In-situ trace element and Fe-isotope studies on

Chemical Geology 453 (2017) 111–127

Contents lists available at ScienceDirect

Chemical Geology journal homepage: www.elsevier.com/locate/chemgeo

In-situ trace element and Fe-isotope studies on magnetite of the volcanic-hosted Zhibo and Chagangnuoer iron ore deposits in the Western Tianshan, NW China T. Günther a,⁎, R. Klemd a, X. Zhang b, I. Horn c, S. Weyer c a b c

GeoZentrum Nordbayern, Friedrich-Alexander-Universität Erlangen-Nürnberg, Schlossgarten 5a, 91054 Erlangen, Germany China Minmetals Corporation, 100010 Beijing, China Institut für Mineralogie, Leibniz Universität Hannover, Callinstr. 3, 30167 Hannover, Germany

a r t i c l e

i n f o

Article history: Received 2 November 2016 Received in revised form 31 January 2017 Accepted 1 February 2017 Available online 3 February 2017 Keywords: Zhibo & Chagangnuoer iron ore deposit Magnetite trace elements Fe-isotopes Skarn IOCG-Kiruna type Tianshan

a b s t r a c t The Carboniferous Zhibo and Chagangnuoer iron deposits are situated within a caldera centre and along the flank of the same volcanic edifice, respectively, in the Awulale Iron Metallogenic Belt of the Western Tianshan orogen. Several stratiform 10 to 100 m large, tabular to lenticular shaped magnetite-dominated ore-bodies occur in (trachy-) andesitic to rhyolitic host rocks. The magnetite mineralization mainly occurs as massive iron ores, partly with columnar-network or flow textures, and as disseminated magnetite ores. Trace element and isotope investigations of the different ore types reveal two major groups of magnetite: Group I, represented by the massive, partly brecciated ores from both deposits, is enriched in Ti, V, Ni, and HFSE such as Y, with concentrations similar to Iron Oxide-Copper-Gold (IOCG) ores. The δ56Fe values (up to 0.4‰) support an ortho-magmatic origin corresponding with an isotopic source calculation at ~800 °C. Positive correlations between Fetotal and δ56Fe (from +0.4‰ to −0.1‰) and incompatible trace element contents (e.g. Si, Al, Nb, Ti and Y) in Group I magnetite are interpreted to be the consequence of a Raleigh-type fractionation process. Decreasing V, Ni and Mn values indicate changing fO2 conditions at the time of ore genesis. Group II, which is represented by the disseminated ores from Chagangnuoer, is - compared to Group I - relatively depleted in elements like Ti, V, Ni and Y and further spans a dominant δ56Fe range from about 0‰ to −0.5‰. These textural and chemical characteristics and the garnet-actinolite-diopside-epidote-carbonate-K-feldspar paragenesis are in accordance with hydrothermal Fe-skarn ores. The similar multi-element patterns of magnetite from all investigated samples, the overlapping δ56Fe ratios of the same massive ore-type from Zhibo and Chagangnuoer and the close proximity of both deposits indicate a common source of Fe-enrichment for the different iron ore types. In contrast to the ortho-magmatic Group I magnetite, reverse trace element trends with decreasing δ56Fe ratios (from 0‰ to −0.5‰) among the disseminated ores cannot simply be explained by a straightforward Raleigh fractionation or alteration processes. Therefore, a bimodal formation model is suggested for the Group II magnetite formation, including a partial remobilization of iron from the proximal, ortho-magmatic ore bodies and a subsequent distal re-precipitation. These processes were driven by late-stage hydrothermal fluids, which originated from deeper- seated granitic/granodioritic intrusions in the immediate vicinity. © 2017 Elsevier B.V. All rights reserved.

1. Introduction The Western Tianshan orogen is a significant metallogenic province in the southwestern part of the Central Asian Orogenic Belt (Fig. 1A) and hosts precious Cu-Mo bearing porphyry-, epithermal Au and volcanichosted Fe-deposits (Pirajno, 2013; Pirajno et al., 2011; Qin et al., 2002; Zhai et al., 2009; Zhang et al., 2010). The latter are concentrated as numerous high-grade iron ore deposits in the Awulale Iron Metallogenic ⁎ Corresponding author. E-mail address: [email protected] (T. Günther).

http://dx.doi.org/10.1016/j.chemgeo.2017.02.001 0009-2541/© 2017 Elsevier B.V. All rights reserved.

Belt (AIMB; Fig. 1B, C) with iron resources of up to 1000 million tons (Dong et al., 2011; Feng et al., 2010). Two of the most investigated iron ore districts of the AIMB are the Zhibo and Chagangnuoer deposits, which are composed of large amounts of magnetite-dominated iron ore with different mineralogical and geochemical characteristics (cf. Jiang et al., 2014; Zhang et al., 2015). Despite their intimate association (Fig. 1C), genetically different models were suggested mainly due to petrographic observations: The iron mineralization at Zhibo was thought to be of volcanic hydrothermal origin (Tian et al., 2009) or of magmatic origin and a late hydrothermal alteration (Feng et al., 2010; Jiang et al., 2014; Jiang et al., 2012b). The

112

T. Günther et al. / Chemical Geology 453 (2017) 111–127

Fig. 1. (A) Tectonic framework of the Central Asian Orogenic Belt (modified after Han et al. (2010); Jahn (2004)). (B) Geological map of the Western Tianshan in NW China showing the major iron deposits in the Awulale Iron Metallogenic Belt (AIMB; black, dashed line); map after from Gao et al. (1998) and Zhang et al. (2012). (C) Geological map of the eastern segment of the Yili block (YB) showing the localities of the Zhibo and Chagangnuoer iron ore deposit (Feng et al., 2010) within an caldera system (caldera margin indicated as white, dashed line; from Zhang et al. (2015)). TB – Tarim Block, NMT – Northern Margin of Tarim Block, CTT – Central Tianshan Arc Terrane, YB – Yili Block, JT – Junggar Terrane.

Chagangnuoer deposit was interpreted as an exhalative VMS-like deposit (Wang et al., 2001) or a complex magmatic deposit with associated Fe-skarn formation (Hong et al., 2012b; Sun et al., 2015; Wang et al., 2011a). In addition, Zhang et al. (2015) suggested a common, “Kirunatype” magmatic-related origin for both deposits, which were overprinted by late hydrothermal alterations. The Zhibo and Chagangnuoer deposits share many geological, mineralogical and geochemical similarities with other alkali-calcic, hydrothermally altered volcanic-hosted iron ore deposits, which are the subject of a long-standing, controversial discussion concerning a possible magmatic or hydrothermal origin, or a combination of both (e.g., Henriquez and Martin, 1978; Nystroem and Henriquez, 1994; Rhodes et al., 1999; Sillitoe and Burrows, 2002; Nystroem et al., 2008; Tornos et al., 2016; Dare et al., 2014, Zhang et al., 2015) In this study, we report for the first time an integrated model using petrographic, textural, Fe-oxide in-situ trace element and Fe-isotope data in order to gain new insight in the formation and enrichment processes of these massive magnetite accumulations. The results have important implications for the discussion on the ‘magmatic versus hydrothermal’ origin of volcanic-hosted iron ore deposits worldwide.

amalgamation of the Junggar-, Yili-, Central Tianshan-, and Tarim microcontinental blocks, which are separated by several major suture zones (e.g. Gao et al., 1998; Gao et al., 2015; Qian et al., 2009), and the associated closure of at least three oceans (Tersky, South- & North Tianshan oceans) during Early Palaeozoic to Early Mesozoic times (e.g. Charvet et al., 2011; Gao et al., 2009; Windley et al., 2007). The Awulale Iron Metallogenic Belt (AIMB) stretches parallel along the NTSZ at the East of the triangular wedge-shaped tail of the Yili Block (Fig. 1B). Its Precambrian basement of mainly granitic gneisses, migmatites, schist and metacarbonates revealed Proterozoic U-Pb zircon ages between ~ 1650 and 800 Ma (Chen et al., 1999; Hu et al., 2010; Li et al., 2009; Wang et al., 2014a). During several periods of the Northern and Southern Tianshan seafloor subduction, especially in the late Devonian and Carbon, arc magmatism led to a voluminous distribution of volcanic and plutonic rocks beside older Silurian and Devonian strata (e.g. Long et al., 2011; Wang et al., 2006; Wang et al., 2007; Zhu et al., 2009). The Carboniferous volcanic rocks are closely related with the AIMB iron deposits and predominantly encompass basaltic to rhyolitic lavas, breccias and tuffs (Qian et al., 2006; Zhu et al., 2005). 2.2. Geology of the Zhibo and Chagangnuoer iron ore deposits

2. Geological setting 2.1. Regional geology The Western and Eastern parts of the Tianshan intercontinental mountain belt (northwestern China) are situated in the southwestern part of the Central Asian Orogen (e.g. Xiao et al., 2008) (Fig. 1A). The Western fold belt was created by complex multistage accretion and

The AIMB comprises several different ore deposit types such as porphyry Cu-Mo, epithermal Au and volcanic-hosted Fe-deposits, and covers an area of about 2500–5000 km2 (Jiang et al., 2014). The Zhibo and Chagangnuoer iron ore deposits belong to an important metal mining district in the southeastern part of the AIMB (Fig. 1C). Both are hosted by hydrothermally altered volcanic-sediments and (trachy-) andesitic to rhyolitic, subvolcanic rocks of the Lower Carboniferous


Dahalajunshan Formation. Andesites, trachy-andesites and rhyolites reveal U-Pb zircon ages of about 350 Ma, 329 Ma and 321 Ma in the immediate vicinity of the Zhibo and Chagangnuoer iron ore deposits (Feng et al., 2010; Jiang et al., 2014; Zhang et al., 2015). Late phases of magmatism such as granitic plutons and dykes with crystallization ages between 323 and 248 Ma crosscut the volcanic host rocks as well as some iron ore bodies (e.g. Sun et al., 2015; Zhang et al., 2015; Zhang et al., 2012). Remote sensing investigations (Feng et al., 2010) revealed a ca. 300 km2 large caldera-like structure hosting the Zhibo and Chagangnuoer deposits (Wang et al., 2014b; Zhang et al., 2015). The Zhibo deposit is situated in the centre of the volcanic edifice (Figs. 1C, 2B) and comprises at least 21 ore bodies, which extend subhorizontally from SE to NW for about 5.5 km (Feng et al., 2010). Most iron ore bodies are massive, tabular to lenticular shaped, partly with tapered extensions and may reach several 100 m in length (Fig. 2C). Often, the trachy-andesitic and trachytic host rocks, either envelope the ore bodies or sometimes occur as lenticular bodies within the mineralized zones. Contact boundaries between the iron ore bodies and volcanic host rocks are usually sharp (Zhang et al., 2015).

113

Along the northwestern caldera flank, subvolcanic rocks, tuffs and marble encompass the Chagangnuoer ore bodies (e.g. Jiang et al., 2012a; Jiang et al., 2014; Wang et al., 2011a; Zhang et al., 2012). The deposit complex consists of two major, lens-like ore bodies (the main body No.1 is illustrated in Fig. 2), which occur in a distance of about 1.5 km (Jiang et al., 2014; Zhang et al., 2015). The stratiform ore bodies, with varying thicknesses between 9 and 125 m, are controlled by an accumulation of NW striking faults and ring fractures (Hong et al., 2012a; Zhang et al., 2015).

2.3. Magnetite-bearing ore types At the Zhibo and Chagangnuoer deposit, four main ore types can generally be distinguished by their textural and mineralogical properties: Breccia (1) and lava flow (2) ores are both located in the upper parts of the deposits. Massive ore (3) occurs in the lowermost part of the deposit while a complex breccia ore (4) is abundant throughout the whole deposit.

Fig. 2. (A) Geological overview of the main iron body at the Chagangnuoer ore deposit showing the sample localities. (B) Geological map of the Zhibo iron ore district with its distributed ore bodies. The location of the cross section A–A′ is marked. (C) Geological profile, constructed from drill cores, across the iron ore body showing the tabular-lenticular shape of the massive NW-SE striking ore bodies. Maps modified after Feng et al. (2010) and Zhang et al. (2015).

