Introduction

Iron oxide-apatite and iron oxide-copper-gold (IOCG) deposits are an important group of economic iron oxide-rich deposits in the world. Some of the well-known districts containing such deposits are northern and central Sweden (Kiruna-type deposits), the Andean Coastal Batholith in Chile and Peru (an example of a paired barren iron oxide-apatite/productive IOCG belt), the El Laco Volcanic Complex in Chile, and the Bafq district, Central Iran, hosting the Chador-Malu deposit.

The Kiruna-type iron oxide-apatite and IOCG deposits may be considered to form the end-members of a continuum of deposit types within one broad class (Hitzman et al. 1992). A common feature that the two end-members share is intracontinental extensional tectonics (Hitzman 2000); however, some districts are associated with orogenic collapse (Duncan et al. 2014). There are also examples of IOCG mineralization associated with porphyry copper deposits in subduction zones (e.g., in the Yerington district, Nevada; see Groves et al. 2010). The iron oxide-apatite deposits of different age and geological settings show remarkably similar alteration and mineralization style, whereas the IOCG deposits are more diverse. While the two end-members may have fundamentally different origins (Hitzman 2000), the presence of both iron oxide-apatite and IOCG systems in the same mineralized belts can suggest their similar genesis, but with different local-scale processes and/or sources.

The genesis of Kiruna-type deposits has been subject to substantial controversy. There is generally close spatial association between Kiruna-type deposits and igneous rocks, the composition and depth of the latter ranging from mafic to felsic and intrusive to volcanic, respectively. There is, however, no consensus among researchers as to the principal source of the fluids possibly responsible for mineralization or the nature of the contribution of associated magmas (e.g., Parak 1975; Hildebrand 1986; Frietsch and Perdahl 1995; Barton and Johnson 1996, 2004; Broman et al. 1999; Sillitoe and Burrows 2002; Williams et al. 2005). Researchers put forward two different genetic models involving magmatic (e.g., Frietsch 1978, 1984; Nyström and Henriquez 1994; Naslund et al. 2002; Jonsson et al. 2013) and hydrothermal processes (e.g., Hitzman et al. 1992; Rhodes and Oreskes 1999; Hitzman 2000; Barton and Johnson 2000; Daliran 2002; Daliran et al. 2007; Smith et al. 2013).

Some researchers have proposed an IOCG affinity for the Bafq district deposits (see, e.g., Daliran 2002; Torab and Lehmann 2007; Torab 2008; Stosch et al. 2011), after the class was defined by Hitzman et al. (1992). Evidence for an IOCG association includes the presence of hydrothermal alteration with the characteristic zoning of the IOCG deposits (as defined by Hitzman et al. 1992) and magnetite chemistry (e.g., Torab 2008).

Jami (2005) developed a three-stage model for the hydrothermal evolution of the Esfordi deposit based on apatite textures, fluid inclusions, and isotopic data; however, the initial stages in the generation of the proto iron oxide-apatite body was largely speculative and involved the immiscible separation of a Fe-P melt (in response to a possible magma mingling in the Esfordi area) and its subsequent equilibration with fluids near 400 °C (see also Jami et al. 2007). From a regional perspective, Torab (2008) considered the evaporites of Central Iran as ligand source for metal leaching and transport (as described by Barton and Johnson 1996), and based on Nd isotopic data proposed that the rare earth elements (REE) and P of the Bafq district deposits were remobilized from phosphorites (which are not documented in the Early Cambrian sedimentary sequence of Central Iran, but are present in the stratigraphically equivalent formations outside the region) and were supplied, to a lesser degree, from the leaching of igneous rocks (see also Torab and Lehmann 2007). They, however, did not discuss the source of such large amounts of iron oxide in the Bafq district. A source for Fe proposed by some researchers has been the pre-existing mafic minerals (iron-bearing amphibole and pyroxene) of the Precambrian rocks of the region which would decompose in the early stages of regional-scale hydrothermal fluid circulation to supply Fe to the system (e.g., Haghipour 1974; Daliran et al. 2007, 2010).

Based principally on flow and quenching textures observed in magnetite, liquid magmatic models have been ascribed a prominent role at Kiruna and El Laco (see, e.g., Foose, and Mclelland 1995; Henriquez and Nyström 1998). This textural evidence was reported from Chador-Malu and certain other Bafq district deposits and interpreted as their direct magmatic origin by Förster and Jafarzadeh (1994). Magmatic models involving immiscibility between silicate and iron oxide-rich melts have widely been proposed for the genesis of Kiruna-type deposits (e.g., Philpotts 1967, 1982; Frietsch 1978, 1984; Naslund et al. 2000). Similar models have been developed for the Bafq district iron oxide-apatite deposits including Chador-Malu (e.g., Williams and Houshmand-Zadeh 1996; Förster and Borumandi 1971; Förster and Jafarzadeh 1984; Jami 2005). Interpreting magnetite lava flows and magnetite-bearing pyroclastic rocks at Chador-Malu, the latter researches supported an immiscible liquid magmatic model of formation, and suggested a carbonatitic affinity for the ore melt based on textural evidence such as xenomorphic calcite and dolomite in the grain boundaries of magnetite, and the absence of relicts of pre-existing carbonates within magnetite grains. However, the later work by Förster and Jafarzadeh (1994) considered the discordant magnetite bodies of the Bafq district, particularly the Chador-Malu deposit, the product of crystallization from an immiscible Fe ± P melt differentiated from an upper mantle carbonated melanephelinitic magma.

