Zircon Chemistry and Oxidation State of Magmas for the Duobaoshan-Tongshan Ore-Bearing Intrusions in the Northeastern Central Asian Orogenic Belt, NE China

The Duobaoshan (DBS)-Tongshan (TS) porphyry Cu–(Mo) deposit (4.4 Mt Cu, 0.15 Mt Mo) is located in the northeastern part of the central Asian orogenic belt (CAOB) in northeastern China. It is hosted by early Ordovician dioritic to granodioritic intrusions which are characterized by the subduction-related geochemical signatures including high concentrations of large ion lithophile elements (LILEs) and light rare earth elements (LREEs), and low concentrations of heavy REEs (HREEs) and high-field -strength elements (HFSEs), such as Nb, Ta, Zr and Ti in bulk rock compositions. Furthermore, they show adakitic geochemical signatures of high Sr/Y ratios (29~55) due to high Sr (290~750 ppm) and low Y (<18 ppm). Zircon trace element abundances and published Sr-Nd-Hf isotope data of these rocks suggest that the parental magmas for these ore-bearing intrusions were rich in H2O and formed by partial melting of a juvenile lower crust/lithospheric mantle or metasomatized mantle wedge during the northwestward subduction of the Paleo-Asian Ocean before the collision of the Songnen block with the Erguna-Xing’an amalgamated block in the early Carboniferous. Values of Ce4+/Ce3+ and Ce/Nd in zircons are 307~461 and 14.1~20.3 for mineralized granodiorites, and 231~350 and 12.4~18.2 for variably altered diorite and granodiorites in DBS, whereas those for DBS-TS microgabbros are 174~357 and 7.4~22, and 45.9~62.6 and 5.0~5.8 for the early Mosozoic Qz-monzonites, respectively. Zircon Eu/Eu* values are high and similar among mineralized granodiorites (~0.6), altered diorite and granodiorites (~0.6) and the Mesozoic Qz-monzonites (~0.8), whereas the values are low and variable for the DBS-TS microgabbros (0.3~0.6). The magma oxidation state calculated from zircon chemistry and whole rock compositions are FMQ +1.0 to +1.5 in mineralized samples, and FMQ +2.4 to +4.2 in altered samples. The values are comparable to those for the fertile intrusions hosting porphyry Cu-Mo-(Au) deposits in the central and western CAOB and elsewhere in the world. Elevated oxidation state is also observed in the TS microgabbros, FMQ +1.4 to +1.9, and the early Mesozoic Qz-monzonites, FMQ +2.4 to +2.5. Comparison of zircon geochemistry data from porphyry deposists elsewhere suggests that positive Ce anomalies are generally associated with fertile intrusions, but not all igneous rocks with high Ce anomalies are Cu fertile. The findings in this study are useful in exploration work and evaluating oxidation state of magmas for porphyry Cu-(Mo) deposits in the region and elsewhere.

The oxidation state of parental magmas is important for the formation of porphyry Cu-Mo-(Au) deposits because oxidized magmas are capable of supplying metals and S from magma sources to ore deposits (e.g., [29,30]). Cerium can be Ce 4+ in oxidized magmas among rare earth elements (REE). Zircon, ZrSiO4, crystallizing in oxidized magmas easily incorporates Ce 4+ to produce positive anomalies of Ce in the abundances of REEs (e.g., [16,[31][32][33]). Trace element geochemistry of zircon has been carried out for many porphyry Cu-Mo deposits in the world in recent years (e.g., [34][35][36][37][38][39][40][41]), but studies on early Paleozoic deposits are not many, and this is the first systematic study on oxidation state of the early Paleozoic DBS-TS porphyry Cu-Mo deposit in eastern CAOB.
Cerium anomaly is generally expressed as Cen/(Lan × Prn) 1/2 . As the concentrations of La in zircon are low, commonly below or around the detection limit of analytical instruments, several different expressions of Ce anomalies are proposed: Ce 4+ /Ce 3+ [31], Ce/Nd [42], and Ce/Ce* = Cen/(Ndn 2 /Smn) [43]. In addition, calculation of Ce/Ce* using all REE was proposed by Zhong et al. [38] and Lee et al. [37]. These parameters are broadly positively correlated [44]). Values of Ce/Ce* and Ce/Nd can be directly obtained from zircon compositions. Recently, Smythe and Brenan [45,46] and Loucks et al. [47] proposed empirical equations to calculate fO2 based on zircon compositions.
Previous studies show that zircons from fertile intrusions associated with porphyry Cu-Mo mineralization contain high concentrations of Ce with variably high Ce 4+ /Ce 3+ values, reflecting oxidized conditions of the parental magmas, and that zircons from barren (not associated with metal mineralization) intrusions contain low Ce 4+ /Ce 3+ values (e.g., [16,31,32,48]). Some workers consider the magnitude of the Ce and/or Eu anomalies to be related to the size (Cu or Mo tonnage) or grade of deposits [16,18,31,44,49]. This paper examines whether the fertility is related to Ce and Eu anomalies of zircon and whether Ce anomalies are associated with highly oxidized magmas based on the data from the DBS-TS deposit.
Regarding the geodynamic setting for the Ordovician magmatism including the hosting intrusions of the DBS-TS Cu-Mo deposit in the study area, three models have been proposed: (1) collision or post-collision of the Erguna and Xing'an block after northwestward subduction of the Xing'an block (Figure 1b; [12,13,28]), (2) post-collision of the Erguna-Xing'an block and Songnen block [19,60], and (3) subduction related island arc or continental arc [21,23].
The area of DBS in the Xing'an block of the eastern CAOB hosts several Cu-Mo-(Au) deposits including two Paleozoic deposits (DBS and TS) and three Mesozoic deposits (Figures 1a,b and 2). Mesozoic deposits are two skarn type Fe-Cu deposits at Sankuanggou and Xiaoduobaoshan and an epithermal Au deposit at Zhengguang (Figure 2 [8]). These Mesozoic deposits are considered to have been formed during the subduction of the Mongolia-Okhotsk Oceanic plate [5][6][7][8] or the Paleo-Pacific Oceanic plate [10,11].
The DBS-TS deposit, with proved reserves of 4.4 Mt Cu and 0.15 Mt Mo, is the third largest porphyry Cu-(Mo) deposit in the eastern part of CAOB after the Oyu Tolgoi and Erdenet (>11 Mt Cu; [61]) deposits. The deposit is hosted by the early Ordovician porphyritic diorite or granodiorite ( Figure 2). U-Pb zircon age for the DBS hosting granodioritic porphyries is 477~482 Ma [23] and 475.9 ± 0.8 Ma [11] for the TS hosting granodiorite. The Re-Os isochron age of molybdenite grains for the DBS deposit is 475.9 ± 7.9 Ma by Zeng et al. [23], and 473 ± 4 Ma by Hao et al. [11] for the TS deposit, confirming the association of granodiorite magmatism with the mineralization. Simplified geological map of the DBS ore field (modified after [23,25]).