114


Briefly summarized, the different magnetite-bearing ore types are: Type (1) brecciated ore consisting of millimetre- to centimetre-sized brecciated host rock fragments (angular to subrounded) in a finegrained, partly disseminated matrix of magnetite. The amount of host rock fragments exceeds 60 vol% within the breccia ore. Type (2) flow or banded iron ore consisting of dense magnetite layers intercalated with silicate host rock minerals. Type (3) massive ore comprising host rock fragments and magnetite. The latter occurs in form of a columnar-network structure, platy aggregates or a dense groundmass. The magnetite content of ore types (1) to (3) ranges between ~ 30 and 60 vol%. Type (4) complex breccia ore comprising centimetre scale, massive magnetite aggregates, which add up to an overall magnetite content between ~40 to 50 vol%. In contrast to Type (1), the magnetite aggregates are brecciated by a late stage, hydrothermal mineral assemblage. For a more detailed description of the different ore types see Feng et al. (2010), Jiang et al. (2014) and Zhang et al. (2015). Subsequent to the primary crystallization, the volcanic host rocks and the iron ore bodies were overprinted by hydrothermal fluids causing different alteration stages such as early pervasive and late Ca-Na-K controlled alterations (e.g. Jiang et al., 2014; Zhang et al., 2015). The immediate host rocks and their fragments, occurring in all ore types, are characterized by a pervasive K-feldspar, epidote ± sulphide alteration. This mineral assemblage also occurs in late hydrothermal veins. However, the vein-type mineral paragenesis of the Chagangnuoer complex breccia ore is more complex and predominantly comprises a skarntype garnet-actinolite-epidote-sulphide ± carbonate assemblage. Furthermore, the minerals of the skarn-type alteration (garnet-actinoliteepidote) at the Chagangnuoer deposit yield an Sm-Nd isochron age of 313 ± 7 Ma (Zhang et al., 2015). 3. Sampling and analytical techniques In this study, six drill core samples from the Zhibo (5 samples) and Chagangnuoer (1 sample) iron ores as well as three samples from outcrops of the ore body of Chagangnuoer were petrographically and geochemically investigated. The samples comprise the flow-type, massive-, complex breccia and disseminated ores with different hydrothermal alteration overprints. Major and some trace elements of magnetite were determined by electron probe microanalysis (EPMA) using a JEOL JXA8200 Superprobe at the GeoZentrum Nordbayern. The quantitative analyses were performed with carbon-coated thin sections in an evacuated chamber with b4.0 ∗ 10− 6 mbar. Standards and samples were analysed with a 20 kV wavelength dispersive signal (WDS) and a beam current of 20 nA with a diameter of b1 μm. The following standards were used: Fe2O3 (Fe), Ca3Si3O9 (Ca, Si), KAlSi3O8 (Al), MnSiO3 (Mn), MgO (Mg), TiO2 (Ti), Cr2O3 (Cr), and vanadium metal (V). Integration times for peaks of Cr and V were 40 s and 30 s, and for the remaining elements 20 s. Each spot analysed by EPMA was marked on a related BSE image for subsequent trace element measurements of the same minerals via laser techniques (see Supplementary material). Trace element concentrations were determined by laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) on the same thin sections as used for EPMA at the GeoZentrum Nordbayern. Spots and lines of interest were ablated using a New Wave Research UP-193 FX laser system coupled to an Agilent 7500c quadrupole spectrometer (1220 W plasma power). Ar (1.05 min−1) and He (0.5 l min−1) acted as carrier gas and Ar was additionally used as plasma (14.9 l min−1) and auxiliary gas (0.9 l min−1). Single spot ablations were carried out with a repetition rate of 15 Hz at an irradiance of 0.76 GW/cm2 and a fluence of 3.79 J/cm2. According to the respective magnetite grain size, a spot diameter of 35 μm was chosen while smaller domains were occasionally analysed using a spot diameter of 20 or 25 μm. The acquisition time for each analysis was 40 s including 20 s for background scanning. Different integration times were 10 ms for 23Na, 29Si and 25 ms for 24 Mg, 26 Mg, 27Al, 31P, 42Ca, 44Ca, 45Sc, 47Ti, 51V, 53Cr, 55Mn, 57Fe, 59Co, 60 Ni, 63Cu, 65Cu, 66Zn, 73Ge, 88Sr, 89Y, 90Zr, 93Nb, 95Mo, 137Ba, 178Hf,

181

Ta, 182W, 208Pb, resulting in 0.9782 s per mass scan. Repeated measurements on the external glass reference NIST SRM 612 (Pearce et al., 1997) were used for external calibration. Analytical precision was monitored by measurements of the international BCR-2 g reference material yielding values better than 10% RSD (1 sd) for most elements except for Ge (18%) and W (28%). To control a potential drift of the device, the reference material was checked by repeated measurements several times during an analytical session. Iron (57Fe) was used as internal standard. Evaluation of the complete data set was carried out with Glitter version 3.0, Online Interactive Data Reduction for the LA-ICP-MS (Van Achterbergh et al., 2000). In-situ Fe-isotope data were obtained using a UV femtosecond laser ablation device coupled to a Thermo Finnigan Neptune Plus multi-collector ICP-MS at the Leibnitz University of Hannover (Horn et al., 2006). The laser system is based on a 100-femtosecond Ti-sapphire regenerative amplifier system (Hurricane I, Spectra Physics, USA) described in detail by Horn and Von Blanckenburg (2007). All Fe-isotope analyses were conducted in accordance with the protocol of (Oeser et al., 2014). Depending on crystal size, the spot diameter and repetition rate of the beam were adapted between 35 and 45 μm and 5 to 2 Hz providing signal intensity between 15 and 25 V on mass 56Fe. The total analysis time for each spot was 180 s including 30 s background scanning. Measurements were performed at high mass resolution (m/Δm ~9000) to resolve mass interferences of argon nitrides and argon oxides on Fe-isotopes and to provide wide plateaus for determining interference-free measuring signals (Weyer and Schwieters, 2003). A deviation of measured δ56Fe and δ57Fe values from the theoretical mass dependent fraction line contributed from interfering argon molecules, was not observed (see Supplementary material). The detector array of the Faraday cups was positioned to acquire simultaneous signals of 52Cr+, 54 Fe+, 56Fe+, 57Fe+, 58Ni+ and 60Ni+. 52Cr+ was monitored to correct for isobaric interference of 54Cr+ on 54Fe+. Via a quartz glass spray chamber, a Ni reference solution (NIST SRM 986) was added into the plasma to monitor instrumental mass bias following the protocol of Oeser et al. (2014). A standard-bracketing sequence with the certified Fe metal standard IRMM-014 was used for sample and standard Fe-isotope measurements. It has previously been shown in several studies that oxides, carbonates and silicates can be analysed with femtosecond Laser ablation MC-ICP-MS relative to a metal standard without a significant matrix effect (Horn et al., 2006; Oeser et al., 2014; Steinhoefel et al., 2009). Accuracy and precision of the method used here were monitored using the in-house JM puratronic (PURA) Fe-standard (99.995% Puratronic, Johnson Matthey, lot No. FE495007IF2). The mean δ56Fe value of 0.09 ± 0.07 (2 sd) ‰ determined in this study agrees with the δ56Fe value of 0.09 ± 0.05 (2 sd) ‰ reported in (Horn et al., 2006). Analytical standard errors (2 sd) of individual PURA analyses were better than 0.05‰ for δ56Fe and b 0.09‰ for δ57Fe. The reference material was analysed several times at the beginning of every analytical session and monitored throughout the session for potential drift. All acquired data lie within their 2 sd uncertainties on the fractionation line (see Supplementary material) and are expressed in delta notation (in ‰) relative to the IRMM-014 reference material calculated as follows: " 56ð57Þ 56ð57Þ

δ

Fesample ¼

# Fe =54 Fe sample −1 56ð57Þ Fe =54 Fe IRMM−14

All measurements were evaluated and corrected for significant variations in δ56Fe and δ57Fe (e.g. caused by micro or nano inclusions) with an adapted template of the LamTool, Data Reduction utility for LA ICPMS (v. 081117). Uncertainties for δ56Fe and δ57Fe values are propagated 2 sd including sample and bracketing standards (sample + 2× standard). Core and rim notation represents the relative point location on an analysed magnetite grain. mag = magnetite; Zb = Zhibo; Ch = Chagangnuoer.


4. Results 4.1. Petrography 4.1.1. Zhibo ore district The flow-type ore (11ZB01-09) is defined by alternating and mingled layers of magnetite and silicate minerals, predominantly actinolite, diopside and alkali feldspar (Fig. 3A). Secondary epidote, chlorite and calcite (ca. 5 vol%) partly replace primary silicate phases such as

115

clinopyroxene and feldspar. The contact between silicate and oxide minerals is predominantly sharp, while the grain size of the equigranular silicates decreases when approaching the magnetite layers. The Fe-oxide domains form irregular, dense masses with subto anhedral magnetite. Some larger and formerly euhedral crystals show corroded rims, holes and are crosscut by several cracks (Fig. 3A3). Furthermore, they appear corroded and deformed within denser magnetite masses. Larger grains are characterized by significant compositional zoning in BSE images or have at least a different rim domain as

Fig. 3. Hand specimens and photomicrographs of the different ore-types from Zhibo and Chagangnuoer. (A1, A2) Equigranular layers of actinolite, K-feldspar, clinopyroxene and epidote mingled with magnetite and pyrite defining a flow texture (11ZB01-09). (A3 – BSE image) Some larger grains show evidence of mechanical deformation in form of cracks and dislocated crystal parts. (B) The complex breccia ore (11ZB04-04) consists of dense magnetite aggregates crosscut by veins filled with mainly actinolite, epidote and calcite. (C1, C2 - BSE image) Typically massive magnetite (zoned) ore (10ZB01-08) has filled interstitial pore-space with mainly K-feldspar and diopside. (D1, D2 - reflected light) Massive ore with a network of columnar-platy shaped magnetite (11ZB05-14). (E1, E2 - reflected light) In the massive ore (10CG16-13), magnetite appears with interstitial fillings of actinolite and pyrite. (F1) Alternating layers of disseminated magnetite (11CG08-02) and diopside and alkali-feldspar have transitional contacts. (F2 - BSE image) Magnetite with compositional zoning. (G) The disseminated magnetite ore (10CG18-07) occurs as layers and clusters of magnetite between a silicate assemblage of diopside, garnet and K-feldspar. Sulphide veins crosscut the whole sample. (H1, H2 - reflected light) Idiomorphic garnet enclosed in a matrix of disseminated magnetite (11CG02-03). Mt = magnetite, Py = pyrite, Epi = epidote, Act = actinolite, Cal = calcite, Kfs = K-feldspar, Di = diopside, Grt = garnet, Bn = bornite.

116


indicated by a brighter reflection (Fig. 3A3). An- to subhedral pyrites, with grain sizes of about 0.3–1.0 cm enclose magnetite as well as silicate minerals and tend to form layer-like aggregates. In hand specimen magnetite from the complex breccia-type ore (11ZB04–04) appears as centimetre-large, massive aggregates, which are replaced and crosscut by an alteration assemblage of diopside, actinolite, alkali feldspar, epidote, calcite and pyrite (Fig. 3B). The alteration assemblage also crosscuts and brecciates larger magnetite aggregates on thin section scale (see Supplementary material). The massive type ore samples (10ZB01-08, 10ZB02-07) consist of 70–85 vol% magnetite (Fig. 3C1) appearing as several 100 μm large, an- to subhedral grains (Fig. 3C2). However, most magnetite forms dense masses and aggregates so that grain boundaries are difficult to identify. BSE images reveal oscillating zoning (Fig. 3C2). Secondary silicate minerals such as K-feldspar, diopside and epidote occur in former pore space and voids (Fig. 3C1). In the massive ore sample 11ZB05-14, magnetite forms a columnar network of sub- to euhedral, platy or lathy grains, which are up to 1 cm sized in length and between 20 and 200 μm in width (Fig. 3D). Many platy crystals are curved or broken by interstitial fillings of actinolite, diopside, alkali-feldspar and granular pyrite (Fig. 3D2). 4.1.2. Chagangnuoer ore district The massive ore sample 10CG16-13 consists of about 60–70 vol% granular to platy-like magnetite (Fig. 3E). Euhedral to subhedral, tabular and columnar-shaped amphibole, clinopyroxene and minor alkali-feldspar are enclosed within or between the Fe-oxide minerals (Fig. 3E2). Pyrite and chalcopyrite occur as interstitial fillings (Fig. 3E2). Sample 11CG08-02 shows a curved-like texture of alternating magnetite and silicate layers (Fig. 3F1). The deformed magnetite layers comprise partly euhedral, zoned grains (Fig. 3F2), which extend into calcitefilled cavities. The contact between massive magnetite and silicate layers, consisting mainly of alkali-feldspar and diopside, is transitional. The massive Fe-oxide layers gradually replace the silicate assemblage (Fig. 3F1). A repetitive layering or banding of fine-grained magnetite characterizes sample 10CG18-07 (Fig. 3G). Most layers, which vary from about 2 to 8 mm in thickness, occur almost parallel throughout the hydrothermally altered silicate assemblage, mainly consisting of secondary garnet, K-feldspar and sulphides and between 20 and 45 vol% disseminated or irregular ca. 10 to 50 μm large, anhedral magnetite grains. The sulphides also occur in veins crosscutting the sample. The intergranular magnetite (20 to 25 vol%) of sample 11CG02-03 forms a groundmass-like texture with 200–800 μm large, euhedral to subhedral garnet phenocrysts (65 to 75 vol%) (Fig. 3H). Grain boundaries between magnetite and garnet are straight to slightly curved, whereby magnetite aggregates enclose individual garnets (Fig. 3H2). K-Feldspar, diopside and sulphides are intimately associated with magnetite and display modal amounts of about 5–10 vol%. Bornite replaces pyrite and chalcopyrite (Fig. 3H2). The BSE images reveal that the magnetite of the different ore types at Chagangnuoer is zoned, especially the rim domains (cf. Fig. 3F2). The investigated samples represent the above described ore types 2, 3 and 4 (Table 3). However, the Chagangnuoer samples 10CG18-07, 11CG02-03 and 11CG08-02 display transitional boundaries between silicate-rich layers and magnetite ore and thus classify as disseminated ore instead as flow-type ore (Zhang et al., 2015). 4.2. Geochemistry 4.2.1. Magnetite chemistry Since zoning is common in the back-scattered images, compositional profiles of magnetite grains were undertaken with EPMA (usually below the detection limit) and LA-ICP-MS. The number of analyses per grain depends on crystal size and mineral zonation. Mean sample values of major and trace elements in magnetite from the different ore-types