Bafq district deposits, of which Chador-Malu is typical, are remarkably similar to the Kiruna-type iron oxide-apatite deposits, considering their mineralogy and associated regional aprons of high temperature sodic-calcic alteration minerals. They consist of massive to brecciated, elongated, and conformable lenses of magnetite-hematite-apatite, which are mainly hosted in felsic volcanics and volcano-sedimentary rocks (e.g., NISCO 1980; Daliran 1990, 2002; Förster and Jafarzadeh 1994; Bonyadi et al. 2011), very similar to the classic Kiirunavaara deposit (northern Sweden), in terms of morphology, ore mineralogy, and host lithology; the latter ranging in composition from felsic to intermediate for the Swedish counterpart (e.g., Hitzman 2000). The volcanic hosts of both districts (Bafq and Kiruna) have undergone intense alteration and formed similar assemblages including actinolite and albite (e.g., Hitzman 2000; Jami 2005; Torab 2008; Bonyadi et al. 2011; Sabet-Mobarhan-Talab 2014).

There is discussion relating Esfordi and Se-Chahun iron oxide-apatite deposits (Bafq District) to the Rapitan-type banded iron-phosphorus formations based on the presence of keratophyres, jaspilites, diamictites, dropstones, and apatite geochemistry (e.g., Aftabi et al. 2009; Mohseni and Aftabi 2012). There have also been other works suggesting a carbonatitic affinity for apatites of the Bafq district deposits based on mineralogical and geochemical studies (e.g., Daliran 2002; Darvishzadeh 1983; Darvishzadeh and Aletaha-Kohbanani 1996).

Although the Chador-Malu deposit is the largest in the Bafq district (400 Mt @ 55 % Fe), it has received the least previous attention to its geochemical characteristics. This paper describes aspects of the geology and geochemistry of this iron oxide-apatite deposit and discusses a possible link between the Cambrian plutonic rocks and mineralization.

Geological setting

The Chador-Malu iron oxide-apatite deposit is located in the Bafq district, which is part of the Central Iranian microplate in the Alpine-Himalayan orogenic system. A series of intersecting regional-scale faults divide the Central Iranian Terrane into three major crustal domains, namely the Lut, Tabas, and Yazd Blocks, among which lies a more than 1,000-km long, and up to 80 km wide, arcuate and structurally complex belt, called the Kashmar-Kerman volcano-plutonic arc (Ramezani and Tucker 2003), which hosts the Bafq district and the Early Cambrian volcanic and plutonic rocks of Central Iran (Fig. 1).

Fig. 1
figure 1

Structural map of eastern Central Iran (modified after Ramezani and Tucker 2003; Torab 2008). The arcuate Kashmar-Kerman volcano-plutonic arc encompasses the Bafq district which covers the area between Bafq and Saghand and hosts the most important iron oxide-apatite deposits

Full size image

A model proposed for the Early Cambrian tectonics and magmatism in the Bafq district is crustal extension associated with intracontinental rifting. It has been proposed that this rifting event formed an aborted rift associated with alkaline magmatism, including alkaline volcanic rocks, alkaline granites, and carbonatites (e.g., Samani 1988, 1998; Daliran 1990, 1999, 2002; Förster and Jafarzadeh 1994). However, attributing the regional-scale arcuate faults of the Bafq district (Fig. 1) to a back-arc extensional setting, Ramezani and Tucker (2003) suggested that the Kashmar-Kerman volcano-plutonic arc represents an active continental-margin environment in relation to subduction under the Central Iranian microplate and closure of a Proto-Tethys ocean in the Early Cambrian.

The Early Cambrian sequence in Central Iran consists of a complex set of volcano-sedimentary rocks which extends from Poshte-Badam to Zarand and hosts the most important iron oxide-apatite, Pb-Zn, and U ore deposits in Iran (Samani 1988) (Fig. 1). The volcanic rocks vary from felsic to mafic in composition and include rhyolite, agglomeratic tuff, splitic lava, and diabase, but dominantly rhyolitic rocks, most of which having been subject to widespread alkaline metasomatism (K-feldsparization and albitization; e.g., Daliran et al. 2007). The sedimentary rocks consist of dolomites, dolomitic limestones, and evaporites (Torab 2008). Ramezani and Tucker (2003) assigned an Early Cambrian age (529–554 Ma) for the volcanic suites of this volcano-sedimentary unit and named this sequential unit as Cambrian Volcano Sedimentary Unit (CVSU).

Intruded into the CVSU, as well as into the Upper Precambrian metamorphic rocks, are plutonic bodies of mostly granitic composition (locally associated with diorite and gabbro-diorite dikes or sills), which are quite common in Central Iran, of which the Zarigan, Narigan, and the Chador-Malu granites are located in the Bafq district (Fig. 2). Based mainly on their high sodium contents, Berberian and Berberian (1981) described these granitic intrusions as alkali granite. The same was considered by later researchers (e.g., Samani 1988; Förster and Jafarzadeh 1994; Daliran 2002; Ghorbani 2013); however, Torab (2008) contends that these rocks are calc-alkaline intrusions affected by widespread sodic alteration (albitization), and suggests an evaporitic source for Na. Both the Zarigan and the Chador-Malu granites are characterized by being conspicuously lacking in mafic minerals (Torab 2008), the latter consisting of orthoclase (51 %), quartz (32 %), plagioclase (14.3 %), and accessory minerals such as zircon, garnet, and titanite (Berberian and Berberian 1981). Ramezani and Tucker (2003) defined an age of 525 ± 7 Ma and 529 ± 16 Ma (U-Pb on zircon) for the Zarigan and the Chador-Malu granites, respectively. Stosch et al. (2011) determined the U-Pb apatite age for major iron oxide-apatite deposits in the Bafq district (527–539 Ma). These ages fell entirely within the age range of the felsic plutonic rocks of the Bafq district (including those mentioned above) dated by Ramezani and Tucker (2003) (525–547 Ma; U-Pb on zircon).