Geology of the DBS-TS Area
The DBS-TS ore-bearing magmas intruded the Duobaoshan Formation and Tonshan Formation, which is composed mainly of tuffaceous siltstone, tuff, andesite and basalt ( [11] Figure 2). The host rocks for the DBS-TS deposit are composed of granodiorite, granodioritic porphyry and diorite with minor microgabbro (Figure 3a-d).
Microgabbro occurs as dykes (10~20 m in width) cutting the granodioritic intrusion in DBS deposit or as rocks intruded by the granodioritic to dioritic rocks in the TS deposit. It shows porphyritic texture with clinopyroxenes as the main phenocrysts (15~20 vol.%) in the matrix of fine-grained granular clinopyroxene and plagioclase (Figure 3d,g). Grains of clinopyroxene are variably altered to chlorite, epidote and magnetite in the TS samples. Primary magnetite grains are present in some microgabbros.
Alteration in the DBS-TS porphyry Cu-Mo deposit is characterized by potassic alteration forming K-feldspar after plagioclase in the inner core of the granodioritic intrusion, pyritic-phyllic and propylitic alterations in the outer part of the intrusion. Pyrite + quartz + white mica represent the pyritic-phyllic alteration, and chlorite + epidote + calcite + quartz the propylitic alteration. The latter alteration types overprint earlier potassic alteration.

Whole-Rock Chemical Composition
After examining over 40 samples collected from the DBS-TS intrusions, we selected 12 representative samples (eight from the DBS intrusions, two samples of the country rock Tonshan Formation, and two from the Mesozoic monzonitic intrusions) for this study. The analyses were performed in the Yanduzhongshi Geological Analysis Laboratories Ltd., Beijing, China. Major elements for whole-rock samples were determined using a Shimadzu XRF-1800 X-ray fluorescence spectrometer. Powered samples, weighing approximately 1.2 g were fused with lithium tetraborate (Li2B4O7, 6 g) at 1050 °C for 20 min. Duplicate analysis of Chinese national standard sample GSR3 shows that the precision is 1% for elements >5 wt %, and 10% for elements <5 wt %. Loss on ignition was measured as weight loss of the samples after 1 h baking at 1000 °C. Trace element concentrations were determined using a Perkin-Elmer Sciex ELAN 6000 inductively coupled plasma mass spectrometer (ICP-MS) after HNO3 + HF digestion of about 40 mg of sample powder in a Teflon vessel at 150 °C. Analytical accuracy and precision were monitored by the Chinese national standard GSR1 (granite) and GSR3 (basalt). The precision is better than 5% of the quoted values for elements present at >1 ppm, and about 10% for elements less than 1 ppm. Accuracy is estimated to be better than 5% for the reported values.