are documented in Table 1 and mean values of every analysed grain are shown in the Supplementary material. Generally, all incompatible trace element contents similarly increase from interior to rim in the different ore-types representing internal growth zoning of magnetite (see profiles in the Supplementary material). Vanadium and Ni contents decrease from interior to rim in magnetite, especially at Zhibo (profile of flow-type and massive ore). At the outer rim areas, the incompatible trace element contents partly appear to decrease, but no consistency for specific elements between the different ore-types was observed. Thus, average whole grain values are used for the different classification diagrams. Trace element contents of the different ore types from Zhibo and Chagangnuoer are plotted on the magnetite discrimination diagrams of Dupuis and Beaudoin (2011) and Dare et al. (2014) (Figs. 4, 5). In the Ca + Al + Mn and Ni/(Cr + Mn) vs. Ti + V diagrams, the majority of ore samples from Zhibo, including the complex breccia ore and the massive ore with network textures, show chemical similarities to ortho-magmatic magnetite of the Kiruna, IOCG and Porphyry copper fields with relatively constant Ti + V contents (Fig. 4A, B). The orthomagmatic iron ores (IOCG, Kiruna, Porphyry copper), which display a high, magmatic temperature range at the time of magnetite crystallization, can be discriminated from low- to moderate temperature (b400 °C), hydrothermal ores (Fe-skarn, BIF, VMS). The term orthomagmatic infers an origin of magnetite by magmatic and/or magmatic-hydrothermal processes (cf. Tornos et al., 2016). The magnetites of the massive ore sample 10ZB01-08 and all Chagangnuoer ore types plot in the hydrothermal field of skarns (Fig. 4A, B). The massive magnetite mineralization at Chagangnuoer, however, has a similar Ti + V content (ca. 0.1 wt%) as most Zhibo samples with IOCG character (Fig. 4A, B). The Ca + Al + Mn, Ti + V and Ni/(Cr + Mn) variations between both deposits span ranges of about one order of magnitude, whereby trace element concentrations of magnetite are relatively constant within one sample (Figs. 4, 5). None of the magnetites plot in the VMS field of the Al/(Zn + Ca) vs. Cu/(Si + Ca) diagram (Figs. 4C). Normalized multi-element diagrams of the magnetite trace elements reveal rather similar patterns of the different ore types from both deposits (Fig. 5). Magnetite data were normalized to magmatic magnetite from Fe-Ti-V oxide deposits (Dare et al., 2014). Low Zr, Hf, Sc, Ta, Nb, Ti, Zn, Co and Cr values of magnetites from Chagangnuoer and Zhibo (Fig. 5A, B) differ from those of “classic” magmatic magnetite (from intermediate melts). However, they correspond in conjunction with other trace elements to an ortho-magmatic signature, similar to magnetites from IOCG, Porphyry copper and skarn deposits (Fig. 5C). Consequently, elements such as Si, Ca, Pb, Ge, W and partly Y, which are incompatible in magmatic systems, are relatively enriched. The normalized trace elements of the different iron ore types from Zhibo display similar trends although with different degrees of enrichment (Fig. 5A). The complex breccia ore has the lowermost trace element contents, except for V and Ni. In contrast, the massive (10ZB0108) and flow-type ores have the highest Si, Ca, Y, Pb, Zr, Hf, Ta, Nb, Mo, Sn and Mg values, but the lowest V and Ni values. The P, Ge, Sc and Cu concentrations reveal no significant differences between the texturally different iron ore types. In general, Zhibo magnetites from ore samples located along the central caldera have normalized trace elements such as W, Mo, Sn, V and Ni, which are relatively enriched and vary about one order of magnitude compared to those from the flank area (Fig. 5). At Chagangnuoer, the massive ore has higher concentrations of trace elements (except Pb, Hf, Cu, Mo and Sn) than the disseminated ore (Fig. 5B). The trace element enrichment of this massive ore, normalized to magmatic magnetite, is in the same magnitude as that observed for the massive Zhibo ore. 4.2.2. Fe-isotope data of magnetite Fe-isotopic compositions are given in Table 2 and are displayed in Fig. 6. The δ56Fe values vary (relative to IRMM-14) between about −0.4‰ and +0.1‰ (~±0.06, 2 sd) for magnetite from Chagangnuoer.


117

Table 1 Average magnetite composition of different ore types from Zhibo and Chagangnuoer. Location

Zhibo

Zhibo

Zhibo

Zhibo

Zhibo

Chagangnuoer

Chagangnuoer

Chagangnuoer

Chagangnuoer

Sample no.

11ZB05-14

11ZB04-04

11ZB01-09

10ZB01-08

10ZB02-07

10CG16-13

11CG08-02

10CG18-07

11CG02-03

Ore type

Massive (network)

Complex brecci

Flow

Massive

Massive

Massive

Disseminated

Disseminated

Disseminated

Analyses

N = 5 n = 10 N = 4 n = 12 N = 5

n = 15 N = 4

n = 14 N = 4 n = 12 N = 5

n = 19 N = 5 n = 20 N = 11 n = 11 N = 5 n = 16

Element

Mean

SD

Mean

SD

Mean

SD

Mean

SD

Mean

SD

Mean

SD

Mean

SD

Mean

SD

Mean

SD

Fe(total) [wt.%] Na Mg Al Si P Ca Sc Ti V Cr Mn Co Ni Cu Zn Ge Sr Y Zr Nb Mo Ba Hf Ta W Pb δ18O [‰]

71.6 257 743 839 4766 28.0 1439 1.54 257 1681 49.5 789 7.54 708 1.23 63.2 8.63 9.28 0.85 0.54 0.11 2.49 3.65 0.10 0.03 41.2 1.25

0.20 100 454 476 1896 4.87 705 0.55 141 260 18.6 47.5 0.77 116 0.19 25.5 1.75 10.8 0.37 0.44 0.00 1.80 2.26 0.03 0.01 23.0 0.94

71.7 36.7 103 417 3044 27.2 236 0.47 420 734 20.2 316 0.41 28.4 1.93 29.0 7.92 0.61 0.13 0.28 0.06 0.14 0.68 b.d.l. 0.02 0.24 0.40 1.2

0.17 7.72 45.6 138 387 4.29 28.0 0.05 48.3 12.0 5.31 6.24 0.11 2.37 0.00 3.56 0.46 0.13 0.01 0.06 0.00 0.01 0.18

68.6 861 4805 2649 17074 40.7 8326 1.82 479 151 144 965 4.73 172 2.16 78.0 7.53 42.1 13.4 5.76 0.30 0.46 83.5 0.34 0.11 14.0 3.45 2.6

0.54 204 478 1094 1981 2.69 879 0.19 146 47.2 57.3 83.5 0.44 39.1 0.14 22.3 1.01 8.88 1.46 2.86 0.18 0.13 45.9 0.15 0.01 1.70 1.79 0.2

69.1 756 2655 5282 10266 21.9 3910 0.61 633 28.4 31.2 1170 9.80 2.29 0.98 64.0 6.40 10.7 1.51 0.65 0.93 10.3 5.47 0.11 0.06 1.21 0.63 0.8

0.70 55.5 159 262 537 4.71 75.9 0.19 21.0 1.14 6.99 51.0 0.66 0.28 0.00 11.3 1.14 1.49 0.35 0.06 0.85 6.81 1.18 0.00 0.03 1.00 0.41 0.2

71.2 181 964 1108 4682 20.8 1497 0.97 505 949 18.2 622 1.56 177 b.d.l. 41.0 5.86 4.90 2.36 0.73 0.12 0.22 6.42 b.d.l. 0.02 3.66 0.63

0.32 12.7 111 194 218 2.93 73.30 0.12 53.1 372 5.30 5.41 0.08 32.6

68.6 1144 1771 6184 14199 23.6 6086 0.8 852 64.0 70.3 1009 16.0 16.3 12.0 87.7 7.01 19.0 0.14 0.60 0.20 0.22 12.7 0.09 0.05 1.36 1.68 -0.4

0.76 144 456 542 682 3.8 521 0.1 87.5 2.8 28.7 47.8 2.89 1.37 2.87 6.20 0.75 2.42 0.04 0.08 0.05 0.02 2.75 0.00 0.01 0.30 1.06 0.2

71.6 24.0 122 287 1530 20.5 211 b.d.l. 36.7 49.2 21.0 536 4.71 20.0 0.58 38.6 6.16 0.36 0.05 b.d.l. 0.03 b.d.l. 0.53 b.d.l. 0.02 0.17 0.57

0.22 10.0 62.3 68.7 153 0.8 122

71.0 308 493 1979 5133 20.4 2132 b.d.l. 134 39.0 36.7 688 17.6 8.6 22.9 85.5 4.40 4.67 b.d.l. 0.14 0.02 0.00 1.53 0.05 b.d.l. 0.05 11.9 3.7

0.34 129 187 752 1623 7.11 1044

70.0 718 2836 7205 7706 23.0 3314 0.41 61.5 50.8 109 706 22.7 12.5 3.99 61.2 5.20 18.0 0.10 0.41 0.07 0.27 7.94 0.12 0.03 0.47 1.87

0.25 27.9 261 304 270 4.30 185 0.14 36.5 6.54 53.5 11.0 7.11 2.93 2.78 12.6 0.77 1.42 0.02 0.07 0.00 0.05 1.56 0.00 0.00 0.08 0.72

0.00 0.00 0.09 0.2

4.82 0.68 1.72 0.81 0.31 0.04 0.00 4.16 0.00 1.22 0.63

7.3 13.6 4.5 31.0 2.51 2.00 0.00 5.53 0.73 0.13 0.01 0.00 0.07 0.00 0.00 0.11

41.1 10.1 14.7 107 2.19 3.12 94.0 14.5 0.92 1.85 0.13 0.00 0.81 0.00 0.15 18.2 0.2

Fe-contents are in wt%, δ18O in ‰, all other element contents in ppm. N = number of analysed grains; n = number of analyses; b.d.l. = below detection limit.

The magnetites from massive- and disseminated ore samples range from −0.08‰ to +0.07‰ and −0.46‰ to −0.15‰, respectively, thus almost all Chagangnuoer magnetites plot in the field of hydrothermal ores with δ56Fe b 0.01‰ (Anbar, 2004; Dauphas and Rouxel, 2006; Dziony et al., 2014; Moeller et al., 2014). In contrast, the δ56Fe ratios of the different Zhibo ore types span a range between about −0.1‰ and + 0.4‰ (~±0.08, 2 sd). Accordingly the majority of the magnetites plot in the field of ortho-magmatic magnetite with δ56Fe N 0.04‰ (Heimann et al., 2008; Weis, 2013). The δ56Fe values of the flow-, massive- and complex breccia ores range from + 0.01‰ to + 0.15‰, −0.09‰ to 0.10‰ and −0.02‰ to 0.22‰, respectively. The δ56Fe ratios of the massive ore sample with the columnar-network texture are about +0.36 ± 0.04‰, representing the heaviest Fe-isotope signatures of the investigated samples. The magnetite δ56Fe variations are between about 0.1 and 0.2‰ within the same sample. Spatially resolved interior and rim analyses of single magnetite grains revealed variations of about 0.02 to 0.05‰ in the flow and massive ores (see Table 2). 4.2.3. Isotopic source modelling Using the Fe-isotope data and the oxygen isotope ratios (Table 1) of the same samples and ore types (Zhang et al., 2015), we recalculated the corresponding δ56Fe and δ18O values of possible magnetite-precipitating sources, according to the approach of Nystroem et al. (2008), Jonsson et al. (2013) and Weis (2013). To determine the δ-ratio of a source fluid or melt, which was in equilibrium with magnetite at the time of precipitation, the degree of fractionation expressed as Δ-notation needs to be calculated. It is defined as:

Δmagnetite ðMtÞ−source ¼ δMt −δsource ≈ 1000 ln αMt−source ðfor δ ≤10‰Þ

αmagnetite ðMtÞ−source ¼

ð1000 þ δMt Þ ðfor δ ≥10‰Þ ð1000 þ δsource Þ

Consequently, the δ-notation of any source can be approximated by: δsource ¼ δMt −ΔMt−source ðfor δ ≤10‰Þ δsource ¼

ð1000 þ δMt Þ −1000 ðfor δ≥10‰Þ αMt−source

The modelling was conducted for magnetite (1) precipitating from a hydrothermal fluid between 200 °C and 800 °C and (2) an intermediate to felsic melt between 650 °C and 950 °C (based on the liquid immiscibility model of Zhang et al., 2015). The high temperature end of the fluid modelling is representative of an origin of ortho-magmatic magnetite comparable to IOCG and Kiruna-like iron ores (Barton, 2014; Nystroem et al., 2008). Temperatures b400 °C cover the range of magnetite precipitated from moderate- to low- temperature, hydrothermal fluids. For the fluid calculation, experimental fractionation factors between magnetite and water (Cole et al., 2004; Frierdich et al., 2014) were used for the O- and Fe-isotopes (Fig. 8A). Further modelling was adapted to calculate the Fe-isotope Δ-values (Heimann et al., 2008) for the system (Fig. 8B). This second calculation is thought to repremagnetite-FeCl2− 4 sent an equivalent to naturally occurring magmatic-hydrothermal fluids derived from igneous rocks (Heimann et al., 2008). Appropriate Fe-isotope fractionation factors for magnetite and melt are unknown to date. Only preliminary experimental- and calculated Δ-values exist and result in a large range of fractionation factors from

118


Fig. 4. Discrimination diagrams for magnetite from Zhibo and Chagangnuoer to identify magnetite from different deposits (after Dupuis and Beaudoin (2011)). Filled symbols are averaged magnetite values (±1 sd) analysed by LA-ICP-MS. Open symbols are the same averaged grains (±1 sd) analysed by EPMA. V and Cr contents, mesured with EPMA, are mostly below the detection limit (see Supplementary material) and thus restrict the comparison between both analytical methods to four iron ores from Zhibo (A). (A) Ti + V versus Ca + Al + Mn and (B) Ni/(Cr + Mn) classification diagrams for magnetite. (C) Al/(Zn + Ca) versus Cu/(Si + Ca) for magnetite from VMS deposits. Red field = melt-originated magnetite, Blue field = orthomagmatic magnetite, grey field = hydrothermal magnetite, B = BIF = banded iron formation, I = IOCG = iron oxide-copper-gold deposits, K = Kiruna = Kiruna type magnetite-apatite deposits, P = Porphyry = porphyry Cu deposits, Skarn = Fe-Cu skarn deposits, Fe-Ti, V = magmatic originated Fe-Ti-V-oxide deposits, VMS = volcanogenic massive sulphide deposits.