Fig. 2
figure 2

Simplified geological map of the Bafq district (modified after Förster and Jafarzadeh 1994; Moghtaderi et al. 2007; Torab 2008). Note the spatial relationship between the Chador-Malu deposit and the Cambrian plutonic rocks (Chador-Malu granite)

Full size image

Chador-Malu is among the Bafq district iron oxide-apatite deposits which have been described as iron oxide melt-filled volcanic vents, a morphological feature interpreted as its magmatic origin (e.g., Jafarzadeh 1981; Förster and Jafarzadeh 1984, 1994). The oxygen isotopic composition (δ18O) of iron oxides from the Chador-Malu deposit (−0.64 to 2.19 ‰; Shamsi-Pour et al. 2008) is also consistent with magma as their direct or indirect source. The northern orebody (this study; Figs. 3 and 4a) has the shape of an upright cylinder with several horizontal extensions and is more than 600 m deep. The southern orebody is a flat-lying lens, interpreted as magnetite lava flow or sill. These two orebodies accompanied by two other magnetic anomalies in the Chador-Malu area are arranged in north-northeast-south-southwest direction along the eastern segment of the ring fracture zone shown in Fig. 2 (Förster and Jafarzadeh 1994).

Fig. 3
figure 3

Simplified geological map of the northern orebody of the Chador-Malu deposit displaying the distribution of ore samples (Modified from Jafarzadeh 1981; Moghtaderi et al. 2007; Torab 2008). The local coordinate system has been established by S.R.K. Consulting Co., and is used in the Chador-Malu mine. The entire map lies in the ring fracture zone shown in Fig. 2 (see the text)

Full size image
Fig. 4
figure 4

The Chador-Malu iron oxide-apatite deposit. a A panorama showing part of the northern orebody (view to the east). b Sharp contact between the massive ore (dominant mineralization style) and the metasomatic host rock

Full size image

According to field observations, the main host rock of the northern orebody is a highly altered actinolite-rich CVSU rhyolite, documented under the name “metasomatite” (Moghtaderi et al. 2007), and locally known as “Green Rock” (Jami 2005; Torab 2008), which displays sharp contacts with the massive orebody (see the next section). Comprising the main host rock of major iron oxide-apatite deposits in the Bafq district (see, e.g., Jami 2005; Torab 2008; Bonyadi et al. 2011), this metasomatic rock is widespread in the region and is thought to be the consequence of a multistage regional metasomatism (Daliran et al. 2007) in relation to the intrusion of the Cambrian plutonic rocks.

Ore types, mineralogy, and host rock alterations

Mineralization at Chador-Malu occurs predominantly within the actinolite-rich metasomatic rocks and rhyolites (Fig. 3), and comprises the northern pipe-like (this study) and the southern sill-like orebodies. The northern orebody, which is currently subject to extensive mining activity (Fig. 4a), is characterized chiefly by massive mineralization style. This orebody shows sharp and linear contacts with the host rocks (Fig. 4b). The contact between the diorites and either rhyolites or metasomatites (Fig. 3) is largely obscured by hydrothermal alteration and sharp contacts are restricted to structurally controlled zones. Breccias with iron oxide cement as well as small ore veins are rare. The iron ore veins are usually 2–3 cm thick, and can be observed at the contact between the massive ore and host rocks (Fig. 5a). The breccias consist of angular fragments of the metasomatic host rock cemented by iron ore (Fig. 5b)

Fig. 5
figure 5

Other styles of mineralization at Chador-Malu. a Small ore vein in the actinolite-rich host rock. b Angular brecciated fragments of the actinolite-rich country rocks cemented by iron oxides

Full size image

The mineralogy of the Chador-Malu deposit is fairly simple compared to the Bafq district deposits further south (Sabet-Mobarhan-Talab and Alinia 2014). The ore minerals consist predominantly of magnetite and apatite (Fig. 6a). Subordinate amounts of pyrite may also be observed (Fig. 6b). Martitization is common in magnetite ores, particularly close to the contacts (Fig. 6c). Primary hematite, however, is less abundant in the ores (Fig. 6d). At least two generations of apatite can be identified in the Chador-Malu deposit: primary coarse-grained subhedral to euhedral (ApI) and secondary fine-grained anhedral (ApII) apatites (Fig. 6e, f). Although rare earth minerals such as allanite, bastnaesite, and monazite have a relatively important presence in some Bafq district deposits (e.g., Esfordi; Jami 2005), they were merely observed as inclusions (predominantly monazite; see Fig. 11) in the present study, and never as distinct rare earth mineralization. Likewise, the gangue minerals (predominantly actinolite) have a paragenetic importance in certain Bafq district deposits and can be present in the main stage of mineralization (e.g., Esfordi; Jami 2005); however, they were never observed in the Chador-Malu ores. Thus the Chador-Malu ores can simply be divided into two types: magnetite ± hematite ore (iron ore) and magnetite ± hematite + apatite ore (iron-apatite ore) (Fig. 7).

Fig. 6
figure 6

Photomicrographs of the Chador-Malu ores. a Magnetite-apatite ore (reflected light). b Pentagonal pyrite in magnetite-hematite ore (reflected light). c Martitized magnetite ore. Note the preserved crystal form of magnetite in secondary hematite (reflected light). d Tabular texture in primary hematite (reflected light). e Iron-apatite ore with primary subhedral to euhedral apatite (ApI) in thin section (XPL). f Brecciated primary apatite (ApI) in a matrix of secondary fine-grained anhedral apatite (ApII) in thin section (XPL) (Mt: magnetite; Hm: hematite; Mr: martite; Py: pyrite; Ap: apatite)

Full size image
Fig. 7
figure 7

The types of ore at Chador-Malu. a, b Hand specimens of rich magnetite (a; photograph from Torab 2008) and iron-apatite (b) ores. c, d Iron-apatite ore with aggregates of apatite (c) and with individual subhedral to euhedral apatite crystals (d) (the apatite crystals in the latter are up to 8 cm long)

Full size image

The actinolite-rich metasomatites comprise the main host rock to the northern orebody of the Chador-Malu deposit (Fig. 3). These rocks are the product of intense actinolitization that replaced most minerals in the pre-alteration protolith (Fig. 8a, b), giving the host rocks a distinctive green appearance (Figs. 4b and 5). Sodic (albitization) and potassic (K-feldsparization and sericitization) alterations can also be observed and are related to the earlier stages of regional metasomatism in the Bafq district (see also Daliran et al. 2007; Fig. 8c, d). Quartz represents the late stage of hydrothermal alteration in the host rock (see also Daliran et al. 2007; Fig. 8e); however, at least part of the quartz (and other minerals) in the groundmass is relict magmatic (Fig. 8a, b, e). The host rocks may have locally undergone Fe-metasomatism which occurred subsequent to the multistage alteration of the host rock, resulting in the growth of magnetite grains on the alteration quartz (late stage) and other alteration minerals (Fig. 8).