Zircon Trace-Element Analyses
Twelve representative samples from the DBS-TS porphyry Cu-Mo deposits were selected to determine trace element contents in zircon. Zircon grains were separated using conventional magnetic and density techniques. Approximately 100 to 200 clear zircon grains were hand-picked under a binocular microscope, mounted in an Epoxy resin 1 inch in diameter, and polished to expose the centers of zircon grains for individual samples. Prior to analytical work, all zircon grains were examined under a transmitted light microscope as well as cathodoluminescence images using a scanning electron microscope (CL-SEM) at the University of Ottawa. Most grains are transparent euhedral crystals ranging in size from 50 to 100 mm.
Zircon analyses were performed at the University of Ottawa with a Photon Machines Analyte G2 excimer laser ablation system attached to an Agilent 7700× ICP-MS with a second interface pump that approximately doubles instrument sensitivity. A spot size of 40 mm and 10 Hz laser repetition rate were employed. Element concentrations were obtained using count ratios of the elements to Si in zircon and those in NIST SRM612G [62] assuming stoichiometric concentration of Si in zircon grains. Data were processed using GLITTER [63]. Six to twelve representative zircon grains were selected from each of the 12 samples of this study. Since sector-zoned zircon grains commonly show a highly heterogeneous distributions of elements [64], CL images of grains were carefully examined to make sure no sector zoning was selected for analysis ( Figure S1).
Minute inclusions of REE-bearing minerals (e.g., apatite, titanite, and monazite) and faint cracks are common in zircon grains and the counts of Ca, Sr, P, Fe, and La were monitored during the analysis ( Figure S1). Analytical data with spikes of these elements were discarded. The analytical method is essentially identical to that described by Kobylinski et al. [44].
Cerium anomalies, Ce 4+ /Ce 3+ , were calculated following the method proposed by Ballard et al. [31], where the abundance of Ce 3+ is determined from the REEs and Y concentrations in zircon and whole-rock data on the basis of a lattice-strain model for mineral-melt partition of elements. The concentration of Y is included in the calculation because the geochemical properties of Y are identical to those of Ho and because the Y concentration is always higher than the latter. The accuracies of zircon trace elements are overall better than 10%, except La and Pr, which are commonly up to 20%.
Compared to those DBS granodiorites or diorites, the TS microgabbro shows variable REE contents and low (La/Yb)n ratios (2.27~3.45) with no marked Eu anomalies (0.91~1.12) (Figure 5a). In addition, they show trace elemental patterns with marked positive Pb, Sr, and Hf anomalies and negative Nb, Ta, Zr and Ti anomalies with a Nb/Ta ratio of 14.0 to 19.8, similar to those DBS granodiorites and diorites (Figure 5b), suggesting that the TS microgabbros are likely co-genetic with those DBS intrusion. Note that the values of chondrite and primitive mantle are from [69], the data of high-Mg basalt and andesite are from [26], and the data of granodiorite porphyry and granodiorite are from [13,24,28].

Crystallization Temperatures of Zircon
Zircon crystallization temperatures (TTi °C) for individual zircon grains are calculated using the Ti-in-zircon geothermometer of Ferry and Watson [70]. The calculation requires the activities of TiO2 (αTiO2) and SiO2 in magmas. Activity of SiO2 is assumed to be 1 considering the abundant igneous quartz in all rocks. The activity of TiO2 should be <1 because there is no igneous rutile. Since all our samples contain primary titanite (Figure 3e,f), we use a value of 0.7 as αTiO2, as suggested by Hayden and Watson [71] and Fu et al. [72].
Zircon saturation temperatures are also calculated using bulk rock compositions based on the equation of Watson and Harrison [73]. Except for microgabbro and one strongly mineralized sample (DBS 12), zircon saturation temperatures are comparable to TTi °C (Table 2), suggesting fairly quick crystallization of zircon.