−0.12 (Schuessler et al., 2006) to +0.07 (Heimann et al., 2008) at temperatures between 700 °C and 1000 °C. Thus, we conducted the melt modelling for O- and Fe-isotopes using experimental fractionation factors between magnetite and olivine at magmatic temperatures (Chiba et al., 1989; Shahar et al., 2008; Fig. 8). This approximation, based on the assumption that olivine is in equilibrium with melt, reflects most likely the primary isotopic composition of the melt (cf. Fe-isotope values for olivine and modelled melt before fractional crystallization (Fig. 4; Teng et al., 2008). For comparison with natural magnetite-precipitating systems, we added the hypothetical melt sources of the different oretypes at 900 °C that were calculated with O- and Fe-isotope fractionation factors between magnetite and an andesitic melt (Fig. 8C; ΔMtmelt = − 2.88‰ for oxygen (Zhao and Zheng, 2003) and ΔMt-melt = 0.03 (Heimann et al., 2008)). The Δ-values for the ortho-magmatic fluid and the magnetite-melt were calculated for the same magmatic system investigated by Heimann et al. (2008). An equilibrium crystallization of the different ore types from a hypothetical, low-temperature hydrothermal fluid (Fig. 8A, B) is generally not in agreement with our source modelling (Beard et al., 2003b; Severmann et al., 2004; Sharma et al., 2001; Weis, 2013). The re-calculated isotopic source values for the different iron ores from Zhibo and Chagangnuoer rather correspond with the isotopic signature of an ortho-magmatic environment at high temperatures (cf. Fig. 8A & B; Heimann et al., 2008; Nystroem et al., 2008). In comparison to an

complexes as iron carriers the comortho-magmatic fluid with FeCl2− 4 plex breccia and flow-type ores also correspond to a hydrothermal origin at temperatures of around 700 °C. Furthermore, an origin from an intermediate melt is not in accordance with the respective isotopic source calculations for the Chagangnuoer and Zhibo iron ores regardless of the approach, i.e. either using experimental or calculated Δ-values. The δ18O ratios of the complex breccia, flow-type and massive ores from Chagangnuoer and Zhibo are far too low to support an origin from an intermediate melt (cf. Fig. 8C; Beard and Johnson, 2004; Bindeman, 2008; Heimann et al., 2008; Taylor, 1968). This is in agreement with the low δ56Fe signatures of the Chagangnuoer disseminated ore (Fig. 8C). 5. Discussion Several different metallogenetic models have been proposed for the Chagangnuoer and Zhibo magnetite mineralization (see above). The contact metasomatic Fe-skarn model (Hong et al., 2012b, 2012c; Sun et al., 2015) is supported for the Chagangnuoer deposit due to the occurrence of a paragenetic mineralization sequence of typical Ca-rich skarn minerals like grossular-andradite-rich garnet, diopside, actinolite and epidote associated with marble in the immediate vicinity of one of the main iron ore bodies (Fig. 2A). In contrast, a volcanic-hosted iron mineralisation occurs at Zhibo. Several observations that have been


119

skarns (Ashley and Plimer, 1989; Karimpour and Shafaroudi, 2005; Liu et al., 2012; Sharpe and Gemmell, 2002; Stanton, 1987). The occurrence of andraditic garnet, partly enclosed by a magnetite groundmass (Fig. 3H), is also in accordance with a hydrothermal origin of this ore type. The immediate country rock and the iron ore of Zhibo and Chagangnuoer have typical Fe-skarn related silicate assemblages such as andradite-grossular-rich garnet, diopside, epidote and actinolite (cf. Meinert, 1992; Meinert et al., 2005). At Zhibo, the flow-type ore and columnar-network ore-type have been interpreted to display primary magmatic melt textures (see above). The dendritic-like, platy to columnar iron ore forms a magnetite network at Zhibo (Fig. 3D). Similar growth textures have been associated with a magmatic origin either by a liquid immiscibility between Fe- and Si-rich melts (Philpotts, 1967; Philpotts and Doyle, 1983) or a supersaturation of a degassing oxide melt (Henriquez and Martin, 1978). The network-like growth textures reflect precipitation during rapidly decreasing temperatures (Rhodes et al., 1999; Sillitoe and Burrows, 2002). However, the dendritic-like and platy magnetite crystal shapes at Zhibo are similar to the magnetite textures reported from the Kiruna iron ore deposits, Sweden (Nystroem and Henriquez, 1994) and El Laco, Chile (Henriquez and Martin, 1978), the origin of which, although highly debated, has been associated with primary magmatic and/or magmatic-hydrothermal processes (e.g. Bilenker et al., 2016; Knipping et al., 2015; Tornos et al., 2016; Velasco et al., 2016).

Fig. 5. Multi-element variation diagrams for magnetite (LA-ICP-MS data) from Zhibo (A) and Chagangnuoer (B) normalized to magnetite from Fe-Ti-V-oxide deposits, which originated from a primitive melt (Dare et al., 2014), in comparison with magmatic magnetite from Fe-Ti-P-V-oxide deposits (evolved melt) and andesite. (C) Average Zhibo and Chagangnuoer magnetite compared to hydrothermal magnetite from high temperature (N500 °C) environments (blueish field = Iron Oxide-Copper-Gold (IOCG) and porphyry Cu deposits) and iron skarn. Data for different magnetite bearing deposits from Dare et al. (2014a; 2014b). Mt = magnetite.

reported so far, such as sharp contacts between the iron ore bodies and the volcanic host rocks, the disseminated occurrence of magnetite in the latter, magmatic hydraulic fracturing as indicated by brecciated magnetite, the occurrence of massive ore bodies with magmatic textures and the dominantly positive oxygen isotope signature of the iron ores are in accordance with an origin from a primary magmatic Fe-oxide melt (Jiang et al., 2014; Zhang et al., 2015). 5.1. Deposit classification – magmatic versus hydrothermal origin 5.1.1. Textural interpretation As already pointed out above, the texturally different magnetite ores at the AIMB volcanic edifice are dominated by brecciated and massive ore types with flow textures, platy magnetite-building network textures or compact, almost pure magnetite ore lenses. Especially, the flow-type and massive ore (with platy network textures) are of particular interest regarding the origin. The stratiform magnetite Chagangnuoer ore bodies Nos. 1 and 2 are parallel to the layering of their intermediate volcanic host rocks (Sun et al., 2015; Zhang et al., 2015). The interlayering of compositionally different silicate-rich rocks and Fe-oxide-rich layers (Fig. 3F1, G) is a common characteristic of many submarine exhalative deposits and stratiform

5.1.2. Trace element systematics In the different discrimination diagrams (Figs. 4, 5), the magnetite composition of all ore types from Zhibo and Chagangnuoer differs significantly from intermediate melt originated magnetite (Dare et al., 2014) and the “classic” magmatic (Fe-Ti-P) magnetite deposits (Dupuis and Beaudoin, 2011). This is due to their unusually low Ti, Zn, Cr and Sc contents as well as partly low Co, V, Ni and Al concentrations, in contrast to the magnetite that originated from a Fe-rich melt (Lester et al., 2013). In fact, the trace element patterns (Fig. 5) show a strong affinity to ortho-magmatic magnetite ore from IOCG deposits and hydrothermal skarns. Physio-chemical conditions such as temperature, oxygen and sulphur fugacites, and host rock buffering of the magnetite-precipitating environment influence this partitioning behaviour (Lester et al., 2013; Nadoll et al., 2014 and references therein). The majority of the massive ore from Zhibo including the platy network magnetite and the complex breccia ore plot in classification fields of IOCG, Kiruna and Porphyry deposits (Fig. 4). This broad range of trace element ratios are not interpreted to be the result of contrasting origins of the different ore types, since their spatial distribution is restricted to a few tens to hundreds of meters within the same ore body (Jiang et al., 2014; Zhang et al., 2015). Therefore, we conclude that the trace element variations of the individual ore types either reflect changes of the physico-chemical conditions of their primary precipitation environment or a secondary re-equilibration by later hydrothermal fluids culminating in a possibly complete dissolution and later re-precipitation of magnetite (see Discussion below). Such changes in the magnetite composition explain the different enrichment degrees of high-field strength elements (e.g. Y, Zr, Nb) and transition metals (e.g. Co, V, Ni), resulting in common trend-like patterns between the different ore types of both deposits (Figs. 4 & 5A). In summary, textural and trace element signatures disagree with an exclusively magmatic origin for magnetite from Zhibo and Chagangnuoer (cf. Zhang et al., 2015) and rather support a magmatichydrothermal formation model, which implies similarities to skarn and IOCG related mechanisms of iron ore enrichment. 5.1.3. Isotopic character of the ore deposits Oxygen isotope analyses of the magnetites from the different iron ore types (massive, breccia, flow-type) mainly reveal values between + 0.9 and + 4.0‰ (Hong et al., 2012c; Zhang et al., 2015), which are consistent with a precipitation from a magma or high-temperature

120


Table 2 Fe-isotope compositions of ablated magnetites. Sample

Loc.

δ56Fe (±2 sd)

δ57Fe (±2 sd)

11ZB05-14, massive ore - columnar network mag 1 - interior mag 1 - rim mag 2 - interior mag 2 - rim mag 3 - interior mag 3 - rim mag 4 - interior mag 4 - rim mag 5 - interior mag 6 - interior mag 6 - rim

Zb Zb Zb Zb Zb Zb Zb Zb Zb Zb Zb

0.40 0.39 0.32 0.32 0.33 0.34 0.34 0.35 0.40 0.36 0.33

0.50 0.52 0.41 0.56 0.44 0.47 0.52 0.47 0.59 0.49 0.44

10ZB01-08, massive ore mag 1 - interior mag 1 - rim mag 2 - interior mag 2 - rim mag 3 - interior mag 3 - rim mag 4 - interior mag 4 - rim mag 5 - interior mag 5 - rim mag 6 - interior

Zb Zb Zb Zb Zb Zb Zb Zb Zb Zb Zb

−0.09 ± 0.06 −0.04 ± 0.07 −0.02 ± 0.07 0.05 ± 0.06 −0.05 ± 0.07 −0.03 ± 0.06 −0.04 ± 0.06 0.02 ± 0.07 0.03 ± 0.07 0.10 ± 0.06 0.27 ± 0.08

−0.10 ± 0.12 −0.08 ± 0.13 −0.08 ± 0.14 0.08 ± 0.12 −0.04 ± 0.14 −0.02 ± 0.12 −0.06 ± 0.11 0.05 ± 0.12 0.02 ± 0.12 0.18 ± 0.11 0.40 ± 0.14

11ZB01-09, flow type ore mag 1 - interior mag 2 - interior mag 2 - rim mag 3 - interior mag 4 - interior mag 4 - rim mag 5 - interior mag 5 - rim

Zb Zb Zb Zb Zb Zb Zb Zb

0.07 0.15 0.14 0.01 0.09 0.11 0.04 0.05

0.15 0.15 0.15 0.06 0.10 0.16 0.05 0.00

11ZB04-04, complex breccia ore mag 1 - interior mag 1 - rim mag 2 - interior mag 2 - rim mag 3 - interior mag 3 - rim mag 4 - interior mag 4 - rim mag 5 - interior mag 5 - rim

Zb Zb Zb Zb Zb Zb Zb Zb Zb Zb

0.18 ± 0.06 0.18 ± 0.06 0.22 ± 0.06 0.20 ± 0.06 0.19 ± 0.06 0.21 ± 0.06 0.14 ± 0.07 -0.02 ± 0.07 0.16 ± 0.06 0.16 ± 0.07

0.29 ± 0.10 0.33 ± 0.10 0.32 ± 0.10 0.33 ± 0.10 0.24 ± 0.10 0.32 ± 0.10 0.22 ± 0.11 -0.06 ± 0.12 0.29 ± 0.11 0.21 ± 0.12

10CG16-13, massive ore mag 1 - interior mag 1 - rim mag 2 - interior mag 2 - rim mag 3 - interior mag 4 - interior mag 5 - interior mag 5 - rim mag 6 - interior mag 6 - rim mag 7 - interior

Ch Ch Ch Ch Ch Ch Ch Ch Ch Ch Ch

0.07 ± 0.05 0.09 ± 0.05 −0.04 ± 0.05 −0.08 ± 0.05 −0.07 ± 0.05 −0.08 ± 0.06 −0.06 ± 0.05 −0.08 ± 0.05 −0.07 ± 0.05 −0.09 ± 0.05 −0.02 ± 0.05

0.13 ± 0.10 0.13 ± 0.10 −0.07 ± 0.09 −0.11 ± 0.08 −0.10 ± 0.10 −0.09 ± 0.09 −0.02 ± 0.09 −0.15 ± 0.10 −0.08 ± 0.09 −0.10 ± 0.09 −0.03 ± 0.10

10CG18-07, disseminated ore mag 1 mag 2 mag 3 mag 4 mag 5 mag 6 mag 7 mag 8

Ch Ch Ch Ch Ch Ch Ch Ch

−0.23 −0.21 −0.28 −0.26 −0.15 −0.20 −0.26 −0.31

± ± ± ± ± ± ± ±

0.06 0.05 0.06 0.06 0.05 0.06 0.06 0.05

−0.37 −0.33 −0.44 −0.44 −0.19 −0.31 −0.42 −0.52

± ± ± ± ± ± ± ±

0.10 0.10 0.11 0.11 0.10 0.09 0.10 0.09

11CG08-02, disseminated ore mag 1 - interior mag 1 - rim mag 2 - interior mag 2 - rim mag 3 - interior

Ch Ch Ch Ch Ch

−0.35 −0.32 −0.33 −0.34 −0.46

± ± ± ± ±

0.06 0.06 0.06 0.06 0.06

−0.55 −0.49 −0.40 −0.48 −0.61

± ± ± ± ±

0.11 0.10 0.10 0.11 0.10

± ± ± ± ± ± ± ± ± ± ±

± ± ± ± ± ± ± ±

0.07 0.06 0.08 0.08 0.08 0.07 0.06 0.07 0.07 0.06 0.06

0.07 0.06 0.08 0.07 0.07 0.06 0.07 0.07

± ± ± ± ± ± ± ± ± ± ±

± ± ± ± ± ± ± ±

0.13 0.13 0.14 0.14 0.16 0.13 0.12 0.13 0.13 0.12 0.12

0.13 0.13 0.14 0.14 0.13 0.11 0.14 0.12


121

Table 2 (continued) Sample

Loc.