Fig. 8
figure 8

Photomicrographs of the metasomatic host rock in thin section (XPL). a, b Actinolitization in the host metasomatite. Note the magnetite grains in b. c K-feldspar and albite in the host metasomatite representing the potassic and sodic alterations, respectively. Quartz represents the late stage of alteration. Note the magnetite growth on actinolite and quartz, and the actinolite crystal on albite, which demonstrate the order of crystallization. d Sericitization in the host metasomatite. Note the magnetite grains on sericitized K-feldspar. e Quartz growth on actinolite crystal occurred during the late stage of alteration (Act: actinolite; Qtz: quartz; Mt: magnetite; K-feld: potassium feldspar; Alb: albite)

Full size image

Samples, analytical methods, and chemical compositions

Thirty-one ore samples, as well as eight magnetite and apatite mineral separates were collected from mined outcrops of the northern orebody of the Chador-Malu iron oxide-apatite deposit. The ore samples included iron ore (magnetite ± hematite) and iron-apatite ore (magnetite ± hematite + apatite). The magnetite samples were separated manually by means of magnet from plattner mortar-crushed rich magnetite ore specimens after removing contaminating fragments under a binocular microscope. The apatite samples were extracted from individual large subhedral to euhedral apatite crystals, as well as from patches of apatite in the ore (Fig. 7c, d). These apatite separates were prepared for analysis after it was confirmed by SEM that they were inclusion-free (4 from 17 examined samples in total). Additionally, the high-grade iron ore concentrate of two lines of the process plant were sampled to be included in iron oxide geochemistry. Five samples were also collected from the actinolite-rich metasomatic host rocks. A portion of the pulverized sample was accurately weighed into a PTFE tube, and a mixture of acids, including hydrofluoric acid, was added. The tube was then placed into a pressure vessel and sealed. The samples underwent digestion at high temperature and pressure (up to 200 °C and 20 Bar). After digestion, the samples were diluted to a predetermined volume, ready for analysis. The sample digest solutions were analyzed for major and trace element concentration by ICP-OES (PerkinElmer Optima 7300DV) and ICP-MS (PerkinElmer Nexion 300Q) at LabWest Minerals Analysis, Malaga, Western Australia. The resulting data were collated and assessed for performance of internal quality control results (blanks, duplicates, and certified reference materials) following LabWest’s MMA-04 multielement analysis technique. The results are presented in Tables 1, 2, 3, and 4. Table 5 lists the locations of the ore samples based on the local coordinate system used in the Chador-Malu mine (see Fig. 3).

Table 1 Chemical composition of the iron ore samples (all concentrations in ppm except Fe)
Full size table
Table 2 P and REE composition of the iron-apatite ore samples (all concentrations in ppm except P)
Full size table
Table 3 Chemical composition of the mineral samples (prepared mineral separates) and the high-grade iron ore concentrate (all concentrations in ppm unless indicated otherwise)
Full size table
Table 4 REE, P, Na, K, and Ca composition of the host rocks (all concentrations in ppm unless indicated otherwise)
Full size table
Table 5 Location of the ore samples based on the local coordinate system used in the Chador-Malu mine
Full size table

Results

Ferride distributions in iron oxides

Tables 1 and 3 present the ferride (Ti, V, Cr, Mn, Fe, and Ni) and some other trace metal (Al, Ca, and Co) contents of the iron oxides (iron ore samples as well as magnetite separates) from the Chador-Malu deposit.

The distribution of these metal concentrations (and associated ratios) in the samples is displayed on a compilation of variation diagrams (Fig. 9a–h). None of these diagrams attribute any sample to a BIF affinity. In spaces composed of various combinations of the ferride metals (diagrams constructed based on whole rock compositions; see Loberg and Horndahl 1983), the geochemical data indicate the Chador-Malu iron oxides to have compositions very similar to those of known Kiruna-type iron oxide-apatite deposits (Fig. 9a–e). Although the discriminant diagrams of Beaudoin et al. (2007), Beaudoin and Dupuis (2010), and Dupuis and Beaudoin (2011) (Fig. 9f–h) are constructed based on microanalytical data, the analytical data from our meticulously prepared magnetite separates from magnetite-rich specimens (see above; Fig. 7a) can yield reliable results on these diagrams. One magnetite sample falls in the reference field of the IOCG deposits in the discriminant diagram of Beaudoin and Dupuis (2010) (Fig. 9g), the one that also plots exactly on the boundary between the Kiruna-type and IOCG deposits in the discriminant diagram of Beaudoin et al. (2007) (Fig. 9f). However, when the average values of the data is plotted on the discriminant diagram of Beaudoin and Dupuis (2010) (as recommended by Dupuis and Beaudoin 2011), the data point (denoted by black cross) falls in the field of the Kiruna-type deposits (Fig. 9g). Likewise, the average value of the magnetite data on the discriminant diagram of Dupuis and Beaudoin (2011) signals a Kiruna-type affinity (Fig. 9h).