Magma Oxidation State
Several methods have been proposed to calculate magma fO2 using zircon compositions (e.g., [74,75]). For our calculations, we used recently proposed methods by Smythe and Brenan [45,46] and Loucks et al. [47]. The method by Smythe and Brenan [46] uses the bulk-rock composition, zircon crystallization temperature and Ce 3+ /Ce 4+ in zircon, water content in magma, and the ratios of non-bridging oxygen to tetrahedrally coordinated cations (NBO/T) of magmas [76]. In this study, we use the NBO/T ratios based on hydrous magmas and zircon crystallization temperatures based on Ti-in-zircon geothermometry. We chose to use the ratios of NBO/T in hydrous melt because the previous studies show that the hydrous-based ratios yield fO2 comparable to the results based on amphibole geothermobarometry [44]. The water content of magma is unknown. Water content in magma for porphyry Cu-(Au) deposits is in the range of 4~6 wt.% at 2 kbars under FMQ +3 to +3.4 based on the published data (e.g., >4 wt.% by Garwin [77]; and 6 wt.% by Rohrlach and Loucks [78]). Kobylinski et al. [44] obtained the H2O content in the parental magmas for tonalite intrusions hosting the Gibraltar porphyry Cu-Mo deposit to be 4.4 ± 0.4 wt.% at a depth of 1.7 km based on amphibole composition. In this study, we assumed H2O content in the parental magma as 4 (lower limit), 5 (median of the published data) and 6 wt.% (upper limit) to calculate the ƒO2 of the DBS-TS deposit.  We consider this assumed water content reasonable considering the solubility of water of a granitic melt is 4.26 wt.% at 800 °C 1 kbar [79]. Use of 5 wt. % H2O yielded fO2 values (expressed as ΔFMQ, logarithmic value from the fayalite-magnetite-quartz buffer) of FMQ +1.0 to +4.2 in individual samples. Using 6 wt. % or 4 wt. % H2O produces a difference in fO2 less than 0.5 logarithmic units ( Table 2). Propagation of error through the fO2 calculation, including the uncertainty of Ce 3+ /Ce 4+ values, crystallization temperatures, analytical errors in the calculated NBO/T, and the estimated amount of water, suggests an estimated uncertainty of <1.5 log units ( Table 2).
The fO2 calculation based on the method of Loucks et al. [47] uses solely zircon chemistry without whole-rock compositions. The values range from FMQ −0.9 to FMQ +0.8 (averaging at FMQ +0.3) for individual samples in the DBS-TS deposit ( Table 2), about 1 logarithmic unit systematically lower than those obtained based on the method of Smythe and Brenan [46]. The fO2 values are also calculated for other deposits elsewhere using the published zircon compositions from the deposits based on the equation of Loucks et al. [47], and the results are summarized in Table S2. The values obtained from the DBS-TS deposit are similar to those of central and western CAOB. Use of the method by Loucks et al. [47] and the method based on hydrous magmas of Smythe and Brenan [46] yielded comparable values for the Mesozoic Mo deposits in NE China (Table S2).

Tectonic Setting Based on Zircon Trace Elements Geochemistry
Zircon grains in DBS-TS show high U and U/Yb, and low Ti and Hf (Table 1). They all fall into the field of continental arc in Figure 7. Especially, the zircons in DBS-TS plot above and close to the upper boundary of mantle-zircon array in Figure 8 Figure S3a). Continuous subduction of the oceanic lithosphere finally led to the collision of the Songnen block with the Erguna-Xing'an amalgamated block at the early Carboniferous ( Figure S3b).  [82]). The field-labeled "Continental Survey" and lower bound (bold line) in Figure 7a were defined by Grimes et al. [83]. The colored symbols in Figure 7b-d are the same as in Figure 7a. The green, grey and white fields circled by pink, grey and blue dashed lines are fields of Cont. Arc-type, mid-oceanic ridge (MOR)-type and ocean island (OI)-type, respectively, based on the compiled dataset of [82]. The thin dashed lines (pink, grey and blue) in Figure 7a refer to the boundary of 80% density distribution. Figure 7c  Our dioritic and granodioritic samples show "adakite-like" geochemical signatures (Table S1; Figure 6). These signatures may be generated by several possible processes: (1) melting of subducted young oceanic lithosphere [68], (2) melting of delaminated ancient lower crust (e.g., [84]), (3) melting of thickened lower crust (e.g., [85][86][87]), and (4) frac-tional crystallization of water-rich arc magmas [44,[88][89][90][91]. Adakites generated by the melting of subducted oceanic crusts are commonly characterized by high Mg# = Mg/(Mg + total Fe) values and high contents of Cr and Ni because they interact with mantle peridotites during their ascent (e.g., [68,92]). However, our diorites and granodiorites samples display relatively low values of Mg# (0.37~0.53, averaging 0.43) and low contents of Cr (10.1~22.2 ppm, averaging 14.9 ppm) and Ni (5.79~8.28 ppm, averaging 7.17 ppm) (Table S1). Therefore, possibility 1 is discounted. Considering that the DBS-TS intrusions formed in a continental arc, possibilities 2 and 3 are ruled out. This leaves possibility 4, fractional crystallization of amphibole in water-rich magma, as the most likely cause for adakitic geochemical signature. This is consistent with the compositional trend of zircon (Figures 7 and 8) and also supported by the spoon-shaped REE patterns of bulk rock compositions because of high partition coefficients of middle REEs between amphibole and melt (e.g., [93][94][95][96]).