δ56Fe (±2 sd)

δ57Fe (±2 sd)

mag 4 - interior mag 5 - interior mag 5 - rim mag 6 - interior mag 6 - rim

Ch Ch Ch Ch Ch

−0.28 −0.38 −0.34 −0.41 −0.40

−0.41 −0.61 −0.47 −0.67 −0.58

± ± ± ± ±

0.06 0.06 0.05 0.06 0.06

± ± ± ± ±

0.10 0.10 0.10 0.12 0.11

δ56Fe and δ57Fe values with ± 2 standard deviation. Core and rim notation represents the relative point location on an analysed magnetite grain. mag = magnetite; Zb = Zhibo; Ch = Chagangnuoer.

magmatic fluids (~800–1000 °C; Jonsson et al., 2013). In contrast, magnetites with δ18O values b+0.9‰ are thought to have originated from low-temperature hydrothermal processes (≤ 400 °C; Jonsson et al., 2013). Thus, the δ18O values between +0.9 and +4.0‰ are interpreted to reflect an associated magma source from which the iron mineralization originated due to processes of liquid immiscibility, phase separation and/or fractionation (Zhang et al., 2015). The Fe-isotope compositional range of the massive Chagangnuoer and Zhibo ores, including the flow-type and complex breccia ore (Fig. 7), is identical with that of melt-originated magnetite (δ56Fe = 0.06 to 0.49‰) of the Bushveld and Skaergaard layered intrusions (Bilenker et al., 2014; Dziony et al., 2014) and high-temperature, ortho-magmatic ore deposits of the IOCG clan from the Pea Ridge, USA (Childress et al., 2016), the Kiruna and Grängesberg districts, Sweden (Weis, 2013) and the Chilean Iron Belt (including El Laco, Bilenker et al., 2016). In addition the geological setting of the Pea Ridge, Kiruna and El Laco deposits is similar to that of the AIMB deposits. All iron ores originated in a subduction-related setting and are hosted by intermediate to felsic volcanic rocks (cf. Childress et al., 2016; Jonsson et al., 2013; Velasco et al., 2016), thereby suggesting genetically similar magmatic conditions for the Chagangnuoer and Zhibo massive ores. However, the predominantly negative δ56Fe isotopic composition (δ56Fe = − 0.46 to + 0.07‰) of the Chagangnuoer magnetite ores is more consistent with a precipitation from an isotopically lighter, hydrothermal fluid (− 0.5‰ to 0‰; Sharma et al., 2001) comparable to Fe-skarns (Fig. 7). The latter usually originate in a tectonic convergent setting and are associated with granitic and granodioritic intrusions (Meinert et al., 2005), which also occur in close proximity of the Zhibo and Chagangnuoer deposits (Figs. 1 & 2). The δ56Fe values of the magnetites, which are interpreted to represent skarn iron ore due to their trace element behaviour (Fig. 4), are identical to those of Fe-skarn magnetites from the Dannemora (−0.43 to +0.01‰; Weis (2013) and Xinqiao deposits (− 0.37 to + 0.13‰; Wang et al., 2011b). However, as shown above the trace element contents and Fe-isotope signatures of the Zhibo ortho-magmatic magnetite are in accordance with those of IOCG deposits. Thus, the Fe-isotope signatures exactly reflect the trace element classification patterns of the different ore-types of both deposits (cf. Figs. 4 & 5). In summary, the positive and negative δ56Fe

Fig. 6. Iron isotope compositions of magnetite. Magnetite δ56Fe ratios of different oretypes from Zhibo (orange symbols) and Chagangnuoer (blue symbols). Error bars represent 2 sd standard error.

values of our investigated iron ores consistently cover the range of ortho-magmatic magnetite and low-temperature hydrothermal originated magnetite ore. On the basis of the Fe- and O- isotope data, the Zhibo massive ore types are interpreted to have preferentially originated from an orthomagmatic process in equilibrium with an FeCl24 − complex-bearing agent at temperatures ~ 800 °C (c.f. orange box in Fig. 8). This agrees with the high-temperature originated network-like magnetite textures and the trace element characteristics (Figs. 4 & 5). The trace element contents and Fe-isotope signatures of magnetite of the massive ore at Chagangnuoer are indistinguishable from those of Zhibo (Figs. 4–9), except for their low oxygen isotope ratio. Hence, we conclude that local disequilibria caused by an alternating hydrothermal fluid lowered the primary δ18O signature of the magnetite. The different re-calculated sources for the disseminated ores at Chagangnuoer are not in accordance with those suggested in the δ56Fe vs. δ18O diagrams (Fig. 8), suggesting a different ore formation process. Summarizing, the different Chagangnuoer and Zhibo ore types form two genetically different groups: Ortho-magmatic iron ores (Group I), the magnetites of which are relatively enriched in typical magmatic trace elements such as V, Ti and Ni and the heavy Fe- (from + 0.4 to ~0.0‰) and O-isotopes. This group encompasses all massive ore types. Group II is exclusively represented by the Chagangnuoer disseminated magnetite ores, which are depleted in magmatic trace elements and have significantly lower Fe-isotope values (b0.1‰) than the Group I ores and are thus in accordance with a hydrothermal origin of Group II magnetite. (See Table 4.) 5.2. Evolution of the magnetite-precipitating agent(s) 5.2.1. The ortho-magmatic ores (Group I) In order to obtain further insight into the evolution of the orthomagmatic iron ores, the different major and trace element compositions of the diverse ore-types were interpreted in conjunction with their Feisotope signatures and magnetite oxygen isotope composition (Zhang et al., 2015). Magnetites with the heavy (ortho-magmatic) δ56Fe values (N0.04‰) show a decrease in Fe(total) (from about 68 wt% to 72 wt%) with decreasing δ56Fe ratios (Fig. 9A). Simultaneously, the incompatible trace elements (e.g. Si, Ca, Al and high-field strength elements (HFSE)) positively correlate with decreasing δ56Fe values and decreasing Fe(total), suggesting Rayleigh fractionation during the magnetite precipitation. This is in accordance with the large size of the massive ore bodies (e.g. one main ore-body with 600 m × 400 m × 110 m) and a widespread distribution of several ore bodies at the volcanic edifice (see Figs. 2B & 4A in Zhang et al., 2015), indicating that not all magnetites had precipitated under the same physico-chemical conditions. When considering the above results, an early-originated magnetite would have the primary, chemical signature of the ore-forming agent with high Fe- and low incompatible trace element contents, and, in addition, an ortho-magmatic, isotopic character (high δ18O and δ56Fe ratios). The more magnetite crystallizes from the primary fluid the more enriched the residual agent becomes in incompatible trace elements, which is also reflected by the later precipitated magnetite. However, more compatible trace elements such as V and Ni, although positively correlating with decreasing δ56Fe among the massive magnetite ores, display a reverse trend

122


Table 3 Ore type classification and magnetite textural characteristics. Ore district

Ore type

Sample no.

Main magnetite textures

Zhibo Zhibo Zhibo Zhibo Zhibo Chagangnuoer Chagangnuoer Chagangnuoer Chagangnuoer

Complex breccia (4) Massive (3) Massive (3) Massive (3) Flow (2) Massive (3) Disseminated Disseminated Disseminated

11ZB04-04 11ZB05-14 10ZB02-07 10ZB01-08 11ZB01-09 10CG16-13 10CG18-07 11CG08-02 11CG02-03

Massive aggregates overprinted by a hydrothermal mineral assemblage eu- to subhedral, platy-shaped columnar network Dense mass Dense mass Irregular, dense and mingled masses/aggregates Dense masses/aggregates Fine-grained, layered, patchy aggregates Irregularly dense and curved layers/striations, euhedral crystals Granular, matrix-forming, dense aggregates

Numbers (2-4) in brackets refer to general textural classification of the texturally different ore types.

compared to the above mentioned more incompatible trace elements (i.e. a decrease with decreasing δ56Fe values and decreasing Fe; Fig. 9). In general, the incorporation of specific trace elements into Fe-Ti-oxides is strongly linked to the prevailing oxidation state of the precipitating agent (e.g. Ghiorso and Sack, 1991; Toplis and Carroll, 1995). The V enrichment in magnetite is highly related to low oxygen fugacity conditions (Bordage et al., 2011; Toplis and Corgne, 2002). The same applies to Ni and Mo (not shown), which are restricted to very low oxygen and sulphur fugacity conditions (Nadoll et al., 2014). Low fO2 conditions are also in accordance with the negative correlation of even more compatible trace elements such as Ti, Cr and Co with decreasing δ56Fe (Figs. 5 & 9). Titanium preferentially incorporates into magnetite with a 4 + charge by a coupled substitution with a divalent cation (Wechsler et al., 1984) and, thus, even higher oxidizing conditions are needed to enrich Ti and most of the other trace elements in magnetite in contrast to V and Ni. The Rayleigh fractionation of the trace elements is accompanied by a fractionation of the Fe-isotopes as displayed in the trace elements vs. δ56Fe diagrams (Fig. 9). Thereby, early-precipitated magnetite has the heaviest Fe-isotope signature due to an early incorporation of 56Fe caused by the relatively higher Fe3+/Fe2+ (2:1) ratio. In general, the heavier isotope is associated with a stronger bond, which is a function of bonding length, and is favoured by a high oxidation-state and a lower coordination site of the element species (e.g. Bigeleisen and Mayer, 1947; Polyakov et al., 2007; Schauble, 2004; Schauble et al., 2009). Ferric iron meets these requirements with its higher ionic potential and a tetrahedral coordination compared to the octahedral coordinated Fe2 +. Thus, the uneven ratio of ferric and ferrous iron in magnetite results in a fractionation of the heavier Fe-isotope and the residual fluid will be enriched in the lighter 54Fe by the massive magnetite precipitation, whereby the stable isotope fractionation increases with decreasing temperatures (e.g. Bigeleisen and Mayer, 1947; Urey, 1947). Thereby, a positive or negative Fe-isotopic signal in Fe-oxides strongly depends on the coordination of Fe in the transporting complexing agent and the bonding partner itself (cf. Hill and Schauble, 2008; Hill et al., 2009; Roe et al., 2003; Skulan et al., 2002). Experimental studies and theoretical predictions show that the heavy Fe-isotopes are

preferentially incorporated from a melt or a hydrothermal fluid into magnetite at a temperatures of ca. 800 °C (Bilenker, 2015; Heimann et al., 2008), which is in accordance with our isotopic source modelling for our ortho-magmatic magnetite samples. In contrast, the oxygen isotope composition of magnetite is lighter than that of a fractionating fluid or a melt (cf. Jonsson et al., 2013; Taylor, 1967, 1968). Consequently, magnetite principally prefers to incorporate the lighter 16O at the time of precipitation and thus, a Rayleigh-type fractionation of a large amount of magnetite would lead to an enrichment of the heavy 18O in the residual liquid, unless equilibrium conditions can be accomplished. Late-precipitated magnetite would then have higher δ18O values than early-crystallized magnetite. The oxygen isotope range of the fractionated, ortho-magmatic ores with δ56Fe ratios between − 0.1‰ and + 0.1‰ reaches values of up to ~ 2.6‰, which are relatively elevated compared to those of the early-precipitated magnetite with δ18O of about 1.0‰ (Fig. 9). However, in accordance with the source modelling the Chagangnuoer massive ore shows lower δ18O signatures (−0.4‰) suggesting local re-equilibration with a low to moderate temperature hydrothermal fluid (Zhang et al., 2015). The fluid source of the ore lies in the corresponding field for hydrothermal fluids at temperatures of around 400 °C (Fig. 8B). However, a hydrothermal alteration overprint cannot have been solely responsible for the lower δ56Fe ratios in some of the massive magnetite ores since the Fe-isotope signature of the magnetite seems relatively unaffected by such a process. For instance, hydrothermally altered magnetite ore from El Laco shows lowered oxygen but an ortho-magmatic, unaffected Fe-isotope composition (Bilenker et al., 2016). Host rock buffering is another possibility to control compositional variations in magnetite by interaction of the ore-forming fluid with surrounding wall rocks (Nadoll et al., 2014). Leaching of unfractionated Fe from silicate rocks (δ56FeIRMM14 = + 0.09‰; Beard et al., 2003a) would produce fluids that are slightly depleted in the heavy Fe-isotopes (Markl et al., 2006). Elevated W, Pb, Mo and Sn concentrations in magnetite from the Neo Dala banded iron formation were attributed to metasomatic processes by fluid/host-rock interaction with a granitic host rock (Bhattacharya et al., 2007). Depending on the host rock composition, specific trace elements are modified in metasomatic

Table 4 Chemical classification of the different ore-types based on trace element and isotope characteristics of magnetite. Ore district

Ore type

Sample no.