Fig. 9
figure 9

Trace element diagrams of iron oxides. The legend included in “a” describes the symbols used in the entire figure. a–e Iron oxide data in various ferride spaces indicating deposit types. All diagrams after Loberg and Horndahl (1983). f Magnetite data points showing a Kiruna-type affinity. Note the one data point exactly on the boundary between the Kiruna-type and IOCG deposits. Diagram after Beaudoin et al. (2007). g Magnetite data points showing deposit types. Average data point is denoted by black cross. Note the one magnetite data point in the IOCG field. See the text for interpretation. Diagram after Beaudoin and Dupuis (2010). h Average data point of magnetite indicating a Kiruna-type affinity. Diagram after Dupuis and Beaudoin (2011)

Full size image

The Ti content of the iron oxide samples from the Chador-Malu deposit ranges from 294 to 6,270 ppm (Tables 1 and 3), which falls in the corresponding range for the Esfordi deposit (Bafq district) (200–15,000 ppm), with a considerable overlap between their samples (see Jami 2005 for Esfordi’s data). Iron oxide samples of both deposits (Chador-Malu and Esfordi) are poor in Ti relative to the Mishdovan and Choghart deposits (Bafq district) which contain up to 2.5 wt% Ti (Daliran 1990) and 15 wt% Ti (Jami 2005), respectively, yet rich in Ti when compared to the classic Kiirunavaara deposit (northern Sweden) with samples containing up to 20 ppm Ti (Nyström and Henriquez 1994). The greater than an order of magnitude variation in the Ti content of the Chador-Malu iron oxide samples (294–6,270 ppm) may suggest the possibility of the contamination of some samples by discrete Ti mineral phases such as titanite; however, microscopic studies has shown that Ti minerals, and even their exsolution in magnetite, are very rare in the Chador-Malu iron oxide-apatite deposit (Sabet-Mobarhan-Talab and Alinia 2014; Sabet-Mobarhan-Talab 2014; Förster and Jafarzadeh 1994).

Apatite Y-Sr compositions and apatite-monazite relationship

Y-Sr composition of apatite samples (mineral separates) from the Chador-Malu deposit indicates a Kiruna-type iron ore and mafic rock genetic association (Fig. 10).

Fig. 10
figure 10

Y-Sr composition of the Chador-Malu apatites. Diagram after Belousova et al. (2002)

Full size image

Several apatite samples were examined by SEM in order to select inclusion-free apatites for chemical analysis (see above). Monazite inclusions were observed in most apatite samples. Figure 11 shows an example of monazite-bearing apatite. Such monazite inclusions (Fig. 11) may nucleate through REE redistribution in apatite through fluid-mineral interaction (e.g., Harlov and Förster 2003; Harlov et al. 2005), and are typical of the classic Swedish Kiruna-type deposits (e.g., Parak 1975; Harlov et al. 2002). Apatite REE leaching and monazite formation in Kiirunavaara deposit was ongoing through infiltration of high-temperature brines (700–800 °C), which were originated from the source magma subsequent to the magmatic emplacement of a proposed Fe-P melt (Harlov et al. 2002). Monazite inclusions have also been observed in the apatites from other Bafq district deposits (e.g., Torab and Lehmann 2007; Stosch et al. 2011; Bonyadi et al. 2011).

Fig. 11
figure 11

Backscattered Electron (BSE) images and elemental maps of the Chador-Malu apatite. a BSE-bright and BSE-dark patterns and monazite inclusions in brecciated apatite indicating REE redistribution during fluid-mineral interaction. b Magnified BSE image of the monazite inclusion circled in “a”. c and d Elemental maps of the inclusion in “b” verifying that the Ce-rich phosphate phase is actually monazite (Ap: apatite; Mnz: monazite)

Full size image

REE concentrations and chondrite-normalized patterns

REE compositions of the samples from the Chador-Malu deposit are listed in Tables 1, 2, 3, and 4. The REE content of the mineral separates of apatite ranges from 0.85 to 1.42 wt%, but is above 1 wt% in three of the them (Table 3). The total REE content of the magnetites (mineral separates) and iron ore concentrate is below 220 ppm, and can be as low as 112 ppm (Table 3). This value ranges from 0.1 to 0.23 wt% for the iron-apatite ore samples, but is commonly lower than 0.15 wt% (Table 2).

The chondrite-normalized REE patterns of the Chador-Malu apatite (mineral separates), magnetite (mineral separates), iron ore, iron-apatite ore, and high-grade iron ore concentrate, as well as the Chador-Malu and the Zarigan granites are represented in Fig. 12a–d. The REE distribution in the Chador-Malu magnetite is similar to those of the apatite (Fig. 12c). The REE patterns of the iron-apatite ore samples also resemble those of the apatites (Fig. 12a). However, the REE patterns of the iron ore samples show two different trends, particularly in the heavy rare earth element (HREE) portion with one branch of elevated curves (Fig. 12b).

Fig. 12
figure 12

Chondrite-normalized REE patterns of the Chador-Malu deposit and granite. a REE patterns of the iron-apatite ore samples. b REE patterns of the iron ore samples. Note that HREE patterns branch into two trends. See the text for interpretation. c Similar apatite, magnetite, and iron ore concentrate REE patterns. d REE pattern of the Chador-Malu granite. Note the similar REE pattern of the Zarigan granite. See the text for interpretation. Analytical data of the granites from Ramezani and Tucker (2003). Chondrite values from Anders and Grevesse (1989)

Full size image

We noticed that the iron ore samples with elevated HREE patterns (Fig. 12b) have higher Ca contents (>0.44 %) than other iron ore samples (<0.26 %). The actinolite-rich host rock of the Chador-Malu deposit has locally undergone intense Fe-metasomatism (Sabet-Mobarhan-Talab 2014). It is therefore very likely to have ore samples contaminated with minor amounts of the actinolite-rich host rock, particularly when collecting samples near the contacts. This might account for the relatively high Ca contents of some of the iron ore samples (Table 1).