Origin of Magmas for the Ordovician Host Intrusions Based on Whole-Rock and Zircon Chemistries
The rocks in this study are characterized by low initial compositions of 87 Sr/ 86 Sr (= 0.7038~0.7040 for TS rocks, and = 0.7044~0.7045 for DBS rocks), high values of εNd(t) (= +3.5~+5.4 for TS rocks, and = +8.1~+8.7 for DBS rocks) and εHf(t) (= +9.0~+14.3 for TS rocks, and +9.73~+13.43 for DBS rocks) as reported by Hu et al. [28] and Zhao et al. [24]. The isotope compositions suggest that the parental magmas were derived from a depleted mantle or juvenile lower crust/lithospheric mantle. The water-rich nature of the magmas and REE patterns of bulk rocks suggest that the source contained abundant amphibole. The three shaded fields (green, pink and orange) are two-dimensional kernel density distributions for compiled datasets of mid-ocean ridge (MOR-type), plume-influenced settings of Iceland and Hawaii (ocean-island (OI)-type), and continental arc (Cont. arc-type) zircon by Grimes et al. [82]. The contours shown are for 50%, 80%, 90%, and 95% levels, which represent the proportion of points inside the contour (after [82]). 2) All the zircons in the study area plot above the "Mantle-zircon array" and close to the upper bound. 3) Zircons in early Paleozoic DBS-TS deposit plot into the contour of 50% Cont. Arc-type, and the zircons in early Mesozoic Qz-monzonite plot in the overlapping area toward MOR-type. 4) The colored symbols are the same as in Figure 7a. The inset (b) shows the effect of fractionation of select minerals on U/Yb and Nb/Yb ratios of the remaining melt (and later-formed zircons), relative to initial ratios = 1. Vectors are shown to scale for the extent of fractional removal (of that mineral only) indicated by the percentages listed (after [82]). Ttn = titanite, amph = amphibole, ap = apatite, and zrn = zircon.

Controlling factors of zircon Ce anomalies
Cerium anomalies of zircon may be expressed as Ce 4+ /Ce 3+ , Ce/Ce* or Ce/Nd. Both Ce/Ce* [37,38,43] and Ce/Nd [42] values are calculated solely based on zircon composition. These three parameters for zircons in the DBS-TS deposit show positive correlations ( Figure S4a), which is consistent with the positive correlations reported from the Gibraltar deposit in Canada [44]. Compilation of published data of zircon trace elements confirms the positive correlations for the porphyry deposits in central and western CAOB ( Figure S4b), the Mesozoic Mo deposits in NE China ( Figure S4c,d), the Tintaya deposit in Peru ( Figure S4e), the EI Teniente deposit in northern Chile (Figure S4f), the deposits in the Yulong belt in eastern Tibet (Figure S4g), and the Dexing deposit in SE China ( Figure  S4h).
The magnitude of Ce anomalies in zircon is affected by the fractionational crystallization of REE-rich minerals [43].  [96]). Crystallization of these minerals causes the melt Ce/Nd (and zircon Ce/Nd) to increase because of the removal of more Nd than Ce. Ratio of Ce 4+ /Ce 3+ in zircon is calculated using D values of REE between zircon and bulk rock [31]. If REE-rich mineral crystallizes, REEs including Ce contents decrease in the melt, leading to low estimates for D[Ce3+] and high estimates for Ce 4+ /Ce 3+ . Therefore, the calculated Ce 4+ /Ce 3+ ratios may vary considerably without any change in fO2.

Zircon Ce anomalies and fertility of intrusions
The high value of Ce 4+ /Ce 3+ in zircon is associated with fertile magmas associated with porphyry Cu deposits (e.g., [16,18,31,32,44]). However, Ce 4+ /Ce 3+ ratios of zircon show large ranges in different deposits and metallogenic belts, and it is difficult to put an absolute Ce 4+ /Ce 3+ value to separate fertile from infertile intrusions. The threshold of Ce 4+ /Ce 3+ in zircons for ore-forming magmas is suggested to be 300 in the Chuquicamata-El Abra porphyry Cu belt in northern Chile by Ballard et al. [31]. This value is >200 for the Gibraltar porphyry Cu-Mo deposit, British Columbia of Canada [44], and >120 for the fertile ore-bearing hosting magmas for the porphyry Cu-Mo-(Au) deposits in the central and western CAOB [16] and the porphyry Cu deposits in the Red River-Ailaoshan zone of eastern Tibet [32].
Elevated Ce 4+ /Ce 3+ ratios are not always associated with fertile magmas, as pointed out by Viala et al. [98] in the Hualgayoc mining district in northern Peru. The TS microgabbro samples are barren and as wall rocks to the TS deposit, which show large range in Ce 4+ /Ce 3+ , from 197 ± 154 (TNS-1) to 357 ± 177 (TNS-3), and sample TNS-3 has similar Ce 4+ /Ce 3+ values to DBS dioritic to granodioritic rocks (231~461; Table 1). The size of the DBS-TS deposit is similar to those of the Bozshakol and Kounrad deposits in western CAOB (Table S2); however, the Ce 4+ /Ce 3+ values in zircons of the DBS-TS samples are higher than that in the Bozshakol (250~270) and Kounrad (200~260) deposits. By contrast, values of Ce 4+ /Ce 3+ in the DBS-TS deposit overlap most of the ranges of Ce 4+ /Ce 3+ for those giant porphyry deposits, such as the Oyu Tolgoi, eastern CAOB (95-554; first quartile to third quartile with mean at 434; Figure 9b; Table S2), EI Teniente, northern Chile (75~588; first quartile to third quartile with mean at 432; Figure 9b; Table S2), and Yulong, eastern Tibet (171~515; first quartile to third quartile with mean at 354; Figure 9b; Table S2), but the size for the DBS-TS deposit is much smaller than these giant deposits. Even within the giant Oyu Tolgoi deposit, there is no significant distinction between zircons from Hugo Dummett ore body (high Cu grade and tonnage) and Heruga ore body (lower Cu grade and tonnage) considering that both systems are hosted by similar rocks and formed during the same magmatic epoch, i.e., under the same geodynamic conditions and probably derived from the same or similar magma source [43]. Thus, the different grade and/or size of these deposits are due to other factors, such as fluid chemistry, and/or fluid focusing [99]. Our data indicate that Ce 4+ /Ce 3+ ratios should not be taken as an only parameter to separate fertile from infertile intrusions and would be unreliable if used to measure the size and/or grade of deposits.
Notwithstanding the problems associated with using Ce anomalies to evaluate the magma redox state, there still exists an empirical relationship between fertile porphyry systems and the compositions of zircons with higher average values for these parameters, as shown in Table S2. In particular, Ce/Nd can be obtained solely from zircon composition, and consequently, zircon may still be useful in exploration, although must be used with caution, and in cooperation with other conventional exploration techniques.  [16], Oyu Tolgoi deposit in eastern CAOB [17,100], Tintaya deposit in Peru [42,101], Dexing deposit in SE China [102], Yulong deposit in eastern Tibet [103,104], EI Teniente deposit in Chile [105,106], Chuquicamata and Radomiro-Tomic deposit in Chile [31,105], Mesozoic Mo deposits in NE China [18], Sn-W deposit [107], and barren intrusions [108].