Group

Classificationa

Classificationb

δ56Fe

δ18O

Zhibo Zhibo Zhibo Zhibo Zhibo Chagangnuoer Chagangnuoer Chagangnuoer Chagangnuoer

Complex breccia Massive Massive Massive Flow Massive Disseminated Disseminated Disseminated

11ZB04-04 11ZB05-14 10ZB02-07 10ZB01-08 11ZB01-09 10CG16-13 10CG18-07 11CG08-02 11CG02-03

I I I I I I II II II

IOCG/Skarn Porphyry IOCG Skarn Skarn Skarn Skarn Skarn Skarn

Or. -mag. (N 500 °C) Or. -mag. (N 500 °C) Or. -mag. (N 500 °C) Or. -mag. (N 500 °C) Or. -mag. (N 500 °C) Or. -mag. (N 500 °C) Or. -mag. (N 500 °C) Or. -mag. (N 500 °C) Or. -mag. (N 500 °C)

Ortho-magmatic Ortho-magmatic Ortho-magmatic/Hydrothermal Ortho-magmatic/Hydrothermal Ortho-magmatic/Hydrothermal Hydrothermal Hydrothermal -

Ortho-magmatic Ortho-magmatic Ortho-magmatic Hydrothermal Ortho-magmatic -

Parameters to distinguish magnetite, based on trace element and isotope composition (see text for details). Minus sign (−) indicates no available data. Or. –mag. = Ortho-magmatic. a After Dupuis and Beaudoin (2011). b After Dare et al. (2014a)


123

Fig. 7. Magnetite δ56Fe values from the Zhibo (orange circles) and Chagangnuoer (blue circles) deposits compared to magnetite of different origins. The grey boxes include the range of the δ56Fe values with analytical standard error (2 sd). The red and blue box show the range of ortho-magmatic (+0.04‰ to +0.80‰ (Heimann et al., 2008; Bilenker et al., 2016)) and hydrothermal magnetites (−0.70‰ to +0.20‰ (Anbar, 2004; Schueßler, 2008; Severmann and Anbar, 2009)). Several magnetite-bearing deposit types are plotted, which encompass magnetite from layered intrusions (from melt), IOCG deposits (from melt or magmatic-hydrothermal fluid) as well as skarn and submarine originated magnetite (precipitated by the influence of sea- or meteoric water). Most δ56Fe values of different ore-types from Zhibo are in accordance with the Fe-isotope range of ortho-magmatic magnetite. Samples from Chagangnuoer preferentially show values in euqilibrium with a hydrothermal fluid.

Fig. 8. Isotopic composition of magnetite-forming magmas and fluids at Zhibo (Zb) and Chagangnuoer (Ch) compared to other sources (colored boxes) at temperatures between 300 °C and 950 °C. δ56Fe values measured in this study, δ18O ratios from Zhang et al. (2015). (A) Hydrothermal fluid calculations (dark grey symbols) with experimental fractionation factors for the system magnetite-water. (B) Source modelling for the system magnetite-FeCl2− 4 (bright grey symbols). This modelling is based on calculated fraction factors for O- and Fe-isotopes to represent an ortho-magmatic environment. (C) Hypothetical, mafic melt compositions modelled with experimental fraction factors of the system magnetite-olivine (red symbols) and a calculated D-value between magnetite and an intermediate melt at 900 °C (yellow symbols). The analytical standard error (2 sd) for O- and Fe isotope data are added to the diagrams. References are cited in the text for isotopic ranges of the coloured boxes and different fractionation factors of the modelling methods. Mt = magnetite; Ol = olivine.

124


Fig. 9. Variation diagrams (A–F) with δ56Fe versus Fe (EMPA), V, Ni, Ti, Nb and Al (LA-ICP-MS) of averaged magnetite grains (±1 sd) of the different ore-types from Zhibo (Zb, orange symbols) and Chagangnuoer (Ch, blue symbols). Trace elements in magnetite define the Groups I and II with distinct compositions, whose relation is best explained by a bimodal formation model (arrows). In the Fe vs δ56Fe plot (A) the oxygen isotope values of the related samples (green values) are shown (δ18O-data from Zhang et al. (2015)).

magnetite, e.g. high Ni for a (ultra-) mafic and high W for a felsic host rock (Angerer et al., 2012; Angerer et al., 2013). However, almost all incompatible trace elements are enriched in the ortho-magmatic magnetite with δ56Fe around 0.00‰. Thus, we conclude that a Rayleigh-type fractionation is responsible for the increasing trace element contents at decreasing Fe-isotopic signatures of the Group I magnetite. Although elevated W, Mo and Sn contents in some ore types from Zhibo may be best explained by local buffering effects with the felsic host rocks (Fig. 5). 5.2.2. The disseminated ores (Group II) The disseminated magnetite of Group II differs from the Group I magnetite due to the occurrence as a matrix-forming mineral (Fig. 3F1–H2), the skarn-like trace element composition (Fig. 4) and the low Fe-isotope values (Figs. 7 & 8). Adapting the above argumentation, a Rayleigh-type fractionation process is the simplest way to explain the low δ56Fe values of the disseminated ores at Chagangnuoer. The heavy oxygen isotope signature (+1.8 to +3.7‰) substantiates this interpretation. Furthermore assuming that a stable isotope fractionation had taken place, the incompatible trace elements should display a similar behaviour. Yet, the Fe-content of the disseminated ores is similar to that of the early-fractionated, massive ores and the incompatible trace

element contents are comparatively low (Figs. 5 & 9). Thus, several bimodal formation models seem to better explain the origin of the disseminated ores: (model 1) Phase separation between vapour and liquid: A kinetically controlled Fe-isotope fractionation depends on the involved phases. For instance, a phase separation of vapour and brine does not necessarily include Fe-isotope fractionation as shown for the Brandon vent at the East Pacific Rise (Beard et al., 2003a). In addition, the oxygen isotope values of the disseminated ores (up to +3.7‰) seem to be too heavy for a vapour-liquid phase separation origin. Degassing of a magmatic system would increase the residual magma or ortho-magmatic fluid in heavier 18 O, because the lower dissociation energy of molecules with the lighter 16 O results in an enrichment of the latter in the escaping vapour phase (e.g. Hoefs, 2009; White, 2015). Thus, vapour-liquid separation alone does not explain the Fe- and O-isotope signatures. (model 2) Leaching/dissolution of the primary, ortho-magmatic ores: The disseminated magnetite ores are thought to present secondary iron ore that was re-deposited by later hydrothermal fluids. Abiotic leaching and dissolution experiments of hematite and goethite (both incorporate only Fe3+) were conducted with HCl and solutions with organic ligands, without revealing a significant Fe-isotope fractionation (Beard et al., 2003a; Brantley et al., 2004; Skulan et al., 2002). However,


the partial dissolution of hornblende, which incorporates ferric and ferrous iron (comparable with magnetite), released, in contrast to only Fe3 + bearing Fe-oxy-hydroxides, fluids with isotopic lighter aqueous Fe (Brantley et al., 2004). Subsequent to the partial iron dissolution, Cl-bearing aqueous fluids have the potential to transport Fe2 + as complexing agent with a preferential incorporation of the lighter Fe-isotopes (Bilenker, 2015; Bilenker et al., 2013; Heimann et al., 2008; Hill and Schauble, 2008). Thus, a Cl-bearing, hydrothermal fluid is able to dissolve and transport iron out of a primary, massive magnetite ore body and precipitate it in form of a isotopically lighter, disseminated iron ore in the immediate vicinity. Such a bimodal formation for instance was suggested for the two generations of the Kiruna-type iron ore (Valley et al., 2011). The low Fe-isotope ratios of the secondary ore from Mineville confirm this hypothesis (Bilenker et al., 2016). Magnetite ores of Group I from Chagangnuoer and Zhibo show crystals with corroded rims, porous surfaces and vesicles (e.g. Figures 3A3, 3C2, 3F2), indicating a partial dissolution of magnetite. Vein-like and partly pervasive alteration occurring in the complex breccia ores suggests an intensive interaction between later hydrothermal fluids and the massive ores. Furthermore, the concentration of several trace elements decreases towards the rims of the magnetite from some massive ores of Group I (see Supplementary material). These element changes indicate a chemical change in the composition of the primary agent or a re-equilibration with a secondary fluid. At Chagangnuoer, the disseminated magnetite, which is thought to have precipitated from a secondary, remobilizing fluid, shows a high iron content similar to that of the Group I magnetite with heavy δ56Fe ratios, which indicates a relatively rapid crystallization with no large influence of a Rayleigh fractionation. The heavy oxygen isotope signature of the disseminated ore (up to +3.7‰) is not in accordance with any source of our modelling calculations (Fig. 8). We suggest that either a magmatic fluid or a hypersaline, connate brine with a high δ18O was responsible for the leaching of the primary ortho-magmatic ore. (Model 3) Replenishment of the magnetite-precipitating system with additional Fe: A replenishment of the Fe-precipitating magma/ fluid of the high Fe- and low incompatible trace element contents of the disseminated ores. In the latter case, the replenishing agent had a different isotopic character regarding the very low δ56Fe and high δ18O signature (assuming a primary oxygen isotope signal) of the disseminated ores. The difference to model (2) is that the iron for the Group II magnetite was provided by the replenishing fluid instead of by the dissolution of Group I magnetite. 5.2.3. Metallogenetic model The processes that lead to such massive iron accumulation in a convergent arc setting with intermediate to felsic magmatism are highly controversial. For instance, Tornos et al. (2016) described the iron mineralization at El Laco (Chile) due to the separation of an iron-rich melt from a parental andesitic magma. On the other hand, Knipping et al. (2015) postulated that a separated high-salinity fluid scavenged Fe in the form of FeCl2 complexes from intermediate melts to form the Kiruna-type deposit of Los Colorados (Chile). Both examples show that ortho-magmatic melts and fluids are important for the Fe accumulation in a volcanic hosted arc setting. We propose that an ortho-magmatic fluid -separated from an intermediate melt- can best explain the Group I magnetite mineralization, supporting the model of Jiang et al. (2014) and Zhang et al. (2015). According to those authors, the iron mineralization at Chagangnuoer and Zhibo is spatially related to the origin of the volcanic host rocks and was interpreted to have resulted from an exsolution between Fe- and Si-rich liquids. It is thought that the differentiation of alkali-calcic magmas (occurring in the AIMB), has the potential to promote such immiscibility processes (Charlier and Grove, 2012; Veksler et al., 2015). Thereby, the high δ56 Fe values (Figs. 6 & 7) of Group I as well as the columnar-network and flowtextures (Fig. 3) in some of these ores support a magnetite origin at

125

high magmatic temperatures. This is in accordance with the trace element composition of Group I magnetites (e.g. low HFSE, see above) and the Fe-isotope source modelling (Fig. 8). Subsequent to the precipitation of the Group I iron ores at Zhibo and Chagangnuoer a hydrothermal process led to the formation of the disseminated Group II ores. Late stage hydrothermal fluids from nearby felsic intrusions (Figs. 1C & 2A) are in accordance with the formation of skarn ores and the a Ca-Na-K alteration assemblage. The low δ 56 Fe values of the disseminated magnetite support such an origin. At low magmatic or sub-solidus temperatures, an exsolution of a hydrothermal fluid from a silica-rich melt generates a kinetic fractionation of the Fe-isotopes so that a hydrothermal fluid preferentially removes the lighter Fe-isotopes in form of Fe2 + complexes (Heimann et al., 2008; Poitrasson and Freydier, 2005). Ferrous iron is more mobile than Fe3 + and preferentially incorporates the lighter Fe-isotopes, such as 54Fe (Anbar, 2004; Beard and Johnson, 2004; Dauphas and Rouxel, 2006). Thus magnetite that originated from a hydrothermal fluid can principally possess relatively low δ56Fe ratios depending on the prevailing aqueous Fe specie(s) and the initial isotopic composition of the source. 6. Conclusions Chemical characteristics of the investigated Chagangnuoer and Zhibo iron ore districts reveal the appearance of two distinctive iron ore groups. The ortho-magmatic magnetite mineralization of Group I occurs as flow-type, massive and brecciated ores forming lenticular ore-bodies, partly with network-like textures of columnar-platy crystals. Chemical signatures are similar to Kiruna- and IOCG-like iron ores with ortho-magmatic trace element signatures. This group occurs in both mining districts. Igneous δ56Fe ratios (up to 0.4‰) are in the ortho-magmatic magnetite range comparable to layered intrusion and iron ore deposits of the IOCG clan, including the Pea Ridge, Kiruna and El Laco deposits. At Chagangnuoer, the second magnetite group appears in form of disseminated iron ores with hydrothermal, skarn-like trace element patterns and δ56Fe signatures b 0.01‰. Calculations using Fe- and O-isotopes of magnetite infer a high-temperature ortho-magmatic origin (~800 °C) of the Group I iron ores. The lower Fe-isotope ratios of the disseminated ores disagree with an origin at such high temperatures. Thus, we conclude a bimodal formation model combining both groups. Positive correlations between decreasing Fetotal, δ56Fe (from + 0.4‰ to − 0.1‰), increasing incompatible (e.g. Si, Al, Nb, Y, Ti) and decreasing compatible trace element contents (e.g. V, Ni) are in agreement with a Raleigh-type fractionation process within the Group I iron ores of the large, massive magnetite ore-bodies. A reverse trace element trend among the Group II iron ores at further decreasing δ56Fe ratios (from − 0.1‰ to − 0.5‰) is thought to be the consequence of a replenishing system and/or Fe-leaching from the ortho-magmatic ore-bodies and re-deposition in form of disseminated magnetite in the immediate vicinity. This process was driven by latestage (Ca-Na-K-rich) hydrothermal fluids, which originated from deeper-seated granitic/granodioritic intrusions in the immediate vicinity. Acknowledgements We acknowledge Christian Abe for his support during work with the electron microprobe and Helene Brätz for help during Laser Ablation ICP-MS analyses. We are most grateful for the constructive and insightful reviews of Kirsten van Zuilen, and an anonymous referee, which clearly helped to improve the manuscript. Furthermore, the manuscript benefitted from the comments of Patrick Nadoll on previous versions of the manuscript. Klaus Mezger is thanked for his comments and editorial help. We also want to thank Jun Gao for his continuous logistical support and discussions on the manuscript.