The REE patterns of the actinolite-rich host rock samples are shown in Fig. 13. Three host rock samples were collected adjacent to the contacts (CHR1a, CHR1b, and CHR2; see Table 4 for coordinates). These samples show elevated HREE patterns. However, two host rock samples were collected at some distance (up to 20 m) from the contacts (OCHR1 and OCHR2; Table 4). The latter have REE patterns similar to most other minerals and ores (Figs. 12a–c and 13). While in magmatic liquids, middle rare earth elements (MREE) are more compatible in amphiboles than HREE (e.g., Rollinson 1993), and the systematics of REE partitioning between fluid and mineral phases in relevant geological conditions to this study are not yet well-established, it can be postulated that the interaction of a fluid with a high HREE/LREE fractionation (see below) with the actinolite-rich host rocks could produce the HREE-elevated chondrite-normalized patterns of the host rock samples (Fig. 13). Thus the hydrothermal mobilization of the apatite-hosted REE during overprinting, apatite leaching and monazite formation (see above; Fig. 11) could be proposed. This process would involve the release of the REE, particularly the HREE, from apatite to produce the fluid with high HREE/LREE fractionation proposed above. The HREE would more readily enter the fluid phase since they would not be significantly partitioned into the crystallizing monazite inclusions, and are more susceptible to complexing (Humphris 1989) due to their relatively higher field strength (Rollinson 1993). The uptake of the HREE by abundant actinolite in the host rocks would be important. This hypothesis, however, needs to be verified by further investigation. This translocation of the REE was most probably restricted to the adjacent host rocks which is suggested by the REE patterns of the host rocks close to (CHR1a, CHR1b, and CHR2) and away from (OCHR1 and OCHR2) the contacts with the ore (Fig. 13). This observation is consistent with the mobility of up to several centimeters for the REE during apatite leaching and monazite formation in the Se-Chahun iron oxide-apatite deposit proposed by Bonyadi et al. (2011). It is also clear that the elevated HREE patterns of the iron ore samples mentioned above (Fig. 12b) are due to the contamination with small amounts of the adjacent actinolite-rich host rocks.

Fig. 13
figure 13

Chondrite-normalized REE patterns of the actinolite-rich metasomatic host rocks. Chondrite values from Anders and Grevesse (1989)

Full size image

The degree of LREE/HREE fractionation, represented by (La/Yb)N ratio, ranges from 31.57 to 47.47 for the apatites, from 19.07 to 39.06 for the magnetites, from 24.90 to 25.46 for the high-grade iron ore concentrate, and from 19.40 to 36.60 for the iron-apatite ore samples and indicates light rare earth element (LREE) enrichment relative to HREE; but it ranges from 3.23 to 27.20 for the iron ore samples, which display two distinct HREE trends. The extent of LREE/HREE fractionation can also be represented by (La/Lu)N ratio, which in the case of actinolite-rich host rocks, ranges from 3.43 to 7.35 for the samples adjacent to the contacts (CHR1a, CHR1b, and CHR2), and for the other two samples away from the contacts (OCHR1 and OCHR2) is 12.33 and 17.17.

Overall, except for the iron ore samples with raised HREE patterns, almost all ore and mineral samples display REE patterns that follow the same path, particularly in their MREE and HREE portions. All these REE patterns can be described as similar, regarding their steep slope with relatively high LREE/HREE fractionation and the negative Eu anomaly (Fig. 12a–c).

The chondrite-normalized REE pattern of the Chador-Malu granite is significantly different from the iron oxide-apatite deposit, mainly in terms of LREE/HREE fractionation and Eu anomaly (Fig. 12d). Interestingly, the REE pattern of the Zarigan granite, which is located further south in the Bafq district, and is associated with iron oxide-apatite mineralization (Fig. 2), is very similar to the Chador-Malu granite (Fig. 12d).

Discussion

Geochemistry of iron oxides

The geochemistry of iron oxides can indicate the mineral deposit type to which they belong (e.g., Loberg and Horndahl 1983; Beaudoin and Dupuis 2010; Dupuis and Beaudoin 2011; Nadoll et al. 2012). While some major Bafq district iron oxide-apatite deposits have been related to the Rapitan-type banded iron-phosphorus formations (e.g., Aftabi et al. 2009; Mohseni and Aftabi 2012), the ferride and trace element geochemistry of the iron oxide samples (iron ores and magnetite separates) from Chador-Malu does not suggest any BIF affinity (Fig. 9). Instead, the whole rock composition of the Chador-Malu iron ores, as well as the composition of the prepared magnetite separates, are similar to the whole rock composition of the Kiruna-type iron ores, including the Swedish ones (see Loberg and Horndahl 1983; Fig. 9a–e), for which direct magmatic genetic models have been proposed by several authors (e.g., Philpotts 1967, 1982; Frietsch 1978, 1984; Nyström and Henriquez 1994; Naslund et al. 2000; Jonsson et al. 2013); however, such models are highly disputed and no longer widely accepted. The composition of the magnetite separates (and/or their average composition) is similar to the mineral composition of the Kiruna-type magnetites (Fig. 9f–h), with one exception which shows an IOCG attribute on the diagrams in Fig. 9f, g. The similar ferride and trace element geochemistry of the iron oxides from Chador-Malu and other Kiruna-type deposits may indicate similar ore-forming processes in their genesis. Additionally, the distribution of Ti concentration in the iron oxide samples from the Chador-Malu and Esfordi deposits is similar (see above). Jami (2005) proposed that the mineralizing fluids at Esfordi were in equilibrium with immiscible Fe-P melts. Similar fluids may have been involved in the formation of the Chador-Malu deposit.

Moreover, the position of the magnetite data points on the diagrams in Fig. 9 can be indicative of some comparable features the geochemistry of the Kiruna-type and Cu-Au-bearing IOCG deposits may have, both of which having been grouped together as the end-members of the IOCG clan according to the classification by Hitzman et al. (1992) (see also Groves et al. 2010).