Zircon Ce anomalies and oxidation state of magma
The ratio of Ce 4+ /Ce 3+ in zircon has been considered to reflect the magmatic oxidation state; i.e., oxidized magma produces zircon with high Ce 4+ /Ce 3+ ratios [31,32], and reduced magma produces zircon with low Ce 4+ /Ce 3+ values [107]. However, this is not the case for the DBS-TS deposit. Individual samples from the ore-bearing intrusions show a wide variation in zircon Ce 4+ /Ce 3+ values (174 ± 82 to 461 ± 290; average at 322; Table 1), but the calculated fO2 values show a rather narrow range, FMQ +1.0 to +1.5, for the mineralized samples and FMQ +1.4 to +4.2 for the variably altered samples (Table 2). Zircon grains from two samples from the Mesozoic Qz-monzonitic intrusion (barren) have much lower Ce 4+ /Ce 3+ values (45.9 ± 18.5 to 62.6 ± 38.8) but yield higher fO2 values (FMQ + 2.4 to +2.5). The lack of correlation between Ce 4+ /Ce 3+ and oxidation conditions (Table 2; Figure  10a) suggests that zircon composition may not reflect the oxidation state of parental magmas.

Europium Anomaly (Eu/Eu*) and Its Controlling Factors
Europium anomalies may possibly reflect the oxidation state of magmas, given Eu is present as divalent cation in reduced magmas and trivalent cation in oxidized magmas. Since plagioclase preferentially incorporates Eu 2+ , crystallization of plagioclase depletes Eu in the melt, whereas crystallization of titanite results in a positive Eu anomaly [43].
The zircon grains in the DBS ore-bearing dioritic to granodioritic intrusion have relatively high Eu/Eu* ratios (from 0.56 ± 0.03 to 0.59 ± 0.05 for mineralized samples and from 0.57 ± 0.06 to 0.62 ± 0.05 for altered samples) (Table 1; Figures S2a,b and 9a), which appears to be consistent with the overall oxidized parental magmas. However, the zircon grains in the TS microgabbro samples have variable and lower Eu/Eu* ratios from 0.37 ± 0.24 to 0.55 ± 0. 30 Figure 9a) for the DBS dioritic and granodioritic samples, but weakly correlate with Ce 4+ /Ce 3+ in zircon (Table 1; Figure 9a) for the TS microgabbro samples. The data suggest zircon Eu anomalies in the DBS dioritic and granodioritic samples are mainly controlled by factors unrelated to the redox conditions of magmas. Our data are consistent with the findings of previous workers (e.g., [43]), who suggested that Eu anomalies are primarily controlled by the crystallization of plagioclase and other phases, such as titanite, monazite and hornblende. Similar conclusions are drawn through studying ratios of other REEs and trace elements in zircons by some recent studies [36,37,40]. Thus, the relatively high and constant zircon Eu/Eu* values for the DBS mineralized and variably altered dioritic and granodioritic rocks suggest H2O-rich parental magma suppressed early plagioclase crystallization, but promoted hornblende crystallization (e.g., [44,49,109]). The proposed interpretation is supported by the presence of plagioclase and variable hornblende throughout the DBS samples, and consistent with the high Sr (average = 425 ppm) and Sr/Y (average = 40) in bulk rocks. On the contrary, the broad correlation of Eu/Eu* with Ce 4+ /Ce 3+ and negative Eu anomalies in REE patterns for the TS microgabbro samples may suggest that minor fractional crystallization of plagioclase occurred prior to or during crystallization of zircon grains. The crystallization of plagioclase thus lowers Eu/Eu* in the residual melt. Large variation in Ce 4+ /Ce 3+ in zircon associated by relatively constant Eu/Eu* is also observed in the Yulong porphyry Cu deposits in eastern Tibet, the Dexing deposit in SE China, the Tintaya deposit in Peru and EI Teniente deposit in Chile, and most of the porphyry Cu-Au-(Mo) deposits in the central and western CAOB (Figure 9b). Thus, compared with Ce 4+ /Ce 3+ or Ce/Nd, as a proxy of magma oxidation state, Eu/Eu* is not robust because it is strongly controlled by the crystallization of other minerals.