126


Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.chemgeo.2017.02.001.

References Anbar, A.D., 2004. Iron stable isotopes: beyond biosignatures. Earth Planet. Sci. Lett. 217 (3), 223–236. Angerer, T., Hagemann, S.G., Danyushevsky, L.V., 2012. Geochemical evolution of the banded iron formation-hosted high-grade iron ore system in the Koolyanobbing Greenstone Belt, Western Australia. Econ. Geol. 107 (4), 599–644. Angerer, T., Hagemann, S.G., Danyushevsky, L.V., 2013. High-grade iron ore at Windarling, Yilgarn Craton: a product of syn-orogenic deformation, hypogene hydrothermal alteration and supergene modification in an Archean BIF-basalt lithostratigraphy. Mineral. Deposita 48 (6), 697–728. Ashley, P.M., Plimer, I.R., 1989. “Stratiform skarns”—a re-evaluation of three eastern Australian deposits. Mineral. Deposita 24 (4), 289–298. Barton, M., 2014. Iron Oxide (–Cu–Au–REE–P–Ag–U–Co) Systems. Treatise on Geochemistry. second ed. 13. Elsevier, Amsterdam, pp. 515–541. Beard, B.L., Johnson, C.M., 2004. Fe isotope variations in the modern and ancient earth and other planetary bodies. Rev. Mineral. Geochem. 55 (1), 319–357. Beard, B.L., et al., 2003a. Application of Fe isotopes to tracing the geochemical and biological cycling of Fe. Chem. Geol. 195 (1), 87–117. Beard, B.L., Johnson, C.M., Von Damm, K.L., Poulson, R.L., 2003b. Iron isotope constraints on Fe cycling and mass balance in oxygenated Earth oceans. Geology 31 (7), 629–632. Bhattacharya, H.N., Chakraborty, I., Ghosh, K.K., 2007. Geochemistry of some banded ironformations of the Archean supracrustals, Jharkhand-Orissa region, India. J. Earth Syst. Sci. 116 (3), 245–259. Bigeleisen, J., Mayer, M.G., 1947. Calculation of equilibrium constants for isotopic exchange reactions. J. Chem. Phys. 15 (5), 261–267. Bilenker, L.D., 2015. Elucidating Igneous and Ore-forming Processes With Iron Isotopes by Using Experimental and Field-based Methods. (Ph.D. thesis). Rutgers University. Bilenker, L.D., Simon, A., Lundstrom, C.C., Gajos, N., Zajacz, Z., 2013. Experimental Constraints on Fe Isotope Fractionation in Fluid-melt-oxide-sulfide Assemblages. Goldschmidt Conference. Bilenker, L.D., VanTongeren, J.A., Lundstrom, C., Simon, A.C., 2014. Fe Isotope Systematics of the Upper and Upper Main Zones of the Bushveld Complex. South Africa, AGU Fall Meeting Abstracts, p. 4797. Bilenker, L.D., et al., 2016. Fe–O stable isotope pairs elucidate a high-temperature origin of Chilean iron oxide-apatite deposits. Geochim. Cosmochim. Acta 177, 94–104. Bindeman, I., 2008. Oxygen isotopes in mantle and crustal magmas as revealed by single crystal analysis. Rev. Mineral. Geochem. 69 (1), 445–478. Bordage, A., et al., 2011. V oxidation state in Fe–Ti oxides by high-energy resolution fluorescence-detected X-ray absorption spectroscopy. Phys. Chem. Miner. 38 (6), 449–458. Brantley, S.L., et al., 2004. Fe isotopic fractionation during mineral dissolution with and without bacteria. Geochim. Cosmochim. Acta 68 (15), 3189–3204. Charlier, B., Grove, T.L., 2012. Experiments on liquid immiscibility along tholeiitic liquid lines of descent. Contrib. Mineral. Petrol. 164 (1), 27–44. Charvet, J., et al., 2011. Palaeozoic tectonic evolution of the Tianshan belt, NW China. Sci. China Earth Sci. 54 (2), 166–184. Chen, Y., Hu, A., Zhang, G., Zhang, Q., 1999. Zircon U-Pb age and Nd-Sr isotopic composition of granitic gneiss and its geological implications from Precambrian window of western Tianshan, NW China [J]. Geochimica 6, 000. Chiba, H., Chacko, T., Clayton, R.N., Goldsmith, J.R., 1989. Oxygen isotope fractionations involving diopside, forsterite, magnetite, and calcite: application to geothermometry. Geochim. Cosmochim. Acta 53 (11), 2985–2995. Childress, T.M., Simon, A.C., Day, W.C., Lundstrom, C.C., Bindeman, I.N., 2016. Iron and oxygen isotope signatures of the Pea Ridge and Pilot Knob magnetite-apatite deposits, southeast Missouri, USA. Econ. Geol. 111 (8), 2033–2044. Cole, D.R., et al., 2004. An experimental and theoretical determination of oxygen isotope fractionation in the system magnetite-H2O from 300 to 800C. Geochim. Cosmochim. Acta 68 (17), 3569–3585. Dare, S.A., et al., 2014. Trace elements in magnetite as petrogenetic indicators. Mineral. Deposita 49 (7), 785–796. Dauphas, N., Rouxel, O., 2006. Mass spectrometry and natural variations of iron isotopes. Mass Spectrom. Rev. 25 (4), 515–550. Dong, L.H., et al., 2011. Discussion of metallogenic models, mineralization characteristic and main type of rich iron ore of Xinjiang. Xinjiang Geol. 29 (4), 416–422. Dupuis, C., Beaudoin, G., 2011. Discriminant diagrams for iron oxide trace element fingerprinting of mineral deposit types. Mineral. Deposita 46 (4), 319–335. Dziony, W., et al., 2014. In-situ Fe isotope ratio determination in Fe–Ti oxides and sulfides from drilled gabbros and basalt from the IODP Hole 1256D in the eastern equatorial Pacific. Chem. Geol. 363, 101–113. Feng, J., et al., 2010. The Syngenetic Volcanogenic Iron Ore Deposits in Awulale Metallogenetic Belt, Western Tianshan Mountains. Geological Publishing House, Beijing. Frierdich, A.J., Beard, B.L., Scherer, M.M., Johnson, C.M., 2014. Determination of the Fe (II) aq–magnetite equilibrium iron isotope fractionation factor using the three-isotope method and a multi-direction approach to equilibrium. Earth Planet. Sci. Lett. 391, 77–86. Gao, J., Li, M., Xiao, X., Tang, Y., He, G., 1998. Paleozoic tectonic evolution of the Tianshan Orogen, northwestern China. Tectonophysics 287 (1), 213–231.

Gao, J., et al., 2009. Tectonic evolution of the South Tianshan orogen and adjacent regions, NW China: geochemical and age constraints of granitoid rocks. Int. J. Earth Sci. 98 (6), 1221–1238. Gao, J., et al., 2015. Record of assembly and breakup of Rodinia in the Southwestern Altaids: Evidence from Neoproterozoic magmatism in the Chinese Western Tianshan Orogen. J. Asian Earth Sci. 113 (1), 173–193. Ghiorso, M.S., Sack, R.O., 1991. Thermochemistry of the oxide minerals. Rev. Mineral. Geochem. 25 (1), 221–264. Han, B.F., Guo, Z.J., Zhang, Z.C., Zhang, L., Chen, J.F., Song, B., 2010. Age, geochemistry, and tectonic implications of a late Paleozoic stitching pluton in the North Tian Shan suture zone, western China. Geol. Soc. Am. Bull. 122 (3-4), 627–640. Heimann, A., Beard, B.L., Johnson, C.M., 2008. The role of volatile exsolution and sub-solidus fluid/rock interactions in producing high 56Fe/54Fe ratios in siliceous igneous rocks. Geochim. Cosmochim. Acta 72 (17), 4379–4396. Henriquez, F., Martin, R.F., 1978. Crystal-growth textures in magnetite flows and feeder dykes, El Laco, Chile. Can. Mineral. 16 (4), 581–589. Hill, P.S., Schauble, E.A., 2008. Modeling the effects of bond environment on equilibrium iron isotope fractionation in ferric aquo-chloro complexes. Geochim. Cosmochim. Acta 72 (8), 1939–1958. Hill, P.S., Schauble, E.A., Shahar, A., Tonui, E., Young, E.D., 2009. Experimental studies of equilibrium iron isotope fractionation in ferric aquo–chloro complexes. Geochim. Cosmochim. Acta 73 (8), 2366–2381. Hoefs, J., 2009. Stable Isotope Geochemistry. Springer-Verlag, Berlin Heidelberg New York. Hong, W., et al., 2012a. Mineralogy of the Chagangnuoer iron deposit in Western Tianshan Mountains, Xinjiang, and its geological significance. Acta Petrol. Mineral. 31 (2), 191–211. Hong, W., Zhang, Z.H., Jiang, Z.S., Li, F.M., Liu, X.Z., 2012b. Magnetite and garnet trace element characteristics from the Chagangnuoer iron deposit in the western Tianshan Mountains, Xinjiang, NW China: constrain for ore genesis. Acta Petrol. Sin. 28 (7), 2089–2102. Hong, W., Zhang, Z.H., Li, F.M., Liu, X.Z., 2012c. Stable isotopic characteristics of the Chagangnuoer iron deposit in Western Tianshan, Xinjiang and its geological significance. Rock Miner. Anal. 6, 031. Horn, I., Von Blanckenburg, F., 2007. Investigation on elemental and isotopic fractionation during 196 nm femtosecond laser ablation multiple collector inductively coupled plasma mass spectrometry. Spectrochim. Acta Part B 62 (4), 410–422. Horn, I., von Blanckenburg, F., Schoenberg, R., Steinhoefel, G., Markl, G., 2006. In situ iron isotope ratio determination using UV-femtosecond laser ablation with application to hydrothermal ore formation processes. Geochim. Cosmochim. Acta 70 (14), 3677–3688. Hu, A.Q., et al., 2010. Formation of the 0.9 Ga Neoproterozoic granitoids in the Tianshan Orogen, NW China: constraints from the SHRIMP zircon age determination and its tectonic significance. Geochimica 39 (3), 197–212. Jahn, B.M., 2004. The Central Asian Orogenic Belt and growth of the continental crust in the Phanerozoic. Geol. Soc. Lond. Spec. Publ. 226 (1), 73–100. Jiang, Z., et al., 2012a. Geochemistry and zircon U-Pb age of volcanic rocks from the Chagangnuoer and Zhibo iron deposits, western Tianshan, and their geological significance. Acta Petrol. Sin. 28 (7), 2074–2088. Jiang, Z., Zhang, Z., Wang, Z., Li, F., Tian, J., 2012b. Alteration mineralogy, mineral chemistry and genesis of Zhibo iron deposit in Western Tianshan Mountains, Xinjiang. Mineral Deposits 31 (5), 1051–1066. Jiang, Z., et al., 2014. Geology, geochemistry, and geochronology of the Zhibo iron deposit in the Western Tianshan, NW China: constraints on metallogenesis and tectonic setting. Ore Geol. Rev. 57, 406–424. Jonsson, E., et al., 2013. Magmatic origin of giant ‘Kiruna-type’ apatite-iron-oxide ores in Central Sweden. Sci. Report. 3. Karimpour, M.H., Shafaroudi, A.M., 2005. Taknar polymetal (Cu-Zn-Au-Ag-Pb) deposit: A new type magnetite-rich VMS deposit, Northeast of Iran. J. Sci. Islam Repub. Iran 16 (3), 239–254. Knipping, J.L., et al., 2015. Giant Kiruna-type deposits form by efficient flotation of magmatic magnetite suspensions. Geology 43 (7), 591–594. Lester, G.W., Clark, A.H., Kyser, T.K., Naslund, H.R., 2013. Experiments on liquid immiscibility in silicate melts with H2O, P, S, F and Cl: implications for natural magmas. Contrib. Mineral. Petrol. 166 (1), 329–349. Li, J.L., Su, W., Zhang, X., Liu, X., 2009. Zircon Cameca U–Pb dating and its significance for granulite-facies gneisses from the western Awulale Mountain, West Tianshan, China. Geol. Bull. China 28 (12), 1852–1862. Liu, X.J., Liu, W., Liu, L.J., 2012. The generation of a stratiform skarn and volcanic exhalative Pb–Zn deposit (Sawusi) in the southern Chinese Altay Mountains: the constraints from petrography, mineral assemblage and chemistry. Gondwana Res. 22 (2), 597–614. Long, L., et al., 2011. Geochemical and geochronological studies of granitoid rocks from the Western Tianshan Orogen: implications for continental growth in the southwestern Central Asian Orogenic Belt. Lithos 126 (3), 321–340. Markl, G., Von Blanckenburg, F., Wagner, T., 2006. Iron isotope fractionation during hydrothermal ore deposition and alteration. Geochim. Cosmochim. Acta 70 (12), 3011–3030. Meinert, L.D., 1992. Skarns and skarn deposits. Geosci. Can. 19 (4). Meinert, L.D., Dipple, G.M., Nicolescu, S., 2005. World Skarn Deposits. Econ. Geol. 100th Anniversary Volume pp. 299–336. Moeller, K., et al., 2014. Comparison of iron isotope variations in modern and Ordovician siliceous Fe oxyhydroxide deposits. Geochim. Cosmochim. Acta 126, 422–440. Nadoll, P., Angerer, T., Mauk, J.L., French, D., Walshe, J., 2014. The chemistry of hydrothermal magnetite: a review. Ore Geol. Rev. 61, 1–32.