Texture and composition of apatites

The Y-Sr composition of the Chador-Malu apatites (Fig. 10) supports the finding of Torab and Lehmann (2007) that the Bafq apatites are chemically very similar to those of other Kiruna-type deposits (see also Bonyadi et al. 2011). However, these results are not consistent with the previous discussions suggesting a carbonatitic affinity for apatites of the Bafq district deposits (e.g., Daliran 2002; Darvishzadeh 1983; Darvishzadeh and Aletaha-Kohbanani 1996). Carbonatitic apatites are characterized by high Sr (commonly more than 2,500 ppm) and relatively low Y contents (commonly less than 400 ppm) (Belousova et al. 2002), which is not the case in the Chador-Malu deposit (Table 3; Fig. 10).

The apatite REE leaching and monazite formation in the Chador-Malu deposit (Fig. 11) can be associated with a high-temperature overprinting event similar to the classic Kiirunavaara deposit (Harlov et al. 2002). The formation of second generation fine-grained anhedral apatite (ApII), which is associated with primary coarse-grained subhedral to euhedral apatite (Fig. 6f), can also be related to the same overprinting event, and is consistent with the restricted mobility of P (millimeters to centimeters) in the high-temperature brines responsible for apatite REE leaching and associated monazite formation (see Bonyadi et al. 2011). These textural observations argue for metasomatism and highlight the importance of high-temperature hydrothermal fluids at Chador-Malu. A magmatic source for similar brines is proposed at Kiruna (Harlov et al. 2002, 2005; Harlov and Förster 2003).

REE patterns and their significance

Similar magnetite and apatite REE patterns have widely been interpreted as an indicator of their common origin (see, e.g., Frietsch and Perdahl 1995; Nabatian et al. 2012; Mokhtari et al. 2013, among others). The REE patterns of the magnetite and apatite samples from Chador-Malu are very similar in terms of their LREE/HREE fractionation and negative Eu anomalies (Fig. 12c; Table 3). This can suggest a common origin for these minerals. However, the magnetites may have inherited their REE patterns (Fig. 12c) from micro inclusions of apatite and/or fluid-mineral interaction during the hydrothermally overprinting events discussed above.

Almost all mineral and ore samples (except the HREE-elevated iron ores) from Chador-Malu have REE patterns similar to the Kiruna-type deposits from northern Sweden (see, e.g., Frietsch and Perdahl 1995; Harlov et al. 2002). There are also other Bafq district iron oxide-apatite deposits which share the same typical REE patterns (see, e.g., Jami 2005; Torab 2008; Bonyadi et al. 2011, among many others). This demonstrates that the regional processes involved in the metallogeny of the Bafq district can be similar to the Kiruna district.

Genetic discussion

A potential source for Fe and P

There is consensus in the literature that the close spatial and temporal association of iron oxide-apatite deposits with igneous rocks is consistent with magma as a direct or indirect source of Fe and P (see the review above). Supporting a magmatic origin, Förster and Jafarzadeh (1994) pointed out melt flow and quench textures in the Bafq district. Such features have also been reported for the Kiirunavaara and El Laco iron oxide-apatite deposits by Harlov et al. (2002) and Naslund et al. (2002), who considered the mineralization to be the product of direct crystallization from immiscible iron oxide melts. Field evidence for an iron oxide-dominated magma (e.g., magnetite-bearing pyroclastic rocks and lava flows) have been reported from Sweden, Chile, and Iran (e.g., Frietsch 1978; Nyström and Henriquez 1994; Förster and Jafarzadeh 1994). Iron ± titanium ± phosphorus melts have also been produced experimentally by immiscible separation from silica-rich liquids (e.g., Philpotts 1967; Naslund 1983). Such immiscible iron oxide melts, into which phosphorus is strongly partitioned (Clark and Kontak 2004), can be generated in various thermodynamic and physicochemical conditions (Hurai et al. 1998), as a late-stage differentiate of the magmatism (e.g., Frietsch and Perdahl 1995). In the Bafq district terms, differentiating alkaline mafic magmas may have generated immiscible Fe-P melts, as suggested by Förster and Jafarzadeh (1994). From a regional point of view, the major granitic intrusions of the Bafq district (i.e., Chador-Malu and Zarigan granites; Fig. 2) could be assumed to be the alkaline silicic residue of this regional-scale immiscible separation of the Fe-P melts. The produced Fe-P melts would be subsequently equilibrated with fluids (temperature?) which would ascend through the zones of ring fracture (Fig. 2; Förster and Jafarzadeh 1994) and would eventually produce the mineralization at Chador-Malu (and similar parts of the Bafq district). This scenario, however, remains largely speculative, but some geochemical discussion in support of this view is provided below.

REE partitioning between the proposed immiscible phases could produce distinctive chondrite-normalized REE patterns which would provide some indication of the separation condition. The positive Eu anomaly in the Chador-Malu granite (Fig. 12d) versus the negative Eu anomaly in the Chador-Malu iron oxide-apatite deposit (Fig. 12a–c) is the most conspicuous attribute of their REE patterns. Due to the compatibility of divalent Eu in plagioclase and K-feldspar structure, Eu anomalies are predominantly controlled by feldspars, particularly in felsic magmas (Rollinson 1993). Additionally, Eu can exist in both the 2+ and 3+ oxidation states in magmatic systems, depending on the redox potential, which is a function of oxygen fugacity (Henderson 1996). Naslund (1983) documented the exsolution of an iron oxide melt in the K(Na)AlSi3O8-FeO-Fe2O3-SiO2 system under strongly oxidizing conditions; however, investigating the REE patterns of the immiscible iron oxide globules in the syenitic and alkali rhyolitic xenoliths in the alkali basalts of the Western Carpathians, and comparing them to the REE patterns of the host K-feldspar-bearing rocks, Hurai et al. (1998) interpreted the positive Eu anomalies in the feldspars and the negative Eu anomalies in the conjugate iron oxide melt as an indicator of the magmatic origin of the latter, and concluded that low oxygen fugacity promoted the immiscibility. Likewise, feldspar crystallization and a reducing condition during immiscibility at the Chador-Malu area might have resulted in the generation of Fe-P melts with a negative Eu anomaly which would correspond to the positive Eu anomaly in the Chador-Malu granite (Fig. 12d). Mineralizing fluids equilibrated with the proposed Fe-P melts would have produced the negative Eu anomalies in the Chador-Malu iron oxide-apatite deposit (Fig. 12a–c).