Values of fO2 and Implications for Magma Source and Mineralization
The mineral assemblage of magnetite, hornblende and titanite in the DBS-TS ore-bearing intrusions suggests that magma fO2 was above FMQ buffer [110]. Our newly obtained fO2 data (FMQ +1 to +4.2) from the DBS porphyry deposit are consistent with the fact that arc magmas are commonly oxidized, FMQ +1 to +4 (e.g., [111][112][113]). The data are also consistent with our proposal that the DBS-TS ore-bearing intrusions were generated in a continental-arc during the subduction of the Paleo-Asian Ocean. It is well documented that metasomatized mantle wedges are more oxidized (e.g., [114][115][116][117]), and that magmas and the arc crust inherit their oxidized nature from the metasomatized mantle wedges [111,112,115]. The redox condition is a critically important factor for magma fertility [118][119][120][121][122][123][124]. An elevated oxygen fugacity will facilitate the extraction of Cu, Mo and Au into the melt during partial melting because metal sulfides are incorporated into the melt owing to the much higher solubility of S as sulfate (SO4 2− ) under oxidized conditions [125]. Copper, Mo and Au in an oxidized magma can then partition into a magmatic-hydrothermal fluid forming porphyry Cu-Mo deposits [29,126,127]. This explains why the igneous rocks associated with porphyry Cu-Mo deposits of variable sizes (tonnage) have modestly high fO2 values in the CAOB, as well as in the NE China (Figure 10a,b).
Therefore, the magmatic oxidation state is an important indicator for forming porphyry deposits [31,49,118,124], but may not be an accurate fertility indicator for porphyry deposits.

Comparison with Other Fertile Granitic Intrusions Hosting Porphyry Cu-Mo-(Au) Deposits in NE China and Elsewhere
Well-known metallogenic belts of porphyry Cu-Mo-(Au) deposits are shown in Figures 1b and 11 (porphyry Mo deposits in the Mesozoic metallogenic belt in NE China) and summarized in Table S2.  Figure 2b, not shown here.

Tectonic Setting and Provenance for Magmatic Zircons
Trace elements in zircon are useful in discriminating tectonic settings and provenances for the porphyry deposits in the well-known metallogenic belts across the world [82]. Compositions of magmatic zircons from the porphyry deposits in central and western CAOB and the Mesozoic Mo deposits in NE China ( Figure S5a; Figure S5b) fall into the contineral arc field, but Mo deposits show higher ratios of U/Yb, suggesting that their magma sources contain more crustal components. The porphyry deposits of the Tethys metallogenic belt according to their different geographical locations show a moderate content of Hf (5000~12,000 ppm) and different U/Yb ratios (up to two orders of magnitude). For example, the porphyry deposits in western Tethys (e.g., Sar Cheshmch and Sungun) and central Tethys (e.g., Qulong, Jiama and Yulong) show similar U/Yb ratios (0.6~1.0) and Hf contents (6000~12,000 ppm), but the porphyry deposits in eastern Tethys (e.g., Tampakan and Batu Hijau) show rather low Hf contents (6500~9500 ppm) and U/Yb ratios (<0.5), which fell into Cont. Arc-type and MOR-type fields, respectively ( Figure  S5a,b), suggesting the ongoing evolution of the Tethys belt through time and that the magmatic zircons in porphyry deposits of western and central Tethys are mainly from the continental arc and the zircons in deposits of eastern Tethys are mainly from mantle source. Compared with other metallogenic belts, magmatic zircons in porphyry deposits of the central Andes metallogenic belt show very large Hf difference (e.g., 2000~4000 ppm for EI Teniente [105,106], >9000 ppm for Tintaya [42,101] and 7000~15,000 ppm for EI Salvador [36]), although they all fall into the Cont. Arc-type field. In addition to the factors of magma sources, the Hf content in zircons is also affected by the fractional crystallization of magma. Hf contents increase during fractional crystallization.  (Table S2). However, for Eu/Eu* values, the order is 1) the Mesozoic porphyry Mo deposit belt in NE China, 2) the porphyry Cu-Mo-(Au) deposits in CAOB, 3) Porphyry Cu deposits in the central Andes belt, and 4) Porphyry Cu-Mo deposits in the Tethys belt (Table S2).
There are weak correlations between Ce 4+ /Ce 3+ (or Ce/Nd) and metal tonnage (Cu or Mo) for the porphyry Cu-Mo-(Au) deposits in Central and Western CAOB, and for the Mesozoic porphyry Mo deposits in NE China. However, no distinct correlations are observed in the Tethys or central Andes metallogenic belts (Table S2).

