with the Weilasituo-Bairendabai vein-type Ag-Pb-Zn deposit (Zhu et al., 2016). Considering identical magmatic sources of the Cretaceous volcanic rocks and mineralization-associated granitic rocks in the south Great Khingan Range, the Sn deposits in this area can be correlated to Sn-Ag ore system related to the subvolcanic rocks, which are similar to those in the Bolivia Sn ore belt (Mao et al., 2018). The Sn deposits in the south Great Khingan Range formed during ca. 149~133Ma (Ouyang et al., 2013; Liu et al., 2016), which are similar to the emplacement of the ore-related granitic rocks at 150~135Ma (Wang et al., 2006, 2017; Zhou et al., 2012; Zhai et al., 2014; Liu et al., 2016). They are mainly distributed along three NE-trending faults, particularly on both sides of the Huanggang-Ganzhuermiao fault (Wang et al., 2017). The NE-trending southeast Great Khingan-northeast Taihang porphyry Cu ore belt, east to the Sn province, has been dated to be 150~140Ma (Fig. 4).

3.2.1 Extensive investigation focused on Sn-related granites, significant progress has been achieved

Tin mineralization is commonly associated with highly fractionated granite (Lehmann, 1990). Due to high degree of fractional crystallization, the Sn-related granites are characterized by the following petrological, mineralogical, and geochemical features: (1) high contents of SiO₂ and alkali elements with K₂O>Na₂O; enrichment of B, F, Cl, Li, Rb, Nb, Ta, Ga, Cs, U, and REE; depletion of Fe₂O₃, MgO, TiO₂, CaO, Ni, Cr, V, Sr, Ba, and Eu; high Rb/Sr, low La/Yb, and significant negative Eu anomaly; (2) volatile-rich minerals such as topaz, fluorite, and/or tourmaline as well as lepidolite, are common; (3) S-type granite (ilmenite-type or crustal remelting type, Ishihara, 1977; Xu et al., 1982; Chappell and White, 1974), which are thought to be derived from remelting of the crust and are characterized by low oxygen fugacity. Based on lacking representative aluminum minerals, such as cordierite and andalusite, as well as negative relationship between the contents of SiO₂ and P₂O₅, Rb and P₂O₅, Li et al. (2007) suggested that the W-Sn-related granites in South China belong to highly evolved I-type, which has been accepted to some extent by many researchers (Cheng and Mao, 2010; Zhou et al., 2012; Guo et al., 2015; Qiu et al., 2017). Recent studies have classified Sn-bearing granites as A-type due to their high alkali contents and Ga/Al ratio (Whalen et al., 1987). Therefore, determining the genesis and reconstructing the formation process of the Sn-bearing granite remains to be resolved.

3.2.2 Development of isotopic dating and quantitative analysis of fluid inclusions have improved our recognition of metallogenic process and regularity

The formation age of most Sn deposits and associated granite can be determined by molybdenite Re-Os, mica Ar-Ar, and zircon U-Pb dating methods. The recently developed and widely applied cassiterite U-Pb dating method allows us to gain mineralization age directly of some Sn deposits (Yuan et al., 2008, 2011). The origin, migration, evolution, and deposition of hydrothermal fluid can be evaluated. In situ quantitative analysis of fluid inclusions helps to gain the accurate composition of the ore-forming fluid and the physical-chemical conditions (Audétat et al., 1998; Wei et al., 2012; Ni et al., 2015). The noble gas isotopic data suggest that the involvement of mantle-derived fluids to some extent in the formation of the W-Sn deposits (Li et al., 2007; Hu et al., 2012b; Zhai et al., 2012). The rapid development of precise analytical methods for dating mineralization, and evaluating physical-chemical condition improves better to understand ore-forming mechanism, regional metallogenic regularity, and geodynamic setting, would help to propose new metallogenic models.

4 Problems for future investigations

Although significant research progresses have been achieved on both tin and copper deposits in the Circum Pacific metallogenic belts, the following key problems remain unresolved.

(1) In East China there are three parallel and coupling copper ore belts and tin or tungsten ore belts, i.e. South Khingan Sn-Ag belt in the west and Nengjiang-northeastern Taihang porphyry copper ore belt in the east, Middle-Lower Yangtze River Valley porphyry-skarn Cu-Au-Mo-Fe ore belt in the north and Jiangnan porphyry-skarn W(Sn) belt in the south, Nanling W-Sn ore belt at west and southeast coast Late Jurassic porphyry-skarn copper ore belt at east. These three coupled ore belts are contemporaneous, but their genesis and related geodynamic settings remain poorly understood.

(2) The Sn-Ag ore system in Bolivia has been well studied and the related ore deposit model proposed has greatly aided in tin exploration. However, the relationship of tin and Ag-Pb-Zn ore deposits in the Southern Khingan area requires further elucidation. The ore genesis and metallogenic regularities as well as available ore deposit model will be helpful for further investigation and prospecting in the area. It is important to identify new Sn deposits to meet the shortage of Sn resource in China.

(3) The >10, 000km long Late Cretaceous ore belt along Eurasian continent margin with Cu-Au-Sn deposits in the basins is another key scientific question. It is essential to answer whether they are post-subduction products, and the reason why these deposits with distinct ore characteristics occur in the same belt or even same basin?

These three scientific problems, if resolved, will provide us with important insights into the regional metallogeny of the Circum Pacific Rim.

Acknowledgments This paper is based on the proposal of the project we applied from the National Science Foundation of China (41820104010). We appreciate Prof. Bernd Lehmann, Prof. Jeremy Richards, Prof. Richard J. Goldfarb and Prof. Larry Meinert for useful discussions and constructive suggestions. We are grateful to both Prof. Jeremy Richards and Prof. Larry Meinert for improving an early draft of the manuscript. Prof. M. Santosh is debt to thank for his kind final edit of language.Editor in chief, and professor Ruizhong Hu and Shaoyong Jiang are thanked for constructive reviews that improved the quality of this paper.