Moreover, Yb and Lu concentrations in the Chador-Malu granite produce an elevated HREE pattern (Fig. 12d). In felsic liquids, accessory phases such as garnet and zircon, may strongly influence a REE pattern, for although they may be present in only small quantities (often much less that 1 % of the rock), their very high partition coefficients mean that they have a disproportionate effect on the whole rock REE pattern. These minerals deplete in HREE so that extreme depletion of HREE relative to LREE is most likely to indicate their presence in the source (Rollinson 1993). Accordingly, the presence of garnet and zircon (documented in the mineral composition of the Chador-Malu granite; Berberian and Berberian 1981) in the silicate magma during immiscibility might have resulted in the generation of Fe-P melts depleted in HREE (relative to LREE), which would correspond to the elevated HREE pattern of the Chador-Malu granite (Fig. 12d). The equilibration of the proposed Fe-P melts with fluids responsible for mineralization would have resulted in the HREE-depleted patterns observed in the Chador-Malu deposit (Fig. 12a–c).

It should be noted that the Zarigan granite, which is associated with certain iron oxide-apatite systems in the Bafq district (Fig. 2), has chondrite-normalized REE patterns very similar to the Chador-Malu granite (Fig. 12d). This suggests that a similar process might have been active in other parts of the Bafq district. The immiscible separation of Fe-P melts can be associated with the mingling of mafic and/or intermediate and felsic magmas (e.g., Clark and Kontak 2004). Although no mafic body has been documented to be present in the Chador-Mau granite, the syn-intrusive character of mafic bodies in the Narigan granite (Jami 2005) located further south in the Bafq district (Fig. 2) suggests that similar immiscibility models may be plausible in the region.

Features in favor of a metasomatic model

There is abundant evidence for the activity of hydrothermal fluids at Chador-Malu. The main host rocks to the iron orebody (metasomatites; Fig. 4b) have undergone intense actinolitization (Fig. 8a, b). These rocks can be interpreted as highly altered CVSU rhyolites based on the spatial relationship between the two rock units (Fig. 3), absence of linear contacts between them, and microscopic studies which suggest that at least some minerals in the groundmass of the metasomatites (predominantly quartz) are relict magmatic (Fig. 8). The host rock alterations include K-feldsparization (associated with sericitization), albitization, actinolitization (dominant type), and late stage silicic alteration (Fig. 8), which are consistent with the multistage model of alteration in relation with the intrusion of the plutonic bodies of the Bafq district proposed by Daliran et al. (2007). The mentioned host rock alterations are overprinted with a Fe-metasomatism at Chador-Malu (Fig. 8) which is consistent with the generation of Fe-P melts as a late-stage differentiate of magmatism proposed by researchers from both Kiruna and Bafq (Förster and Jafarzadeh 1994; Frietsch and Perdahl 1995).

The northern orebody of the Chador-Malu deposit (this study; Figs. 3 and 4a) is characterized chiefly by massive mineralization style (Fig. 4b) and a cylindrical morphology, and has been interpreted as a Fe-P melt-filled pipe-like diatreme (Jafarzadeh 1981; Förster and Jafarzadeh 1984, 1994); however, these features (i.e., massive orebody cross-cutting the wall rock) may also be described as a product of pervasive metasomatic replacement in the CVSU. Evidence for high-temperature hydrothermal fluids at Chador-Malu includes the intense actinolitization in the host rocks (Fig. 8a, b) and the nucleation of apatite-hosted monazite inclusions (typically occurs at temperatures >700 °C; Harlov et al. 2002; Fig. 11). Another evidence for hydrothermal mineralization is the presence of the breccias consisting of the angular fragments of the metasomatic host rocks cemented by iron oxides as well as the observation of the small ore veins (Fig. 5). The dolomites in the ore zone (Fig. 3) might also suggest that at least part of the rocks replaced by mineralizing fluids were sedimentary rocks of the CVSU. This is consistent with the precipitation of magnetite and apatite in response to the reaction of mineralizing fluids with dolomitic country rocks and changes in pH and Eh proposed by Torab (2008).

The Bafq mining district represents a volcanic field, with extensive activity associated with cauldron subsidence in Early Cambrian (Förster and Jafarzadeh 1994). Hydrothermal media could link the magmatism (and possible Fe-P melt differentiates, see above), through the zones of ring fracture (Fig. 2), to the iron oxide-apatite mineralization at Chador-Malu (and similar parts of the Bafq district). Some evidence for the effects of high-temperature hydrothermal fluids has been presented here; however, further work is required to study the principal source and characteristics of these fluids.

Conclusions

The geology and geochemistry of the Chador-Malu iron oxide-apatite deposit demonstrates its similar characteristics to the Kiruna-type deposits. There is circumstantial evidence for REE mobilization during the leaching of apatite by high-temperature fluids and the associated nucleation of monazite inclusions (Fig. 11). Another evidence for high-temperature fluids at Chador-Malu includes the pervasive replacement of the CVSU rhyolites by actinolite (actinolitization; Fig. 8a, b), which resulted in the formation of the actinolite-rich metasomatic host rocks with a distinctive green appearance (metasomatites; Fig. 4b). The small ore veins and the breccias cemented by iron ore (Fig. 5), as well as the Fe-metasomatism which overprints other types of alteration in the host rock (Fig. 8) support hydrothermal mineralization. Based on the REE geochemistry and spatial relationships, it is proposed that a potential source for metals and P would be late stage Fe-P melt differentiates of the Cambrian magmatism, which is consistent with the late Fe-metasomatism in the host rocks. The proposed Fe-P melts and the mineralization would be linked by hydrothermal media through the zones of ring fracture at Chador-Malu and similar parts of the Bafq district.