Oxidation State of Magmas
In terms of oxidation state (calculated based on the model of Loucks et al. [39]), the Mesozoic Mo deposits in NE China are more oxidized with a mean value of FMQ +2.0, followed by the deposits in Tethys belt (FMQ +1.4), central Andes belt (FMQ +0.96), and the CAOB (FMQ +0.74), as shown in Table S2. However, no correlations are observed between values of Ce 4+ /Ce 3+ and ΔFMQ for deposits either in a single belt or among all the belts ( Figure 12). The relatively oxidized nature of the Mesozoic granitic intrusions hosting porphyry Mo deposits in NE China is likely the result of the prolonged subduction of the oceanic plate from the Pacific Ocean below the eastern part of CAOB since the early Jurassic. The Mesozoic Qz-monzonites (e.g., sample DBSA1 and DBSA2) in the DBS-TS deposit also show an elevated oxidation state (FMQ +1.2), supporting this interpretation. Figure 12. Plots of ΔFMQ versus Ce 4+ /Ce 3+ for intrusions host to the porphyry Cu-Mo deposits in well-known metallogenic belts. Note: 1) solid circles with numbers refer to porphyry Cu-Mo deposits, which are the same as listed in Table S2; 2). ΔFMQ values present in this figure are calculated by the method of Loucks et al. [47].

Conclusions
1) The DBS-TS ore-bearing intrusions were formed from water-rich magmas. The hydrous magmas formed by partial melting of a juvenile lower crust/lithospheric mantle, or metasomatized mantle wedge in a continental-arc setting during northwestward subduction of the Paleo-Asian Oceanic plate below the Xing'an block.
2) The dioritic and granodioritic intrusions show an adakitic signature of high Sr/Y and low HREE, due to fractional crystallization of amphibole in water-rich magmas. 3) Ratios of Ce 4+ /Ce 3+ and Ce/Nd in zircons range from 174 to 461 (mean 322 ± 203) and from 7.4 to 20.3 (mean 14.9 ± 7.8) for samples from DBS-TS deposit, which confirms a positive correlation between the two parameters. 4) The magma oxidation state is calculated to be FMQ +1.0 (±1.2) to +1.5 (±1.2) for mineralized samples and of FMQ +1.4 (±1.2) to +4.2 (±1.4) for variably altered samples using the hydrous-based equation by Smythe and Brenan [1]. The values are comparable to recalculated values using the same equation for fertile intrusions hosting the porphyry Cu-Mo-(Au) deposits in the central and western CAOB and elsewhere. 5) Although the magnitude of Ce anomalies in zircons is affected by the magmatic compositions and fractional crystallization of REE-rich minerals, positive anomalies of Ce are still associated with fertile porphyry systems. The values, however, may not be a valid fertility indicator of magmas.

Materials:
The following are available online at www.mdpi.com/2075-163X/11/5/503/s1, Figure S1: SEM-cathodoluminescence images of zircon grains from the DBS-TS ore-bearing microgabbroic, dioritic and granodioritic intrusions. Figure S2: Chondrite-normalized REE patterns of zircon for a (a) representative sample (DBS16) of mineralized granodiorites, (b) representative sample (DBS4) of altered granodiorites, (c) representative sample (DBS7) of microgabbros, and (d) representative sample (DBSA1) of Qz-monzonites. Figure  S3: Evolution model for the Paleo-Asian Ocean and magmatism in NE China from (a) early Paleozoic to (b) early Carboniferous of the eastern CAOB (modified after [1,2]). Figure S4: (a) Plot of Ce 4+ /Ce 3+ versus Ce/Nd for zircons in the gabbroic, dioritic and granodioritic intrusions of the DBS-TS porphyry Cu-Mo deposit. Shown also for comparison are (b) deposits in central and western CAOB [3], (c1 and c2) Mesozoic Mo deposits in NE China [4], (d) Yulong Cu deposit belt in eastern Tibet [5,6], (e) Dexing Cu deposit in SE China [7], (f) Tintaya Cu deposit in Peru [8,9], and (g) EI Teniente Cu deposit in Chile [10,11]. Figure S5: Tectono-magmatic setting discrimination diagrams of (a) U/Yb versus Hf, and (b) Ti versus U/Yb based on trace elements of zircons for porphyry deposits in the world-famous metallogenic belts. Note that the references are listed in Supplementary Figures. Table S1: Bulk-rock major and trace element contents of samples from the DBS-TS ore-bearing intrusions. Table S2: Summary of the porphyry Cu-(Mo) deposits in other metallogenic belts (areas). Table S3. Average trace element contents of zircons and recalculated Ti-in zircon temperature (°C) and fO2 values in porphyry Cu-Mo-(Au) deposits from the central and western CAOB.