T. Günther et al. / Chemical Geology 453 (2017) 111–127 Nystroem, J.O., Henriquez, F., 1994. Magmatic features of iron ores of the Kiruna type in Chile and Sweden; ore textures and magnetite geochemistry. Econ. Geol. 89 (4), 820–839. Nystroem, J.O., Billström, K., Henríquez, F., Fallick, A.E., Naslund, H.R., 2008. Oxygen isotope composition of magnetite in iron ores of the Kiruna type in Chile and Sweden. GFF 130 (4), 177–188. Oeser, M., Weyer, S., Horn, I., Schuth, S., 2014. High-precision Fe and Mg isotope ratios of silicate reference glasses determined in situ by femtosecond LA-MC-ICP-MS and by solution nebulisation MC-ICP-MS. Geostand. Geoanal. Res. 38 (3), 311–328. Pearce, N.J., et al., 1997. A compilation of new and published major and trace element data for NIST SRM 610 and NIST SRM 612 glass reference materials. Geostand. Geoanal. Res. 21 (1), 115–144. Philpotts, A.R., 1967. Origin of certain iron-titanium oxide and apatite rocks. Econ. Geol. 62 (3), 303–315. Philpotts, A.R., Doyle, C.D., 1983. Effect of magma oxidation state on the extent of silicate liquid immiscibility in a tholeiitic basalt. Am. J. Sci. 283 (9), 967–986. Pirajno, F., 2013. The Geology and Tectonic Settings of China's Mineral Deposits. 688. Springer, New York. Pirajno, F., Seltmann, R., Yang, Y., 2011. A review of mineral systems and associated tectonic settings of northern Xinjiang, NW China. Geosci. Front. 2 (2), 157–185. Poitrasson, F., Freydier, R., 2005. Heavy iron isotope composition of granites determined by high resolution MC-ICP-MS. Chem. Geol. 222 (1), 132–147. Polyakov, V., Clayton, R., Horita, J., Mineev, S., 2007. Equilibrium iron isotope fractionation factors of minerals: reevaluation from the data of nuclear inelastic resonant X-ray scattering and Mössbauer spectroscopy. Geochim. Cosmochim. Acta 71 (15), 3833–3846. Qian, Q., Gao, J., Xiong, X.M., Long, L.L., Huang, D.Z., 2006. Petrogenesis and tectonic settings of Carboniferous volcanic rocks from north Zhaosu, western Tianshan Mountains: constraints from petrology and geochemistry. Acta Petrol. Sin. 22 (5), 1307–1323. Qian, Q., et al., 2009. Early Paleozoic tectonic evolution of the Chinese South Tianshan Orogen: constraints from SHRIMP zircon U–Pb geochronology and geochemistry of basaltic and dioritic rocks from Xiate, NW China. Int. J. Earth Sci. 98 (3), 551–569. Qin, K., et al., 2002. Paleozoic epithermal Au and porphyry Cu deposits in North Xinjiang, China: epochs, features, tectonic linkage and exploration significance. Resour. Geol. 52 (4), 291–300. Rhodes, A.L., Oreskes, N., Sheets, S., 1999. Geology and Rare Earth Element Geochemistry of Magnetite Deposits at El Laco, Chile. 7. Society of Economic Geology, Special Publication, pp. 299–332. Roe, J.E., Anbar, A.D., Barling, J., 2003. Nonbiological fractionation of Fe isotopes: evidence of an equilibrium isotope effect. Chem. Geol. 195 (1), 69–85. Schauble, E.A., 2004. Applying stable isotope fractionation theory to new systems. Rev. Mineral. Geochem. 55 (1), 65–111. Schauble, E.A., Méheut, M., Hill, P.S., 2009. Combining metal stable isotope fractionation theory with experiments. Elements 5 (6), 369–374. Schuessler, J., Schoenberg, R., Behrens, H., von Blanckenburg, F., 2006. Quantification of Iron Isotope Fractionation Between Sulphide and Oxide Minerals and Silicate Melts. AGU Fall Meeting Abstracts p. 0574. Schüßler, J.A., 2008. Controls on Stable Iron Isotope Variations in Magmatic Systems: Significance of Mineral-melt Isotopic Fractionation in Experiments and Nature (Doctoral dissertation). Severmann, S., et al., 2004. The effect of plume processes on the Fe isotope composition of hydrothermally derived Fe in the deep ocean as inferred from the Rainbow vent site, Mid-Atlantic Ridge, 36 14′ N. Earth Planet. Sci. Lett. 225 (1), 63–76. Severmann, S., Anbar, A.D., 2009. Reconstructing paleoredox conditions through a multitracer approach: the key to the past is the present. Elements 5 (6), 359–364. Shahar, A., Young, E.D., Manning, C.E., 2008. Equilibrium high-temperature Fe isotope fractionation between fayalite and magnetite: an experimental calibration. Earth Planet. Sci. Lett. 268 (3), 330–338. Sharma, M., Polizzotto, M., Anbar, A.D., 2001. Iron isotopes in hot springs along the Juan de Fuca Ridge. Earth Planet. Sci. Lett. 194 (1), 39–51. Sharpe, R., Gemmell, J.B., 2002. The Archean Cu-Zn magnetite-rich Gossan Hill volcanichosted massive sulfide deposit, Western Australia: genesis of a multistage hydrothermal system. Econ. Geol. 97 (3), 517–539. Sillitoe, R.H., Burrows, D.R., 2002. New field evidence bearing on the origin of the El Laco magnetite deposit, northern Chile. Econ. Geol. 97 (5), 1101–1109. Skulan, J.L., Beard, B.L., Johnson, C.M., 2002. Kinetic and equilibrium Fe isotope fractionation between aqueous Fe (III) and hematite. Geochim. Cosmochim. Acta 66 (17), 2995–3015. Stanton, R.L., 1987. Constitutional features, and some exploration implications, of three zinc-bearing stratiform skarns of eastern Australia. Trans. Inst. Min. Metall., Sect. B 96. Steinhoefel, G., Horn, I., von Blanckenburg, F., 2009. Matrix-independent Fe isotope ratio determination in silicates using UV femtosecond laser ablation. Chem. Geol. 268 (1), 67–73. Sun, W., et al., 2015. Petrogenesis of the Chagangnuoer deposit, NW China: a general model for submarine volcanic-hosted skarn iron deposits. Sci. Bull. 60 (3), 363–379. Taylor, H.P., 1967. Oxygen Isotope Studies of Hydrothermal Mineral Deposits. Geochemistry of Hydrothermal Ore Deposits. pp. 109–142. Taylor Jr., H.P., 1968. The oxygen isotope geochemistry of igneous rocks. Contrib. Mineral. Petrol. 19 (1), 1–71. Teng, F.-Z., Dauphas, N., Helz, R.T., 2008. Iron isotope fractionation during magmatic differentiation in Kilauea Iki lava lake. Science 320 (5883), 1620–1622.

127

Tian, J., et al., 2009. Metallogenetic Condition and Prospection Prediction Along Chagangnuor and Beizhan Iron Ore Deposits, Western Tianshan. West-China Exploration Engineering. pp. 88–91. Toplis, M.J., Carroll, M.R., 1995. An experimental study of the influence of oxygen fugacity on Fe-Ti oxide stability, phase relations, and mineral—melt equilibria in ferro-basaltic systems. J. Petrol. 36 (5), 1137–1170. Toplis, M.J., Corgne, A., 2002. An experimental study of element partitioning between magnetite, clinopyroxene and iron-bearing silicate liquids with particular emphasis on vanadium. Contrib. Mineral. Petrol. 144 (1), 22–37. Tornos, F., Velasco, F., Hanchar, J.M., 2016. Iron-rich melts, magmatic magnetite, and superheated hydrothermal systems: the El Laco deposit, Chile. Geology 44 (6), 427–430. Urey, H.C., 1947. The thermodynamic properties of isotopic substances. J. Chem. Soc. 562–581. Valley, P.M., Hanchar, J.M., Whitehouse, M.J., 2011. New insights on the evolution of the Lyon Mountain Granite and associated Kiruna-type magnetite-apatite deposits, Adirondack Mountains, New York State. Geosphere 7 (2), 357–389. Van Achterbergh, E., Ryan, C., Griffin, W., 2000. GLITTER (Version 3.0, On-line Interactive Data Reduction for LA-ICPMS). Maquarie Research Ltd. Veksler, I.V., et al., 2015. Electrochemical processes in a crystal mush: cyclic units in the upper critical zone of the Bushveld Complex, South Africa. J. Petrol. 56 (6), 1229–1250. Velasco, F., Tornos, F., Hanchar, J.M., 2016. Immiscible iron-and silica-rich melts and magnetite geochemistry at the El Laco volcano (northern Chile): evidence for a magmatic origin for the magnetite deposits. Ore Geol. Rev. 79, 346–366. Wang, Q., Lin, Z., Huang, C., Huang, Q., Zhen, J., 2001. Metallogenic series and prospecting target of ore deposits a Chagannur Region of the Western Tianshan. Xinjiang Geol. 4, 008. Wang, B., et al., 2006. Late Paleozoic tectonic evolution of the northern west Chinese Tianshan belt. Geodin. Acta 19 (3–4), 237–247. Wang, B., Shu, L.S., Cluzel, D., Faure, M., Charvet, J., 2007. Geochemical constraints on Carboniferous volcanic rocks of the Yili Block (Xinjiang, NW China): implication for the tectonic evolution of Western Tianshan. J. Asian Earth Sci. 29 (1), 148–159. Wang, B.Y., et al., 2011a. Geological characteristics and genesis of Chagannur iron deposit in Western Tianshan, Xinjiang. Mineral Deposits 30 (3), 385–402. Wang, Y., Zhu, X., Mao, J., Li, Z.-O., Cheng, Y., 2011b. Iron isotope fractionation during skarn-type metallogeny: a case study of Xinqiao Cu–S–Fe–Au deposit in the Middle–Lower Yangtze valley. Ore Geol. Rev. 43 (1), 194–202. Wang, B., et al., 2014a. Early Neoproterozoic crustal evolution in northern Yili Block: insights from migmatite, orthogneiss and leucogranite of the Wenquan metamorphic complex in the NW Chinese Tianshan. Precambrian Res. 242, 58–81. Wang, B., Zhang, B., Jing, D., 2014b. Geological characteristics and genesis of Chagangnuoer and Zhibo Volcanogenic Iron Deposit, in Western Tianshan, Xinjiang. Acta Geol. Sin. 88 (s2), 1532–1534. Wechsler, B.A., Lindsley, D.H., Prewitt, C.T., 1984. Crystal structure and cation distribution in titanomagnetites (Fe3 – xTi × O4). Am. Mineral. 69 (7–8), 754–770. Weis, F., 2013. Oxygen and Iron Isotope Systematics of the Grängesberg Mining District (GMD). Ph.D. dissertation, Uppsala, Sweden, Uppsala Universitet, Central Sweden (77 p). Weyer, S., Schwieters, J.B., 2003. High precision Fe isotope measurements with high mass resolution MC-ICPMS. Int. J. Mass Spectrom. 226 (3), 355–368. White, W.M., 2015. Isotope Geochemistry. 478. Wiley Blackwell. Windley, B.F., Alexeiev, D., Xiao, W., Kröner, A., Badarch, G., 2007. Tectonic models for accretion of the Central Asian Orogenic Belt. J. Geol. Soc. (London, U.K.) 164 (1), 31–47. Xiao, W., et al., 2008. Middle Cambrian to Permian subduction-related accretionary orogenesis of Northern Xinjiang, NW China: implications for the tectonic evolution of central Asia. J. Asian Earth Sci. 32 (2), 102–117. Zhai, W., et al., 2009. Geology, geochemistry, and genesis of Axi: a Paleozoic lowsulfidation type epithermal gold deposit in Xinjiang, China. Ore Geol. Rev. 36 (4), 265–281. Zhang, D., Zhang, Z., Xue, C., Zhao, Z., Liu, J., 2010. Geochronology and geochemistry of the ore-forming porphyries in the Lailisigao'er-Lamasu region of the Western Tianshan Mountains, Xinjiang, NW China: implications for petrogenesis, metallogenesis, and tectonic setting. J. Geol. 118 (5), 543–563. Zhang, X., et al., 2012. Geochronology and geochemistry of granitoid rocks from the Zhibo syngenetic volcanogenic iron ore deposit in the Western Tianshan Mountains (NWChina): Constraints on the age of mineralization and tectonic setting. Gondwana Res. 22 (2), 585–596. Zhang, X., et al., 2015. Metallogenesis of the Zhibo and Chagangnuoer Volcanic Iron Oxide Deposits in the Awulale Iron Metallogenic Belt, Western Tianshan Orogen. China, Journal of Asian Earth Sciences. Zhao, Z.-F., Zheng, Y.-F., 2003. Calculation of oxygen isotope fractionation in magmatic rocks. Chem. Geol. 193 (1), 59–80. Zhu, Y., Zhang, L., Gu, L., Guo, X., Zhou, J., 2005. The zircon SHRIMP chronology and trace element geochemistry of the Carboniferous volcanic rocks in western Tianshan Mountains. Chin. Sci. Bull. 50 (19), 2201–2212. Zhu, Y., Guo, X., Song, B., Zhang, L., Gu, L., 2009. Petrology, Sr–Nd–Hf isotopic geochemistry and zircon chronology of the Late Palaeozoic volcanic rocks in the southwestern Tianshan Mountains, Xinjiang, NW China. J. Geol. Soc. (London, U. K.) 166 (6), 1085–1099.